Aristotle’s Physics

Aristotle’s study of the natural world plays a tremendously important part

in his philosophical thought. He was very interested in the phenomena of
motion, causation, place and time, and teleology, and his theoretical
materials in this area are collected in his Physics, a treatise of eight books
which has been very influential on later thinkers. This volume of new
essays provides cutting-edge research on Aristotle’s Physics, taking into
account recent changes in the field of Aristotle in terms of its under-
standing of key concepts and preferred methodology. The contributions
reassess the key concepts of the treatise (including nature, chance, teleol-
ogy, art, and motion), reconstruct Aristotle’s methods for the study of
nature, and determine the boundaries of his natural philosophy. Because of
the foundational nature of Aristotle’s Physics itself, the volume will be a
must-read for all scholars working on Aristotle.

mariska leunissen is Assistant Professor of Philosophy at the

University of North Carolina, Chapel Hill. Her recent publications
include Explanation and Teleology in Aristotle’s Science of Nature (2010).
 cambridge critical guides

Titles published in this series

Hegel’s Phenomenology of Spirit
edited by dean moyar and michael quante
Mill’s On Liberty
edited by c. l. ten
Kant’s Idea for a Universal History with a Cosmopolitan Aim
edited by amélie oksenberg rorty and james schmidt
Kant’s Groundwork of the Metaphysics of Morals
edited by jens timmermann
Kant’s Critique of Practical Reason
edited by andrews reath and jens timmermann
Wittgenstein’s Philosophical Investigations
edited by arif ahmed
Kierkegaard’s Concluding Unscientific Postscript
edited by rick anthony furtak
Plato’s Republic
edited by mark l. mcpherran
Plato’s Laws
edited by christopher bobonich
Spinoza’s Theological-Political Treatise
edited by yitzhak y. melamed and michael a. rosenthal
Aristotle’s Nicomachean Ethics
edited by jon miller
Kant’s Metaphysics of Morals
edited by lara denis
Nietzsche’s On the Genealogy of Morality
edited by simon may
Kant’s Observations and Remarks
edited by richard velkley and susan shell
 Augustine’s City of God
edited by james wetzel
Descartes’ Meditations
edited by karen detlefsen
Kant’s Religion within the Boundaries of Mere Reason
edited by gordon michalson
Kant’s Lectures on Anthropology
edited by alix cohen
Kierkegaard’s Fear and Trembling
edited by daniel conway
Kant’s Lectures on Ethics
edited by lara denis and oliver sensen
Aristotle’s Physics
edited by mariska leunissen
Aristotle’s Politics
edited by thornton lockwood and thanassis samaras
ARISTOTLE’S
Physics
A Critical Guide

edited by

mariska leunissen
 University Printing House, Cambridge cb2 8bs, United Kingdom

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781107031463
© Cambridge University Press 2015
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2015
A catalogue record for this publication is available from the British Library
Library of Congress Cataloguing in Publication data
Aristotle’s physics : a critical guide / edited by Mariska Leunissen.
pages cm. – (Cambridge critical guides)
isbn 978-1-107-03146-3
Includes bibliographical references.
1. Aristotle. Physics. 2. Physics – Early works to 1800. 3. Philosophy,
Ancient. 4. Science – Philosophy. I. Leunissen, Mariska, 1979– editor.
q151.a8a76 2015
530–dc23
2015014691
978-1-107-03146-3 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy
of URLs for external or third-party internet websites referred to in this publication,
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.
 Contents

Notes on contributors page ix

Introduction 1
mariska leunissen

1 How to study natural bodies: Aristotle’s μέθοδος 10

james g. lennox
2 Aristotle on interpreting nature 31
sean kelsey

3 Nature as a principle of change 46

stasinos stavrianeas

4 Aristotle on chance as an accidental cause 66

james allen
5 Man from man but not bed from bed: Nature, art
and chance in Physics ii 88
margaret scharle
6 In defense of the craft analogy: Artifacts and
natural teleology 107
charlotte witt
7 The origins of Aristotle’s natural teleology in Physics ii 121
robert bolton

8 Substantial generation in Physics i.5–7 144

devin henry

9 A dynamic ontology: On how Aristotle arrived at

the conclusion that eternal change accomplishes ousia 162
diana quarantotto

vii
viii Contents
10 Aristotle’s processes 186
david charles

11 Physics v–vi versus viii: Unity of change and disunity

in the Physics 206
jacob rosen

12 Perfection and the physiology of habituation according

to Physics vii.3 225
mariska leunissen

13 Self-motion as other-motion in Aristotle’s Physics 245

ursula coope
14 The argument of Physics viii 265
andrea falcon

Bibliography 284
Index 296
 Contributors

james allen is Professor of Philosophy at the University of Toronto,

Canada. His principal interests are in ancient Greek and Roman philoso-
phy. He is the author of articles concerning ancient conceptions of exper-
tise, ancient scepticism, ancient medicine, Aristotelian logic,
Epicureanism, Stoicism, and Cicero, as well as of the monograph
Inference from Signs: Ancient Debates about the Nature of Evidence (2001).
robert bolton is Professor of Philosophy at Rutgers University, USA.
He is the author of Science, dialectique et éthique chez Aristote (2010), and
author or editor of numerous books and articles on Aristotle’s methodol-
ogy, epistemology, and psychology, and on other aspects of ancient philo-
sophy. He is a former Rhodes Scholar and Fellow of the Centre National
de la Recherche Scientifique, Paris.
david charles is Professor of Philosophy at Yale University, USA. He is
the author of Aristotle’s Philosophy of Action (1984) and Aristotle on Meaning and
Essence (2000), and editor of Definition in Greek Philosophy (2010). He has also
written many articles on Aristotle’s discussion of action and teleology.
ursula coope is Professor of Ancient Philosophy and Fellow of Corpus
Christi College, Oxford University, UK. She is the author of Time for
Aristotle: Physics iv.11–14 (2005), as well as of articles on Aristotle’s physics
and philosophy of action.
andrea falcon is Associate Professor in the Department of Philosophy at
Concordia University, Montreal, Canada. He is a specialist on Aristotle and
the Aristotelian tradition in antiquity. He was a member of the Institute for
Advanced Study in Princeton in 2008. He is the author of Corpi e movimenti:
Il De caelo di Aristotele e la sua tradizione nel mondo antico (2001), Aristotle
and the Science of Nature: Unity without Uniformity (2005), and
Aristotelianism in the First Century bce: Xenarchus of Seleucia (2011).

ix
x Notes on contributors
devin henry is an Associate Professor of Philosophy at the University of
Western Ontario, Canada. He received his PhD from King’s College
London, where he wrote a dissertation on the metaphysical foundations
of Aristotle’s account of biological generation. He is the author of several
articles on topics in Aristotle’s philosophy of science (including classifica-
tion, teleology, inheritance, and sexism), as well as Plato’s late
epistemology.
sean kelsey is Associate Professor of Philosophy at the University of
Notre Dame, USA. He is a specialist on Aristotle’s natural science and
metaphysics, and his papers have appeared in Ancient Philosophy, Phronesis,
Pacific Philosophical Quarterly, and Oxford Studies in Ancient Philosophy.
james g. lennox is Professor of History and Philosophy of Science at the
University of Pittsburgh, USA. He is the author of a translation, with
introduction and commentary, of Aristotle: On the Parts of Animals (2001)
and of Aristotle’s Philosophy of Biology: Studies in the Origins of Life Science
(2001), as well as of several edited volumes and numerous articles on
Aristotle’s science and biology. He has held fellowships at the Center for
Hellenic Studies (1983–4); Clare Hall, University of Cambridge (1986–7); and
the Istituto di Studi Avanzati, University of Bologna (2006).
mariska leunissen is Associate Professor of Philosophy at the
University of North Carolina at Chapel Hill, USA. She is the author of
Explanation and Teleology in Aristotle’s Science of Nature (2010), co-editor of
Interpreting Aristotle’s Posterior Analytics in Late Antiquity and Beyond
(2010), and has written several articles on Aristotle’s theory of demonstra-
tion, on topics in his natural science, and on the relation between natural
character and virtue. She has held fellowships at the Center for Hellenic
Studies (2010) and at Leiden University (2013).
diana quarantotto is Professore Aggregato at the University of Rome
Sapienza, Italy. She is the author of Causa finale, sostanza, essenza in
Aristotele (2005), as well as of articles on Aristotle’s science of nature,
biology, metaphysics, philosophy of language and ethics.
jacob rosen is Assistant Professor of Philosophy at Harvard University.
Previously he was the Academic Coordinator in the Graduate Program in
Philosophy and Researcher in the TOPOI Excellence cluster ‘Place, Space,
and Motion’ at the Humboldt University, Berlin, Germany. His papers on
Aristotle’s logic and physics have appeared in Phronesis, Oxford Studies in
Ancient Philosophy, and Mind.
 Notes on contributors xi
margaret scharle is Associate Professor of Philosophy at Reed
College, USA. She works in ancient philosophy, with special interest in
Aristotelian natural philosophy. Her papers have appeared in Oxford
Studies in Ancient Philosophy, Phronesis, and Apeiron, as well as in edited
volumes.
stasinos stavrianeas is Lecturer at the University of Patras, Greece.
He has published papers on Aristotle’s natural science and biology, and a
modern Greek translation and commentary of Aristotle’s Parts of Animals.
charlotte witt is Professor of Philosophy and Humanities at the
University of New Hampshire, USA. She is the author of two books on
Aristotle’s metaphysics, Ways of Being: Potentiality and Actuality in
Aristotle’s Metaphysics (2003) and Substance and Essence in Aristotle (1989).
She has also written many articles on Aristotle’s philosophy, and on
feminist theory and feminist history of philosophy.
 Introduction
Mariska Leunissen

Throughout his life, Aristotle was deeply committed to the study of natural
phenomena: About one-third of the surviving Corpus Aristotelicum inves-
tigates and explains the motions and attributes of things that have a nature,
that is, of things that have an internal principle of change and rest. The
Physics – an intellectual masterpiece in itself and one of the most widely
read Aristotelian treatises – forms Aristotle’s most fundamental treatise in
his studies of natural philosophy.
In this treatise, Aristotle investigates the principles and causes of all
natural things in general, and, in the course of doing so, defines a large
number of key concepts of his natural philosophy, such as motion and
change, space and time, matter and form, causal explanation, luck and
spontaneity, teleology, and necessity. In addition, Aristotle specifies in the
Physics the methodological guidelines for how one should study natural
entities and their properties if one wants to gain scientific knowledge of
them, which includes the famous – but still ill-understood – recommenda-
tion to start from things that are “more known and clearer to us” and to
work from there to what is “more known and clearer by nature.” In this
way, the Physics lays out Aristotle’s conceptual apparatus and methodolo-
gical framework for all of his natural philosophy, including his psychology,
biology, and other inquiries into the more specific and more complex
segments of the natural world preserved in Aristotle’s remaining natural
treatises.
The Physics is relevant not just for Aristotle’s natural philosophy, how-
ever. For, since the objects of metaphysics do not, for the most part, exist
independently of the objects of physics and can thus only be studied
through those, the science of metaphysics often has to rely on the same
concepts, definitions, and approaches as are presented in the Physics. And
the same might be said for certain aspects of Aristotle’s political science:
For instance, the very notion of the perfection of human nature builds on
Aristotle’s on natural teleology and involves a type of change (“perfection”)
1
2 mariska leunissen
that is introduced in the Physics. In this way, the Physics forms the
conceptual entry-way into much of the Aristotelian Corpus.
Despite what its ancient title might suggest, Aristotle’s “Lecture
concerning Nature” (Φυσικὴ ἀκρόασις) barely counts as a unified treatise.
The eight books naturally divide into five separate sections or investiga-
tions: Book i identifies the principles of change at a very general level;
Books ii–iv inquire into nature and natural motion and its preconditions
more properly; Books v–vi provide a technical discussion of how motion
works; and Book viii establishes rather independently of the rest of the
Physics the existence of an unmoved mover. The place of Book vii in the
treatise is disputed: Parts of the book – dealing mostly with the relation
between the mover and what is moved – are handed over in two quite
different versions, and when Eudemus produced his paraphrase of
Aristotle’s Physics he did not include it, suggesting that it was either not
part of the edition of the Physics he possessed or that he did not consider it
part of the same investigation.1
The current volume is not overly concerned with the history, unity, or
structure of the edition of the Physics as we have it today, nor does it try to
provide a comprehensive treatment of the rich materials presented in it.
The existing scholarship on Aristotle’s Physics is wide-ranging and volumi-
nous, and it would be impossible to engage with it all. Instead, as is the case
with the other volumes in the Cambridge Critical Guide series, the fourteen
chapters collected in this volume all attempt to make optimal use of the
recent changes in the field of Aristotle studies – changes both in terms of its
understanding of key concepts in Aristotle’s philosophy and in terms of its
preferred methods for gaining such understanding – and thereby to push
forward the scholarship on Aristotle’s Physics. Each of the chapters engages
with these changed perspectives on Aristotle in at least one of the following
three ways:
(1) Reassessing the key concepts of Aristotle’s natural philosophy.
First, most of the chapters provide a challenge to existing interpretations
of some of the key concepts of Aristotle’s Physics and argue for alternative
understandings: The notions of nature, chance, teleology, and art are
discussed in Chapters 2 to 7, while traditional understandings of the notion
of kinêsis (translated as change, process, or motion) are revised in
Chapters 8 to 13. These chapters not only draw from the latest research
in the field, but also exhibit a greater sensitivity towards the richness and

1
The best discussion of the structure, unity, and transmission history of Aristotle’s Physics is
Brunschwig 1991.
 Introduction 3
complexity of Aristotelian concepts, as well as towards the extent to which
Aristotle builds on and reshapes his concepts in different explanatory
contexts. Instead of just studying the main concepts of Aristotle’s natural
philosophy in the immediate context in which they are first introduced in
the Physics and trying to provide a unified account of their definitions and
roles, the chapters presented in this volume also pay attention to some of
the (methodologically) later uses to which the concepts are put in other
(natural) treatises, provide explanations of why these other uses require
conceptual changes, and answer the meta-question about why Aristotle
needs the specific understandings of, for instance, his concepts of nature
and cause for his natural philosophy as a whole.
(2) Reconstructing Aristotle’s methods for the study of nature. Second,
several of the chapters provide a reconstruction of the methods Aristotle
uses and/or describes for the study of nature, and do so either directly, as in
Chapter 1, or as part of a reassessment of one of Aristotle’s key notions in
his Physics, as for instance in Chapters 3, 8, 9, and 14. Scholars working on
Aristotle have become increasingly interested in the relationship between
Aristotle’s “geometric-style” theory of scientific demonstration and inves-
tigation as presented in the Posterior Analytics and his practice in the
natural treatises, and although much work has been done on the metho-
dological connections between the Posterior Analytics and Aristotle’s biol-
ogy (i.e. his study of living nature), the connections with his Physics remain
largely unexplored territory. The chapters in this volume aim to work
towards closing this gap in the existing scholarship by offering interpreta-
tions of (a) what it means according to Aristotle to investigate things
physikôs – i.e., in the manner of a natural scientist; (b) how this method
relates to other methods available to a philosopher (such as conceptual
analysis or dialectic), as well as to the scientific method outlined in the
Posterior Analytics; and (c) how methodological concerns stemming from
the Posterior Analytics drive the investigations in the Physics.
(3) Determining the boundaries of Aristotle’s natural philosophy.
Third, some chapters – most notably Chapter 14, but also Chapters 1, 9,
11, 12, and 13 – concern themselves with the boundaries (and the extent to
which these boundaries are crossed) between Aristotle’s natural philosophy
and his metaphysics or even his ethics. In recent years, there has been a
growing awareness among scholars of Aristotle that Aristotle’s full views
can rarely be plucked out of single passages, or even single treatises. This
has led to modifications in how we believe one should conduct conceptual
analysis in Aristotle (e.g. as taking into account the different uses to which
concepts are put in different parts of a science, as described under item
4 mariska leunissen
number one above), but it has also highlighted problems concerning
Aristotle’s division of the sciences and his requirement that principles
that belong properly to one science cannot be used in the generation of
knowledge in another science (unless the two sciences are subordinated to
each other). Several of the chapters in this volume examine the ways in
which Aristotle demarcates the science of nature and sets it apart from
other theoretical sciences, as well as how his accounts in the Physics relate to
or are fundamental for Aristotelian views that belong properly to other
sciences, such as his metaphysics, but also his ethics.
The opening chapter by James Lennox addresses the three methodolo-
gical questions concerning the science of nature head on, while also
providing a reconsideration of several of Aristotle’s key notions in the
Physics. Lennox argues that even though the order of and method for the
investigation of nature Aristotle presents in the Physics reflect in important
ways the recommendations for scientific investigation he had already laid
out in the Posterior Analytics, ultimately Aristotle believes that every
scientific domain is governed by norms for inquiry that are quite specific
to that domain. This means that an important task of Aristotle in the
Physics is to specify how natural entities – which undergo change and are
always enmattered – ought to be studied, and thereby to establish the
norms for scientific inquiry at a general level for the whole domain of
natural science. These norms will then have to be specified even further in
the other natural treatises which deal with particular kinds of natures. In
his chapter, Lennox identifies these “local” norms for natural science and
also shows how Aristotle’s concern for establishing these norms for natural
inquiry drives much of the conceptual analysis that can be found in Books
ii and iii regarding especially the notions of change, nature, and motion.
Chapters 2 to 7 offer reassessments of some of the most fundamental
notions Aristotle introduces in Physics ii, such as nature, chance, teleology,
and art, and often do so by drawing from other Aristotelian texts or from
the pre-Socratic and Platonic traditions.
Sean Kelsey, in Chapter 2, turns to Aristotle’s definition of nature in
Physics ii.1 with the idea of tracing its role in the remainder of Book ii,
where explicit appeals to the definition are surprisingly scant. According to
Kelsey, the definition ultimately functions as a kind of instruction for how
to interpret the phenomena of nature. Aristotle defines nature as a princi-
ple of motion and rest; taken together with the idea that nature is above all
form, this implies that form is a principle of motion. Kelsey argues that this
is a result that Aristotle expects us to find difficult: How can form, which is
immobile, be a principle of movement? This difficulty, Kelsey suggests, sets
 Introduction 5
a tacit agenda for the rest of Physics ii, and the key to unraveling it lies in
realizing that the way form is a “principle” (archê) is by being an “end”
(telos). Put another way, the suggestion is that the definition of nature,
together with the identification of nature with form, instructs us (in effect)
to put a certain interpretation on natural phenomena; it tells us to read the
“behaviour” of things due to nature – their characteristic ways of moving
and staying at rest – as for the sake of perpetuating their distinctive forms of
being. The moral is that if nature as form stands at the beginning of natural
phenomena, it must do so by being its end.
Chapter 3 also engages with Aristotle’s definition of nature in Physics
ii.1, but, where Kelsey focused on its role, Stasinos Stavrianeas provides a
reassessment of the content and meaning of this definition. However,
rather than arguing that the definition yields just one criterion for judging
what counts as a nature for Aristotle (such as the standard criterion that is
offered in the scholarship on this topic, namely the possession of an
internal principle of motion and rest), Stavrianeas believes that the defini-
tion provided in the Physics is left intentionally vague and general (in much
the same way as Aristotle leaves his definition of soul vague and general in
On the Soul), thereby allowing for further specifications and filling-in
within the individual natural treatises. By treating Aristotle’s definition
of nature in this way, Stavrianeas is able to handle difficult cases in other
natural treatises, such as the motion of the elements, and even hypothe-
tical, miraculous automata. The aim of Physics ii.1, then, is not to identify
the one thing that characterizes all natural things, but rather to demarcate
the realm – and thereby the science – of nature at the most general level
possible.
In Chapter 4, James Allen reinterprets Aristotle’s notions of luck and
spontaneity (which are both a type of chance) in Physics ii.4–6 by
drawing not just on these chapters, but also on Aristotle’s treatment
of these notions in the ethical treatises. Allen shows that chance for
Aristotle is not an alternative to teleological explanation, which he is
determined to reject, as most standard readings hold, but rather that
chance is an inevitable byproduct of final causation and unintelligible
apart from it. Accordingly, the account of chance in Physics ii.4–6
should be read as a complement to Aristotle’s defense of natural tele-
ology in Physics ii.8. Thus, under the interpretation of Aristotle’s own
account of chance as a “cause by accident” as Allen defends it, chance
events are a marginal exception to the teleological rule that prevails in
nature, and thereby in fact constitute themselves proofs of the existence
of natural teleology.
6 mariska leunissen
Margaret Scharle, in Chapter 5, offers a new interpretation of Aristotle’s
rainfall-example in Physics ii.8. By building on Aristotle’s account of the
teleology of the elements in especially the Meteorology, On Generation and
Corruption, and On the Heavens, Scharle shows that Aristotle understands
winter rain as a natural phenomenon that is teleological in its own right,
and that this teleological view of winter rain is indeed required by the
dialectic Aristotle engages in with the Empedoclean opponent. As a result
of this, Scharle is also able to offer a unified interpretation of the dialectic in
Book ii of the Physics as a whole: The underlying agenda that drives
Aristotle’s discussions in this book is his attempt to carefully carve out
his own position that the natural world owes its order and regularity to
teleology, and he does this against the pre-Socratic tradition that fails to
properly distinguish nature from chance and against the Platonic tradition
that fails to properly distinguish nature from art.
In Chapter 6, Charlotte Witt continues the focus on Aristotle’s defense
of natural teleology in Physics ii.8, this time by turning to his analogy
between art and nature. Against standard readings which have downplayed
the importance of Aristotle’s analogy between art and nature for his
argument in favor of natural teleology, and which have assumed a radical
ontological difference between the two (which sometimes lead them to
conclude that artifacts are not substances), Witt argues that artifacts have
intrinsic ends and proper functions just like natural beings (and are thus
substances), and that this similarity is of crucial importance to Aristotle’s
appeal to the craft analogy in order to argue for natural teleology.
Moreover, she shows that Aristotle is very careful to bracket off the
question of the origin of change causing a creation or generation whenever
he appeals to the analogy between art and nature as part of his argument for
natural teleology. She thereby counters the often-heard objection that
Aristotle’s analogy between art and nature is mistaken or misleading, and
re-establishes its importance for Aristotle’s defense of natural teleology in
the Physics.
While Chapters 5 and 6 focused on Aristotle’s defense of natural tele-
ology, Robert Bolton in Chapter 7 offers an account of its origins, and
thereby corrects some longtime misunderstandings of the nature of his
theory. Crucial to his innovative account is that we should appreciate the
differences between Aristotle and his pre-Socratic predecessors such as
Empedocles, as well as the differences between him and Plato, especially
as these differences are emphasized by Aristotle himself. Aristotelian final
causality, as Bolton understands it, is not – as in Plato – reducible to any
other kind of causality such as efficient causality, but is instead based on the
 Introduction 7
notions of fitness and natural regularity. And in its turn, this notion of
fitness is not something that holds of the natural world by chance – as
Empedocles thinks – but rather something that holds of it by nature. In this
way, Bolton introduces a thorough reassessment of Aristotle’s natural
teleology.
Chapters 8 to 13 offer reinterpretations of Aristotle’s notion of kinêsis –
variously translated as change, process, or motion – while also exploring
issues related to Aristotle’s natural scientific method broadly construed and
offering suggestions about the demarcation of (the science of) nature.
In Chapter 8, Devin Henry provides a critique of the traditional view
that according to Physics i.5–7 Aristotle holds that every change – including
substantial change – requires a persistent subject of change. Instead, Henry
argues that, even though the evidence in Physics i.7 is silent on the matter of
substantial change, evidence from other natural treatises (most promi-
nently from On Generation and Corruption and Generation of Animals)
suggests that Aristotle only believes that substantial change requires a
subject from which the change proceeds, not one that persists as a con-
stituent of the substance that comes into being. Henry thereby reshapes
our understanding of change, while also showing the importance of read-
ing the Physics in the context of Aristotle’s broader natural scientific views.
In Chapter 9, Diana Quarantotto also analyzes Aristotle’s concept of
change and examines the relationship between change and substantial
being (ousia). She observes that Aristotle’s treatment of this issue exhibits
a development from what is more familiar to us to what is “closer to
nature” and claims that the result of this inquiry is a major innovation
by Aristotle within the Greek tradition of natural philosophy. She identi-
fies this development as mostly taking place within the Physics, especially in
Books i–iii and viii. In Books i–iii, where Aristotle presents the view
that is more familiar to us, there is a clear-cut distinction between sub-
stantial being and change. However, in Book viii (as well as in passages
from other natural treatises), Aristotle presents views that are “closer to
nature,” and this is where Aristotle presents his innovative idea that being is
itself dynamic. Critical to this inquiry is the relationship between the
definition of change given in iii.1–3 and the treatment of eternal change
in Book viii.
David Charles, in Chapter 10, analyzes three aspects of Aristotle’s
account of processes (kinêseis): namely, (1) Aristotle’s definition of process
in Physics iii.1–3 and what exactly this definition entails when read in the
broader context of the Physics and Metaphysics Theta; (2) Aristotle’s account
of the individuation of processes in Physics iii.3; and (3) the nature and role
8 mariska leunissen
of the concept of actuality in Aristotle’s definition of process, especially as it
is used in Metaphysics Theta. One major upshot of Charles’ account is that
it is a mistake to think of kinêseis as events rather than processes; if he is
right, this has major repercussions for our understanding of Aristotle’s
notions of action, time, and causation.
In the next chapter, Chapter 11, Jacob Rosen explores the relationship
between Aristotle’s account of the basic structure of motions and other
continuants in Physics v–vi and his arguments in Physics viii.8 that there
can be no eternal motion along a straight line, and that therefore eternal
motion (which he thinks must exist) must be circular. Rather than trying
to read these various accounts as presenting one unified theory of motion,
Rosen zooms in on the tensions between these accounts, especially regard-
ing Aristotle’s thesis about continua as formulated in Physics viii.8,
according to which continua do not possess any actually existing proper
parts or middle-points, the defense of which conflicts with an important
theorem presented in Physics vi.5. From this he concludes that Books v, vi,
and viii – even though not completely unrelated – are, most likely, not
part of one single lecture that was composed in one sitting, but also, and
more importantly, that Aristotle’s physics and cosmology face some diffi-
culties establishing the priority of circular motion over rectilinear motion if
some of his strongest arguments from Physics viii.8 in favor of this thesis
are already pre-emptively defeated by claims made in Physics v–vi.
In my own chapter, Chapter 12, I turn to Aristotle’s account in Physics
vii.3 of the changes one undergoes when acquiring virtues of character. In
this chapter, Aristotle argues, somewhat surprisingly, that “conditions” –
such as the virtues of the body or soul – do not belong to the category of
“quality” and do not come to be as a result of qualitative change, but that
they are rather “some kind of perfections” that “exist in virtue of a
particular relation,” thereby seemingly introducing a fifth type of change
in addition to the traditional four (substantial change, qualitative change,
quantitative change, and locomotion). This chapter analyzes Aristotle’s
account of the type of change involved in the acquisition of virtues in the
Physics and shows how it offers a physiological and naturalistic explanation
for his account of habituation or “the perfection of human nature” in the
Nicomachean Ethics. In a way, the Physics thus provides a physical ground-
ing for Aristotle’s political science, as only in the Physics do we learn what
kind of change is involved in moral development.
In Chapter 13, Ursula Coope turns to Aristotle’s account of self-motion
in Physics viii.5. In this chapter, Aristotle characterizes self-movers as
involving two components – a part that produces the movement while
 Introduction 9
being itself unmoved, and a part that is moved. Although Coope also offers
an analysis of Aristotle’s notion of self-motion in her chapter – in particular
of how it forms a response to Plato’s suggestion that all motion must
ultimately be grounded in self-motion – and solves some problems that
seem particular to Aristotle’s alternative, she is mostly interested in deter-
mining the role of Aristotle’s appeal to self-motion in the overall project of
establishing the existence of an unmoved mover (an appeal that seems
surprising, given Aristotle’s arguments in Metaphysics L, which establish its
existence without such an appeal to self-motion). According to Coope,
Aristotle’s overall aim is to show how motion must ultimately be caused by
something that is not itself moving. She argues that his account of self-
movers in Physics viii.5 gives necessary but not sufficient conditions for
being a self-mover properly speaking. As such, it is able to apply both to
genuine self-movers, such as animals, and also to the conjunction of the
first unmoved mover together with the thing it causes to be in eternal
motion. Providing a single account that covers both these cases helps
Aristotle to show how physics, in so far as it is the study of motion, is a
single unified science.
For this volume, the issue concerning the relation between the different
sciences arises most prominently with regard to Aristotle’s concept of the
unmoved mover – the origin of all motion in the universe – which he
introduces in Physics viii. Since this unmoved mover is a divine being that
is pure form and is without matter, it is technically speaking not part of the
physical world, but rather belongs to Aristotle’s “first philosophy” or
metaphysics. Aristotle’s treatment of the unmoved mover in the argument
of Physics viii and the questions it raises for the boundaries between
Aristotle’s physics and metaphysics are the topics of the fourteenth and
final chapter, by Andrea Falcon. Falcon argues that, against traditional
readings of Physics viii, Aristotle’s treatment of the unmoved mover there
does not, and need not, go beyond the boundaries of natural philosophy,
but that instead he offers a single extended natural scientific argument
concerned with eternal motion. This argument proceeds in two stages that
follow the two stages of inquiry as presented in the Posterior Analytics:
Aristotle first sets out to determine whether there is eternal motion and then
tries to identify what eternal motion is. The unmoved mover comes into
play as the efficient causal factor that ultimately needs to be picked out by
the definition of eternal motion. In this way, Falcon’s chapter also con-
tributes to our understanding of Aristotle’s scientific methods in the
Physics.
 chapter 1

How to study natural bodies: Aristotle’s μέθοδος

James G. Lennox

Introduction
If, as the opening sentences of the Physics strongly suggest, the basic
framework for inquiry in the pursuit of scientific knowledge presented in
Posterior Analytics ii is shaping Aristotle’s views about natural inquiry, then
there are a number of features one would expect to find. In light of Posterior
Analytics ii.1–2, there ought to be two intimately related goals of natural
inquiry: knowledge of what natural beings are, ideally to be formulated in
definitions; and knowledge of causes used to demonstrate why natural
beings have the necessary but non-essential attributes that they do; and
fact-establishing stages on the way to those goals.1 The intimate relationship
between these two lines of inquiry stems from Aristotle’s conviction that
the essences of things are, in various ways, causally responsible for their
non-essential features – if one has scientific knowledge of what something
is, one will thereby be able to explain why it has the other non-accidental
features it has.
Nevertheless, the conviction motivating this chapter is that, notwith-
standing the generality of this framework, Aristotle sees different subject
matters or domains as governed by norms that are quite specific to them.2
This specificity derives from (at least) three sources: (1) differences in the
objects being investigated; (2) differences in our epistemic access to those
objects; and (3) differences in the perspective we take on those objects.3

This chapter has benefitted greatly from comments on an earlier draft by Marko Malink, Tom
Ainsworth, and our editor, Mariska Leunissen; and from discussions with Christopher Shields,
David Charles, Michael Peramatzis, and Alan Code about this chapter’s concerns. I am reasonably
confident none of them will fully agree with the conclusions I have reached.
1
See Posterior Analytics ii.1–2. The view is, of course, much more complicated than this brief
summation suggests. For detailed discussion of the complications, see Charles 2000, 2010b;
Lennox 2004.
2
This is the primary thrust of my forthcoming Seeking and Knowing: Aristotle on Norms of Inquiry.
3
For concrete examples of how each of these sources might affect the way in which inquiry is carried
out, consider Aristotle’s remarks in Parts of Animals i.1 on the differences between studying natural

10
 How to study natural bodies 11
Thus, while the Posterior Analytics provides what I will call an erotetic
framework for scientific inquiry, it does not provide the domain-specific
norms that are needed to guide actual research. This chapter aims to
articulate those domain-specific norms that Aristotle thinks guide natural
inquiry. A key concept in Aristotle’s discussion of this issue is μέθοδος.
Tellingly, it is a concept completely absent from the Posterior Analytics, and
yet extremely prominent whenever Aristotle is discussing how to proceed
successfully in a specific inquiry – including, as we will see, many of the
opening sentences of his treatises in various fields of inquiry. It is with this
concept that I begin.

I A discourse on μέθοδος
Physics iii opens by noting that there has been a serious lacuna in the
discussion up to that point:
Since the nature [of a natural thing] is a source of motion and change, and
our μέθοδος is concerned with nature, [the question] what is motion must
not escape our notice; for necessarily when we are ignorant of this we are also
ignorant of nature. (Phys. iii.1, 200b12–15)
Our μέθοδος is concerned with nature – but what is a μέθοδος, and what is
distinctive about one concerned with nature? I begin by attending to the
concept of μέθοδος in Aristotle generally, and return to the question of
what is distinctive about a natural μέθοδος in the next section.
Based on how often Aristotle uses this concept in the introductions to
his treatises,4 this is a concept worthy of more attention than it has
received.
The opening lines of the Physics can be used to illustrate the problems
people have in deciding how to translate μέθοδος. The relevant Greek
phrase at 184a10 is περὶ πάσας τὰς μεθόδους, which Charlton translates
“in all disciplines,” Pellegrin “toutes les recherches,” Waterfield “in any

objects that undergo generation and perishing, and those that do not for (1); for (2), consider
Aristotle’s remarks about how limited our access is to phenomena about the objects in the heavens
compared with “the animals and plants around us” in Parts of Animals i.5; for (3), notice that
Aristotle’s study of the rainbow in Meteorology iii involves two different perspectives – that of natural
science and that of mathematical optics. Different norms are engaged for these different perspectives
on the same natural phenomenon.
4
E.g. Phys. i.1, 184a10–11, PA i.1, 639a1–2, NE i.1, 1094a1–3, Meteor. i.1, 338a25–6, Top. i.1, 100a18. One
could argue that this opening passage of Phys. iii should be included, on grounds that the lists of
Aristotle’s works in Diogenes Laertius and Hesychius include two references to a book entitled Περὶ
κινήσεως (v. 23). For a judicious discussion of these references and what they might imply about the
structure of our Physics, see Ross 1936: 2–7 and, more recently, Menn (in press).
12 james g. lennox
subject,” Hardie and Gaye “in any department,” and Irwin and Fine “in
every line of inquiry.” There appears to be general agreement that the
term refers to some sort of human endeavor, but there agreement ends.
To make matters more difficult, Aristotle also deploys the concept in a
different (though, I will argue, closely related) way, to refer to a method, a
way of proceeding. Its use in the first chapter of On the Soul provides a rich
illustration of this usage.
Its opening lines tell us that inquiry into the soul (τὴν περὶ τῆς ψυχῆς
ἱστορίαν) will contribute greatly to all truth, but especially truth related to
nature. But we are warned of the extreme difficulty of this investigation,
because we are inquiring not merely into the soul, but into its “substantial
being and essence,” which raises questions about how to proceed
(402a10–14). Is the way forward, Aristotle asks, “demonstration, division
or some other method” (τις ἄλλη μέθοδος)? The μέθοδοι on offer bring to
mind the dialectical development of Posterior Analytics ii.1–10,5 which
explores the relationship between definition and demonstration. Given
that we suppose we have scientific knowledge of something when we have
a demonstration of it, he considers whether we can come to know
definitions, i.e. statements that identify what something is, through
demonstration. This possibility is discussed, and apparently rejected, in
chapters 3 and 7. The possibility that we can come to know definitions
through division is considered and rejected in Posterior Analytics ii.5.6
Chapters 8–10 articulate another option, which he summarizes by relat-
ing three kinds of definition to demonstration: one states the conclusion of
a demonstration (and thus can be known through demonstration), one
can be reconfigured as a demonstration, and one is the starting-point of a
demonstration.7
Here, then, μέθοδος refers to different possible ways of pursuing a specific
goal of inquiry, in this case knowledge of what the soul is; the translitera-
tion “method” seems to be a perfectly reasonable translation.
The term is deployed in the opening lines of Parts of Animals in a
manner similar to the opening of the Physics.
Regarding every study [theoria] and every methodos, the more humble and
more valuable alike, there appear to be two sorts of state, one of which may

5
The background is Academic: Dialogues such as the Phaedrus, Sophist, and Statesman view division as
the primary tool for inquiry into being and essence.
6
Division can be helpful in hunting for the essence: see Posterior Analytics ii.13.
7
There is controversy about whether Posterior Analytics ii.10 discusses three or four sorts of definition:
see Barnes 1993: 222–223; Bolton 1976; Charles 2000: 23–56; and Ross 1949: 634–636.
 How to study natural bodies 13
properly be called scientific knowledge of the subject matter, the other a
certain sort of educatedness. (PA i.1, 639a1–5)8
There are two cognitive states associated with every methodos, and Parts
of Animals i is a guide to acquiring the second, a special form of paideia,
which provides one with the ability to make critical judgments about
what is well or poorly expressed.9 Aristotle goes on to distinguish one
who has a very general form of this skill, and one who has it about a
specific discipline (say, the art of medicine), and transitions to the
business at hand with the following words: “So it is clear, for natural
inquiry too (καὶ τῆς περὶ φύσιν ἱστορίας), that there is need of some such
standards (ὅροι)” (639a12–13). The ability to make such judgments
requires certain standards, and if one’s paideia is about a specific field,
then it will be standards appropriate to that field that one needs to
acquire. This will be the topic of discussion for the remainder of Parts of
Animals i.
It is notable that θεωρίαν and μέθοδον are conjoined not simply by
“and” (καὶ), but by “both and” (τε καὶ), which makes it quite clear that
these terms are conveying different ideas, and are not merely synonyms.
It could be that Aristotle has two different categories of cognitive
endeavor in mind. However, there is an attractive alternative – that he
is referring to two different aspects of a cognitive endeavor, aspects
which are immediately identified as two different states associated
with them – scientific knowledge (ἐπιστήμη), on the one hand, and
that general critical judgment he identifies as a certain sort of paideia, on
the other. Support for this reading comes from the end of chapter 4,
where Aristotle summarizes what has been accomplished up to that
point:
We have said, then, how the methodos of nature (τὴν περὶ φύσεως μέθοδον)
should be appraised, and in what way (καὶ τίνα τρόπον) the study
(ἡ θεωρία) of these things might come about methodically and with greatest
ease (ὁδῷ καί ῥᾷστα). (PA i.4, 644b15–17)
This summation echoes the use of μέθοδος and θεωρία in the opening lines,
and here one can see that the two words emphasize different aspects of a
single study: μέθοδος emphasizes the way in which the investigation is to be

8
Cf. also NE i.1, 1094a1–3 and Phys. i.1, 184a10–11.
9
Cf. Pol. iii.6, 1282a1–8. Although medicine is the subject being considered, just as in the Parts of
Animals i.1 passage, there is a contrast between those with ability to judge and those with knowledge
of the subject matter.
14 james g. lennox
carried out; θεωρία emphasizes the object towards which the investigation
is directed.10
The other use of μέθοδος is on display in the concluding lines of Parts of
Animals i, looking forward to Book ii:
Enough said about our manner of proceeding (περὶ μὲν τοῦ τρόπου τῆς
μεθόδου); we must now attempt to discuss the causes concerning both the
common and distinctive [parts], beginning first, as we have determined,
from the primaries. (646a2–4)
The combination of τρόπος and μεθόδος here, which is fairly unusual, is
difficult to capture in translation, since the idea of proceeding in a
certain manner seems already implicit in μέθοδος itself.11 The combina-
tion stresses, however, that what has been defended in Parts of Animals i
is one among a number of possible ways of proceeding, one distinctive
(in this case) to the study of animals. The discussion in Parts of
Animals i, then, supports a hypothesis about the connection between
the two uses of the term μέθοδος: A μέθοδος is an inquiry that is carried
out according to a distinctive set of norms or standards that are appro-
priate to a specific domain of investigation. That is, to refer to an
inquiry as a μέθοδος is to emphasize that the inquiry is characterized
by a specific μέθοδος, i.e. by specific norm-governed procedures that are
appropriate for that subject matter.
Note, however, that Parts of Animals i does not explicitly restrict its
scope to zoology; the norms outlined are said to be appropriate to
natural inquiry (639a12, 644b16).12 Having thus explored the semantic
landscape of the concept of μέθοδος in that discussion, let us now return
to our original question – what sort of μέθοδος is the μέθοδος concerning
nature?

10
It is the use of ὁδῷ adverbially to characterize the way in which the θεωρία stresses the
connection to which I am drawing attention. The noun ἡ ὁδός refers to a road, path, or
track, and is used metaphorically about inquiry in much the same way as those English
expressions are. The dative form used here often has adverbial force, conveying the idea of
staying on the road to your destination, thus my “on course.” Μέθοδος is formed from that noun
and a prepositional prefix, μετά, which when used as a prefix carries the sense of “going after”
and “in quest of.” The basic idea, then, is a path taken in quest or in pursuit of something. It is
already used in Plato’s Sophist (218d, 235c, 243d) to refer to an inquiry, and Republic vii 533c
refers to “the dialectical method” (ἡ διαλεκτικὴ μέθοδος) as the only way to advance to first
principles.
11
Indeed at PA i.4, 644b17–21 Aristotle uses τίνα τρόπον twice as a sort of substitute for μέθοδος in this
second sense – perhaps because of the conceptual confusion he would have caused by using the term
in both its senses in the same passage!
12
For a discussion of why this might be so, see Lennox 2006.
 How to study natural bodies 15
II “Our methodos concerns nature”
The Physics opens, recall, in a manner strikingly similar to Parts of Animals.
Both start with a general claim about every μέθοδος and then apply what is
said generally to the same specific inquiry, the inquiry into nature:
Since, concerning every methodos of which there are starting points, causes
or elements, knowing, i.e. scientific knowing (τὸ εἰδέναι καὶ τὸ ἐπίστασθαι)
comes about from knowing (γνωρίζειν) these things . . . it is clear that,
concerning the scientific knowledge (ἐπιστήμη) of nature too we must first
attempt to determine the facts about the starting-points. (184a10–16)
Aristotle then reminds us that we are here focused on the proper way to
proceed:
And the path is by nature (ἡ ὁδὸς . . . πέφυκε) from things more knowable
(γνωριμωτέρον) and clear to us to the things that are by nature (φύσει)
clearer and more knowable; for what is clear to us and what is clear without
qualification are not the same. (184a16–18)
It is a path towards knowledge of the starting-points, causes, and
elements regarding nature that Aristotle is characterizing – it is by
achieving these that we acquire scientific knowledge of nature. And
the natural way to proceed is to start with what is better known and
clearer to us and proceed towards what is better known and clearer “by
nature” or without qualification, that is, without epistemic relativiza-
tion to the state of the knower. Up to this point, however, we have been
told nothing specific about natural inquiry, since Aristotle considers
this the proper and “natural” way to proceed in any discipline
whatsoever.13
So what is the proper μέθοδος for the μέθοδος concerning nature? For
Aristotle’s answer to that question I will explore a variety of texts, all
concerned with φυσική, but not all of them from the Physics. I begin
with a discussion in Metaphysics E.1.

Metaphysics E.1
In Metaphysics E.1, Aristotle aims to clarify the special sort of wisdom that
has been the object of his inquiry from the beginning by differentiating it
from other kinds of inquiry. Yet the chapter has a great deal to say about
what is distinctive about scientific knowledge of nature (1025b18–19:

13
Cf. APo. i.2, 71b33–72a5; EN i.4, 1095b2–4.
16 james g. lennox
ἡ φυσικὴ ἐπιστήμη), and the reason why becomes clear near the end of the
discussion:
If then there is no other substantial being apart from those constituted
naturally, the science of nature would be primary; but if there is some
changeless substantial being, this would be prior and <the science of it>
would be primary philosophy, and universal in this way, because primary.
And it would then be for this science to study being qua being, both what it
is and the things that belong to it qua being. (1026a27–32)
That is, one of Aristotle’s chief concerns is to determine whether the
science of nature is primary, or whether there is a study prior to it.
Yet this is a puzzling concern.14 The chapter opens by reiterating the
distinction, introduced at the beginning of Metaphysics Γ.1, between the
investigation of being qua being and investigations focused on some one
part of being (1003a21–25); and the science of nature is introduced in
Metaphysics Ε.1 within that very framework as the science of the kind of
being in which there is an inherent source of change and rest (1025b18–21).
This would appear to disqualify natural science immediately as a candidate
for first philosophy, since, in the language of this chapter, it is one of those
disciplines that “circumscribe a certain being and a certain kind, and treat
of it, and not of being without qualification nor qua being” (1025b8–10).
After looking carefully at the ways in which Aristotle demarcates natural
investigation, we will return to this puzzle.
One clear mark of φυσικὴ ἐπιστήμη has already been identified in the
process of distinguishing it as a theoretical rather than a practical or
productive activity: The particular kind of being which the natural
investigator takes as his object consists of objects which have inherent
sources of change and rest. This was already alluded to in the passage from
Physics iii.1 with which we began – we need to investigate what change is
because natures are sources of change in natural things. He next argues that
if all scientific knowledge is theoretical, or productive, or practical,
scientific knowledge of nature would be a particular kind of theoretical
knowledge. Which kind?: “about the sort of being that is capable of
undergoing change, and for the most part about substantial being accord-
ing to account (οὐσίαν τὴν κατὰ τὸν λὸγον), only not separable.”15

14
See Shields 2012b: 343–371 on this puzzle, and Frede 1996: 81–95 (following Patzig) for one sort of
solution.
15
Cf. Ross 1924: 354 at 1025b27n. Ross prints ὡς οὐ χωριστὴν μόνον following mss. E and T, and omits
the preceding comma of the OCT. I don’t think the addition of ὡς is necessary. I’m following
Alexander’s reading, taking μόνον adverbially (LSJ, 1145 cites this passage for this use of the neuter
μόνον). On its own it is difficult to understand, but the next few lines make Aristotle’s point clear.
 How to study natural bodies 17
Commentators typically assume that οὐσίαν τὴν κατὰ τὸν λὸγον is a way
of referring to substantial being as form.16 But caution is called for: In light
of what follows, it is likely that Aristotle is leaving the question of what to
substitute for οὐσία here open. For immediately after this very puzzling
sentence, Aristotle issues a caution:
It is necessary not to lose sight of how the-being-what-it-is and the account
is (τὸ τί ἦν εἶναι καὶ τὸν λόγον), because without this the inquiry is destined
to be unproductive. (1025b28-30)
What is significant and suggestive about this caution is that it introduces
Aristotle’s characterization of the objects to be investigated by the natural
scientist. It strongly suggests that what one identifies as “the-being-what-it-
is” of something and what an account of that thing should include will
differ between mathematics and natural science.
Among things defined and among the “what-it-is” of things, some are in the
manner of the snub and others are in the manner of the concave. These
differ in that the snub is comprehended with the matter (for the snub is
concave nose) while the concave is <comprehended> without perceptible
matter. And if all natural things are spoken of in like manner to the snub, for
example, nose, eye, face, flesh, bone – generally, animal – leaf, root, bark –
generally, plant (for the account of none of these things is without change,
but always has matter), then it is clear in what way one ought to inquire into
and define the what-it-is in natural things, and the reason why it is also for
the natural inquirer to study certain things about the soul, as many as are not
without matter. (1025b30–1026a5)
Thus the reference to “substantial being in accordance with the
account” in the initial specification of the objects to be investigated
by the natural scientist may well have been intentionally left open –
and, after being cautioned about attending to the ways that different
inquiries approach “the account” and “the what-it-is-to-be,” Aristotle
introduces us, conditionally, to the thought that all natural things are
spoken of like “the snub.” These are things the accounts of which are
“not without change and always include matter.” Unfortunately,
although he claims that on the assumption that natural things are
spoken of like the snub, “it is clear in what way one ought to inquire
into and define the what-it-is in natural things,” he doesn’t actually
spell out the way of inquiry and definition at all. Why would he – it’s
clear!

16
Ross 1924: 354 at 1025b27n; Kirwan 1971: 185; Peramatzis 2011: 100.
18 james g. lennox
A number of different options have been suggested, since Aristotle’s
views about what the definition of a natural composite should look like,
and even whether there are such definitions, are anything but clear. Indeed,
scholarship on Metaphysics Ζ–Η has focused on a tension, especially appar-
ent in Ζ.10–11, which arises when one asks whether, in Aristotle’s view, the
material aspect of a natural substance should be part of its definition or not.
Determining the answer to this question is directly relevant to the question
of what a μέθοδος of nature will look like, since arriving at definitions that
identify the what-it-is of natural substances must surely be a central part of
the goal of such an inquiry.
What then, are the options?
Option 1. Natural substances are matter/form composites, but a proper
account of such composites will only identify the form of the compo-
site, not the matter.
Option 2. Natural substances are matter/form composites, but a proper
account of such composites will refer to matter only as a part of the
form (or conversely to form as matter-involving).
Option 3. Natural substances are matter/form composites, and a proper
account of such composites will be an account of a universal composite,
i.e. will refer to the material and formal aspects of the natural
substance.17
Finally, although I won’t label this as a distinct option, it is possible to
combine options 1 and 3, by arguing that there are primary and derivative
definitions of natural substances, definitions that only refer to their form
and definitions that refer to both form and matter.
In this chapter I will assume the truth of a version of option 3, one which
I have defended elsewhere18 and for which I present additional evidence in
its favor by indicating how this option coheres best with Aristotle’s meth-
odological recommendations for natural inquiry. But before turning to
those recommendations, I will highlight two features of the above passage
that are rarely mentioned: First, all the examples of snub-like natural things
are biological, both non-uniform19 and uniform parts of animals and
plants, as well as animals and plants themselves; second, the fact that all
natural things are snub-like not only has implications for how one ought to

17
Option (1) Devereux 2011, Frede 1990; (2) Charles 2010b, Peramatzis 2011; (3) Balme 1987, Gill 1989,
Whiting 1986.
18
Lennox 2008, 2010.
19
The fact that he mentions noses first in his list of parts that are “snub-like” may be an instance of
Aristotelian humor.
 How to study natural bodies 19
define them, but (Aristotle says) also for how one ought to inquire into what
they are.20 The last line quoted above also has significant implications for
how one interprets the opening chapter of DA and for how one under-
stands the relationship of DA to the Smaller Natural Treatises (Parva
Naturalia [PN]). For that line makes it clear that it is by no means a
foregone conclusion that the natural philosopher should study the soul,
and it leaves open, as does DA i.1, the possibility that there is some part of
the soul that is without matter. The mere fact that all animals, plants, and
their parts are spoken of like the snub and are not without matter does not
imply that all soul is to be thought of in the same way.21
What then does this passage tell us about the μέθοδος of ἡ μέθοδος ἡμῖν
περὶ φύσεως? It quite straightforwardly says that the account we are aiming
for in natural inquiry must include reference to matter and change; and
that therefore the natural scientist, while acknowledging the priority of
form, must make the material make-up of natural things and the changes
they undergo central to his investigations. Thus, the sort of abstraction that
is appropriate in mathematics, where the proper account of (say) a paralle-
logram does not mention perceptible material or change at all, would be
inappropriate in natural science. However, little or nothing is said in this
chapter about the definition referring as well to form, nor to the nature of
the relationship between matter and form in natural objects, and how that
relationship should be studied or represented in a proper definition or
causal demonstration. Similarly, nothing is said about the implications of
this characterization of the objects of natural science for how one ought to
inquire into them. However, that is the central focus of Physics ii.2, and the
“snub-like” character of the objects of natural investigation is Aristotle’s
starting-point.

Physics ii.2 and snub-like inquiries

22
Physics ii.2 begins with the assumption of one major result of the
previous chapter, namely the number of ways in which nature is spoken,23
and turns to a question he apparently thinks follows more or less directly

20
. . . δῆλον πῶς δεῖ ἐν τοῖς φυσικοῖς τὸ τί ἐστι ζητεῖν καὶ ὁρίζεσθαι . . .
21
As discussed in Lennox 2010, Parts of Animals i.1 states unequivocally that nous and dianoia are not
proper objects of inquiry for the natural scientist.
22
Some results of Lennox 2008 are assumed here.
23
Reading λέγεται at 193b22. Later in the chapter, at 194a12 we are told that nature is διχῶς τό τε εἶδος
καὶ ἡ ὕλη, a result Aristotle argues for from 193a9–b5, after which he makes the case for the form
being a composite’s nature “more than” the matter.
20 james g. lennox
from that result – in what way does the mathematician differ from the
person who investigates nature? Aristotle generates a concern about this by
highlighting two related but distinct points: First, the objects investigated
by the mathematician, such as points, lines, planes, and solids are features
of natural bodies; and second, it is apparently a goal of the natural scientist
to grasp the nature of the sun and moon, and to determine whether the
earth or the cosmos is spherical or not. But these would appear to be the
concerns of the astronomer, who is a type of mathematician (193b22–30).
From his responses to these problems it becomes clear that at this point
in the Physics Aristotle is concerned to make a case for the unity of the
objects being investigated by the natural scientist, as a precondition for
there being a single science of nature. This becomes critical because chapter
1 has argued that the nature of a natural thing is, and is spoken of, in two
ways, as matter and as form. Chapter 2 points to the errors of “those who
speak of Ideas” (193b36), indicating his concern clearly to be about what it
means to say that the natural investigator should be pursuing knowledge of
the forms of natural things. He precisely criticizes them for attempting to
separate the natural things, which are “less separable than the mathema-
ticals” (193b36–37). He explains:
This would become clear if one were to attempt to state the definitions of
each – both of the objects and of their attributes. For the even and odd and
the straight and curved, as well as number, line, and figure will be [defined]
without change, while flesh, bone, and human being will not – these are
spoken of as is snub nose (ῥίς σιμὴ), not as the concave. (194a1–7)24
As in Metaphysics E.1, here the contrast with mathematical definitions is
intended to highlight the fact that the definitions of natural things refer to
change – and, once again, they are said to be referred to in the way one
refers to a nose as snub, not as concave. But what exactly are we referring to
when we refer to a natural object in that way?25
It is of the very essence of a living thing, and more generally of any
natural object, to move, behave, act, and change in specific ways; in fact, in
specifying what it is to be an eye or a leaf – or even air – one must mention
capacities to function or change in specific ways.26 But it is also the case
that those capacities are the capacities of specific materials, and thus any
such definition must refer to a material structure constituted in precisely
the way it must be in order to have the capacities to move and change as an
eye or a leaf.
24 25
Cf. Metaph. Z.10, 1035a4–5, Z.11, 1036b21–31. Cf. Metaph. Z.11, 1036b24–31.
26
Cf. Meteor. iv.12, 390a10–12.
 How to study natural bodies 21
And this has implications for how one carries out a natural inquiry, as
Aristotle indicates in Metaphysics Z.11, while explaining why first philoso-
phy bothers to discuss natural composites:
the study of perceptible substances is in a way the task of the study of nature,
or second philosophy. For the natural investigator must have knowledge not
only of the matter of things but also, and more especially, of the substance
given by the formula. (1037a15–17)
To stress this point in Physics ii.2, Aristotle once again invokes the analogy
of natural substances and the snub, but now as an implication of the fact
that natural things have a twofold nature:
Since the nature [of natural substances] is in two ways (διχῶς), both the
form and the matter, we should study as if we were investigating, about
snubness, what it is; that is, we should study such things neither without
matter nor in accordance with matter. (194a12–15)
The message appears to be this: Natural things have a nature in two ways –
a material nature and a formal nature.27 For that reason, we need to
investigate the “what is it” of natural things while keeping their twofold
nature in focus; we cannot ignore their matter, but we cannot focus only on
their matter, for they have a formal nature as well – indeed, Physics ii.1 gives
a series of (more or less compelling) arguments for giving more attention to
the formal nature. It is somewhat surprising, then, that Aristotle immedi-
ately raises an aporia about which of the two natures the natural scientist
should study! Have we not just been told that he should study both? The
full statement of the aporia is worth reviewing:
For indeed one might also raise a problem about the following: since the
natures of things are two, which of the two is it for the natural scientist to
study – or is it rather for him to study that which derives from both natures?
But if [it is for him to study] that which is derived from both, [then it is] also
[for him to study] each of them. (194a15–17)
A provisional solution to the aporia, which flows from the idea that the
natural scientist should study the hylomorphic composite, is immediately
challenged, on grounds that one would still be studying the two natures,
though not treating form as separate from matter. The question at issue is:
27
Aristotle identifies the two natures in relation to one another in various ways: at Phys. ii.1, 193a28–31,
the contrast is of the primary underlying matter with the shape and form in respect of the account
(cf. Phys. ii.1, 193b3–5); at PA i.1, 640b28–29 it is of the nature in respect of form with the material
nature; at PA iii.2, 663b22–24 the nature in respect of the account is contrasted with the necessary
nature; and at GA iv.4, 770b9–18 the nature in respect of the form is contrasted with the nature in
respect of the matter.
22 james g. lennox
Even if not ontologically separate, can these two very different natures be
studied by a single science (ἐπιστημή)? That question has purchase
because, within an Aristotelian framework, although the objects of
mathematics are not ontologically separate from natural objects, they are
“separated in thought” and studied by distinct sciences, sciences which use
very different methods guided by very different norms from those of the
natural scientist.
To this point he has claimed that natural substances have two natures,
that they and their attributes are spoken of like snub nose and are not
defined without reference to change, and that we are to study them “not
without matter yet not according to matter.” He is now prepared to present
his solution to the above aporia, albeit in conditional form:
If art imitates nature, and it is for the same science (ἐπιστημή)28 to know the
form and the matter to a certain extent (e.g. for the doctor to know health
and bile and phlegm, in which health is present, and similarly also for the
house builder to know both the form of the house and the matter, that it is
bricks and wood, and likewise in other cases), then it would also be for
natural science to know both natures. And again: it is for the same science to
know that for the sake of which and the goal and as many things as are for the
sake of these. (194a21–27; emphasis added)
The knowledge the natural scientist seeks is of both the matter and the
form; and they can be the subject of one science, on the model of the crafts,
if the materials are both present for the sake of, and defined by reference to,
the formal nature. That is the reason why, although natural things have
two natures, they are properly the subject matter of a single science.
Aristotle picks up on the qualification “to a certain extent” in the passage
I have just quoted, near the end of the chapter:
To what extent is it necessary for the natural scientist to know the form
and the what-it-is? Is it just as [it is necessary] for the doctor [to know]
sinew or the sculptor bronze, to the extent of knowing what each thing is
for the sake of, and concerning those things which are separable in form,
yet in matter? (194b9–13)
It appears that the issue of the extent to which the two natures are to be
studied by the natural scientist is only explicitly addressed with respect to
form, a point I will return to shortly. First we should recall the wider
context of this question, specifically issues concerning the demarcation of
natural philosophy from first philosophy. Charlton translates the chapter’s

28
Here ἐπιστημή is being used in a “big-tent” sense to include the crafts.
 How to study natural bodies 23
closing sentence thus: “What it is which is separable, and how things are
with it (πῶς δ’ ἔχει τὸ χωριστὸν καὶ τί ἐστι), it is the work of first
philosophy to determine” (194b14–15). Whatever Aristotle exactly means
by the Greek here, one thing is certain: He is making this point because he
has just made a claim about the limited nature of the separability of the
forms to be investigated by the natural scientist.29 And this passage surely
looks back to the end of Book i, where Aristotle says that the discussion to
follow (in Book ii) will investigate the forms of natural and perishable
things,30 while it will be “the work of first philosophy to discuss with
precision the principle in respect of form, whether it is one or many and
what it is or they are” (192a34–36). Together, these passages are saying that
the natural philosopher’s investigation of form is distinct from that of the
metaphysician. Specifically the natural philosopher is to study the formal
nature in so far as it is that for the sake of which the materials that make up the
material nature are present. Thus, the natural scientist should study matter
to the extent that it is for the sake of the form. This chapter, then, looks
forward to chapter 8 and its defense of the teleological dependence of
material nature on formal nature, succinctly summed up in the following
sentence:
And since the nature [of a thing] is twofold, on the one hand as matter and
on the other as form, and the nature as form is an end, while the others are
for the sake of the end, this [nature as form] would be the cause for-the-sake-
of-which. (199a30–32)31
This passage, moreover, looks forward to the discussion of the relationship of
conditional necessity between materials and what they are for in chapter 9.
We now begin to make contact with norms that govern Aristotle’s actual
practice as a natural investigator, such as the following, during a critical
review of his predecessors’ attempts to explain respiration:
The most significant reason for [previous thinkers] not discussing these
[questions about respiration] well is a combination of (τε) their being

29
Cf. Waterfield 1996: “Questions remain – in what sense is anything separable? What is it that is
separable? – but it is the job of first philosophy to answer them”; Hardie and Gaye 1930: “The mode
of existence and essence of the separable it is the business of the primary type of philosophy to
define”; Pellegrin 2000: “Ce qu’il en est du separable et ce qu’il est, c’est l’affaire de la philosophie
première de le determiner”; Angioni 2010: “Mas delimiter como se comporta e o que é o separável é
tarefa da filosofia primeira.” Certainly the nature of the separability of the objects of mathematics,
Platonic Forms, the forms of natural substances, and a first unmoved mover is a theme that runs
through the entirety of the Metaphysics.
30
I defend this reading (which follows the Λ mss.) in Lennox 2008.
31
καὶ ἐπεὶ ἡ φύσις διττή, ἡ μὲν ὡς ὕλη ἡ δ’ ὡς μορφή, τέλος αὕτη, τοῦ τέλους δὲ ἕνεκα τἆλλα, αὕτη ἂν
εἵη ἡ αἰτία, ἡ οὗ ἕνεκα.
24 james g. lennox
inexperienced with the internal parts and (καὶ) not grasping that nature
makes them all for the sake of something; for seeking what respiration
belongs to animals for, and examining this question in the presence of the
parts, e.g. in the presence of gills and lungs, they would have discovered the
cause more easily. (Resp. 3. 471b23–29)
That is, one must carry out a natural investigation of a living activity like
respiration on the assumption that it and the structures that perform it are
present in animals for the sake of something; and one must carry out such
an investigation making use of comparative dissection so that one knows
exactly which internal parts are involved in performing the activity of
respiration. Recalling an oft-cited passage from the beginning of
Aristotle’s investigation of animal locomotion, we should not be surprised
by what Aristotle says about the study of respiration, for he there claims
that this is an assumed starting-point of natural inquiry generally:
The starting-point of our investigation is achieved by supposing principles
we are accustomed to use often in natural inquiry (πρὸς τὴν μέθοδον τὴν
φυσικήν) – assuming this is the way things stand in all the works of nature.
One of these principles is that nature does nothing in vain, but always, given
the possibilities, does what is best for the substantial being of each kind of
animal (τῇ οὐσίᾳ περὶ ἕκαστον γένος ζῴου τὸ ἄριστον). (704b11–17)
The assumed starting-points Aristotle introduces here are starting-points
of investigation (not of proof) and are said to be used often in the μέθοδος of
nature. Yet the first one that is stated, the nature-does-in-vain principle, is
stated in a way that is specific to animals. There is, of course, absolutely
nothing inconsistent about that – a principle that is often used in the
investigation of animals will, by that very fact, be used often in natural
inquiry. But he also says that it is appropriate to assume that these
principles (he lists two others) are applicable to “all the works of nature.”
He could mean that they collectively apply to all works of nature, or he
could simply be allowing that, at the beginning of an investigation of
nature, we have no grounds for restricting these principles to a limited part
of nature.32
In Physics ii.2, the primary focus was on the question of whether, given
that natural substances have both material and formal natures, they could
be investigated by a single science – where the goal of such an investigation
tended to be stated in terms of defining and knowing the essence of such
things. However, the possibility of a unified science of natural, “snub-like”

32
Cf. Lennox 2009a: 187–214, and Leunissen 2010: 215–237.
 How to study natural bodies 25
beings turned on there being a teleological dependence of their material
nature on their formal nature. Indeed, the natural scientist’s knowledge of
the matter and form of a natural substance was restricted to knowing
what the material constituents of a natural being are for, which means
knowing the form in so far as it is that for the sake of which the matter is as
it is. Chapter 2 thus concludes by implicitly grounding the “what is it?”
inquiry in a causal inquiry, the sort that is in focus in chapter 9, namely an
inquiry into the causal relationship between ends, on the one hand, and
materials and their movements, on the other. But that is precisely what the
erotetic framework of Posterior Analytics ii would have led us to expect –
that the search for an answer to the “what is it?” question will be inter-
twined with causal inquiry in precisely this way.
That is how Physics ii.2 ends – but recall how chapter 3 begins, echoing
the opening sentence of the Physics (which in turn echoes Posterior
Analytics i.2 and ii.11).
Having determined these things, we must investigate concerning the causes,
both of what sort and how many in number they are. For since our study is
for the sake of knowing, and we do not think we know each thing until we
grasp that-on-account-of-which (τὸ διὰ τί) about each thing (and this is to
grasp the primary cause), it is clear that we must do this concerning
generation, destruction, and all natural change, so that, knowing their
starting-points, we may try to bring each of the objects of inquiry back to
them. (194b16–22)
However, though introduced as important for knowledge of nature, the
discussion of the four types of cause in this chapter (as well as in the
next three chapters on chance) says little that is specific to the inquiry
into nature. In fact, chapter 3 is, as Ross puts it, “a doublet which
corresponds in all except small details with Met. Δ.2.”33 It is not until
chapter 7 that Aristotle explicitly discusses the idea that there are four
distinct kinds of cause in the context of natural inquiry. It begins by
reminding us of the four kinds of answer to the “on account of what?”
question that were discussed in chapter 3, still without particular
specificity to nature (198a14–21). He then proceeds to make the
connection:
So then, that the causes are these and this many is apparent; and since the
causes are four, it is for the natural scientist to know about them all, and by
referring to all of them – the matter, the form, the mover, and that for the

33
Ross 1936: 511.
26 james g. lennox
sake of which – he will display that-on-account-of-which (τὸ διὰ τί) in a way
appropriate to natural science (φυσικῶς). (198a21–24; emphasis added)34
Notoriously, he goes on to say the last three often come to one – and then,
substituting τὸ τί ἐστι for τὸ εἶδος, he explains what he has in mind. In
cases of substantial generation, the what-it-is and that-for-the-sake-of-
which are one, while the primary source of motion is the same as these
in form (198a24–26). The goal of generation is the form, so in these cases
there is referential (and numerical) identity between them; while the parent
and the off-spring are not numerically identical, they are the same in form.
The remainder of the chapter is, though puzzling in a number of respects,
equally important, since its focus is, as with chapter 2, on distinguishing
natural inquiry from other sorts of inquiry. The mention of the primary
source of motion, he realizes, might well be misunderstood – thus, he
immediately restricts the set of such movers to those that “move while
being moved” (a25) and goes on:
for as many as are not like this are not natural; for they move <other things>
without having a change or a source of change in themselves, but unmoved;
for which reason there are three subjects of study – that concerned with the
immoveable, that concerned with what moves but is imperishable, and that
concerned with perishable things. (198a28–31)
This way of dividing up theoretical investigations is quite different from
what we found in Metaphysics E.1; assuming that the last two subjects
mentioned correspond to a study of the heavens (which, while imperish-
able, partake of eternal circular motion) and a study of sublunary natural
objects, these are both appropriate objects of natural inquiry.35 And the
third subject area is described in such general terms that it could refer to a
number of things in Aristotle’s universe, including souls, objects of desire,
and the heavenly unmoved movers, all of which move other things while
being themselves unmoved.36 It is best, I think, to read this as having the

34
Cf. Metaph. H.4, 1044a32–b20, where, however, after citing the four causes and saying that in the
study of generated natural substances it is necessary to discuss all the causes, Aristotle states that we
should proceed by “another logos” when discussing the natural and eternal substances (1044b2–6),
and that certain natural processes, such as eclipses, perhaps don’t have final causes (1044b9–13); no
final cause is ever provided in the Meteorology for any of the cases discussed here.
35
Cf. PA i.5, 644b22–25: “Among the substantial beings constituted by nature, some are ungenerated
and imperishable for all eternity, while others participate in generation and perishing.” See, too,
Cael. i.1, 268a1–6 and Metaph. H.4, 1044b3–6.
36
The discussion in Metaphysics E.1 depends on comparing the objects of mathematics, natural
science, and a still hypothetical first philosophy along two axes, separability and changeability,
whereas here separability is not discussed. Recall that in Physics ii.2 he refers that topic to first
philosophy.
 How to study natural bodies 27
limited purpose of clarifying what sort of first mover the natural scientist
can investigate, for he goes on to state that “the sources of things moving
naturally (φυσικῶς) are of two sorts, one of which is not natural; for it does
not have a source of motion in itself” (198a35–b1).
Things that are by nature have a source of change or rest within
themselves (192b13–15, b20–23, b32–34). He postulates two sorts of things
that fall into the category of things that are sources of natural motion
without being natural in that sense: One of these is not to be investigated
by the natural scientist, the prime unmoved mover; the other is the form
viewed as final cause, “the what-it-is and form, for it is an end and for the
sake of which” (198b2–3). He goes on:
So since nature is for the sake of something, and it is necessary to know
this, it is also necessary to display that-on-account-of-which in all ways,
e.g. that this comes from this necessarily (either without qualification or
for the most part); <what must be> if this is to be (as the conclusion from
premises); that this is what it is to be; and for what reason it is better in this
way, not better without qualification but <better> in relation to the being
of each thing. (198b1–9)37
These concluding remarks build on the idea that there are two different
kinds of thing that are sources of natural change while not themselves
changing, and that one of those, while not itself a nature, is an appropriate
subject of natural inquiry, namely that which nature, understood as an
inherent source of change, acts for the sake of. Note that both here, and in
the opening sentence of chapter 8, Aristotle says that nature is “among the
causes for the sake of something” (198b4, b10), not that nature is “that for
the sake of which.” The concept of nature is here being used strictly to refer
to the inherent source of change delineated in Physics ii.1–2. And here we
see the importance of the distinction between form understood as that for
the sake of which and form understood as source of change.38 For the
defense of natural science as a unitary science in Physics ii.2 depended on
the idea that the two natures of a natural substance are teleologically
unified – and that unity depended on the formal nature of a natural
substance being that for the sake of which its material nature comes to be
and persists. This conception of natural science depends on nature not only
being among the causes that act for the sake of something – it must also be
“the cause for the sake of which.” This claim is asserted in Physics ii.2, but

37
On the many puzzling aspects of this remark, see Ross 1936: 527–528; Philoponus In Ar. Phys. 303.8–
306.12; and Simplicius In Ar. Phys. 366.30–369.13.
38
Cf. DA ii.4, 415b8–28.
28 james g. lennox
only comes to the fore in chapters 8 and 9, beginning with the following
(already quoted) passage.
And since the nature [of a thing] is twofold, on the one hand as matter and
on the other as form, and the nature as form is an end, while the others are
for the sake of the end, this [nature as form] would be the cause for-the-sake-
of-which. (Phys. ii.8, 199a30–32)
This teleological way of conceiving the relationship between the materials
that constitute a natural being and the form of the thing so constituted
provides the ontological underpinnings for a new, conditionalized, way of
conceiving of necessity, the subject of Physics ii.9: Given that a certain goal
is to be realized, certain materials, with appropriate dispositional proper-
ties, must come to be in a certain temporal and spatial order.39 And this, in
turn, provides yet another norm that is specific to natural inquiry, to the
μέθοδος of nature.
It is apparent, then, that the necessary in natural things is that which is
spoken of as matter and its motions. And it is for the natural scientist to
discuss both causes, but more the cause for the sake of which; for this is a
cause of the matter, rather than the matter being the cause of the end; and
the end is that for the sake of which, and the starting-point is from the
definition and the account. (200a30–35; emphasis added)40
Again, questions of causation and definition are integrated tightly. As so
often when teleological causation is under consideration, Aristotle stresses
that the goal, that for the sake of which, has priority because in these cases
the account that specifies the identity of the goal must be understood in
order to properly understand the materials and processes that are needed to
achieve that goal. That “that for the sake of which” will be prioritized in
accounts of natural substances is clear; however, Physics ii ends by
considering whether “the necessary” will as well:
Perhaps the necessary is also in the account. For in defining the work of
sawing, that it is a certain sort of dividing, surely this will not exist unless the
saw has teeth of a certain sort; and such teeth will not be, unless they are
made of iron. Thus, even some parts of the account are in the account as
matter. (200b4–8)41

39
Cf. Charles 2012: 230–232. 40 Cf. PA i.1, 639b12–16 and 642a14–18.
41
This example should be compared to PA i.1, 642a10–13, where Aristotle makes the analogy of the axe
to the living body explicit; whereas in Physics ii.9 the analogy to living bodies is mentioned (rather,
the act of sawing is being defined, perhaps analogous to Aristotle’s study of respiration or sleep); the
Parts of Animals example also makes clear that it is the dispositional property that is conditionally
necessary, not a specific uniform material.
 How to study natural bodies 29
Once again, then, form, understood as the end for the sake of which
materials are as they are, is prioritized, but matter is not ignored. Indeed,
it is argued that the necessary materials should be brought into the account
by way of conditional necessity.
As with the teeth of saws, so too with the teeth of animals. In discussing
the uniform materials out of which hard parts are made in Parts of Animals
ii, Aristotle considers the material out of which teeth are made:
The nature of teeth is . . . in some cases present for a single function, the
preparation of nourishment, while in other cases it is present both for this
and defense, e.g. in all those with saw-like teeth . . . . Of necessity all of these
parts have an earthen and hard nature; for this is the defensive potential
(ὅπλου δύναμις). (PA ii.9, 655b8–13)
When he moves on to the non-uniform parts to discuss the teeth he
assumes the material nature of teeth is of this character, and focuses on
differences in their shape and organization in relationship to the different
lives that animals lead:
Some animals have teeth . . . for the sake of nutrition alone. But those which
have them for protection as well as for strength in some cases have tusks . . .
and in other cases have sharp, interlocking teeth, for which reason they are
called saw-toothed. For since their strength lies in their teeth, and this comes
about because of their sharpness, teeth which are useful for strength fit
together in an alternating pattern, so as not to be worn down by being
rubbed against one another. (PA iii.1, 661b17–22)
Here we see in his actual biological practice exactly the sort of teleological
account of teeth that are intended both for nutrition and defense, from
which it follows that they must have a certain structure and organization,
and be made of a suitably hard material.

III Conclusion: One natural μέθοδος or many?

We were provoked to explore the question of what a μέθοδος of nature
might be by Aristotle’s comment, at the beginning of Physics iii, that
“[s]ince the nature of a natural thing is a source of motion and change, and our
μέθοδος is concerned with nature, the question ‘what is motion?’ must not
escape our notice.”42 And indeed, for the next three chapters of Book iii he

42
The concept μέθοδος appears four more times in the Physics – once in iii.5, 204a34–b11, and the
remainder in Book viii (251a5–8, 253a32–b10, 261a27–31). As one would expect if the position I have
been defending is correct, all are concerned with delimiting the approach appropriate to natural
inquiry, as opposed to mathematics or first philosophy.
30 james g. lennox
pursues that question and comes up with a notoriously puzzling answer.
But before doing so, he notes that, since motion seems to be continuous,
there are a number of other questions – about the unlimited, place, void,
and time – that must be taken up after the question of what motion is.
Before turning to these topics, which will occupy him for the next two
books, he makes a comment that sheds light on how this treatise and its
method differ from the rest of Aristotle’s natural investigations. The lines I
am especially interested in might be translated as follows:
It is clear, then, that, both on account of these things and because the being
of all these things is common and universal (κοινὰ καὶ καθόλου), we must
investigate by getting hold of each of them (for the study of the proper
things (τῶν ἰδίων) follows the study of the common things (τῶν κοινῶν)).
(200b21–25)
In any particular investigation of nature, whether of meteors, rainbows,
minerals, animal parts or activities, or the motions of the heavens, one
needs to have a clear conception of motion and change in general, as well as
of the unlimited, place, void, and time in order to proceed. These concepts
are, to borrow Waterfield’s translation of καθόλου here, pervasive to the
study of nature. The “Physics”43, then, is a work in the conceptual
foundations of natural science; not in the sense that its methodology is
any more “conceptual” than any other, but in the sense that the concepts
about which it aims to get clear – nature, cause, chance, motion, place,
time, and so on – are presupposed by any of the more specialized natural
inquiries in which Aristotle was engaged, each of which may have distinc-
tive norms of inquiry. In Physics i.2–3 Aristotle makes it clear that he
doesn’t think it is appropriate for the natural scientist, qua natural scientist,
to question whether change exists – obviously it does. At the beginning of
Physics iii, however, he makes it equally clear that what change is is
anything but obvious, and to fail to answer that question is paramount
to being ignorant of nature itself. Moreover, in the case of the unlimited,
void, time, and place, it is even legitimate to consider the question of
whether they exist, and part of the exploration of what they are involves
exploring in what sense they exist.

43
I remain neutral on how the various books of our Physics came to be related to one another (but see
Menn [in press]), but note that this opening of Book iii both looks back to the earlier books, in the
sense that it notes a lacuna in that discussion which he now intends to fill, and also previews all of the
topics covered in Books iii and iv. Book v can be viewed as a natural appendix to the general
discussion of change; and since all of the topics in Books iii and iv arose out of the undefended
assumption that motion “seems” (δοκεῖ, 200b16) to be continuous, the discussion of continuity in
Book vi fills an obvious gap in the earlier discussion as well.
 chapter 2

Aristotle on interpreting nature

Sean Kelsey

Physics ii begins (more or less) with a definition of nature. I want to

consider how this definition is used in the rest of Book ii. I don’t think
this has been done before, and in any case a first look turns up a surprise: It
seems as though it is hardly used at all.
A closer look suggests that this initial appearance, though resilient, is
ultimately misleading. Although the definition seldom functions as an
explicit premise in an important argument, it does influence the course of
the subsequent discussion. A little more fully, it serves as both source and
signal of an important theme that sets a tacit agenda for the book as a whole.
It is not hard to see that the case for such a suggestion must rely on what
we might call “the argument from salient omission” – always perilous. But
although I do not think such arguments establish the suggestions they are
invoked to defend, I do think it quick to suppose they simply refute them.
A case in point: Aristotle does not say, in introducing his discussion of
chance and luck, that its purpose is to supply the backbone of an argument
that nature is a final cause (Physics ii.4–6, 8). On the contrary, he intro-
duces that discussion as almost an afterthought – a kind of appendix to the
discussion of causes in ii.3 (195b31–36, 196b6–9). Yet it is impossible to
believe that the opening of ii.4 gives a fair representation of the real
purpose of this material. Despite the absence of any explicit statement to
this effect, it is clear that the function of these chapters – their place in the
design of the book as a whole – is to support Aristotle’s first and principal
argument that nature is a final cause. I mention the case to exhibit a
principle – in this text, at least, silence is not probative against.1

I would like to thank David Ebrey, Joseph Karbowski, Mariska Leunissen, Yannig Luthra, Gretchen
Reydams-Schils, Wiebke Marie Stock, and especially David O’Connor for conversation and comments
on earlier drafts of this chapter.
1
Another example is Physics i, where it emerges in retrospect that the problem raised at the end of i.6
was staged: Its solution requires certain distinctions made in i.7, just the ones (lucky chance!)
necessary to solve “the” puzzle discussed in i.8. Or again in On Generation and Corruption i.3,

31
32 sean kelsey
I
It will be useful to have the passages in which the definition is used
collected in one place, if only to show how little appears to be gotten
from it in the body of Physics ii.
So first, in ii.1, having said what nature is (an internal source of move-
ment and rest), and what sorts of things have natures (substances), and
what sorts are “by” or “in accordance with” nature (the substances that
have natures, and whatever belongs to them per se), and that it would be
ridiculous to try to show that there are natures, Aristotle turns to identify
the natures of those beings that have them – to say what “in” those things is
their nature (193a9ff.). Here is a place we might expect him to use the
definition. Having just established (as he thinks) that nature is an internal
source of movement and rest, we might expect him to assess something’s
claim to be a nature by considering whether it is such a source. But in fact
he refers to the definition but twice in the ensuing discussion, both times in
stating a result already arrived at, and even then only to specify that nature,
whether it be “the first underlying matter,” or “the shape and form,” is that,
not just of any old thing, but of things “having in themselves a principle of
motion” (that is, nature is the “matter” or “form” of things that have
natures) (193a29–30, b3–4).
So much for ii.1; the definition is not referred to again until ii.6, in a
passage discussing the difference between chance and luck (197b32–37).
Having said that this difference is most visible in the case of things that
happen “contrary to nature” – we ascribe such things to chance, not to
luck – Aristotle enters a qualification: This is not quite right, inasmuch as
the cause of things due to chance is “external,” whereas the cause of things
that happen contrary to nature is “internal” (namely, to the extent that
even such things are caused in a way by nature) (197b36–37). This remark
does make an appeal to the definition, but by way of establishing a point
that (at least on its face) could hardly be less important.
The definition is next appealed to in ii.7, where it is used twice to make
the point that certain causes are not “natural” (φυσική) or “within the
purview of natural science” (φυσικῆς). The idea is that these causes do not
themselves have natures, because, being immobile, they do not themselves
have an internal principle of motion (198a28–31, 198a35–b1). Last come two
passages towards the end of ii.8, about which it suffices for now just to say

where the point of the indirection is made explicit: Aristotle introduces a new puzzle, about why
generation is everlasting, in order to solve an old puzzle, about the very possibility of unqualified
generation and corruption (317b33–35, 318a10–13).
 Aristotle on interpreting nature 33
that they occur long after the chapter’s main business has ostensibly been
completed.
A total of seven passages, in none of which is the definition obviously the
crux of a matter of any consequence.

II
I now want to revisit these passages, starting with ii.1. This chapter high-
lights two results in particular: that nature is a principle of motion (and
rest), and that form is nature (“more so” or “rather than” matter).
Although these results imply that form is a principle of motion and rest,
this is nowhere stated in ii.1. Indeed, the very idea of a principle of “motion
and rest” – that is, of a thing’s characteristic behavior – is more or less idle in
the latter half of ii.1; as we will see, Aristotle’s focus in identifying the
natures of things is not on their behavior, but on their “substance” or
“being” or “essence” (οὐσία) – not on what they “do,” but on what they
most fundamentally “are.” Being and behavior are importantly related, of
course, but this only makes Aristotle’s disregard of behavior in this context
all the more striking. The effect, I submit, is to make the (unstated)
implication of the results he does highlight – the implication that form is
a principle of motion – something of an elephant in the living-room.
The point that wants showing is that in Physics ii.1, in identifying
the nature(s) of things due to nature, Aristotle doesn’t consider their
behavior – their characteristic forms of motion and rest. Consider first
his handling of the idea that matter is nature (193a9–28). He begins with
the remark that “the nature and substance of things due to nature” is
thought by some to be “the first [thing] present in it, which is in and of
itself unformed” (193a9–12). Significant here is the tacit identification of
“nature” with “substance”; although it was said earlier that things that have
natures are substances (192b33–34), this is the first indication that a thing’s
nature is its substance. Doubtless one reason this identification is made
here is that it is useful for understanding the view being considered; this is
borne out by the remarks that follow, in which that view is represented as
resting on the thought that “the first thing present in” each thing is, not the
source of its characteristic motion and rest, but its real substance. It is true
that the thought experiment attributed to Antiphon might suggest a focus
on behavior: “if you planted a bed and the rotting wood acquired the
power of sending up a shoot, it would not be a bed that would come up,
but wood” (193a12–14). But this suggestion is canceled by Aristotle’s state-
ment of the experiment’s moral: “the one, the conventional arrangement
34 sean kelsey
and the art, belongs but incidentally; the substance is that other, which
persists continuously while undergoing these things” (193a14–17). As this
statement makes clear, Antiphon, in asking after a thing’s “nature,” is
focused on its substance, which he conceives of not as the source of a thing’s
movement and rest – of its behavior while it exists – but as what survives its
demise. Indeed, this approach betrays an attitude in which the nature of
things is positively concealed by their behavior, being manifested instead in
their destruction.
This reading is strengthened by the lines that follow, in which Aristotle
extends Antiphon’s reasoning and, in so doing, widens his net to take in
other thinkers:
And if each of these experiences the same thing in relation to something else
(e.g. bronze and gold in relation to water, bones and wood in relation to
earth, and so on), that (they say) would be their nature and substance. That’s
why some say fire, and others earth, and some air, and others water, and
some [say] some of these and others all of them are the nature of the things
that are. For whatever any one of them supposed to have this
character – whether one thing or more than one thing – this or these he
declared to be the whole of substance, all else being its affections, states, or
dispositions of these. Every such thing they held to be eternal (for it could
not pass into anything else), but other things [they held] to come into being
and cease to be again and again interminably. (193a17–27)
Aristotle represents these thinkers as conceiving of the nature of things as
something that is above all permanent, un-generated, and indestructible.
But on this way of thinking, it would be a mistake to look for the nature
of something in the source of its characteristic behavior, which is as
ephemeral as the thing itself is; on the contrary, the place to look would
be in its remains – or, equivalently, in its antecedents or origins.2
In giving reasons for thinking that nature is matter, then, Aristotle makes
no appeal to the influence of a thing’s matter on its characteristic behavior.
This is at least partly explained by his conducting the discussion in a
historical mode; the reasons he gives are represented as reasons that actuated

2
To be clear, Aristotle does not think these theorists are wrong to identify a thing’s nature with its
substance; this is an identification he himself would endorse. The mistake must lie elsewhere, and
apparently (if Parts of Animals i.1 is any guide) in different places: e.g. in simply not considering the
substance of things (640a10–12), or in considering the substance of the wrong things (640b4–6, 17–22),
or of the right things, but in the wrong way (640b22–25, 640b29–31, 641a14–17) – in any case and in a
word, in not treating the substance of things by reference to their characteristic behavior or work
(640b35–641a3). (Parts of Animals i.1 is a guide to Physics ii, and vice versa, as is shown by the careful and
illuminating discussion of the relationship between these two texts in a recent unpublished conference
paper by Stephen Menn [in press].)
 Aristotle on interpreting nature 35
his predecessors. But the case is different in his discussion of the idea that
nature is form; as we will see, although there too he makes no appeal to
behavior, the reason he gives in support of the view is his own (193a30–b5).
The case for the view that nature is form is made to rest on two points:
First, that things that have natures are all “natural” or “due to nature” or “in
accordance with nature,” and, second, that nothing is any of these things
before it lays hold of the form whereby its type is defined.3 These two points
are recommended by the fact that analogous points hold for works of art,
and taken together they imply that nothing “has” its nature before it “has”
this form; this in turn implies (or anyhow suggests) that the nature of things
that have natures just is this form. Notice that this line of reasoning makes no
reference at all to what things that have natures “do” – to their characteristic
ways of moving and resting (let alone to the sources thereof). It focuses
instead on what they are: or rather, on when they first properly are called that
(a moment here specified just as coeval with the advent of form).4
Things are much the same for the reasons Aristotle gives for thinking
that form is “more” nature than matter (193b6–21). The first simply repeats
the line of argument we were just considering (193b7–8). The second re-
appropriates “Antiphon’s bed” (193b8–12) and thus should probably be
read in line with Aristotle’s earlier use of that example (193a12–14).5 The
third is based on a use of the word “nature” as the name for a kind of
process or “road” (a road “towards nature”); since this “road” towards
nature is also a road towards form, the fact that it is a road towards nature is
taken to show that nature is form (193b12–13, 18). Whatever we make of this

3
Aristotle takes pains to make clear that the “form” he has in mind is associated with definition;
though twice referring to it as “the shape” (ἡ μορφή), he offers several accompanying clarifications: it
is not just “the shape,” but “the shape i.e. the form” (ἡ μορφὴ καὶ τὸ εἶδος), “the one in accordance
with the definition” (τὸ κατὰ τὸν λόγον), “by which in defining we say what a thing is,” “not being
separable except in account” (193a30–31, b1–2, 4–5). This specification excludes certain perceptible
forms, of the sort whereby finished works of nature and art are recognized (one might think) for the
kinds of things they are. (Cf. PA i.1, 640b32–33, where Democritus is referred to as holding that “it is
evident to everyone what form (μορφή) it is that makes the man, seeing that he is recognizable by his
shape (σχῆμα) and color.”) It also excludes the forms posited by “those who talk of ideas,” who are
said to “separate the objects of natural science, though they are less separable than those of
mathematics” (Phys. ii.2, 193b36–194a1).
4
This needn’t imply that a thing’s matter is not its nature at all. After all, Aristotle does call the nature
a thing doesn’t yet have, namely when the thing is [what it is] only potentially, its nature (193a36–b1:
οὔτ’ ἔχει πω τὴν ἑαυτοῦ φύσιν). This suggests that he thinks of such things as somehow already being
what as yet they are not, in which case they might already somehow “have” what as yet they have not,
i.e. their own nature (which perhaps they would not, were matter not nature).
5
The principle underlying Aristotle’s reasoning here is admittedly difficult to make out. Most
conservatively, it is that a thing’s nature is revealed by “what comes to be from it.” But it is left
unclear why that should be: whether because what comes to be from a thing reveals its substance or
essence, or because it reveals the principle of its characteristic behavior.
36 sean kelsey
reasoning, it hardly appeals to the idea of nature as a kind of “principle” or
“beginning” (ἀρχή) of behavior; on the contrary, the key point is that
nature is not “that from which” (ἀφ᾿ οὗ or ἐξ οὗ), but “that towards which”
(εἰς ὅ) – not the beginning, but the end.6
In this chapter, then, in identifying the nature(s) of things, Aristotle
makes no appeal to the sources of their behavior; this is contrary to the
expectation created by his definition of nature, especially in the case of his
treatment of form, where he argues in propria persona. The effect (if not the
design) is to draw our attention to the idea that, if his discussion is in good
order, form must be a principle of motion too.

III
I have suggested that, in ii.1, the idea that form is a principle of motion is
conspicuous by its absence. But it is not just its absence that makes the idea
conspicuous; there is the further point that, as Aristotle is very much aware,
the idea presents difficulties of a kind that are salient in this context.
The definition of nature, in conjunction with the idea that form is
nature, implies that form is a principle of motion. The reason Aristotle will
expect this result to be difficult is that his predecessors (he thinks) were
unanimous in the view that nothing can be a source of movement that is not
itself in motion. (This is clear from his account of earlier theories of soul in
On the Soul i.) Surely Aristotle realizes that the obstacles in the way of
thinking correctly about how soul is a principle of motion will likewise
stand in the way of thinking that form is a principle of motion. This is not
merely because, in Aristotle’s view, soul is form. There is the further point
that the definition, in making nature a principle of motion, is evidently
meant to correct the Platonic idea that soul is a principle of motion.7 Given
this, Aristotle will almost certainly have had the topic of soul on his mind
in writing Physics ii.1; not only that, but he will have expected his defini-
tion to put that topic in the mind of his audience. But in that case it is very
unlikely that the difficulties involved in making form a principle of

6
This point depends on a contrast between nature and art: Unlike “doctoring,” which is not named
after what it is “to” – it is a road to health, not to medicine – the word “nature” or “growth” (φύσις) is
different. (It is true that this contrast is rooted in an underlying similarity: just as “doctoring” [ἡ
ἰάτρευσις] takes its name from doctors [or more precisely their art, “medicine” (ἰατρική)], because
[after all] it is an undertaking of doctors, so too “growth” [φύσις] takes its name from “growing
things” [τὰ φυόμενα], or more precisely from “nature” [φύσις], and that for just the same reason,
namely that it is an undertaking of them [193b17–18].)
7
On the Platonic background to Aristotle’s definition of nature, see e.g. Mansion 1946: 83ff.; Solmsen
1960: 95ff.; and Wieland 1970: 234, 240ff.
 Aristotle on interpreting nature 37
movement would simply not have occurred to him, or would have
occurred to him but struck him as not salient in this context.
Here it might be objected that surely these difficulties are easily side-
stepped by reflection on Aristotle’s well-known comparison of nature to art
(which is also very salient in this context). This comparison invites a
picture of nature-as-form as being (like art) a kind of capacity or power
(δύναμις); on such a picture, it seems straightforward enough just to say
that form is a “principle” of movement by being a capacity for movement
(perhaps better, a whole suite of capacities, of which a whole ensemble of
movements and rests are the manifestation or exercise (ἐνέργεια)). Simply
put, why should the idea that form is a principle of motion be any more
problematic than the idea that art is?
Plausible as this may seem, it is not the way Aristotle actually goes in
Physics ii; he does not highlight the idea that form is a capacity, and that
perhaps with good reason. Certainly it is no general solution to his
problem – how can something altogether immobile be a principle of
movement? – to reply that the issue presents no difficulties when the
something in question is a capacity. There will remain particular questions
about the prime mover, not to mention more general questions about
whether there must not be something else, a prior principle, whereby this
capacity is “energized.” Note too that Aristotle’s principal ground for
saying that form is nature is that form is “actuality” (ἐντελέχεια); although
this doesn’t rule out that form is also capacity, it hardly puts the point in
relief. What is more, as we will see, where Aristotle does emphasize form’s
causality, what he emphasizes is its role as an end.8
I return to the idea that form is an “end” (τέλος) below. My point now is
just that the idea that form is a “principle” (ἀρχή), and that precisely of
motion, stands out in ii.1, not just from Aristotle’s failure to mention it, but
also from the fact that it presents difficulties he will have expected his
audience to be very much aware of.

IV
I turn now to the next text on our agenda, Physics ii.7.9 This chapter’s
official business is to conclude the discussion of causes begun in ii.3,

8
This is forecast already at 193b12–18 (discussed briefly above, where it was noted that in this respect –
in its role as an end – nature is precisely contrasted with art).
9
The definition is used in passing in ii.6 (197b32–37), but the issues there are too slight to warrant
discussion in the main text. (The distinction between chance and luck is of some interest, considering
that Aristotle often speaks of nature as a cause to be ranged alongside of “reason” or “thought” [νοῦς,
38 sean kelsey
drawing its lesson for the study of nature, which is that it is concerned with
all four (198a22–24, b5–9). This is the conclusion of ii.7, and it is in the
course of arguing for it that Aristotle appeals to the definition of nature
(twice, at 198a28–29 and 198a36–b1). My ultimate focus here is on the
definition’s role in this reasoning, but it will take a little time to get there;
the basic movement of thought in this relatively neglected chapter is not
transparent.
I start with some summary, taking the chapter in two pieces (198a14–31,
a31–b9). First, after a brief restatement of the main results of ii.3 (198a14–
21), Aristotle begins by announcing his conclusion: The student of nature
will know all four causes (198a22–24). For reasons that are not immediately
apparent, he next remarks that three of them (form, mover, and end – τὸ
εἶδος, τὸ κινῆσαν, τὸ οὗ ἕνεκα) “often come to one”; he supports this by
claiming that form (τί ἐστι) and end (οὗ ἕνεκα) flat-out are one and the
same, while the first mover (ὅθεν ἡ κίνησις πρῶτον) is the same as these “in
kind” (τῷ εἴδει), defending this latter point by appeal to an example
familiar from ii.1 (“man begets man”), which example he says generalizes
to all moved movers (198a24–27). This generalization prompts an apparent
aside about unmoved movers, to the effect that they lie outside the study of
nature; it is here that Aristotle makes his first appeal to the definition “such
as are not of this kind are no longer inside the province of physics, for they
cause motion (κινεῖ), not possessing motion or a principle of motion in
themselves, but being themselves incapable of motion” (198a28–29).
Second, though making as if drawing a consequence from all this (ὥστε
κτλ.), Aristotle continues by picking up a different thread entirely, drop-
ping the three causes he has just said are one and picking up another three –
matter, form (τί ἐστιν), and first mover (τὸ πρῶτον κινῆσαν). He starts by
saying that it is by referring to these three that (as it is) people explain things
(198a31–33); this is his own interpretation of what people do, which he
defends by pointing to the questions that actually preoccupy them (198a33–
35).10 With this description of the status quo in hand, he then proceeds to
enlarge it so as to encompass his preferred ideal. His first step is to “divide”

διάνοια] and contrasted with “chance” or “luck” [196a30; cf. 196b22, 198a4, 6, 10, 12]. The contrast is
marked by the presence or absence of order, as is clearest in a passage from ii.4, in which Aristotle
writes as if it were accepted by all parties that “nature, reason, and the like” differ from chance and
luck in that the proceedings ascribed to them unfold in a definite order [196a28–35; cf. PA i.1,
641b10–28]. Such passages make clear that Aristotle regards the association of nature with order, not
as an innovation, but as a piece of common ground [on this point, see also Scharle, Chapter 5 in this
volume].)
10
“For in respect of coming to be it is mostly in this last way that causes are investigated – ‘what comes
to be after what? what was the primary agent or patient?’ and so at each step of the series.” Aristotle
 Aristotle on interpreting nature 39
the moving cause – what he here calls “natural kinetic principles” (αἱ ἀρχαὶ
αἱ κινοῦσαι φυσικῶς) – into two kinds (198a35–36). Noting that one of
these, although it causes motion “naturally” (φυσικῶς), is not itself “nat-
ural” (φυσική), because it lacks an internal principle of motion (this is his
second appeal to the definition), he then claims that this non-natural sort
of cause encompasses all unmoved movers, giving as examples the prime
mover and form, and defending the latter (or is it both?) on the grounds
that it is an end (198b1–4). Only then does he finally draw the conclusion
announced at the outset, that the student of nature should explain in terms
of all four causes, and this on the grounds that such a one (of all people)
should know nature, which is for something (198b4–5ff.).11
Although these two pieces of text are apparently meant to comprise a
single discussion, aimed at a single conclusion, it is not obvious how to put
them together; certainly their several contentions are not simply so many
premises of a single deductive argument. I will work by first just identifying
some overlapping threads, and then trying, with the help of a little friction
and spit, to twist these together into a single strand.
So, first, both pieces at least tacitly identify formal and final cause (the
first identifies them outright, while the second does so implicitly, in
claiming that form is a “kinetic principle” (ἀρχὴ κινοῦσα) because it is an
end) (198b1–4). Second, both pieces associate moving cause with formal/
final cause (again, the first does this outright, while the second does it
implicitly, offering form as an example of an (unmoved) kinetic principle)
(198b1–3). Third, both pieces at least tacitly divide moving causes into
“moved” and “unmoved,” and, fourth, both appeal to the definition of
nature to mark the latter as not “natural” (φυσική) or “belonging to the
study of nature” (φυσικῆς).
These common elements identified, I propose to read the second piece
of text as conducting the argument proper, drawing on materials supplied
by the first and refining them as necessary along the way. Read this way, the

means to align these three questions with his three causes, namely (respectively) form, mover, and
matter. Note the concluding phrase, “and so at each step of the series,” i.e. until we come to the first.
11
“Hence since nature is for the sake of something, and this we must know, we must explain the ‘why’
in every sense of the term (ὥστε ἐπεὶ ἡ φύσις ἕνεκά του, καὶ ταύτην εἰδέναι δεῖ, καὶ πάντως
ἀποδοτέον τὸ διὰ τί).” I take the words “and this we must know” (καὶ ταύτην εἰδέναι δεῖ) as part
of the dependent clause introduced by “since” (ἐπεί). This is because I think that (1) the antecedent
of “this” (ταύτην) must be “nature” (φύσις), in which case I think that (2) the words “and this we
must know” must give a premise, not a conclusion. So too (on both points) Philoponus 305.11–13
(Vitelli) and (on (1)) Simplicius 368.12 (Diels). (Simplicius considers (2), although he does not prefer
it.) Apropos (1), many translate as though the antecedent of “this” (ταύτην) were “cause” (αἰτία,
which occurs several lines back at 198a34) (so e.g. Charlton 1970, Hardie and Gaye 1930, Irwin and
Fine 1995), but I think this is difficult.
40 sean kelsey
basic movement of thought is from a statement of established scientific
practice to a recommendation of Aristotle’s own ideal – from explaining
things in terms of three causes to explaining them in terms of four (198a31–
33, b5). The argument’s basic strategy, encapsulated in the idea that “often
the three come to one,” is to bring out how a commitment to explanation
in terms of the missing fourth is already implicit in a commitment to
explanation in terms of the other three; in particular, the argument brings
out how a commitment to asking “for what end?” is already implicit in a
commitment to asking “what is it?,” which in turn is already implicit in a
commitment to asking after “the first moving cause” (τὸ πρῶτον κινῆσαν).
In appreciating this series of commitments, it is a help to realize that final,
formal, and moving cause “often coincide,” at least in kind. This is a help,
but it is not enough: In addition, one must also appreciate not only that
kinetic principles are of two sorts, mobile and immobile, but also that the
latter, though not themselves “natural” (φυσικαί), are indeed among the
principles of natural science, and this because they do indeed cause motion,
and do so “naturally” (κινεῖν φυσικῶς) (198a36) – not by “having” natures
(as Aristotle concedes), but by being natures (as he leaves us to infer).12
Once we appreciate this – that these principles are natures – we will
likewise appreciate their claim on the professional attention of the natural
scientist.
This anyhow is how I propose to read “the argument of ii.7” – the
“official” argument for the chapter’s stated conclusion. I turn now to
the role of the definition of nature in this argument and begin with the
following observations. First, although the definition’s ostensible function
is to exclude certain principles – to show they are not “natural” – it is clear
that Aristotle also holds, and indeed must hold, if his reasoning is to
support his conclusion, that these principles are also included in the
study of nature: not as a department of its subject matter, but as among
its explanatory principles. Second, although Aristotle’s stated conclusion is
that the student of nature will be concerned with all four causes, his
“argument” for this is very indirect. Forget about the present account of
it and just consider some of its raw materials: the association of formal,
final, and motive causes, the division of the last into moved and unmoved,
or the division of the latter into God and form. Is all this necessary? Is it
helpful, or at least more helpful than obfuscatory? Is this really the best
way of arguing that the student of nature will be concerned with all four

12
The distinction between “having” a nature and “being” a nature is an important theme in Scharle
2008.
 Aristotle on interpreting nature 41
causes – i.e. final causes too, and not merely the other three? Surely not.
Third, not only does the chapter’s argument proceed circuitously, but
Aristotle appears to take up the controversial element in its conclusion –
that the student of nature must seek final causes too – again in the very next
chapter, in a way that is vastly more direct and more satisfying.
These observations invite the thought that if there is a method to this
madness, it is not to be read off the chapter’s explicit signposts. With that
thought in mind, I propose that we read ii.7 in light of the problem
thrown up by ii.1, namely, the problem of how form can be a principle of
movement. We know that Aristotle thinks this a difficult problem, and that
he thinks the answer to it lies in the point that “ends” (τέλη) are genuine
“beginnings” (ἀρχαί) – just the point left uncovered by the meandering
argument of ii.7. Read this way, against the backdrop of ii.1 – which ii.7’s
use of the definition will remind us of anyway – perhaps there is some sense
in a proceeding that seems otherwise rather pointless.

V
So far I have suggested that one purpose of the definition is to highlight the
idea that form is a principle of movement, that this idea will be felt to create
difficulties, and that resolving these difficulties requires an appreciation of
the fact, mentioned in ii.7, that ends are kinetic principles too. Putting
these points together, the result is that, for Aristotle, nature is indeed a
principle of movement, not only by being a capacity for movement, but
also and primarily by being an end. Before turning to the last set of passages
on our agenda, I want to pause briefly to develop this result.
The chief interest of this idea – that form is a principle of movement,
primarily by being an end – lies in the shape it gives to the definition’s
implications for how to interpret the phenomena of nature. The definition
makes nature a principle of “movement and rest”; in so doing, it places
nature at the origin of behavior (as distinct from genesis) – that is, it tells us
that the substance of things due to nature is implicated in, and thus
revealed by, what they do. But to add to this, first that the nature and
substance of things is their form, and then that the way form is implicated
in behavior is primarily by being its end, is now to say something really
striking about things due to nature. It is to say that the point of their
characteristic behavior is to perpetuate their distinctive form of being.
I do not expect that this result will be surprising, considered just as a thesis
Aristotle subscribes to; on the contrary, I hope it will be recognized as
confirmed from elsewhere – perhaps most strikingly in On the Soul ii.4,
42 sean kelsey
where Aristotle puts precisely this interpretation on reproduction, going on
to claim that for just this reason reproduction is the most “natural”
(φυσικώτατον) (415a26) of all the works of living things. The novelty (such
as it is) lies in seeing not that this is something Aristotle holds, but that it is a
quiet, controlling theme of Physics ii as a whole. The lesson of Aristotle’s
definition of nature is that the key to deciphering natural phenomena – to
seeing them for what they are – is to interpret the movement and rest of
things due to nature as for the sake of immortalizing their being and form.13

VI
I have observed that ii.1’s definition of nature, in conjunction with the
identification of nature with form, implies that form is a principle of
movement, and I have suggested, first, that we are expected to find this
difficult – how can something immobile be a principle of movement? –
and, second, that we are to recognize in ii.7 the elements of a solution,
inasmuch as that chapter goes out of its way to say that forms are ends and
that ends are immobile principles of movement. Notice that this solution is
no good unless the particular forms we call natures are ends, and that just
this is the burden of ii.8: that nature (in the sense of form) is a final cause.14
So far as that goes, then, ii.8 as a whole may be read as advancing the line
of thought we’ve been following. But it remains to consider the two
passages in ii.8 in which Aristotle actually uses the definition of nature
(199b9–26, 28–32). Both passages occur towards the very end of the
chapter; I take them in turn.
In the first, Aristotle is arguing against an implication of a view he finds
in Empedocles, namely that “coming to be among seeds too would have to
have been as luck would have it” (199b13–14).15 Aristotle’s complaint is not

13
This sits well with the recurring example (in ii.1 and ii.7): “man begets man.”
14
It is true that, in formulating the chapter’s conclusion, Aristotle sometimes uses the phrase “for
something” (ἕνεκά του) (rather than “that for the sake of which” [οὗ ἕνεκα]) (notably ad init. and
ad fin.). But it is plain that in these passages the phrase “for something” functions to specify which kind
of cause Aristotle is here claiming nature to be. If that is right, then when Aristotle says that nature is
“among the for-something causes” (τῶν ἕνεκά του αἰτίων), or that it is a “cause in the for-something
way” (αἰτία οὕτως ὡς ἕνεκά του), what he means is that it is an end (as opposed to for an end). If doubts
remain, they are removed by another, fuller formulation, in which it is clarified that the “nature” this
conclusion holds of is form, seeing as form is an end:
And since “nature” means two things, the matter and the form, of which the latter is the end, all
the rest being for the sake of the end, the form must be the cause in the sense of “that for the
sake of which” (καὶ ἐπεὶ ἡ φύσις διττή, ἡ μὲν ὡς ὕλη ἡ δ’ ὡς μορφή, τέλος δ’ αὕτη, τοῦ τέλους δὲ
ἕνεκα τἆλλα, αὕτη ἂν εἴη ἡ αἰτία, ἡ οὗ ἕνεκα). (199a30–32)
15
Trans. Charlton 1970.
 Aristotle on interpreting nature 43
simply that this is false, or even impossible, but (stronger) that “a person
who says that does away with nature and things due to it altogether (ὅλως δ’
ἀναιρεῖ ὁ οὕτως λέγων τὰ φύσει τε καὶ φύσιν; 198b14–15).”16 His reasoning
draws explicitly on the definition of nature and goes something like this: If
seeds grow by nature, then (unless something interferes) they grow into the
same thing every time,17 in which case it can hardly be a coincidence that they
grow into that – which it would be if they grew “as luck would have it,”
which they would if their growth were not for an end (199b14–26). Of
interest in this reasoning is the broader moral: that nature will be a principle
(ἀρχή) only if it is an end (τέλος). For in condensed form the argument is
simply this: If nature really does stand at the beginning of the
proceedings whereby seeds grow into plants, then the result (if nothing
interferes), namely nature as form, will be precisely what those proceedings
are for.18 If that is right, then Aristotle’s use of the definition here comple-
ments the use he makes of it in ii.7. There the moral was that nature (as
form) can be a motive principle if it is an end; here the moral is that nature
will be such a principle only if it is an end. That is why to deny that nature is
an end is “[to] do away with nature and the things due to it altogether”
(198b14–15).
The second passage I want to consider occurs at the very end of ii.8:
If the ship-building art were in the wood, it would produce the same results
by nature (ὁμοίως ἂν τῇ φύσει ἐποίει). If, therefore, purpose (τὸ ἕνεκά του)
is present in art, it is present also in nature. The best illustration is a doctor
doctoring himself: Nature is like that (μάλιστα δὲ δῆλον, ὅταν τις ἰατρεύῃ
αὐτὸς ἑαυτόν· τούτῳ γὰρ ἔοικεν ἡ φύσις). (199b28–32)
This illustration evokes a passage from ii.1:
[A] man who is a doctor might cure himself. Nevertheless it is not in so far as
he is a patient that he possesses the art of medicine: It merely has happened
that the same man is doctor and patient – and that is why these attributes are
not always found together. So it is with all other artificial products. None of
them has in itself the source of its own production. But while in some cases
(for instance, houses and everything else wrought by hand) that principle is

16
Trans. Charlton 1970.
17
It is here that Aristotle appeals to the definition: “for ‘by nature’ is everything that, moved
continuously from some principle in itself, arrives at some end” (φύσει γὰρ, ὅσα ἀπό τινος ἐν
αὑτοῖς ἀρχῆς συνεχῶς κινούμενα ἀφικνεῖται εἴς τι τέλος; 199b15–17). (As he clarifies immediately, he
does not mean that all arrive at the same end, nor just that all arrive at some end or other, but that
each arrives at the same end every time, unless something interferes: ἀφ᾿ ἑκάστης δὲ οὐ τὸ αὐτὸ
ἑκάστοις οὐδὲ τὸ τυχόν, ἀεὶ μέντοι ἐπὶ τὸ αὐτό, ἂν μή τι ἐμποδίσῃ [199b17–18].)
18
Aristotle’s reasoning in this passage may be usefully compared with that in Phys. ii.8, 199a8–12 and
PA i.1, 641b23–642a1.
44 sean kelsey
in something else external to the thing, in others those which may cause a
change in themselves in virtue of a concomitant attribute – it lies in the
things themselves, but not in virtue of what they are. (192b23–32)
Significant in both of these passages is the idea that nature’s work is
reflexive: As Aristotle puts it in ii.1, unlike artifacts, which (being wrought
by hand) have “the principle of their making outside them,” things due to
nature have this principle “in themselves.” The claim Aristotle makes here
is arresting; how could anything contain within itself “the principle of its
own making” – this seems to require that it pre-exist its own coming into
being. The solution, I submit, lies in seeing that, for things due to nature,
the “principle of making” and “principle of movement” are one and the
same: that is, although their “making” and “moving” – that is to say, their
genesis and behavior – are distinct proceedings, they have the same
principle, and that because they have the same end.19
Earlier I suggested that one of the definition’s important morals is that
the phenomena of nature are for the sake of perpetuating being or form.
This moral is reinforced by Aristotle’s use of the definition at the end of
ii.8, and it makes Aristotle’s closing remark in that chapter highly apposite:
“the best illustration is a doctor doctoring himself: nature is like that.” In a
word, nature mends itself.

VII
My suggestion, then, is that the definition of nature holds the “key” to
interpreting both the phenomena of nature (as Aristotle understands them)
and Physics ii itself. The “key” is the idea that form is a principle of motion
and rest. The definition “holds” this key by setting off the following
sequence of ideas: that forms are natures, which means that they must
also be principles of movement, which they can be if and only if some
principles of movement – indeed, the truest ones, i.e. the ones that are
most of all “first,” and thus most of all “principles” (ἀρχαί) – are immobile,
causing motion naturally by being its end. This sequence of ideas shows us
how to interpret the phenomena of nature, because its meaning is that
movement and rest are a striving for being (and thus an imitating of the
“one ruler” (εἷς κοίρανος), the true “principle” (ἀρχή), “which is comple-
tely unchangeable and the first of all things (τὸ παντελῶς ἀκίνητον καὶ

19
As Aristotle puts it in ii.7, things behave as they do “because it is better thus, not without
qualification, but in relation to their substance or essence” (διότι βέλτιον οὕτως, οὐχ ἁπλῶς,
ἀλλὰ τὸ πρὸς τὴν ἑκάστου οὐσίαν; 198b8–9).
 Aristotle on interpreting nature 45
πάντων πρῶτον)” (Metaph. Λ.10, 1076a4, Phys. ii.7, 198b2–3)). It shows us
how to interpret Physics ii by revealing that one part of that book’s agenda
is to clear away the obstacles to the idea that form is a principle of move-
ment. We notice this by keeping track of and interrogating Aristotle’s use
of the definition in ii.1, ii.7, and ii.8.
This last suggestion, about how to interpret Physics ii itself, is of course
very tentative. Even if it were accepted, the question would remain: Why
does Aristotle proceed so obliquely? It is maybe a step towards answering
this question to consider the alternative, which would have been to argue
for the idea directly: Since form is nature, and nature a principle of motion,
form must be a principle of motion too. What is striking about this
argument is that, even allowing that its premises are in proper order,
each resting on a decent set of more or less independent considerations,
it is powerless to persuade; the obstacles to the idea that form is a principle
of motion, not arising from the lack of an argument that would demon-
strate it, are not removed by the supply of one. We might also ask exactly
what Aristotle is trying to accomplish in Physics ii: Is it to demonstrate
theorems, for the books and for the ages, or rather something different, e.g.
to help make a particular audience “capable of judgment” (δυνάμενος
κρίνειν) (193a4–5). The latter is arguably the more difficult undertaking,
in which indirect methods and even hidden themes are perhaps sometimes
the better course.
 chapter 3

Nature as a principle of change

Stasinos Stavrianeas

Introduction
The study of nature, says Aristotle, aims at the understanding of the
principles, causes, and elements of the natural world. This is the fruit of
an investigation that starts from things more familiar to us, yet mixed with
much confusion as to their true nature, and proceeds towards things less
familiar to us but more intelligible by nature, i.e. to the discovery of this
true nature (Phys. i.1, 184a14–23). The things that are familiar to us are
compounded, comprehensive, and closer to perception, and we under-
stand them in a universal and undifferentiated manner; further investiga-
tion is needed in order to grasp our subject matter in a more detailed
manner. Our first conception does not have the status of scientific under-
standing, but it has the advantage of being based on some sort of inductive
reasoning that lends to it the status of an obvious and fundamental
assumption.1 One such assumption is that the natural world of changing
things exists (Phys. i.2, 185a12–13). Starting from this, Physics i investigates
the principles necessary for change, natural or otherwise. This investigation
meets the programmatic thesis that our study should move from what is
more intelligible to us towards what is more intelligible by nature, but
merely identifying the principles of change does not mean that the analysis
of change is complete. Other principles and causes need to be added as
well, such as most prominently the efficient cause. In other words, nothing
in the Physics i analysis of change provides in itself sufficient material for
understanding the complexity of – say – the changes involved in animal
generation.

I am grateful to Mariska Leunissen for stimulating and extensive comments that helped me to elaborate
and clarify many of the ideas included here, to Pavlos Kontos and Lampros Spiliopoulos for generously
commenting on earlier drafts, and to Andreas Anagnostopoulos, Panos Dimas, Vassilis Kalfas, and
Spyros Rangos for comments and discussion.
1
See Bolton 1991: 27–29.

46
 Nature as a principle of change 47
Physics ii turns to the investigation of the principles of natural change,
and first to what nature is, i.e. to a definition that demarcates the entities
that possess a nature. Still, further investigation will be needed in order to
determine in detail that which is characteristic of each kind of natural
being. In that sense, the definition, or so it may be argued, no more gives us
an articulated picture of what nature is than the analysis of change in Book
i, in terms of the principles of form, subject, and privation, gives us a
complete account of what change is. So although Physics ii.1 establishes
more clearly what nature is, this is by no means the complete account. It is
understandable, then, that the proposed definition of nature as an inner
principle of change and rest is often found wanting. Both this definition
and the ideas associated with it require further elaboration before they can
yield a complete understanding of nature.
My aim in this chapter is to examine the conditions that Aristotle’s
preliminary definition of nature in Physics ii.1 introduces. I argue that the
claim that nature is a principle of change need not be read as offering a
complete and final articulation of what nature is, but rather should be
interpreted as a general account or preliminary sketch that will be qualified
in order to accommodate the nature of various natural kinds. For, ulti-
mately, it is the detailed study of natural kinds that Aristotle presents in other
works that provides a much richer network of distinctions and a more precise
determination of what is characteristic and distinctive of natural beings. In
section I, I specify the conditions included in Aristotle’s definition of nature.
In sections II and III, I explain the way in which natural generation satisfies
these general conditions, while artificial production does not. Section IV
shows how even the locomotion of the elements counts as natural given these
conditions, despite the fact that the elements are not self-movers.

I Aristotle’s definition of nature in Physics ii.1

The list of entities that exist by nature includes animals and plants, parts
thereof, the four simple bodies, and, most likely, their mixtures. The
reason these entities exist by nature is, according to Aristotle, that “each
of them has within itself a principle of change and rest, some in respect of
place, some in respect of growth and decline, some in respect of alteration”
(Phys. ii.1, 192b13–15).2
2
Animal parts and elements are excluded in other works, because they lack the kind of unity necessary
for being substances, let alone natural ones (Metaph. Z.16, 1040b5–10). Animal parts are, I think,
included here only for the sake of completeness. The case of elements is dubious, but at least
according to the Physics they possess a principle of change in themselves, though in a special way
48 stasinos stavrianeas
Two things deserve comment here. First, it is uncertain how we should
interpret the expression “principle (archē) of change.” A natural way to
understand it, and the way in which it has been traditionally read, is as
denoting the “efficient cause” of change.3 This reading, however, is
questionable for several reasons: (1) Nothing in what precedes in Physics i
indicates that “principle” refers merely to the efficient cause;4 (2) this
reading seems to exclude several types of natural changes of entities such
as the elements, their mixtures, and possibly some of the animate beings,
because they are not produced by an internal efficient cause; (3) the
arguments offered in Physics ii.1 supporting the identification of nature
with either the matter or the form of entities that are considered to exist by
nature do not turn exclusively on the activity of matter or form as efficient
causes (Kelsey 2003: 61). Rather, the ways in which something can be the
nature of an entity includes (in many cases) all four Aristotelian causes.
Second, it is worth asking whether the definition of nature Aristotle
offers here is supposed to count as a scientific definition in the strict
Posterior Analytics’ sense. I do not believe it is; it is rather a first approxima-
tion or nominal definition that states a common, general characteristic of
natural things. More precise definitions are to be provided elsewhere, and
they will track the different ways in which various kinds of entity possess a
nature. Starting an investigation with a general definition is not unprece-
dented in Aristotle. In On the Soul, he argues that it would be absurd to try
to provide a common definition of the soul without also working towards
the definitions of particular kinds of soul that correspond to the proper
indivisible species of ensouled beings,5 as a common definition of the soul
will not be able to fully track the latter. Even though Aristotle is not explicit
about this, I believe that the same methodological considerations that
pertain to soul also pertain to nature, especially given that nature, in
animate beings, is in fact their soul (PA i.1, 641b9; GA ii.4, 740b37–38).

(see section IV). This way may not, in the end, be sufficient for attributing to them a nature, thus
excluding them from the list.
3
See, for instance, Simplicius In Ar. Phys. 264.9–10 (cf. Phys. ii.3, 194b29–30). However, all four causes
are principles (Metaph. Δ.1, 1013a17).
4
The word is used extensively in Physics i (forty-one occurrences), but only in one of them does it refer
to efficient causes (Phys. i.6, 189a24–26, referring to the Empedoclean powers of Love and Strife).
The word usually refers to the principles of change, such as the subject of change, its form, and its
privation (this is also the case when it is used in conjunction with the word aitia [cause]: see e.g. Phys.
i.7, 190b17–20 and i.9 192a13).
5
According to De Anima ii.3, there is a single logos (account) of soul only in the sense in which there is
a single account of the genus “rectilinear figure.” Since this account is common to all rectilinear
figures but not unique to any one of them, Aristotle says it is the definition of none (414b20–28, cf.
DA i.1, 402b6–9).
 Nature as a principle of change 49
Given Aristotle’s remarks about the need to provide more specific defini-
tions of the soul to fill in the more general one already provided in On the
Soul, we should expect the same to be the case regarding Aristotle’s
definition of nature in the Physics: It will be satisfied by different natural
beings in different, yet related, ways.
The fullest formulation of the definition introduces two further condi-
tions on nature being a principle of change and rest: namely, that nature is
such a principle in that in which it inheres (a) primarily and (b) per se, i.e.
not per accidens (192b21–23). Condition (a) exploits the distinction between
something belonging to a subject primarily or immediately and it belong-
ing to a subject non-primarily or non-immediately. An attribute,
F, belongs to a subject, s, non-primarily if it belongs to s, because there is
a further subject, s’, such that F belongs to s’ and s’ belongs to s. By contrast,
an attribute, F, belongs to a subject, s, primarily if it belongs only to s and to
no other subject (or, if it belongs to other subjects as well, it belongs to
them because s is attributed to them). Condition (b), on the other hand,
exploits the familiar distinction between belonging to a subject per se, i.e.
necessarily, and belonging to a subject accidentally, i.e. contingently.6
The two distinctions are independent. First, an attribute may belong to
a subject per se, but not primarily, in those cases where it belongs to it in
virtue of belonging to some other property which in its turn belongs to the
subject per se: Having the sum of its three angles equal to two right angles
belongs per se, though not primarily, to an isosceles triangle, because it
belongs primarily and per se to triangle and triangularity in its turn belongs
per se to isosceles. Second, an attribute may belong primarily to a subject,
though not per se, in those cases where the relation is not mediated by a
third term and yet the attribute is not essential to the subject. This is the
way white, for instance, belongs to a surface. Aristotle provides illustrations
for both conditions in his text, namely essentiality and immediacy.
The condition of essentiality is used in order to block one way in which
things that do not exist by nature may be thought of as possessing an inner
principle of change. Aristotle’s example is a physician who happens to be
ill, and who can thus cure himself. The restoration of health is the product
of his craft, i.e. medicine, so in this case the principle directing the change
belongs to the person who suffers the change. They are one in number, and
this may be taken as suggesting that products of craft may possess an inner
principle of change. Hence, Aristotle needs to explain why the definition
does not apply in such cases. He does that by noting that here the agent and

6
See Simplicius In Ar. Phys. 267.21–268.3, followed by Ross 1936: 501.
50 stasinos stavrianeas
the patient, i.e. the source and the subject of the change, may belong to one
and the same entity but they belong to it under different causally relevant
descriptions. Hence, not only can source and subject be separate in number
(no physician needs to be a patient and vice versa), but they also must be
separate in definition or essence (GA i.18, 724a43–45; ii.1, 735a2; ii.4,
745b28–29). So, in craft, the principle may be internal, accidentally, but
needs to be external, essentially.
The condition, of immediacy, is plausibly at work in the following
passage:
Οn the other hand, a bed and a coat and anything else of that sort do not in
so far as they happen to receive these predicates – i.e. in so far as they are
products of craft – have an innate impulse to change. But in so far as they
happen (sumbebēken)7 to be composed of stone or of earth or of a mixture
of the two, they do have such an impulse, and [they have it] just to that
extent. (Phys. ii.1, 192b16–20)
The examples here illustrate a different way in which artifacts may be
thought of as possessing a principle of change, compared with the self-
doctoring doctor. The latter exemplifies the coincidence of active and
passive powers in one entity. The former focuses on a principle of change
belonging to artifacts in virtue of the material they are made of. Indeed,
artifacts may be thought of as possessing an internal principle, because their
matter possesses such a principle (or is itself composed out of matter that
possesses one). For instance, a coat will fall to the ground if not hung from
a nail; a balloon will fly upwards. However, such motions are not due to
formal properties of the artifacts, but rather to the material constituting
them. Thus, the corresponding principle belongs primarily or immediately
to the matter. Since the matter constitutes the composite, its principle
belongs to the composite as well, but only in virtue of belonging to the
matter, thus non-immediately; therefore it does not belong to the artifact
properly speaking. We may take then the two examples as illustrating
two independent conditions for possessing a nature: essentiality and
immediacy.
7
The use of sumbebēken for characterizing the way artifacts relate to their matter may suggest that what
is at issue here is, again, the distinction between per accidens and per se belonging. However, this
distinction does not block the claim that the matter of artifacts is their nature, since the latter is
conditionally necessary and not accidental to them. Hence, it is arguable that they possess per se the
principle of change in their matter. Further, Phys. i.3, 186b18–20 introduced two senses of sumbebēkos
(attribute): (a) accidental attributes (the principle of medicine belongs in this way to the patient); and
(b) attributes in whose account the subject to which they belong is included (the principle of the
matter belongs in this way to the artifact as a compound of matter and form). Cf. Phys. viii.4,
254b8–10 and Simplicius In Ar. Phys. 1207.46–56.
 Nature as a principle of change 51
The need for two independent conditions becomes relevant if we turn to
the second part of Physics ii.1, where Aristotle rehearses the competing
arguments for identifying nature with either matter or form (193a9–b21).
Τhe argument in favor of matter suggests that the matter of natural beings
belongs both essentially and immediately to them, while the matter of
artifacts satisfies only the former condition. If so, then both conditions are
needed for understanding the special role of matter in animate beings and its
contribution to their nature. The argument proceeds from the claim that the
nature of any material being is its first constituent, which in itself does not
possess the arrangement of what it constitutes, as wood does not possess the
arrangement of bed. Two characteristics of this constituent matter support
the claim: First, matter, by being able to survive the loss of what it consti-
tutes, has some ontological priority over the compound; second, if matter
had a power to generate anything, it would generate something similar to
itself rather than something similar to the compound, suggesting again that
the principle of change belongs to it rather than to the arrangement of the
compound (193a9–11, cf. 193a29). The first aspect turns on a criterion of
survivability, while the second attributes a generative power to what survives.
Satisfaction of both criteria grounds immediacy, but it is the second one that
reflects a radical difference between artifacts and natural beings.8
The first criterion, survivability, does not seem to be a necessary
condition concerning the contribution of matter that Aristotle endorses.
For there are cases where Aristotle claims that the matter that survives the
perishing of a compound is the same only in name or “homonymously.”9
This principle of homonymy seems to go against survivability, if the latter
means that the matter persists as the same type of thing throughout in all
cases. And even if Aristotle only endorses a weaker version of survivability,
it seems that both artifacts and natural beings can satisfy it. Aristotle
therefore needs a second criterion to further distinguish natural beings
from artifacts.

8
Even if we grant that wood, perhaps qua plant, generates wood, it seems weird to make the same
claim for earth. Still, Aristotle argues in On Generation and Corruption that generation by reproduc-
tion of a specifically similar individual, as when human generates human, is on a par with the
production of something generically the same, as when fire generates fire (i.5, 320b19–21). The
argument here, however, does not require that natural entities literally reproduce themselves: It aims
to show that if the matter of artifacts could generate something, what will be common between
“parent” and “offspring” will be the material and not the arrangement. The fact that artifacts made
out of, say, stone do not generate stone does not contradict this point.
9
This principle applies to the matter of animate beings (Metaph. Z.10, 1035b25, PA i.1, 641a1), but
possibly also, in a weaker way, to artifacts (see Pol. i. 2,1253a22, DA ii.1, 412b15, Metaph. Z.10,
1034a22–1034b1, Meteor. iv.12, 390a12).
52 stasinos stavrianeas
Aristotle’s example of the wooden bed shows that the matter of artifacts
fails to satisfy this second criterion, while the fact that the matter of natural
beings does satisfy it was already suggested by Aristotle’s claim that the
underlying thing out of which a compound is constituted is the first thing
out of which the constitution proceeds, which in animate beings is identi-
fied with their seed (or embryo: Phys. i.7, 190b2–3; i.9, 191a25–31). Now,
the seed’s power is analogous to the imagined generative power in the
wood: It generates another being similar to the one it itself comes from.
Thus, the nature of what gets generated can be partly identified with this
matter that is its first constituent. Given this identity, the principle that
immediately belongs to the first constituent will immediately belong to
what it is the first constituent of. And, if so, the matter of natural things
satisfies the criterion of immediacy as well, although properly speaking the
generative principle merely resides in the matter (as we learn from
Aristotle’s Generation of Animals, it is the formal principle in the matter
that generates; this specification is not pursued here). However, all this
shows is that the matter of artifacts fails to satisfy the immediacy criterion,
and Aristotle also wants to claim that artifacts also do not possess a nature
because their efficient cause fails to satisfy the essentiality criterion (that is,
their active and passive causes of changes are always separate).

II Aristotle’s account of why artifacts cannot produce themselves

Artifacts and natural beings are contrasted in terms of the efficient cause
involved in their production: “None of them [that is, artifacts] has in itself
the principle of its own production” (Phys. ii.1, 192a27–28). Since the
production referred to here is a substantial change, the intended contrast
seems to be that artifacts do not generate themselves, while natural beings
do. However, even if this is in some sense true of natural beings, it is true
only in a qualified sense. Strictly speaking, animate beings do not generate
themselves. Thus, Aristotle’s reason for denying that artifacts have a nature
cannot merely be that they do not generate themselves. So what can it be?
And whatever this reason is, will it also be able to exclude artifacts such as
miraculous automata that can reproduce themselves as natural beings do?
The reason lies in the way Aristotle understands art as an exclusively
human activity. This view surfaces in Physics ii, but is formulated explicitly
in other texts. For instance, in the Nicomachean Ethics Aristotle distin-
guishes art (poiesis) from action (praxis) by defining the former as “a
productive disposition accompanied by true rational prescription (logoi)”
(vi.4, 1140a10–11). Artifacts are produced by changes that stem from a
 Nature as a principle of change 53
disposition, i.e. some sort of psychological state in an agent that involves a
capacity to attend to true logoi. And one can attend to such logoi in virtue of
having a soul that is receptive to them in some way or another (Int. 1, 16a9–
11). This in its turn suggests, first, that any craft contains an epistemic
component, and, second, that to the extent that only human souls possess
such an intellectual capacity, it is only humans that are able to invent, or
internalize through learning, the principles that define a certain craft.
Consequently, only humans can act as principles and efficient causes of
artificial production.10 This feature then differentiates artificial production
from natural generation, in a way that puts the source of production in
crafts necessarily in something separate from what is produced.
Aristotle comes to this view because of a conviction he shares with the
materialist, namely that matter cannot organize itself so as to take the form
of an artifact. The principles that reside in matter are not capable of
effecting changes that will result in something that qualifies as the product
of craft. At least not intrinsically and regularly so, for in some cases, such as
the restoration of health, this may happen by accident. Aristotle makes this
point in Phys. ii.8, 198b15–17 when he claims that art generally either (a)
completes what nature cannot bring to a finish, or (b) imitates nature. In
cases under (a), what seems to be missing is an appropriate natural
principle that can effect the necessary changes resulting in the correspond-
ing end. Craft is described in Metaphysics A.1 as the result of the discovery
of such principles that will realize specific ends, promoting utility or
recreation.11 What is essential for such a discovery is a piece of knowledge
that is universal and explanatory, explaining what such and such material
changes are good for. Cases under (b), on the other hand, are cases where
the corresponding principle can be found in the domain of nature. A
typical example is the restoration of health. Health can be the result of a
spontaneous, accidental change, but it is more regularly and intrinsically
produced by medicine. What differentiates these two cases is not the
material changes in the body, but rather that in the latter case these changes
are conceived, planned, and effected as something good for restoring the
healthy state. Thus, medicine is described as comprising an intellectual,
deliberative component, as well as a productive one (Metaph. Z.7, 1032b6–
10, b15–23). The first component works towards discovering the principles

10
See Phys. ii.8, 199a20–23 and Metaph. A.1, 980b25–28. For the dependence of craft on practical
wisdom and the tension this generates for Aristotle’s analogy of craft with nature, see Broadie 1990.
11
See Metaph. A.1, 981b13–17. The first inventors were admired because they were thought to be wiser,
since they discovered something new, beyond the common perceptions. The object of their
discovery must have been some universal and explanatory principle. See Cambiano 2012: 34.
54 stasinos stavrianeas
that govern the production of health, while the second follows the opposite
course by producing the healthy bodily state. What differentiates the two
components is the intellectual grasp of the principles that govern the
artistic process (absent in the spontaneous restoration of health) and that
produce standardly beneficial results.12 If this is so, then the dependence of
crafts on knowledge of principles in the craftsman divorces the efficient
cause of artificial production from what it produces, and entails that
artifacts do not possess their principle per se or essentially (NE vi.5,
1140b6). Further, any other changes they can undergo qua products of
such a principle will not belong to them essentially, for these changes will
ultimately be explained by reference to the principle of their production.
To conclude, craft is primarily an intellectual capacity. The manual
work necessary for artificial production is not a sufficient condition for the
product to qualify as craft.13 Master craft, i.e. art proper, must include, and
is primarily identified with, universal and explanatory knowledge. In that
sense it is necessarily separate (a) from the material changes directed
towards the production of an artifact, and (b) from that particular artifact.
For the case of miraculous automata this means that even if some craft
could indeed produce artifacts that are able to effect their own changes,
including reproducing themselves, these automata would not possess the
principle of their changes in themselves per se, for their source is a principle
that ultimately lies in human intelligence.

III Differentiating nature as a principle from the efficient cause

Artifacts, it was proposed, do not possess a nature because their principle of
change is necessarily external. The principle of their matter does not belong
immediately to them. And the principle producing their form and func-
tions does not belong essentially, and need not belong immediately, to
them (except accidentally), since it is something external. This may suggest

12
Physics ii.2 distinguishes two kinds of craft that involve knowledge: (1) one that uses the product;
and (2) one that directs its production. (1) is directive in that it involves knowledge of the use and the
form, while (2) is directive in that it involves knowledge of how the matter can acquire the qualities
necessary for producing something appropriate for this use (194a36–b7). (2) also directs a further
component, (3), the executing side of the craft, the actual manual labor needed. On this, see
Pellegrin 2000: 1–26; cf. also Pol. iii.11, 1282a2–5. Metaphysics A.1 distinguishes (1) from (3), which
is relegated to the level of mere experience (empeiria) (980b5–12; b15–16), as lower-level manual
workers act mechanically, like inanimate tools (981b2–5).
13
This is confirmed by the fact that productive activity in other animals does not qualify as craft
because it is not the result of inquiry or deliberation: see Phys. ii.8, 199a20–21; cf. HA vii (viii).1,
588a29.
 Nature as a principle of change 55
again that in natural bodies the internal principle must be an internal
efficient cause of their changes. However, several changes of natural bodies
are caused by external efficient causes. For instance, the locomotion of
simple bodies is initiated by external causes, and simple bodies and their
mixtures suffer a number of natural qualitative or quantitative changes that
are triggered by external efficient causes. Furthermore, animate beings
seem to rely on external bodies for exercising some of their own natural
capacities, e.g. nutrition, perception, and locomotion, and are thus at least
in some sense moved by external efficient causes.14 Of course, one could
deny either (a) that the movements in question have per se and external
efficient causes, or (b) that they are natural, i.e. depend on the nature of the
being in question in an absolute sense.15 Neither of these options, though,
seems to fit with Aristotle’s text.16 A third option (c), which has recently
been defended by Sean Kelsey, would be to deny the claim that possessing
an internal efficient cause is necessary for possessing a nature that qualifies
as a principle of change. I believe that option (c) is the most promising one,
so let us explore Kelsey’s interpretation further.
Kelsey’s suggestion is to understand principle of change in terms of the
following definition offered in Metaphysics Δ.1 (1013a10–13): “that in accor-
dance with whose will what is moved is moved and what is changed changes.”
This definition of principle points to some kind of authority the principle has
over what it is a principle of. In the Politics we find two types of authority
relation between ruler and ruled, and Kelsey exploits them for distinguishing
nature from craft, in a way that can accommodate the claim that some
natural bodies are not efficient causes of their motion. The first authority
relation, which Kelsey labels “despotic,” is exemplified in the relation between
master and slave, while the second, the “non-despotic” one, is exemplified in
the relation between teacher and student (Pol. iii.6, 1278b30–1279a8). The
relevant difference between them concerns the location of the good that the
authority relation aims at, what Kelsey calls the “proper subject of the
good produced”: Despotic authority relations aim at the good of the ruler,

14
Physics viii.4 argues that elements are not self-movers. For animate beings, see viii.6, 259b6–20. See
Johansen 2012b: 128–145 for a detailed defense of the claim that soul is part of the efficient causal
story in nutrition, perception and locomotion in a way that does not threaten their possessing a
nature.
15
For the former view, see Waterlow 1982 and Graham 1999; for the latter, see Matthen 2001.
16
Aristotle’s claim that the movement of simple bodies is by nature and according to nature seems to
rule out (b) (Phys. ii.1, 193a1); the conclusion of Phys. viii.4, 255b33–34, that they require external
efficient causes of their movement, rules out (a). Alternative (c) is also followed by Scharle 2008:
171–172, who argues that elemental natures are principles of being changed and thus not principles as
an efficient cause.
56 stasinos stavrianeas
while non-despotic ones aim at the good of the ruled (Kelsey 2003: 76–77).
Thus, even though changes involving non-despotic authority relations
have an efficient cause, and this efficient cause is located in the ruler
(which is the active component of the change and which is external to
the proper subject of the change, namely the ruled or the passive compo-
nent of the change), the ruled is still the subject and source of the changes
undertaken, in the sense that the good of the relation is located in the
ruled. Hence, a body can suffer a natural change, e.g. an elemental body
can move naturally, just because the movement it suffers and the good
produced is located in it, without possessing the efficient cause of that
change internally.17 A sufficient condition, then, for possessing a nature is
to be the passive subject of a change undertaken in the context of a non-
despotic authority relation where the good is located on the passive side of
the change. In this way, Kelsey’s interpretation can accommodate natural
motions due to external efficient causes.
This proposal remedies the problem of natural changes that are not self-
motions. Further, by interpreting nature in terms of a relation (and thereby
classifying it in the category of relatives), it is easier to accommodate
changes that are due to nature and at the same time necessarily involve
external objects. For each of the capacities constituting the nature of a
body, whether active or passive ones, will be related to a counterpart
principle either in the material of the compound or to an external object.18
However, I would like to amend the proposal in two respects.
First, the proposal seems rather restrictive by tying natural change to the
subject of a good produced. There are natural changes that have external
efficient causes, especially in inanimate bodies, where the result is not some
good for the passive subject of the change. Rather, the result produced is a
quality in the compound, the passive capacity for which belongs to the
matter, but is ultimately determined by the form of the compound
(Meteor. iv.2, 379b25–32; Aristotle mentions tears becoming rheum as an
example). Aristotle explains that changes in inanimate bodies (such as
solidification or liquefaction) involve three causes: the matter, the efficient
cause, and the quality in the sense of form (iv.5, 382a28–30). The efficient

17
Kelsey (2003: 79) writes:
The suggestion then is that when Aristotle says that natural things have an internal principle of
movement, he means that they are the proper subjects of their movements; by contrast, to say
that artifacts have an external principle of movement would be to say that the proper subject of
their movements is precisely not they themselves but something else distinct from them.
18
On the classification of the natures of living beings under the category of relatives, see Johansen
2012b: 81–83.
 Nature as a principle of change 57
cause is associated with the powers of heat and cold, while the dry and the
moist are the passive powers that constitute composite bodies as matter,
either directly or through the primary bodies, earth and water (iv.4,
382a1–7; iv.5, 382a31–32). Indeed, the passive and active factors of a change
must belong to one genus, as the pairs wet–dry and heat–cold do (GC i.7,
323b31–32). True, we can also refer to what possesses these material proper-
ties, the compound constituted by the wet and dry, as the passive side of
the change (GC i.7, 324a15–21). Indeed, an affection of its material proper-
ties brings about a change in the compound: For instance, fiery heat affects
solid bodies, such as horn or iron, consisting of mostly earth and some
moisture, and makes them soft (Meteor. iv.9, 385b6–11). Softness (or
hardness) itself is a quality of the elemental material (GC ii.2, 329b8–12),
as well as of what this material constitutes. But the result of the change in
the compound is the realization of a dispositional property, which is itself a
characteristic differentia of bodies of its kind (Meteor. iv.4, 381b24–26; 8,
385a1–8), and the matter does not determine the limits within which the
dispositional property is expressed. Otherwise, there would be no limit to
the kind of change each body could suffer under the influence of heat or
cold. Furthermore, it is questionable whether the expression of such
dispositional properties is in every case something good for the subject.
For although Aristotle seems to hold that even inanimate compounds have
an end and function (Meteor. iv.12, 390a4–15), and this must be something
good, not every change will result in something that accords with such a
function. And yet such changes are due to the nature of the corresponding
body. Thus, inanimate bodies seem to possess a number of dispositional
properties as part of their formal natures, and such formal natures have
priority as principles of the corresponding changes. If so, the inanimate
body is the passive source of the change due to its material constitution, but
the result of the change ultimately depends on its form, and this form can
thus be identified as the internal principle.
Second, the proposal is also too generous. For it seems that being the
proper subject of change, in the intended sense, applies not only to subjects
of natural changes, but also to subjects of spontaneous changes. Typically,
changes that result spontaneously resemble natural ones. They bring about
an end-state, standardly achieved by an efficient cause aiming at that end-
state, but in their case the efficient cause was not aiming at this end. So, for
instance, the healing of a patient, standardly the result of the exercise of
medical craft, may also be the product of a spontaneous process caused by
environmental conditions. The difference between natural and sponta-
neous cases lies in the relation of the efficient cause producing that change
58 stasinos stavrianeas
and the good that results for the entity suffering the change. In sponta-
neous cases, the efficient cause and the end produced are related per
accidens, while in standard, natural cases, the relation is per se. Now, if an
entity suffers a natural change only in virtue of being the proper subject of
the good the change produces, then it is not clear how to differentiate
between causes that relate per se and causes that relate per accidens to the
good produced, and hence between standard and spontaneous cases, for in
both cases, the proper subject of the good produced is the entity that
constitutes the passive side of the change, e.g. a generated organism. Both
types of process, it seems, will count as natural.19
For an account of nature that can accommodate the above difference we
need to turn to Physics ii.1, 193a30–b18, where Aristotle gives three argu-
ments for the claim that nature must be identified with form rather than
matter. Aristotle there specifies three different aspects of forms, being an
actuality, an efficient, and a final cause, but we only need to focus on the
latter two, since they apply exclusively to the forms of living beings.
The second argument revisits the claim that nature must be identified
with what persists throughout generation and corruption. But the thought
implicit earlier, and equally present here, is that the nature of an entity is
the element common between it and what generated it, since its persistence
might be taken as evidence for locating the efficient cause of the change in
it. What is common between generating and generated animal is that they
belong to the same kind. And the nature of each kind is identified, sc.
193b3–4, with the corresponding form. So the nature of an animate being
includes its form in its role as an efficient cause in reproduction. However,
this nature acts as an efficient cause for generating a different specimen,
and thus it is an external principle of what is generated and no part of its
own nature. It seems, though, that the third argument qualifies this claim.
In this last argument Aristotle contrasts changes produced by nature with
those produced by craft. The latter do not lead to the craft itself but to a
different form in the subject suffering that change, e.g. medicine starts a
change that leads not to medicine but to the recovery of health. Natural
generations, by contrast, proceed to nature, and thus what grows should be
identified with what it will become, its form. The exact reference of
Aristotle’s mention of the thing that grows (phuomenon) is not certain,20

19
The proposal creates problems for other changes as well: For instance, any change of place that an
animal undergoes, where it is the proper subject of the good produced by that change, e.g. Daedalus
flying from Crete to Icaria, would qualify as natural.
20
The word is used primarily for plants (e.g. DA ii.1, 413a25, GA i.23, 731a8, and HA viii.19, 601b12),
but could also refer to animal fetuses (e.g. GA ii.3, 736a34–35).
 Nature as a principle of change 59
but Metaph. Δ.4, 1014b20–22 informs us of the following: Those things are
said to grow that derive augmentation from something else by contact and
organic unity, or by organic adhesion, as in the case of embryos. It is thus
plausible that Aristotle here has in mind animal embryos formed out of
semen and catamenia (as well as the seeds of plants). What defines organic
unities is that there is something common to both parts of the unity that
makes them grow together and makes them become one with respect to
continuity (Metaph. Δ.4, 1014b23–25, cf. Phys. v.3, 227a10–16).
If this is the stage of generation Aristotle has in mind here, it is one he
touches upon in his detailed account in Generation of Animals, where we
are told that seeds and fetations not yet separated from the parent possess
nutritive soul only potentially. They do so actually when they start drawing
nourishment to themselves (ii.3, 736b8–10), and this happens once their
first principle has become distinct, in the form of the heart in blooded
animals or in the form of shoot and root in plants (ii.4, 739b34–740a4).
Previously, the growing thing was only potentially separate, while actually
it was a unity with the parent, as the comment on growth from Metaphysics
Δ suggests. But now that comment is connected with the following sense of
nature: “the source from which the primary movement in each natural
object is present in it in virtue of itself” (1014b18–20). If the embryo that is
not yet actually separated is a unity with what generates it, and, addition-
ally, if the principle of its primary movement belongs to what generates it
in virtue of itself, it follows that the principle must, somehow, belong to
the embryo. Further, the principle of this movement is, by the lights of the
second argument, the form in the generating substance. So the embryo has
as its principle a form that is also the form of the unity in which it
participates. This should not lead us to the thought that the growing
thing possesses its own internal efficient cause. It indicates, though, that
by forming a unity with what generates it, the embryo shares in the form of
that entity. And in that sense it, arguably, possesses an internal principle of
change. To conclude, although the third argument aims to establish that
the growing thing’s nature is its final cause, the association of the growing
thing in conception with its efficient cause as the form in the parent also
suggests, together with the second argument, why the growing thing’s
nature is its end, and also why it is always an internal principle that
determines the growing thing.21
21
See PA i.1, 641b32–35 (and Lennox 2001b comments ad loc.) on a similar account of why the end can
be called a nature, building on the dual nature of the seed involved in natural generation: “for the
seed is a seed in two ways, as that from which and that of which; that is, it is a seed both of what it
came from . . . and it is a seed of what will be from it.”
60 stasinos stavrianeas
Therefore, in order to differentiate natural from spontaneous genera-
tion, we must require that the parent’s form effecting the change is shared
by the growing thing in the relevant way. They are not mere passive
subjects of the change, but they share the same form with that which acts
as their efficient cause. To be sure, even if we amend the proposal in this
way, the proposal results in a merely sufficient condition for possessing a
nature, not in a necessary one. For, as argued earlier, inanimate beings are
not, at least not universally, efficient causes of the changes they undergo.
From this we should conclude, not that for all natural beings efficient
causation must be excluded from our understanding of their principle of
change, but rather that nature must be understood non-uniformly. The
remaining question, of course, is whether, despite their differences, the
principles in question share sufficient similarity, such that they can legiti-
mately be grouped together under a general conception of what it means to
possess a nature. The most difficult case our interpretation needs to
accommodate as a case of natural motion is the motion of the elements.

IV The natural motion of the four elements and their mixtures

Aristotle denies that elements are self-movers, and holds that their local
motion has an external efficient cause (i.e. the active power that moves
them is always external), while they themselves possess merely a passive
principle of change (Phys. vii.4, 255a6–18; 256a1–2).22 Despite this, he
holds that their motion is natural, and this leads to the following aporia
(viii.4, 254b33): The elements move naturally in one direction on the axis
of up and down, and yet do not qualify as self-movers.
The resolution of this aporia comes in two stages. The first introduces
a distinction between causes, the second a distinction between senses of
potentiality. At the first stage, Aristotle suggests that, as movers may
move naturally or move unnaturally, so with things moved, in some cases
they are moved naturally, while in others they do so unnaturally. We
should ask, though: What causes does Aristotle have in mind here?
The point about natural versus unnatural movers concerns efficient
causation; but the point about naturally versus unnaturally movable
things concerns, it seems, formal causation: Movables move naturally
when they are “potentially in their own actuality” (viii.5, 255a24–30).

22
The interpretation offered here is indebted to Bodnár 1997 and Gill 2009. For other interpretations,
see Matthen 2001, Scharle 2008, Katayama 2011. I offer a more elaborate defense of my reading in
Stavrianeas, “The nature of the elements” (in press).
 Nature as a principle of change 61
This potentiality cannot be their potentiality to move (intransitive), for
they are actually moving. Rather, it is a potentiality that belongs to
them while they are moving to their proper places. And to be in their
proper place is – at least part of – their form. However, even if we grant
that the actuality they are potentially in is their principle in the sense of
form, each moving portion is not in its proper place yet. So in what
sense does it possess this principle already? The second stage of the
resolution may be read as addressing this point, as Aristotle suggests that
in order to grasp what moves the elements, we need to specify the
particular sense in which the potentiality/actuality distinction applies
to their movement:
(1) One who is learning a science knows potentially in a different way from
(2) one who while already possessing the knowledge is not actually exercis-
ing it. Wherever (3) something capable of acting and something capable of
being acted on are together, what is potential becomes actual, e.g. the learner
becomes from one potential something another potential something (for
one who possesses knowledge of a science but is not actually exercising it
knows the science potentially in a sense, though not in the same sense as
before he learnt it). And (4) when he is in this condition, if something does
not prevent him, he actively exercises his knowledge: otherwise he would be
in the contradictory state of not knowing. (Phys. viii.4, 255a33–b5)
Sections (1) and (2) divide two senses of potentiality, while (3) and
(4) two corresponding senses of actuality. For instance, the first-level
potentiality to become hot (1) belongs to a portion of an element that is
not hot, but cold, such as earth. Once a cold portion is transformed into
something hot, such as fire, the first-level potentiality becomes actual
(3), and from then on the generated portion of fire possesses the
actuality of being hot. By the same token it also possesses a second-
level potentiality (2) to heat, if not impeded. The exercise of this
second-level potentiality constitutes a second-level actuality, on a par
with the exercise of the power to theorize (4), i.e. with the exercise of
a capacity (255b6–7). It is in this latter sense that the movable is
potentially in its actuality, so let us see how this schema applies to
elemental motion and resolves the aporia.
The power of fire to be light, i.e. to be up, belongs as a first-level
potentiality (1) to something that is actually heavy, e.g. earth. Once the
latter is transformed into something light, i.e. earth into fire, this potenti-
ality is fulfilled and the element must, according to this schema, possess a
capacity (2) that is similar to its capacity to heat. However, the two
capacities behave differently with respect to their transition to their
62 stasinos stavrianeas
respective actualities (4).23 The capacity of fire to rise higher whenever
exercised includes a local change and resembles the actuality of a first-level
potential, as potential, i.e. as something incomplete, while the exercise of
its capacity to heat does not involve such a change. In the light of this
difference, it remains puzzling why Aristotle classifies the exercise of the
capacity for locomotion as a transition from a second-level potentiality to
the corresponding actuality (a transition that standardly does not involve
change), rather than to the transition from a first-level potentiality.
Part of the answer is found in what differentiates the two levels of
potentiality–actuality pairs: The learner becomes from one potential some-
thing another potential something, while the knower will exercise his
capacity, if nothing impedes, i.e. whenever he wills to exercise it, for other-
wise he would be in the contradictory state of not knowing. Transitions at
the first level consist in moving from one state to another state (possibly
its contradictory state).24 Now, in one sense, this is true of elemental
locomotion, which is a change to the natural place from its contrary one.
However, unlike other local changes – say, a stone thrown upwards –
movement to the natural place is not accidental to the elements. Rather, it
is a constant and regular feature, and it expresses what each element is.25 If
this change, then, consisted in the destruction of a contrary, its result would
be the annihilation of what is essential to the moving entity. Far from this
being the case, the element in moving to its proper place remains unchanged
and gets united with its like. So its potentiality to be in its proper place by
becoming active preserves its nature by bringing it closer to its form. In this
respect, the locomotion of the elements resembles cases at the second level of
the actuality–potentiality distinction.26 Therefore, an activity of this kind
may be attributed to a principle that something possesses in itself per se, even
if it has an external efficient cause.27
23
Indeed, the distinction concerning causes in the first stage of the resolution of the aporia prepares us
to expect just that: The capacity of fire to heat is a capacity of fire as a natural mover, while its
capacity to move up is a capacity of fire to be moved naturally.
24 25
Cf. DA ii.6, 417a20ff. See Gill 2009: 151 on this.
26
On the other hand, the fact that this motion is an activity of some kind justifies, to some extent, the
thought that it is not a mere capacity for a corresponding actuality or activity, but that it is rather
itself an actuality.
27
This is confirmed by Aristotle’s illustration in On the Heavens: “to ask why fire moves upward and
earth downward is the same as to ask why the healable, when moved and changed qua healable,
attains health and not whiteness” (Cael. iv.3, 310b15–17; cf. 310b27–30). The healable that is already
moved qua healable, i.e. as something that has the potentiality to be healthy, is already on its way to
being healthy, in the sense that it already possesses the form of health (Metaph. Z.7, 1032b25–27; see
also DA ii.2, 414a10). Similarly, the fire moving up is partly realizing its form in moving towards it,
as it occupies successively different locations that are gradually nearer to its proper place, even
though none of them is its proper place.
 Nature as a principle of change 63
And there is a further reason for holding that elemental motion is similar
to second-level activities and to the exercise of natural capacities. This
becomes clear if we consider the qualification Aristotle adds concerning
such capacities, namely that they are active if not impeded, and its applica-
tion to the respective capacities of fire to heat and to be up. All that is
needed for impeding the capacity of fire to heat is that no suitable subject is
found in its surroundings; then its capacity is inactive. In any other
situation, fire is active and heats the proximate bodies. No change in the
fire is involved in its transition from inactivity to its exercise. The capacity
of fire to be up, on the other hand, is impeded, we are told, whenever fire is
not in its proper place. This may be so either because it is held there by an
external force, or simply because it was generated there. So as soon as the
impediment is removed or the portion generated, the corresponding
capacity is active. True, one may be tempted to think that the portion is
in some sense impeded while still moving, but this is not quite the sense in
which it is impeded when it is held immobile by force. Further, it does not
seem right to say that fire is inactive when moving to its proper place in the
way the capacity of fire to heat may be inactive. In the former case it rises
higher, while in the latter there is no heating going on.28 What is common
in the two cases, and similarly to the exercise of other natural capacities, is
that there is no internal obstruction that can form an obstacle to their
exercise; no process within the element can stop its capacity from being
constantly active. In other words, Aristotle’s qualification “if nothing
impedes” should be understood as “if nothing external impedes.” And
this seems to be a standard feature of capacities.29 To be sure, natural
capacities in animate beings form part of a complex network of potential-
ities whose exercises may interfere with each other. Moreover, they belong
to ensouled bodies, where a formal and a material side may be antagonizing
each other. Regardless, Aristotle’s point is that, considered on its own, each
capacity is internally indefeasible. No impediment can come from a nor-
mally functioning capacity such as to block its being exercised. This

28
Indeed, Aristotle concludes the aporia by saying that the portion of fire is already light and will
realize its proper activity as soon as it is generated (viii.5, 255b10–11). And he states as much a few
lines below: “if what hinders it [that is, the portion of air] is removed, it realizes its activity and
continues to rise higher” (255b20–21).
29
The exercise of a capacity seems to be compromised only by something external – see Metaph. Θ.7,
1049a7; 1049a13–14; and MA 8, 702a17. This means that when the external conditions are right, the
capacity itself is sufficient for explaining the changes its exercise will be involved in. Thus, it makes
sense to say that what possesses such a capacity possesses in itself a principle for the corresponding
changes, which can be understood as its internal principle of change, its nature.
64 stasinos stavrianeas
feature, then, offers one more reason to view elemental motion as governed
by an internal principle.
At this point, however, the similarity between capacities in animate and
inanimate beings breaks down. The former are capacities for complex
activities, while the latter are for simple and uniform changes. The transi-
tion to activity does not depend on a complex process involving changes in
other material beings (such as in the activity of nourishing oneself) or in
parts of the entity itself (such as in the activity of moving one’s own body
locally). We could qualify such activities as homogeneous or non-plastic. For
this reason, too, it may be that a passive principle is sufficient for governing
elemental motion, as well as changes of inanimate beings in other
categories.30
In sum, elemental motion is crucially similar to transitions from capa-
cities to their corresponding activities. It is not a standard local change, but
part of the realization of elemental formal nature. Since, as we just saw, this
capacity differs from capacities in animate beings by being active from the
start, its per se efficient cause must be something external.31 However, this
does not make it a capacity to be changed by something else, for it requires
no external stimulus for becoming active, nor is it a capacity to be changed
by itself qua other, due to the simplicity of elemental body. What it is is a
capacity to be changed in itself qua itself, and hence it satisfies in one way
Physics ii.1’s definition of nature.

V Conclusion
We may conclude, then, that we should not insist on an exclusive under-
standing of what a principle of change is, or on constraining our inter-
pretation of nature as a principle to the role of efficient causation in all
natural phenomena. The absence of an internal efficient cause does not
disqualify a change as natural, as in the case of elemental locomotion, nor
does its presence alone qualify a change as natural, as in the case of
miraculous automata. Rather, what matters are the specific capacities
that characterize natural kinds and whose exercise regularly produces
changes that preserve the nature of the entities of those kinds. In these
changes, entities can play active or passive roles in conformity to their
form.

30
Note that the model applied to the motion of the elements holds similarly of their qualitative and
quantitative alterations: Phys. viii.5, 255b12–13.
31
Simplicius In Ar. Phys. 1220.20.
 Nature as a principle of change 65
With respect to Physics ii.1, then, it seems neither necessary nor desirable
to read it as proposing a complete and finished definition of what nature is.
Rather, Aristotle offers up an initial account of what differentiates natural
beings as a first conception or general description, while providing further
developments and qualifications in the remaining, more specific, natural
treatises concerning what is distinctive of each natural kind. The main aim
of Physics ii.1 is to delineate the domain that natural bodies populate and
that natural science investigates.
 chapter 4

Aristotle on chance as an accidental cause

James Allen

I
“Chance” stands in for the genus designated αὐτόματον by Aristotle,
which he divides into two species, τύχη, luck or fortune, which is operative
in the realm of practice, and its counterpart in the natural world, the
automatic or spontaneous, called, like the genus, αὐτόματον. The termi-
nological distinction is stipulative and corresponds neither to Aristotle’s
own usage nor that of other authors. He defines luck and the automatic as
causes by accident, and of the efficient type (Phys. ii.6 198a2–7; cf. ii.5,
196b23, 30, 197a5–6, 12–14; ii.8, 199b23–25). The definition and the
distinction both come from his fullest discussion of chance, which occupies
chapters 4–6 of Physics ii.
He has good reasons for tackling the subject here, in the treatise he
dedicates to fundamental questions of natural philosophy. Chance is
accounted a cause, in the sense that people appeal to it to explain what
happens, not only in the practical sphere, but in the natural world as well.
His point of departure in Physics ii.4 is the question whether and how
chance can be accommodated in the framework of the four causes
(ii.4, 195b31–36; cf. EE viii.2, 1247b1–9). Chance is also regarded – rightly,
in Aristotle’s view – as indefinite (ἀόριστος), unpredictable (παράλογος),
obscure (ἄδηλος), and uncertain (ἀβέβαιος),1 and he aims to supply an
account that will show how chance occurrences can have causes while
retaining the distinctive character that sets them apart from their non-
chance counterparts. His other main object is to combat on the basis of this
account the errors of the natural philosophers who would exclude teleology
from physics, replacing explanation in terms of final causes with appeals to
chance. And, although he touches on it only briefly in the Physics, Aristotle

1
παράλογος (197a18; EE vii.14, 1247a33); ἀόριστος (199a20; EE vii.14, 1247b12; Metaph. D.30,
1025a25; E.2, 1027a6); ἄδηλος ἀνθρώπῳ (197a10); ἄδηλος (Metaph. K.8, 1065a32); ἀβέβαιος
(197a30–31).

66
 Aristotle on chance as an accidental cause 67
expects that his account will also undermine another widely held view,
namely that lucky occurrences are the effects of unseen divine agency,
so that they are mysterious only from a limited human perspective
(ii.4, 196b5).2
Of course, if chance did not exist, there would be no problem to solve,
but Aristotle has little patience with the well-known argument that there is
no such thing as chance because it is always possible to specify a definite
cause for everything that has happened (196a5–6). He observes that,
though familiar to everyone, this argument has not prevented anyone
from regarding some things as due to chance, others as not (196a11–17).
His brisk way with this argument is a clue that an issue that one might have
expected to loom large in an examination of chance will receive little
attention. On a certain conception of chance, i.e. roughly what we mean
when we speak of “objective chance,” events are by chance because they are
not determined. No specification of the conditions in which they occur,
however complete, is incompatible with their failure to occur, so that, one
might suppose, there is no cause because of which they occur when they do
(as opposed perhaps to an explanation of the conditions that make it
possible or likely for them to occur). The argument that “to be caused” is
thereby “not to be by chance,” with the converse implication that “to be by
chance” is “to lack a cause,” suggests that this idea was not inconceivable,
but Aristotle nowhere confronts it directly.
One reason is because Aristotle is taking sides in a controversy that was
already under way, and – for all the differences between them – on the side
of Plato.3 In this controversy the opposed alternatives are not by chance or
by necessity, but by chance or for the sake of something; chance and necessity
can even be combined.4 Their adversaries are philosophers who would
explain the beneficial arrangements in which the natural world abounds as
the result of chance, which they present as an alternative to the form of
explanation favored by Aristotle’s party, according to whom it is for the
sake of the good they do or make possible that matters are so arranged.5
One opponent identified by Aristotle is Empedocles, who, he says, held

2
It receives more attention in EE viii.2, where, in the course of discussing the popular view that
happiness is good fortune (εὐτυχία), Aristotle rejects the idea that lucky people are the beneficiaries of
divine favor (1247a23–29). Epicurus too treats this view as widespread, maintaining that the person
who has taken his philosophy to heart will not believe, as the many do, that luck is a goddess
(Men. 134). It is credited to the Pythagoreans by Aristoxenus (DK i 478, 23–33).
3
And Socrates, on whom see Sedley 2007.
4
At Int. 9, 18b5–6, Aristotle says, without mentioning causation, that if everything is by necessity,
nothing is by chance.
5
On beneficial arrangements in nature, see Cooper 1982.
68 james allen
that animal organs are mostly products of chance (Phys. ii.4, 196a23–24; cf.
ii.8, 198b23–32). Others, whom he does not name, have, he says, made
chance the cause of both this and other worlds (ii.4, 196a24ff.). They
resemble the “impious” people, likewise unnamed, against whom Plato has
the Athenian argue in a passage in Laws x that prepares the way for a proof
of the gods’ existence (888eff.). They maintain that the greatest and most
beautiful things in the world, including the whole heaven, everything in it,
animals, plants, and seasons, are products of chance and nature (889a), and
that they arise in accordance with chance from necessity (κατὰ τύχην ἐξ
ἀνάγκης). By appealing to chance and nature, they deny that these things
are products of mind (νοῦς), god (θέος), or art (τέχνη) (889c). According to
Plato’s Athenian, they have inverted prior and posterior; it is the soul
(ψύχη), its faculties, and motions that are prior, while nature – if by this are
meant bodies, their powers, and motions – is posterior. But the way in
which those who couple nature with chance conceive of nature is mis-
guided. The Athenian’s opponents want to describe as nature the genesis of
the primary beings, but this honor, he maintains, belongs rather to soul,
which can best and most correctly be said to be by nature (892c).
As it figures in this controversy, then, chance is the random factor
which, by bringing together bodies in the right conditions and propor-
tions, supplies the initial conditions that allow them, acting in accordance
with the necessities imposed by their material natures, to give rise (among
other things) to the greatest and most beautiful of existing things – a kind
of cosmic luck. The question whether the random element introduced by
chance is due to indeterminism or simply has (determining) causes that are
beyond the scope and ambition of natural philosophy appears not to arise
in the controversy, in which Plato and Aristotle’s object is simply to prove
that chance, conceived as the not-further-to-be-explained source of the
material combinations that give rise to beneficial arrangements in nature,
cannot replace teleology.
To be sure, Aristotle and Plato are on the same side only with substantial
qualifications. Plato’s preferred account of the natural world assigns prior-
ity to soul or mind, ultimately divine mind. Aristotle’s god neither creates
nor administers the world. The things that are and come to be for the sake
of something in the realm of nature do so without the benefit of delibera-
tion or design. What is more, the intended scope of final causal explanation
in Aristotelian natural philosophy remains a matter of controversy.
Nonetheless, on the fundamental point that at least some natural beings
and their doings are for the sake of something, he agrees with Plato.
 Aristotle on chance as an accidental cause 69
Although Aristotle parts ways with his ally when it comes to the priority
of art and mind, his account of chance pays at least as much attention to
luck and the practical sphere, where mind is operative and which he tackles
first in Physics ii.5, as it does to the automatic, which he takes up in ii.6.
This may be an instance of proceeding from what is prior and more
familiar to us to what is less so, but there is a risk in adapting an account
of luck tailored to the realm of agency, where it is easy to contrast beneficial
occurrences from which the for the sake of something is absent with the
those in which it is present. On my interpretation, Aristotle regards chance
occurrences as something like byproducts of final causation, which are only
intelligible as exceptions to the teleological rule. As a result, his account of
chance threatens to beg the question in dispute between him and his
predecessors, according to whom nothing in nature ever occurs for the
sake of anything. This is especially plain in Aristotle’s own argument from
priority, with which his account ends and in which it seems it was meant to
culminate.
But since the automatic and luck are causes of things of which mind or
nature might be causes, when something comes to be as a cause of these
same things per accidens, and since nothing that is per accidens is prior to that
which is per se, it is clear that no cause per accidens is prior to that which is a
cause per se. Hence the automatic and luck are posterior to both mind and
nature; so however much the automatic may be the cause of the heavens,
mind and nature are necessarily prior causes both of many other things and
of this universe. (198a5–13; trans. Charlton modified)
It is clear from the broader context that, unlike those who couple nature
with chance, Aristotle takes nature, which he puts on a level with mind
here, to be, like it, operating for the sake of an end. The burden of his
argument is that automatic outcomes are possible in nature only if it
contains occurrences that are for the sake of their results as well.6
In what follows I shall be chiefly concerned to examine Aristotle’s
account of chance, on the basis of which he takes himself to be entitled
to the argument for the priority of per se causation for the sake of some-
thing and the consequent impossibility of chance causation per accidens
apart from it. But I shall conclude by returning briefly to the question
whether Aristotle begs the question against his opponents. I conclude that
he does, but that he implicitly acknowledges this when he returns to

6
This is how the passage is understood when it is summarized in Metaphysics K.8 (1065b26ff.) and by a
number of commentators. Aristotle has already made this view explicit in chapter 2 and will do so
again in chapter 7 (194a25ff., 198b4). Cf. M. R. Johnson 2005: 102, 104; Sedley 2007: 192–194.
70 james allen
confront their appeal to chance as an alternative to final causal explanation
in Physics ii.8, whose argument does not beg the same question.

II
Aristotle begins his treatment of luck in ii.5 by drawing attention to two
features, both of which belong likewise to the automatic (196b10–22).
1) Luck is not a cause of things that come to be always or for the most
part, but is, in a sense still to be specified, infrequent.
2) It is actions done for the sake of something that have lucky outcomes.7
These features are presumably intended to illuminate the notion of accident
as it applies to luck. Though not the only available interpretative strategies,
two opposed approaches can be distinguished by which of the two features
they emphasize. One takes being an accident, or an accident of the relevant
kind, to be at bottom a matter of the (low) relative frequency with which
the initiating event, under some specification, is followed by its lucky or
automatic outcome.8 In my view it is the second feature that will help us get a
better fix on how the notion of accident applies to chance, while the
infrequency condition is best viewed as an additional constraint on chance
accidents, whose more fundamental nature is specified in other terms.
The second feature has proved to be a puzzle, however. Both in Physics ii
and elsewhere, not coming about for the sake of what results is said to be a
mark of luck or chance.9 For example, in a passage from Posterior Analytics
ii.11 to which we shall return, Aristotle roundly asserts that “nothing that is
by chance comes to be for the sake of something”:10
Among things that come to be from thought some never arise from chance
or from necessity, e.g. a house or a statue, but for the sake of something,
while others such as health or safety also arise from luck . . . but nothing by
chance comes to be for the sake of something. (95a3–6)

7
Note that both outcomes and the actions that bring them about, i.e. what it was lucky that one did,
are said to be “by luck” (cf. Phys. ii.8, 199b18–22).
8
Judson 1991b: 81, 90 takes Aristotle to hold that the conditions that chance events come to be
accidentally and that they do rarely are the “same” or “equivalent.” My disagreement has not
prevented me from profiting from his article, especially his account of how the notion of the “for the
most part” is brought to bear on chance (92–93).
9
Phys. ii.5, 196b34, 197a16; ii.6, 197b19, 23; ii.8, 199b21–22; APo. ii.11, 95a5–9; Rh. i.10, 1369a32–34.
Protrepticus fr.11 Ross. Cf. Lennox 1984: 250–251 and Judson 1991b: 77 with n.8.
10
ἐν δὲ τοῖς ἀπὸ διανοίας τὰ μὲν οὐδέποτε ἀπὸ τοῦ αὐτομάτου ὑπάρχει, οἷον οἰκία ἢ ἀνδριάς, οὐδ’ ἐξ
ἀνάγκης, ἀλλ’ ἕνεκά του, τὰ δὲ καὶ ἀπὸ τύχης, οἷον ὑγiεια καὶ σωτηρία . . . ἀπὸ τύχης δ’ οὐδὲν
ἕνεκά του γίνεται.
 Aristotle on chance as an accidental cause 71
Yet Aristotle places chance occurrences squarely among those that are for
the sake of something both at the start of and throughout the account in
Physics ii.
Anything which might be done as the outcome of thought or nature is for
the sake of something. Whenever something like this comes to be by
accident, we say that is by luck. (196b17–24)
A little further on, in connection with his favorite example of a man who
went to the market with another end in view but succeeded in collecting a
debt by luck, Aristotle says:
Plainly then luck is a cause by accident among those things that are for the
sake of something that are according to choice. (197a5–6; cf. 5, 196b29–31)
And other passages say much the same thing (cf. 196b23–24, 32–33, 197a6,
197b18). Since the philosophers against whom Aristotle is chiefly concerned
to argue, at least as he understands them, maintain that nothing in nature
comes to be for the sake of anything and propose coming to be by chance as
an alternative to coming to be for the sake of something, we should expect
the fact, which Aristotle takes it to be, that everything that happens by
chance is – somehow – for the sake of something to be more than a merely
verbal coup: It should deliver a real sting.One solution that is favored by a
number of ancient commentators and modern scholars holds that a lucky
occurrence must be for the sake of something, to be sure, but something
other than the outcome by bringing about which it qualifies as lucky.11
That is, for the action A whose lucky outcome is X, there is a Y not identical
with X for whose sake A was done. Although this may be true according to
Aristotle, it is not his point, which is rather that the action is somehow for
the sake of the effect by bringing about which it qualifies as lucky. Consider
a passage in Physics ii.8 that looks back to the doctrine of chapters 4–6.
That for whose sake and what is for the sake of this may come about by luck;
for example, we say that a stranger came by luck, paid the ransom and went
away, when he did so as if he had come for the sake of this though he did not
come for the sake of this – and this by accident. (199b18–22)12

11
Porphyry reported by Simplicius In Ar. Phys. 336, 28–29; Philoponus In Ar. Phys. 268, 30ff., 274,
10ff.; Themistius In Ar. Phys. 51, 16–20, 52, 10ff.; Bonitz 1969: 58; Torstrik 1875: 429, 445–448, 466,
468; Zeller 1879: ii.2 335–336, 947–948. But cf. Irwin 1990: 520–521, who thinks the solution works
for luck but not the automatic.
12
τὸ δὲ οὗ ἕνεκα, καὶ ὃ τούτου ἕνεκα, γένοιτο ἂν καὶ ἀπὸ τύχης, οἷον λέγομεν ὅτι ἀπὸ τύχης ἦλθεν ὁ
ξένος καὶ λυσάμενος ἀπῆλθεν, ὅταν ὥσπερ ἕνεκα τούτου ἐλθὼν πράξῃ, μὴ ἕνεκα δὲ τούτου ἔλθῃ. καὶ
τοῦτο κατὰ συμβεβηκός . . .
72 james allen
Here paying the ransom is “that for whose sake” and coming to the
marketplace is “what is for the sake of this,” even though the stranger did
not come to the marketplace for the sake of paying the ransom. When
Aristotle takes lucky actions to be actions for the sake of something, then,
he means that every lucky action A whose lucky outcome is X belongs
among actions that are, somehow, done for the sake of X.
The more recent tendency has been to distinguish senses of “for the sake
of X,” so that a chance event may be for the sake of X in one sense, but fail
to be for its sake in another. This is on the right lines so far as it goes, but I
shall argue that one of the ways of distinguishing senses goes astray from
the outset, while the other can be better grounded and pursued further.
The proposal with which I disagree takes the sense of “for the sake of X” in
which an event comes to be for the sake of its automatic or lucky outcome
to be that it actually has that outcome.13 This seems to me less like another
sense of the expression than its reference or extension in that sense,
however. According to the second proposal, the other sense of “for the
sake of X” is something like “such as might have come to be for the sake
of X.”14 This is better, but it is still misleading, I maintain, to speak of
“senses” of the expression “for the sake of something.” Rather, I suggest,
the difference in meaning that Aristotle exploits here is, at bottom, a
difference in verbal aspect.
As we know, in Greek and English, among other languages, the
simple present is used for purposes other than saying that something
is going on at the time of the speaker’s utterance. In English, where one
typically uses the present progressive to do this, it would often simply be
incorrect. “Paul drives to work” means that Paul drives to work usually
or habitually. But “Paul drives to work now” is incorrect, unless “now”
means something like these days. One says “Paul is driving to work
now.” But the simple present is used not only to represent an action as
habitual in this way, but also to express another kind of general propo-
sition as well. If, on a hike in the woods, one of us idly observes to the
other that “trees burn,” the effect should be different not only from the
one elicited by “trees are burning,” but also from the response to an
assertion that trees habitually burn. Statements like these are sometimes

13
Judson 1991: 77 with n.11, who adopts this interpretation, calls this the non-explanatory sense. He
takes himself to disagree with Ross 1936, but it is not clear that Ross, who believes that, in the relevant
sense, heneka tou means actually having a result that is end-like (39, 517, 518, 520, 523), is not of the
same opinion. Cf., however, Ross 1924: ii 187 (ad Metaph. Z.7, 1032a29).
14
Cf. Simplicius In Ar. Phys. 335, 20–26 and 336, 27–30. Cf. Lennox 1984: 251–252.
 Aristotle on chance as an accidental cause 73
called generic propositions.15 This way of speaking is not confined to
assertions about independently specified subjects. It can also be used to
specify a subject with the aid of a relative phrase: e.g. “things which
burn” or “things that fly” do not mean things that are burning or are
flying (or even do so usually).
This use of the present tense, which is familiar to us in propositions
about substances and natural kinds, is not confined to them. Teachers of
philosophy should have no difficulty imagining a situation in which it
would be appropriate to say “propositions are true and arguments valid” or
“the things that are valid are arguments, not propositions,” although there
is, of course, no dearth of invalid arguments or false propositions. It is also
available for use in assertions about actions and occurrences. Consider
statements like the following, which would be out of place in an ordinary
conversation. “Paul did something that is done in order to hurt others (or
alternatively ‘that people do,’ ‘that one does’), but he did not do it in order
to hurt anyone.” Plainly we are talking about an agent, a particular action
of his, and a kind of action to which the action belongs. What is asserted is,
roughly, that Paul performed an action (or, if you prefer, an action of a
kind) that is such as to be done in order to hurt someone without doing it
in order to hurt anyone on this occasion. If different senses are in play here,
they are not different senses of “in order to hurt someone” but of “is done”
or the whole verbal phrase “be done to hurt someone.” The same action
can be (something that is) done to hurt someone, viewed as an instance of a
type, and not done in order to hurt someone, regarded in all its concrete
particularity, as done by this agent, on this occasion, out of these motives.16
It is a point of this kind, I suggest, that Aristotle is making about

15
Cf. Moravcsik 1994: 229–244. The terminology is somewhat fluid: Sometimes “generic” designates
the genus of which the habitual is a species.
16
“That for the sake of which” (τὸ οὗ ἕνεκα) is “one of the items in which or about which an action is”
(EN iii.1, 1110b33–a1), and ignorance regarding one or more of “the particular circumstances of the
action and the objects with which it is concerned” can render an action involuntary (1110b30–a2). All
interpreters agree that Aristotle is not describing a self-deceived agent ignorant of his own intentions.
To avoid this, some interpreters have taken τὸ οὗ ἕνεκα to mean the result or issue of the action in
three relevant passages from Aristotle’s ethical works (EN iii.1, 1111a2–19; v.8, 1135b12–16; EE ii.9,
1223b1–5). Gauthier and Jolif 1968: ii.1, 185 ad EN 1111a5; Judson 1991b: 77–78; and Stewart 1892: i 240.
This sense is otherwise unattested (except perhaps in the Physics) and goes unrecognized by lexico-
graphical authority. To judge by their translations and commentaries, others are unpersuaded: see
Crisp 2000: 39–40; Dirlmeier 1956; Rowe and Broadie 2002: 125; explicitly Grant 1866: ii 13n.18 and
Taylor 2006: 148. I suspect that the kind of ignorance Aristotle has in mind is about the actions by
means of which to bring about the agent’s end. His meaning would have been clearer had he written:
“the agent was in error regarding that for the sake of which in supposing that so acting was a way of
bringing it about.” The culprit would then be the extreme grammatical compression that results from
Aristotle’s determination to force all the possible kinds of error into a single telegraphic construction.
74 james allen
chance events, most explicitly when he summarizes his results in
chapter 6.
It is clear, then, that among the things that come to be for the sake of
something without qualification, when they come to be not for the sake of
what results, the cause being external, we then use the expression “by
chance” of them.17 (197b18–20)
The meaning of “without qualification” (ἁπλῶς) here has long been a
puzzle.18 I suggest that Aristotle is using the term in the well-attested sense
of “in abstraction from the particular qualifications or circumstances, or with
those circumstances or qualifications mentally removed” (cf. Top. ii.11,
115b29–35). He means occurrences or actions that one would say are for
the sake of X when they are viewed in abstraction from the particular
circumstances in which they are done or take place. A good parallel is
furnished by a passage from the discussion of the voluntary in
Nicomachean Ethics iii.1, where Aristotle describes actions like throwing
goods overboard as involuntary without qualification, but voluntary in
certain circumstances (1110a18, b5; cf. 1110a10). Translators have turned to
phrases like “in the abstract” to bring out the fact that it is actions that are not
worthy of choice when viewed in isolation from the circumstances in which
they are performed that are involuntary without qualification, even though
there are circumstances in which they are performed voluntarily.19 So, too,
actions and processes that come to be or are done for the sake of X, when
viewed in abstraction from the circumstances in which they actually occur,
can take place without doing so for the sake of X. When they do, Aristotle is
saying, and all goes well, they and their results come about by chance.
Aristotle sometimes conveys the idea of things that come to be or are
done for the sake of something in the abstract by means of another
construction, in which the adverbial phrase “for the sake of something”
is joined to the definite article, which has the effect of turning it into an
adjective and allows the whole phrase to serve as a substantive (196b35–36,

17
ὥστε φανερὸν ὅτι ἐν τοῖς ἁπλῶς ἕνεκά του γιγνομένοις, ὅταν μὴ τοῦ συμβάντος ἕνεκα γένηται ὧν
ἔξω τὸ αἴτιον, τότε ἀπὸ τοῦ αὐτομάτου λέγομεν. The point of calling the cause “external” is a
puzzle. See Judson 1991b: 94 with n.52.
18
The commentators tend to treat ἁπλῶς as equivalent to καθόλου (universally) and take Aristotle to
be talking about things that come about for the sake of something both in the sphere of luck and in
that of the automatic (Philoponus In Ar. Phys. 288, 17–20; probably Themistius In Ar. Phys. 55, 8–9;
Ross 1936: 522). Judson (1991b: 93) takes ἁπλῶς to mean “in the most general sense” and this to be
the non-explanatory sense, which means actually having the result in question.
19
“Voluntary in the abstract” (Ross); “an sich betrachtet” (Dirlmeier 1956); “abstraction faite de telles
circonstances” (Gauthier and Jolif 1968). Cf. Irwin and Fine 1995: 111n.91, who connect the use of
ἁπλῶς here with that in Nichomachean Ethics iii.1.
 Aristotle on chance as an accidental cause 75
197a6; cf. 199b19). In one, admittedly difficult and much disputed, passage
in chapter 6, we find “when what is for the sake of another thing does not
come to be for the sake of this” (ὅταν μὴ γένηται τὸ ἕνεκα ἄλλου ἐκείνου
ἕνεκα), which, I suggest, is once again meant to express the idea that chance
occurs when an occurrence that takes place for the sake of something in the
abstract takes place without on this occasion doing so for its sake
(197b23).20 Although there is no verb, it is natural to understand one,
and when translating to supply it, as I have done in the less controversial
passage that I already cited from Posterior Analytics:
Among things that are from thought some never arise from chance or from
necessity, e.g., a house or a statue, but for the sake of something, while
others such as health or safety also arise from luck . . . but nothing by chance
comes to be for the sake of something. (APo. ii.11, 995a3–6)
Here, being for the sake of something is a characteristic said to be absent
from every episode of chance. Yet some things that come about from
thought – the Greek simply has the definite article joined to “from
thought” – also come about by chance.21 If we assume, as we are surely
entitled to do, that on every occasion when something comes about from
thought in the sense intended, it comes about as the result of an action
undertaken for its sake, then it might appear that we are faced with a
contradiction. Yet I doubt that readers have ever been troubled by this
passage (cf. Metaph. Z.9, 1034a9ff.). It should now be plain that when
Aristotle classes chance occurrences among “things that come to be for the
sake of something” (196b17) he does not mean that individual episodes or
particular occurrences each take place for the sake of something.

III
How does the idea that lucky agents do something that is done for the sake
of an end (in the abstract) without doing it for the sake of that end

20
The context is an obscure etymological argument that turns on the phonetic similarity between
μάτην, “in vain,” and αὐτόματον. Charlton 1970: 48, followed by Judson 1991b: 77n.9, emends
196b34–36 to yield another example: ἦλθε δ’ οὐ τούτου ἕνεκα, ἀλλὰ συνέβη αὐτῷ ἐλθεῖν, καὶ
ποιῆσαι τοῦτο τὸ τοῦ κομίσασθαι ἕνεκα· Bonitz 1961: 58, excises the phrase τοῦ κομίσασθαι ἕνεκα,
with which there is already a problem: see Ross 1936: 520 and Irwin 1990: 520n.12.
21
The widespread view about affinity between art and chance, technê and tuchê, that Aristotle shares is
an expression of the same outlook. These are in a way concerned with the same things (EN vi.4,
1140a17; Phys. ii.5, 197a7). Aristotle can, without contradiction, treat art, deliberation, and the like,
on the one hand, and chance, on the other, as excluding each other because an instance or episode of
one cannot at the same time be an instance or episode of the other, and also as productive of the same
things by the same means (cf. EN iii.3, 1112a22; vi.4, 1140a17).
76 james allen
contribute to the understanding of luck and chance? The point, as I have
already suggested, is to help us get a fix on how the notion of accident
applies to chance, and the sting will come in the argument from priority
that is based on the opposition between the per accidens, so conceived, and
the per se, conceived in the corresponding way. Very roughly, Aristotle’s
formulation serves to situate luck in a world in which there are things – call
them “ends” – for the sake of which beings – call them “agents” or
“people” – do things, and things that they do for the sake of these ends.
There is, if you will, a repertoire of actions that are done for the sake of ends
and that are available to agents acting for the sake of those ends. Again,
speaking very roughly, when someone performs an action drawn from this
repertoire, but does not, on this occasion, perform it for the sake of the end
for the sake of which it is done (in the abstract), he performs the action and
achieves this end by accident and by luck. Were he to perform the action
for the sake of the end, however, and succeed, it would not be by accident
or by luck.22
This help is welcome because of the way in which Aristotle introduces
the notion of accident and the corresponding idea of the per se (ii.5,
196b25–27, 197a14–15). He observes that, just as there is both being per se
and being per accidens, something may be a cause of, or be responsible for,
something per se or per accidens (196b24ff.). E.g. of a house the cause per se is
a house builder, but per accidens the pale or the musical. This does, to be
sure, illustrate what it means to be a cause by accident. It is because the
builder, who is the cause per se of the house, happens to be or is per accidens,

22
Aristotle’s examples are all of inadvertently lucky agents, e.g. someone who goes to the market with
another end in view and by chance meets and is paid by a debtor. Yet other cases of luck will be like
gambling and winning, where the agent wants the outcome and is trying to bring it about. Aristotle
considers cases like these in the discussion of good fortune in the Eudemian Ethics, where he
mentions throwing dice (viii.2, 1247a27, b17) and people who want something good act with a
view to succeed, but do so despite reasoning poorly (1247b28–32). A full discussion of this kind of
luck would require an essay of its own, but two observations are apposite. First, to avoid being by
accident, the non-chance achievement of an end may require more than the presence of a wish to
bring it about, resulting in an action that does in fact bring it about. Paradigm examples of per se
activity of the relevant kind involve the exercise of a faculty for the sake of an end by the selection of
an action suited to bring it about, which is at the opposite extreme from doing something –
anything – at random in the hope of hitting upon the means to secure one’s end. (There will, of
course, be many – harder – cases between the two extremes.) It may be possible to understand the
relevant notion of accident so that, in addition to cases of inadvertent luck, it also covers cases of this
latter kind. Roughly speaking, to avoid doing something that is done for the sake of an end (in the
abstract) by accident it is not enough to do it out of a wish to attain that end: one must also do it
because it is suited to the end, and this will require that one somehow recognizes what makes it so.
Second, perhaps Aristotle is less concerned with this kind of luck in Physics ii because he does not
think there is a simple analogue in the natural world to the untethered intentions, wishes, or hopes
found in the practical realm.
 Aristotle on chance as an accidental cause 77
say, a flute player that the flute player can be said to be the cause of the
house per accidens, but in this case, one assumes, the house was not built by
accident, and its building owes nothing to chance. It seems that it will
always be possible to find a description under which an event that is plainly
not due to chance has an accidental cause.23 To see how chance occurrences
are by accident, we need to direct our attention away from substances or
objects and towards their faculties or powers and the actions or processes in
which these are realized or displayed.
In Nichomachean Ethics v.8 Aristotle says that when people perform
actions that are just or unjust involuntarily, they do not act justly or
unjustly, except per accidens, because they perform actions that happen
to be, or are by accident, just or unjust (1135a15–19; cf. Metaph. E.2,
1027a3–5). The construction rendered “happen to be” contains the verb
sumbainein from which the expression that we translate as “accident” is
formed. In a more literal translation, which also preserves more of the
original grammar, Aristotle says: “they perform actions to which being
just or unjust happens” or “with which being just or unjust coincides.”
We should draw two lessons. First, it is possible to perform an action in
two ways, accidentally and non-accidentally – sometimes Aristotle says
“simply” or “without qualification” (ἁπλῶς). In this passage, the person
who does what is unjust voluntarily does it simply; the person who does it
involuntarily does it by accident. And, second, the person who performs
an action by accident does so by performing another action simply or
without qualification, the one to which the first action – the one he or she
does by accident – happens to be identical or with which it is per accidens
the same. Here, for example, the involuntarily and accidentally unjust
actor does something simply, not knowing that in so doing he is inflicting
undeserved harm by accident if he is acting out of ignorance; or non-
accidentally does something that prevents a greater evil by doing which
he unavoidably but accidentally inflicts undeserved harm if he is acting
under compulsion.
It is also important to note that what counts as accidental for one
purpose need not do so for others. So, for example, elsewhere Aristotle is
willing to say that someone who thinks that it is necessary to have the
virtues for the sake of external goods does noble deeds per accidens
(EE viii.3, 1249a14–16). Since he does them because advancing his own
selfish interests requires doing noble deeds, what he is doing simply,
properly speaking or non-accidentally, is advancing his own selfish

23
Cf. Judson 1991b: 90.
78 james allen
interests. However, there is no doubt that he does the noble deeds
intentionally (if not with the intention of acting nobly) and voluntarily.
What of the per se with which the per accidens character of luck and the
automatic is contrasted? Here are the relevant passages from Physics ii.5 in
which the expression “cause per se” is introduced. Note the shift from
agents to the same agents in activity and the actions they perform, and that
all the examples are of agents acting with a view to an end (cf. ii.3, 195b5).
This is not because per se causation is confined to cases that evidently
involve final causation. A hot or a cold body could per se be the cause of
the heating or cooling of another body. But the cases in which we are
interested, where luck is involved, have outcomes that are ends, the per se
causes of which, if any, would act for their sake.
E.g., of a house the cause per se is a house builder, but per accidens the pale or
the musical. The per se is determinate, the per accidens indeterminate.
(196b26)
And, twenty-two lines later:
E.g., of a house the cause [per se, without qualification] is a house builder, but
per accidens a flute player, and of someone having come collecting the money,
not having come for the sake of this, [the causes by accident are] infinite
in number, e.g., wishing to see someone, or prosecuting or defending or in
order to watch. (197a14ff.)
Collecting the money plays the part of “that for whose sake,” and the
action of having come (to where the debtor is) that of “what is for its sake.”
The “se” in “per se,” the itself in virtue of which, refers to the agent, but not
simply the agent, or even the agent in so far as he or she possesses a certain
power or faculty for bringing about the result at issue, but to the agent as
actually using or exercising it to that end.
Normally when we discover that, e.g., a flute-player who is the acciden-
tal cause of, say, a cure, is also a doctor, we are entitled to rule out luck or
chance as the cause, but not always. Suppose a doctor is throwing darts to
relax after hours and, quite unintentionally, hits someone suffering from
fever, as it happens drawing precisely the amount of blood that he would
have prescribed be drawn had he been practicing the medical art, and, as a
result, cures the person struck by the dart. A doctor, exercising the medical
art, would be the per se cause of the corresponding action and outcome,
those of which and for which the medical faculty or power is a faculty or
power. When, as in the present case, the doctor does this action and causes
this outcome without so acting – by performing simply another action,
throwing darts, with which being something that is done for the sake of a
 Aristotle on chance as an accidental cause 79
cure in the abstract, drawing such and such an amount of blood, happens
to coincide or is by accident the same – he is not their per se cause, not their
cause as a doctor, and so the person with the fever is cured by accident and
as a matter of luck.
To judge by these examples, then, the agent is the cause per se or in virtue
of itself of an outcome when it exercises one of its faculties – the faculty of
bringing about results like this – with a view to that end, by selecting from
the repertoire of actions open to it one that is for its sake in the abstract.24 It
is this action that the agent does simply or without qualification. The
result, when all goes well, is not by accident, and hence not by luck. By
contrast, the agent brings about the same result by luck and is the cause of it
by accident when, as we should say, he does the same thing with the same
result, not, however, simply, but by accident. Although not the per se cause
of the lucky outcome this time, the agent is the per se cause of something
else, namely of the action that he performs simply or without qualification
for the sake of another end, possibly exercising another faculty, by per-
forming which he performs by accident the action that is for the sake of the
lucky outcome in the abstract, which happens to be or is by accident the
same as the first action.
Cases of luck and chance more generally are, Aristotle insists, also char-
acterized by low relative frequency (ii.5, 196b10–13, 36; ii.8, 199b24–25). In
Aristotle’s example, by going to the market, say, to buy fish, I place myself in
proximity to my debtor only rarely. But are the two characterizations of luck,
being accidental, as I have interpreted it, and being rare necessarily
coextensive? Aristotle’s language is compatible with more than one inter-
pretation. It seems conceivable that there could be cases in which by doing
one thing per se one could hardly fail to do something else per accidens, as a
concomitant. Elsewhere, of course, Aristotle is happy to allow that a
substance is always and necessarily attended by some of its accidents
(Metaph. Δ.30, 1024a30–34; Apo i.7, 75b1); and some cases where the items
accidentally unified are actions are also like this – doing noble deeds by
accident in the case of the agent acting for the sake of external goods cited
above, for instance. Perhaps, then, the rarity requirement functions as an
additional constraint, so that for an occurrence to be by luck, it must be both
by accident in the specified way and rare. Alternatively, rarity might be a
further requirement on being an accident of the type relevant to chance
(cf. Phys. ii.8, 199b24–25). And the gap between these two alternatives could
be reduced by further reflection on the place of concomitant effects. To a

24
Cf. Judson 1991b: 92–93, whose conclusion, reached by a different route, mine resembles.
80 james allen
well-informed agent they come as no surprise, and they are part of what he
does intentionally even if not with the intention of bringing them about.
If we restrict our attention to the practical realm, Aristotle’s account
promises to do justice to the desiderata that he set up at the outset. Luck is
neither a break in the causal order nor a case of divine meddling in mortal
affairs by the gods on behalf of their favorites. Every lucky outcome has a
cause, namely, it seems, an agent acting, and acting for the sake of an end,
albeit one different from the one by bringing about which he is lucky. It is
merely that in the Aristotelian universe everything, including every action,
is what it is and, unavoidably, an indefinitely large number of other things
by accident as well. As a result, in doing what we do simply or that of which
we are the cause per se we cannot fail to do by accident and be the cause of
an indefinitely large number of other things as well (cf. ii.5, 196b28–29).
Sometimes, one of these will be one that (in the abstract) is done for the
sake of an end that we value and brings it about. This, roughly speaking, is
luck. As an inescapable byproduct of agents acting with a view to ends,
then, there is nothing at all surprising about luck in general; it is, rather, to
be expected. At the same time, individual episodes of luck will be unex-
pected, obscure to human understanding, and of an indefinite character
that sets them apart. Whereas the explanation for a non-lucky outcome,
say, a cure, will be the exercise by an agent of a faculty for curing directed to
this end, that of a lucky cure will now be the exercise of this unrelated
faculty to this unrelated end, now to that end (cf. EE vii.14, 1247b12ff.).
The priority argument at the end of Physics ii.6, which concludes that
mind and nature are necessarily prior causes both of many other things and
of this universe, plainly applies to the natural world, but we will be in a
better position to understand it if we continue, for a little while longer, to
restrict our attention to the realm of practice by considering a version
applying to luck. To do something by luck and to be lucky as a result is to
be the cause per accidens of an action that is done or that people do for the
sake of that result. Speaking in this way is only intelligible if the practical
realm contains agents who are per se causes of many of their actions and
their results by performing the former for the sake of the latter. Without
this, luck so conceived would be unintelligible, and, whatever happened
and however it came about, it would not be lucky or by luck.25

25
I think this is the kind of explanatory priority that Aristotle has in mind, although every action for
the sake of an end in the abstract of which an agent is the cause per accidens is itself per accidens the
same as an action of which the agent is the per se cause and which could, perhaps, be seen as prior to
the other.
 Aristotle on chance as an accidental cause 81
What, then, of automatic occurrences in the realm of nature? Much
that is at home in the practical realm will have no place there –
intention, deliberation, choice, art – but if Aristotle’s account applies
mutatis mutandis to nature as well, for there to be per accidens automatic
outcomes there will have to be in the natural world a repertoire of
processes that occur for the sake of ends (in the abstract), but that can
on occasion take place without doing so for the sake of those ends. And
this can be the case, it would seem, only if there are episodes of per se
causation understood along the same lines as in the practical sphere,
i.e. episodes in which beings function as per se causes for the sake of
ends. Plainly, on this account, luck and mutatis mutandis the automatic
have a place only in a domain where explanation in terms of final causes
is at home. But any such assertion about the domain of nature would be
vehemently rejected by Aristotle’s opponents, who are nevertheless
happy to appeal to luck or chance. I conclude that the account of
chance on which the priority argument is based does beg the question
in dispute.

IV
Were the argument of Physics ii.8 based on the same account without
modification, it would suffer from the same defect and be open to the same
objection. But on closer inspection, it becomes clear that it is not. Aristotle
begins with the example of rainfall, which, by falling, makes the crops grow
(198b16–23). Although it remains a matter of controversy whether he is
expressing his own opinion or making a concession to his opponents for
the sake of argument, Aristotle presents this process as one that does not
take place for the sake of its beneficial consequences, but is rather explicable
entirely in terms of material necessity (198b18). It just so happens, or is an
accident (συμβαίνειν, 198b21, 23), that they coincide with it. The point is to
make available the idea of an accidental unity whose two terms are, on the
one hand, processes stemming from material necessity – heating, cooling,
condensing, and the like – and, on the other, those bringing about
beneficial outcomes like plant growth. When they say that things like
bodily organs, e.g. the jaws, are products of chance, Aristotle now takes
opponents to be committed only to the view that these processes come to
be or take place as if, or just as they would if, for the sake of the good they
do (198b29ff.). To give the notion of chance a possible purchase on the
occurrences whose explanation is disputed, an as if teleology is sufficient;
the alternative to accident now is merely the notional possibility of per se
82 james allen
causation for the sake of an end.26 In effect, Aristotle has conceded, if only
for the sake of argument, that chance is conceivable even in a domain from
which final causation is completely absent. He does not argue, as he did
before, that if some occurrences in the realm of nature are by chance, and
therefore caused per accidens as lucky events in the sphere of agency are,
there must be others that are caused per se by end-directed powers acting
for the sake of their results.
Instead, he relies for his main argument on the infrequency require-
ment, lent support by the sheer implausibility of supposing that complex
beneficial arrangements like the organs of an animal body could be the
product of accident. The argument is framed by a stark alternative. Either
the processes that give rise to, e.g., the jaws, do so by coincidence, i.e. occur
only as if for the sake of the outcome and, it is implied, by chance, or they
come to be for its sake. The infrequency requirement is used to eliminate
the first alternative. Animal parts like the jaws arise as they do always or for
the most part, therefore not by chance, and therefore, since this is the one
remaining alternative, for the sake of something. The argument, which has
been the object of much study, is hardly free from problems. The rainfall
example with which the chapter begins seems to have left room for at least
the notional possibility of regular coincidences in nature – whether this is
Aristotle’s settled opinion or not – where the beneficial result, e.g. plant
growth, and the processes that give rise to it, e.g. heating, cooling,
condensation, are not related as that for the sake of which and what is
for its sake respectively (except with the qualification as if), and all the
explanatory work is done by material necessity. An argument that, what-
ever might be said of rainfall and crop growth, bodily parts like the jaws
cannot be the products of such coincidences is conceivable, but none is
supplied. And Aristotle does not seem to pay adequate attention to the fact
that those thinking along the same lines as Empedocles, as he is viewed
here, seem to have envisaged original, one-in-a-million accidents (which
would satisfy the requirement for infrequency) that give rise to arrange-
ments that are not only beneficial in the short term, but are also self-
sustaining or self-replicating, and that, once in place, can be relied upon to
work always or for the most part. Nonetheless, the argument does not beg
the question begged by the priority argument of Physics ii.6.
There follows a series of supporting arguments based on the resemblance
between natural and artistic processes. In both, a set of events that must
occur when and in the precise sequence that they do if the benefit that

26
I owe the idea of an as if teleology to Wieland 1970: 259 and passim.
 Aristotle on chance as an accidental cause 83
results is to be produced can be observed occurring regularly and predic-
tably (199a8ff.). Aristotle invites us to appreciate, e.g., that if a house were
to come to be by nature, it would come about just as it does now by art
(199a12; cf. 199b28).
It is only now, after he has established to his satisfaction that nature is a
cause for the sake of something, that Aristotle alludes, in a passage that I
have already cited, to the idea that played such an important part in
chapters 4–6:
That for whose sake and what is for the sake of this may come about by luck;
for example, we say that a stranger came by luck, paid the ransom and went
away, when he did so as if he had come for the sake of this though he did not
come for the sake of this – and this by accident. (199b18)
Although it is possible that Physics ii.8 contains Aristotle’s second
thoughts, I suspect that the two arguments are complementary. The
priority argument belongs to an exposition of Aristotelian doctrine. The
main argument of Physics ii.8 meets Aristotle’s opponents half way, and,
by showing, if it succeeds, that they cannot be right even so, provides
crucial support for one of the assumptions on which the account of
chance already expounded rests, namely that there are final causes in
nature. With this principle secured, Aristotle now re-states and re-affirms
his account. An outcome comes about by chance when an action or process
that is for its sake, i.e. belongs to a repertoire of actions that are done, or
processes that are set in train, for its sake (in the abstract), is done or set in
train by accident, meaning without being done or set in train for the sake of
the lucky or automatic outcome that is its issue.

Appendix: The etymological argument of Physics ii.6, 197b22–31

The point that Aristotle is trying to extract by etymological means, I
suggest, resembles that about the relation between thought or art and
chance. Just as the same things come about by thought or art as come
about by chance and vice versa, so occurrences (of the kind) that take place
in vain when they are done for the sake of an end but without success come
about by chance when they are done by accident but with success, and of
course neither by chance nor in vain when they are done non-accidentally
for the sake of the outcome at issue and with success. It is meant to lend
support to the idea expressed in the immediately preceding passage, that
when something that is for the sake of something in the abstract comes
about without coming about for its sake in actual fact, we speak of chance
84 james allen
(197b18–20). The next remark (197b20–22), that those chance or automatic
events that are objects of choice are by luck, is a parenthesis (cf. ii.5,
196b18–19). The main topic remains the conditions under which it is
proper to speak of the automatic.
Aristotle now turns to the expression μάτην, “in vain.” The ὅτι in the
next sentence is usually taken to mean because:27
The expression “in vain” is a sign because it is said when what is for
the sake of another thing does not come to be for the sake of that
thing. (197b22–23)28
But these are not the conditions under which the expression “in vain” is
correctly applied. They restate the conditions under which the expression
“by chance” is used. Attempts have been made to emend the passage so
that it means: “the expression ‘in vain’ is used when something is for the
sake of something else and what it is for the sake of does not come
about.”29 I suggest that instead we take ὅτι to mean “that” and to
introduce a clause specifying that of which the expression “in vain” is a
sign:30
And the expression “in vain” (μάτην) is a sign that it [namely, ἀπὸ τοῦ
αὐτομάτου] is said when what is for the sake of another thing does not come
to be for the sake of that thing.
Aristotle explains the meaning of μάτην with two examples (197b23–29).31
If one walks for the sake of relaxation and relaxation does not ensue, one
has walked in vain. It is correct to say that something occurs in vain, he
continues, when it is by nature for the sake of something else, and that for
whose sake it was and which it is by nature such as to bring about does not
take place (197b25–27). To occur in vain, something must:

27
Trans. Lacey 1997.
28
σημεῖον δὲ τὸ μάτην, ὅτι λέγεται ὅταν μὴ γένηται τὸ ἕνεκα ἄλλου ἐκείνου ἕνεκα.
29
Simplicius knew of an alternative reading that yields this sense (In Ar. Phys. 349.4–6), which is
adopted by Pacius 1596: 58; Torstrik 1875: 464; Carteron 1926: i.31; and Wicksteed and Cornford
1929: 34. Prantl 1854: 82 proposed an emendation to the same effect, adopted by Bonitz 1969: 39;
Hamelin 1931: 134; and Charlton 1970: 49. Ross 1936: 523 believes the received text can yield the
required sense: “the MS reading offers no great difficulty if we remember that ἕνεκα του may mean
‘producing an end-like result.’ The phrase means ‘when that which was intended to produce an end-
like result does not produce it.’” He is followed by Wagner 1989: 48; cf. 472.
30
Although committed to a different account of why “in vain” is a sign, Philoponus construes this
sentence as I do (289: 28–29 Vitelli). Ross 1936 also treats it as a possibility (523 ad ll. 22–23).
31
οἷον εἰ τὸ βαδίσαι λαπάξεως ἕνεκά ἐστιν, εἰ δὲ μὴ ἐγένετο βαδίσαντι, μάτην φαμὲν βαδίσαι καὶ ἡ
βάδισις ματαία, ὡς τοῦτο ὂν τὸ μάτην, τὸ πεφυκὸς ἄλλου ἕνεκα, ὅταν μὴ περαίνῃ ἐκεῖνο οὗ ἕνεκα ἦν
καὶ ἐπεφύκει, ἐπεὶ εἴ τις λούσασθαι φαίη μάτην ὅτι οὐκ ἐξέλιπεν ὁ ἥλιος, γελοῖος ἂν εἴη· οὐ γὰρ ἦν
τοῦτο ἐκείνου ἕνεκα.
 Aristotle on chance as an accidental cause 85
(i) be for the sake of X in the abstract
(ii) actually be done for the sake of X
(iii) fail to bring X about.
Aristotle emphasizes condition (i), as the next example, of the man who
bathes with a view to bringing about an eclipse, makes plain. It would be
absurd, he says, for this man to say that he had bathed in vain, absurd
because bathing is not for the sake of eclipses, meaning not that the agent
did not want an eclipse, but that (in my terminology) bathing is not
something done for the sake of eclipses in the abstract.
Now comes the phrase that is supposed to bring what we have learned
about μάτην (in vain) into relation with the automatic (τὸ αὐτόματον)
(197b29–30).
So then the αὐτόματον even according to its name when it (αὐτό) should
come to be in vain (μάτην).32
That the expressions αὐτόματον and αὐτό μάτην resemble each other is
plain; but the remark is otherwise obscure in the extreme. The absence of a
main verb in the translation is not an oversight; it is also missing from the
text, and Aristotle’s next remark does not help (197b30–32):
For (γὰρ) a stone fell not for the sake of hitting someone, but it fell
automatically because it might have been made to fall by someone and for
the sake of hitting.33
This is not an example of something happening in vain, but seems to
illustrate the contention for which the etymology is supposed to provide
evidence, that is, once again, I suggest, that the expression αὐτόματον is
used when an occurrence that is for the sake of X in the abstract does not
occur for its sake (197b22–23).34
Two explanations have been proposed, each with serious defects.
According to the traditional favorite, an automatic occurrence brings
about an automatic outcome instead of the outcome for whose sake it
was, and therefore occurs in vain in relation to the latter but not in relation
to the former, its chance result.35 The problem is that, though true of some

32
οὕτω δὴ τὸ αὐτόματον καὶ κατὰ τὸ ὄνομα ὅταν αὐτὸ μάτην γένηται·
33
κατέπεσεν γὰρ οὐ τοῦ πατάξαι ἕνεκεν ὁ λίθος· ἀπὸ τοῦ αὐτομάτου ἄρα κατέπεσεν ὁ λίθος, ὅτι
πέσοι ἂν ὑπὸ τινὸς καὶ τοῦ πατάξαι ἕνεκα.
34
Noticed by Bonitz 1969: 39, who, however, did not think it was a sufficient ground for retaining the
ms. reading at 197b22–23.
35
Alexander, On the Soul ii, 178.24–29 (in Bruns 1887); Alexander apud Simplicium In Ar. Phys.
349.13–35; Themistius In Ar. Phys. 55.19–56.7; Pacius, 468; Ross 1936: 524; Guthrie 1946: 70–76;
Waterfield 1996: 242.
86 james allen
episodes of chance, it is not true of others, in which the chance outcome
occurs in addition to the outcome for whose sake the event takes place.36
According to the alternative, μάτην in the crucial phrase αὐτό μάτην (at
197b30) means not “in vain” but, as it also can, “without a goal or
purpose”.37 Automatic outcomes would then be the result of events that
took place without a goal, or at least without the goal of producing this
result. The problem is that, in the context of the bathing and walking
examples, μάτην does mean “in vain.”38 What is more, Aristotle means his
etymological digression to furnish support for a contention about the
automatic. Which contention is especially clear if one retains the manu-
script reading at 197b22–23 and understands it, as I do, to mean the
expression “in vain” is a sign that, but, even if one does not, the most
likely contention remains the same, namely that X is an automatic out-
come when an event that is for its sake in the abstract brings it about
without, on this occasion, occurring for its sake. On neither proposal does
the etymological excursus obviously support this or any other relevant
contention.
I suggest that just as the same occurrences come about by art as come
about by chance and vice versa, so occurrences that take place in vain come
about by chance and vice versa. Take an event that is for the sake of X in the
abstract: Have someone perform it for the sake of X, but without bringing
X about, and it was done in vain. Change the story by making it bring X
about without being done for the sake of X, and it brings X about by
chance.39 If this interpretation is on the right lines, Aristotle’s point is
sound, even if his etymology is not. It is not clear that the phrase as it
appears in the ms. at 197b29–30 can bear this weight as it stands, but there
are plausible emendations that would allow it to do so. Although it would
hardly make matters crystal clear, an original present subjunctive (ὅταν

36
Cf. Guthrie 1946, who, although he favors this explanation as the best available solution, sets out this
problem very clearly. It had already been noticed by Philoponus In Ar. Phys. 290.17–20 and 290.25,
and Simplicius In Ar. Phys. 348.27–31.
37
Philoponus In Ar. Phys. 290.9–25; Prantl 1854: 82, Wicksteed and Cornford 1929: 34, Wagner
1989: 472.
38
Torstrik 1875: 463 and esp. Guthrie are very clear on this point. Even Wagner (1989: 472), who favors
this interpretation, acknowledges that it requires μάτην (“zwecklos”) first to mean “vergeblich” and
then to mean “ohne Zweckbestimmung.”
39
Torstrik 1875: 463–464, who thinks much of the passage about the relation between the expressions
μάτην and αὐτόματον is nonsense and very likely not due to Aristotle himself, nevertheless comes
closest to this interpretation. He notes that one of the points that the passage could have made but
does not is that both vain and chance events are by nature for the sake of something else τὸ πεφυκὸς
ἄλλου ἕνεκα. In my view, the passage does make this point by taking it to underlie the similarity
between αὐτό μάτην and αὐτόματον.
 Aristotle on chance as an accidental cause 87
αὐτό μάτην γίνηται), altered under the influence of the nearby aorist
subjunctives at 197b19, 197b23, and 197b34, might make it easier to view
the temporal clause a filter for event types at 197b29, so that Aristotle’s
meaning would be, roughly, “So then the αὐτόματον even according to its
name [is found] when it (αὐτό) comes to be in vain (μάτην) [i.e. is the sort
of thing that comes to be in vain or is such as to come to be in vain].”
Alternatives worth considering include: an original ὅτι ἂν or ὅσ’ ἂν with
γένοιτο (in the latter case with the singular αὐτό μάτην understood as
though in inverted commas and explained as the result of Aristotle’s desire
to exhibit the phonetic similarity between αὐτόματον and αὐτό μάτην).40
In its favor, the point this interpretation would have Aristotle make here
is one that he has already been at pains to emphasize, namely that X is an
automatic outcome when an event that is for its sake in the abstract brings
it about without, on this occasion, occurring for its sake.

40
Cf. Phys. ii.5, 196b22; ii.6, 198a6; and esp. ii.5, 197a35: ὅσ’ ἂν γένοιτο ἕνεκά του.
 chapter 5

Man from man but not bed from bed:

Nature, art and chance in Physics ii
Margaret Scharle

The first argument in Physics ii.8 serves as the foundational text for
understanding the domain, extent, and character of Aristotle’s natural
teleology. On the basis of this text, most scholars think that Aristotle’s
natural teleology applies exclusively to biological things (plants and ani-
mals) and that the elements (earth, air, fire, and water) either are not
teleological or are teleological only in so far as they play a role in biological
processes. In addition, some scholars think this text shows natural teleology
to operate not only within an individual living thing, but also to extend
throughout the scala naturae, with lower things (like elements) existing for
the sake of higher things (like animals and plants, and ultimately humans).
With what they take to be the domain and extent of natural teleology
confirmed by this text, scholars look outside the Physics to deepen their
understanding of the character of natural teleology (as well as related
concepts such as cause, end, nature, chance, and necessity) through careful
consideration of its application in particular explanatory contexts. Those
convinced of the restriction of natural teleology to individual biological
things seek clarification predominantly in the biological works, such as
Generation of Animals and Parts of Animals, while those seeking, in addi-
tion, better understanding of a supposed commitment to an overarching
teleology across the scala naturae turn also to such works as Metaphysics xii
and even the Politics.1
In a previous paper I argued that the role of nature in Physics ii.8’s first
argument for natural teleology has been widely misunderstood, and as a

I thank Sean Kelsey and Devin Henry for written comments on an early draft of the chapter, and
Mariska Leunissen for her comments on the penultimate draft. I also thank Kellyn Bardeen for her
excellent editorial assistance and philosophical insight.
1
Gotthelf and Lennox 1987 and Lennox 2001a are paradigmatic examples of the former approach,
while Sedley 1991 is such of the latter. Notable exceptions include M. R. Johnson 2005, Leunissen
2010, and Quarantotto 2005.

88
 Man from man but not bed from bed 89
result Aristotle has been interpreted with an overly biological focus.2 I
suggested a new reading of the winter-rain example that appears in the
argument and argued that water is teleological on its own, independent of
biological processes. If I am correct in my interpretation of the text, we
should be looking not only at the biological works, but also at the
elemental works such as the Meteorology, On Generation and Corruption,
and On the Heavens to understand the character of natural teleology and
related concepts.
In general, there are two desiderata for a proper interpretation of the first
argument in Physics ii.8: First, the interpretation must show that the
premises and conclusion are ones Aristotle himself would accept; and
second, since the argument is meant to engage an opponent, its interpreta-
tion must offer a satisfying account of the dialectic in the passage.3 In my
previous work, I showed my new interpretation to best satisfy the first
desideratum, but did not speak to the second. In this chapter, I argue that
my interpretation best satisfies the second desideratum as well and, more
importantly, suggests a unified interpretation of the dialectic across the
whole of Physics ii.

I The argument of Physics ii.8

In Physics ii.8, Aristotle considers a puzzle from an opponent to his natural
teleology. The challenge makes use of some of Empedocles’ views, yet, as
Alan Code points out, the problem is not posed by Empedocles and may
not be one that he would have endorsed.4 I follow Code in referring to the
opponent of Physics ii.8 as the “Empedoclean” opponent, who presents
what I call the “statement of the problem”:
There is the difficulty: what prevents nature from acting neither for some-
thing nor because it is better, but as Zeus rains – not in order that the corn
may grow, but of necessity. (For what was taken up must become cold, and
what has become cold, having become water, must come down. When this
has happened, it turns out that the corn grows.) Similarly also, if someone’s
corn on the threshing floor is ruined it does not rain for the sake of this, so
that the corn may be ruined, but this simply results. Why then should it not

2
Scharle 2008.
3
In arguing for the importance of the dialectic in Physics ii.8, I do not mean to suggest that Aristotle’s
method in the Physics is dialectical as opposed to scientific. I see the Physics’ use of dialectical and
scientific methods as complementary and integrated, but I will not pursue the complex issue here. For
the purposes of my interpretation, “dialectic” need not be understood in any more technical a sense
than engaging an interlocutor, as opposed to talking past him.
4
Code 1997: 127.
90 margaret scharle
be the same with the parts in nature, e.g. that our teeth should come up of
necessity – the front teeth sharp, fitted for tearing, the molars broad and
useful for grinding down the food – since they did not arise for this end, but
it was merely a coincidental result; and so with all other parts in which we
suppose there is purpose? Wherever then all the parts came about just what
they would have been if they had come to be for an end, such things
survived, being organized by chance in a fitting way; whereas those which
grow otherwise perished and continue to perish, as Empedocles says his
“man-faced ox-progeny” did. (Phys. ii.8, 198b17–34)5
In the course of the chapter I will discuss what this challenge amounts to.
But first I will focus on Aristotle’s response to this challenge with the
passage I call “the winter-rain argument”:
Such are the arguments (and others of the kind) which may cause difficulty
on this point. Yet it is impossible that things are this way. For these things
[e.g. animals] and all things that are by nature, come to be in this way either
always or for the most part, and nothing from luck or chance does. For it
does not seem to be from luck or from coincidence that it rains often in
winter, but if in the dog-days; nor that there are heatwaves in the dog-days,
but in winter. If, then, things seem to be either from coincidence or for the
sake of something, and if these things are not able to be from coincidence or
from chance, they would be for the sake of something. But clearly all such
things are by nature, as these speakers themselves would say. The “for the
sake of something,” then, is in things which are and come to be by nature.
(Phys. ii.8, 198b35–199a8)
Many commentators suggest that Aristotle presents nothing in the winter-
rain argument that does anything more than baldly beg the question –
asserting, more than arguing, the position of natural teleology against his
Empedoclean opponent. I believe a more compelling argument can be
constructed from this passage by bringing in texts from outside the Physics
to sharpen our understanding of the shared ground between Aristotle and
his Empedoclean opponent. The structure of the argument, I suggest, is as
follows:
1. (Accepted Premise): “Things [e.g. animals] seem to [come to] be
either from coincidence or for the sake of something.”
2. (Disputed Premise): “These things [e.g. animals] are not able to [come
to] be from coincidence or from chance.”
3. (Sub-Conclusion): “These things [e.g. animals] would [come to] be
for the sake of something.” (1, 2)

5
Trans. Barnes 1984; modified. All translations are from Barnes 1984, unless otherwise noted.
 Man from man but not bed from bed 91
4. (Accepted Premise): “All such things are by nature, as these speakers
themselves would say.”
5. (Conclusion): “The ‘for the sake of something,’ then, is in things
which are and come to be by nature.” (3, 4)
As presented, the argument is valid. In order to establish his conclusion,
however, Aristotle must be able to convince the Empedoclean opponent to
accept each of the premises, so that the conclusion will follow. After
considering Premises (1) and (4), I will focus on what is, perhaps, the
most problematic premise, Premise (2).
As Leunissen notes, Premise (1)’s use of δοκεῖ suggests mutual agree-
ment.6 There is a trivial reading of the agreement, argued by Code and
Charles, in which the disjunction is trivially true because the opponent
agrees to one of the two disjuncts.7 On my view, Aristotle positively moves
his opponent to accept Sub-Conclusion (3), that “these things [e.g. ani-
mals] come to be for the sake of something,” by asserting the disjunction in
the strong sense required for the entailment of (3) (that the rejection of one
disjunct entails the acceptance of the one remaining). If this were the case,
he would not be simply arguing against the disjunct that the opponent
accepts and asserting his own: He would be using the rejection of the
opponent’s position to drive the acceptance of his own. By moving the
Empedoclean opponent to accept Sub-Conclusion (3), Aristotle can make
use of Accepted Premise (4) to reach the final Conclusion (5). If Charles
and Code are correct that all Aristotle can do is show one of the disjuncts to
be false, then the fanfare that Aristotle makes over Premise (4) as shared
ground – “as these speakers themselves would say” – would be for naught;
without the opponent’s acceptance of Sub-Conclusion (3), the mutual
agreement on Premise (4) would serve no purpose in the dialectic.
On my view, the Empedoclean opponent maintains that tooth growth
and the generation of whole animals appear to be teleological: We suppose
that there is purpose in these cases (ἐν ὅσοιϛ δοκεῖ ὑπάρχειν τὸ ἕνεκά του;
Phys. ii.8, 198b28–29).8 The opponent suggests that things might not be as
they appear and offers his own alternative: Maybe they just came to be that
way by chance, in a similar way to things that are not, in fact, teleological
(like rain resulting in corn growth or rain resulting in corn rot). The
opponent thereby concedes that phenomena like tooth development in
animals and the generation of whole animals might be teleological, and the

6
Leunissen 2010: 29. 7 Charles 1991: 113; Code 1997: 129.
8
I acknowledge Sean Kelsey for suggesting this view to me.
92 margaret scharle
question on the table is whether they, in fact, are – either they are
teleological (as they appear to be) or they are not (because the opponent
has presented a successful alternative). This is what Premise (1) formalizes.
In admitting that the phenomena appear teleological, the opponent
accepts that the onus is on him to unseat the presumption in favor of
teleology.
So interpreted, Premise (1) does not claim that for any occurrence
whatsoever, it happens either by chance or for the sake of something. Not
even Aristotle thinks that.9 Rather, Premise (1) states that the phenomena in
question happen either for the sake of something (as they appear to) or by
chance (because the opponent has offered a successful counter).10
Let us now consider Premise (4). Aristotle directly states that the
Empedoclean opponent would accept the truth of this premise, that all
such things that are at issue in the dialectic are “by nature” (φύσει). This is
important because Aristotle and his Empedoclean opponent disagree as to
which types of natures exist, so the meaning of the term “by nature” must
be interpreted in a way that is neutral to this philosophical disagreement.
The Empedoclean opponent denies the existence of natures above the level
of the four elements: earth, air, fire, and water. Homogenizing the views of
Aristotle’s materialist predecessors – including Empedocles – Physics ii.1
attributes to them the view that the four elements are “the whole of
substance, all else being its affections, states, or dispositions” (193a25).11
Aristotle’s own view, by contrast, is that each natural thing – each animal,
in this case – has a nature of its own, where nature is an inner source of
movement and rest that belongs to the thing “primarily in virtue of itself”
(πρώτωϛ καθ’ αὑτὸ) (Phys. ii.1, 192b22), and that animal generation
(propagation of the species) comes about caused by the source of move-
ment and rest that is the animal nature.

9
The production of bile is a case in point: Bile is produced neither for the sake of something nor
coincidentally, but simply as a necessary byproduct of the teleological operations of the liver (PA
iv.2, 677a12–18). The nature of the animal does not aim to produce bile: In fact, the healthier the
liver and the more pure the blood, the less bile is produced.
10
This premise is suggested by the Empedoclean opponent himself in the statement of the problem:
When it comes to the phenomena under consideration, the two options are “either from coin-
cidence or for the sake of something.” Sauvé Meyer 1992: 796–797 points out that although the
initial disjunction is misleadingly stated as “not in order that . . . but of necessity” (198b18–19), the
gloss of that disjunction (198b19–23) does not mention necessity. The reason the opponent gives for
denying that the phenomena are teleological in the restatement of the position at 198b27 is that they
are coincidental (198b27), even though necessity is mentioned in this passage. Ultimately, the final
statement of the opponent’s position (198b27–32) does not mention necessity at all, thereby
suggesting that the opponent agrees to the disjunction for the phenomena in question.
11
See also Metaph. v.4, 1014b35–1015a3.
 Man from man but not bed from bed 93
Without begging the question, Aristotle cannot build into (4) the
assumption that animal generation is by the nature of the animal, because
in the Empedoclean opponent’s view animal generation is by the nature of
the elements. However, in Physics ii.1, Aristotle carefully marks off the
distinctions between something that “has a nature” and something that
is “by nature” (φύσει). The locution “by nature” is introduced as a descrip-
tion of that which is by an inner source of movement and rest, which has a
much wider scope than “has a nature,” for it includes not only things that
have a nature, but, more generally, that which is the result of nature (e.g.
the natural activities of things that have nature), whether the nature of an
animal or the nature of the elements.
Premise (4) therefore remains neutral between Aristotle’s own view and
that of the Empedoclean opponent, who denies that animals “have” a
nature (for they do not have a nature that belongs to it primarily in virtue
of itself), yet admits that animals are “by” nature in the sense that they are
by the nature of the elements. In order to reach his conclusion, Aristotle
needs nothing stronger than the neutrally stated Premise (4).
The greatest difficulty in interpreting Aristotle’s argument lies in show-
ing why he thinks the Empedoclean opponent could be moved to accept
the truth of (2) – so I will spend the bulk of the chapter explaining and
defending this premise. The text seems to offer the following claims in
support of (2):
A. (Accepted Premise): “It does not seem to be from luck or from
coincidence that it rains often in winter, but if in the dog-days; nor
that there are heatwaves in the dog-days, but in winter.” (There is a
regularity in winter rain, and in summer heatwaves, which cannot be due
to coincidence.)
B. (Disputed Premise): “These things [e.g. animals] and all things that
are by nature come to be in this way either always or for the most
part.” (Animals and things that are “by nature” come to be with the same
type of regularity as winter rain and summer heatwaves.)
C. (Disputed Premise): “Nothing from luck or chance does [i.e. nothing
from luck or chance comes to be in this way either always or for the
most part].” (Nothing that comes by chance comes to be with the same
type of regularity as winter rain and summer heatwaves.)
Premise (2) (Conclusion): “These things [e.g. animals] are not able to
[come to] be from coincidence or from chance.” (Therefore, animals
and things that happen “by nature” cannot come to be from chance.) (A,
B, C)
94 margaret scharle
The question of how A, B, and C are meant to support Premise (2) is not
obvious, although the gloss after each quotation offers a preview of the
interpretation I will argue. To begin, I want to consider whether the
Empedoclean opponent would accept the truth of these claims. Let us
consider Premise (A). Most scholars now think Aristotle means that while
summer rain produces corn growth coincidentally, winter rain produces
corn growth teleologically.12 In my previous paper, I offered a detailed
argument against this interpretation by showing that it would lead to a
conclusion Aristotle himself would not accept.13 Importantly, this inter-
pretation also fails to be dialectically satisfying. Again, Aristotle’s use of
δοκεῖ here, as in his statement of Premise (1), suggests mutual agreement.14
But the case of rain’s production of corn growth was originally intended by
the Empedoclean opponent as an obvious example of chance, and it is not
at all clear why pointing out this seasonal connection would move the
opponent to change his mind. After all, the opponent has already said that
“cold” is responsible for the rain (Phys. ii.8, 198b14), which is close to
acknowledging its seasonality.
On my reading, Premise (A) presents winter rain, taken on its own, as an
example of a non-coincidental phenomenon, and this is simply a restate-
ment of the opponent’s assertion: “For what was taken up must become
cold, and what has become cold, having become water, must come down”
(Phys. ii.8, 198b19–20). Premise (A) simply makes this mutual agreement
precise by clarifying that the process takes place when it is cold, and that it
is typically cold in the winter, and not in the summer.15 Premise (A)
additionally notes that we would not say that winter rain happens by
coincidence, but only summer rain.
Let us now consider Premises (B) and (C). The statement of the problem
shows the Empedoclean opponent to reject either or both: Although
nowadays species reproduce true to type, and thus nowadays come to be,
for instance, with a set of teeth with molars in back and sharp teeth in front
(either always or for the most part), nonetheless it was not always so, the
Empedoclean opponent would say. The species we see now were once less

12
I argue against the full range of alternative interpretations in Scharle 2008: 148–167. Most recently,
Leunissen 2010: 10–48 has suggested that Premise (A) refers to the fact that farmers use winter rain to
grow their crops. Although I do not have the space to pursue the point, I think her otherwise
persuasive interpretation strains the text in requiring that winter rain not be among the things whose
nature is for the sake of something.
13
Scharle 2008: 151–167. 14 Leunissen 2010: 29.
15
Pace Leunissen, who argues, “If the argument is to be rhetorically effective, it seems that there must be
some non-accidental way in which Aristotle thinks winter rain (even if not itself caused teleologi-
cally) serves the growth of crops” (2010: 30; my emphasis).
 Man from man but not bed from bed 95
common and only became common because the arrangement of their parts
was conducive to survival. So, if (B) is interpreted as a claim about what has
always or for the most part happened throughout all time, the Empedoclean
opponent will simply deny its truth: One would only think (B) is true if
one were focused myopically on the current era and ignored the fact that
things did not always or for the most part come to be as they do nowadays.
However, if (B) is interpreted as a claim about what nowadays comes to be
always or for the most part, (B) will be true, but (C) will be false. For the
Empedoclean opponent claims that what nowadays happens always or for
the most part nonetheless comes to be by coincidence. In putting forward
B and C in support of (2), it is hard to see how Aristotle does not just baldly
beg the very question at issue.
Some commentators simply concede that he does so. For example,
Cooper holds that Aristotle’s argument relies on his view that the species
of organisms are eternal and thus did not come to be as Empedocles
hypothesizes.16 Other commentators push Aristotle’s question-begging
back a step. For example, Judson holds that Aristotle’s argument relies
on the claim that the proper account of the generation of animals must
make reference to the fact that the arrangement of teeth, for example,
“serves the life of the organism.”17 Thus, Aristotle shows that the
Empedoclean opponent’s account of the arrangement of teeth is impover-
ished because it maintains that the arrangement of teeth is merely coin-
cidentally beneficial. But, again, this issue lies squarely within the disputed
ground. Similarly pushing the question-begging back one step, Code
maintains that “all the opponent must admit is that it is always or for the
most part the case that if in human development a front tooth is formed,
then it is suitable for biting.”18 But, as we have seen when examining
Premise (4), the Empedoclean opponent denies the existence of natures
above the level of the four elements, and thus thinks that an animal is
simply a coincidental arrangement of earth, air, fire, and water, and there is
no robust, non-arbitrary kind “Human” by which to distinguish certain
occurrences of tooth formation from others. Further, the Empedoclean
opponent might press that even if the designation of a kind is not wholly
arbitrary, the designation is completely ad hoc: If you can designate the
kind “Human” in part by reference to the fact that things in this category
have teeth suitable for biting and chewing, then it will be true, in a trivial
sense, that tooth formation in a human regularly leads to teeth suitable for

16
Cooper 1987: 246–253. See also Charlton 1970: 123, who claims the argument is “inconclusive.”
17
Judson 2005: 352. 18 Code 1997: 131.
96 margaret scharle
biting and chewing. But the kind of regularity within the “kind” desig-
nated in this ad hoc fashion is not the kind of regularity that (C) claims is
not the result of coincidence.
In what follows, I will suggest that Aristotle uses the uncontroversial
shared ground of (A) as the fulcrum of his argument against the
Empedoclean opponent: Because the Empedoclean opponent will uncon-
troversially agree to (A), he will have to concede (C) and (B), and thus (2).
In order to appreciate the impact of this strategy, we first have to under-
stand the origin of the challenge posed in the statement of the problem.

II The origin of the problem

In Physics ii.4–6, Aristotle argues that both thought and nature enjoy a
priority over chance: “spontaneity and chance . . . are posterior (ὕστερον)
to intelligence and nature” (Phys. ii.6, 198a9–10), in the sense that move-
ments brought about by either thought or by nature are teleological, while
those that might have been brought about by either, but in fact were not,
are coincidental (Phys. ii.5, 196b22–23). For example, “the stone that struck
the man did not fall for the sake of striking him; therefore it fell sponta-
neously, because it might have fallen by the action of an agent and for the
sake of striking” (Phys. ii.6, 197b29–32). Here the fall is coincidental,
because it might have fallen by thought – for the sake of striking the
man – but in fact did not.
It is just this priority of nature and thought over chance that misleads the
Empedoclean opponent in the statement-of-the-problem passage. The
opponent’s case of rain accidentally resulting in corn growth – “Zeus
rains, not in order to make the corn grow” (Phys. ii.8, 198b18–19) – is
perfectly analogous to the ii.6 stone case: replacing “striking the man” with
“corn growth,” Aristotle can give the same description: “it did not fall for
the sake of [growing corn]; therefore it fell spontaneously, because it might
have fallen by the action of an agent [i.e. Zeus] and for the sake of [growing
corn].” Coincidence for Aristotle is the lack of either thought or nature, and
in the case of both the stone and the rain, the coincidence is because of the
lack of thought, emphasized by the opponent’s appeal to “Ζεὺϛ” in stating
the problem. The opponent seizes on what we might call the “lack-of-
agency” model of coincidence (that is, coincidence by lack of thought) to
formulate his understanding of the “lack-of-nature” coincidence suppo-
sedly exhibited in the case of animals. The opponent asks: “Why
then should it not be the same with the parts in nature?” (Phys. ii.8,
199b24), and ultimately with the generation of whole animals (Phys. ii.8,
 Man from man but not bed from bed 97
198b29–32)? On my reading, the winter-rain argument responds to this
question by claiming that the “lack-of-agency” model is the wrong model
of coincidence.19 In the winter-rain argument, Aristotle re-orients the
Empedoclean opponent by offering a “lack-of-nature” case of coincidence
(summer rain) and a natural case of teleology (winter rain) to serve as the
exemplars of the model on which to reconsider the opponent’s case of
animal generation. Given that the case of animal generation fits the model
exemplified by winter rain, not summer rain, the opponent should con-
clude that animal generation (and thus the development of functional
teeth as part of that generative process) is similarly teleological.
This reading demands a distinction between what I am calling the
“natural” model of teleology and what I am calling the “agency” model of
teleology. Physics ii.4–6 claims that regularity is a hallmark of both models:
Both things that come to be by nature and things that come to be by thought
come to be regularly, while things that come to be by coincidence do not
(Phys. ii.5, 196b10–11, 20–21; 197a3–4, 20, 31–32; 197a33–35). Agents, by the
teleological direction of their thought, and natural things, by the teleological
direction of their natures, both reliably produce their respective ends, unless
something impedes: The builder regularly builds houses, and the plant
regularly grows roots. But there is something distinctive about the type of
regularity exhibited by natural things at the level of generated wholes, a
distinction Aristotle makes in his argument in Physics ii.1 against his materi-
alist predecessors (including Empedocles): “man is born from man but not
bed from bed” (193b9–10). The type of regularity exhibited by natural things
is that of continuous generation, for nature is an internal principle of
“production” (ποιήσεωϛ) (Phys. ii.1, 192b30). And this is the type of
regularity with which winter rain comes to be.

III Regularity on the natural model of teleology

Aristotle’s refined view, as my previous paper argues in depth, is that water
moves into its natural place naturally and teleologically only upon being
generated by the sun.20 Although it is true that water moves into its natural

19
Pace Cooper 1987: 245n.5, who maintains that the art–nature analogy plays no role in the winter-rain
argument, and therefore claims, given the fundamental importance of the winter-rain argument,
that one must “reject the suggestion” that the analogy is “central and fundamental to Aristotelian
natural teleology.”
20
For the full argument for this view, see Scharle 2008: 150–181. As I mention there (n.70), the sun is
the efficient cause of three teleological cycles of evaporation and condensation, and water’s coming
to be and movement as part of these cycles is imitative of the divine and teleological.
98 margaret scharle
place whenever it falls unobstructed (whether as rain or otherwise), winter
rain is the only rainfall that forms a cycle that imitates the divine.
Metaphysics xii.10 (1075a19–22) suggests that elemental movements are
teleologically directed towards the prime mover via their imitation of the
heavenly bodies: Water’s rectilinear motions can imitate the circular
motion of the heavenly bodies, which in turn imitates the activity of the
prime mover. The only way for a rectilinear motion to imitate circular
motion is for it to be part of a cycle that “reverts again to the beginning . . .
Hence it is by imitating circular motion that rectilinear motion too is
continuous” (Cael. ii.10, 337a7). The rectilinear movement of water in
winter rain, then, is imitative because (expressed by ὥστε) it occurs on the
heels of another rectilinear movement (of air) with which it composes a
cycle. Meteorology i.9 confirms that winter rain composes part of an
imitative cycle of generation: “[W]e get a circular process that follows the
course of the sun . . . When the sun is near [i.e. in the summer] the stream
of vapor flows upwards; when it recedes [i.e. in the winter], the stream of
water flows down [as winter rain]” (346b35–347a1; see also Meteor. ii.4,
359b34–360a3).21 The fact that water’s natural movement is efficiently
caused by the sun ensures that the natural movement of water will occur
on the heels of air’s upward rectilinear movement. Only in this way does
water’s movement imitate circular movement, which imitates the prime
mover. In contrast, whereas winter rain is properly caused by the recession
of the sun, which ensures its coordination with air to form an imitative
cycle, summer rain is caused by the “recoil” (ἀντιπερίστασιϛ) of hot
and cold (Meteor. i.12, 348b8–10, 349a5–9) and is thus “violent” (ὕδατα
λαβρότερα, Meteor. i.12, 348b11, 348b23, or ῥαγδαῖα, 349a7). Summer rain
is therefore neither imitative nor teleological – a rectilinear dead-end.22
In general, Metaphysics xii.10 shows that things are in “joint-
arrangement” (συντέτακται) (Metaph. xii.10, 1075a15) with one another
to the extent to which their activities approximate that of the prime mover
by their imitating the circular motion of the heavenly bodies: Winter rain

21
See also Metaph. ix.8, 1050b28ff.
22
Most recently, Leunissen 2010: 30n.57 has resisted my interpretation of winter rain in Physics ii.8
because she “take[s] it that the crux for Aristotle in Ph ii.8 is to show that regular natural phenomena
have regular beneficial outcomes due to the fact that nature is an efficient cause that acts for the sake of
something. Under Scharle’s interpretation, however, it is the retraction of the sun in the winter that is
the efficient cause that makes the water return to its natural place, but this efficient cause itself
never – neither in the winter, nor in the summer – acts for the sake of this outcome.” My response is
to deny that natures are always efficient causes (see Scharle 2008: 171–173) and to maintain that the
stated conclusion of Physics ii.8 is that nature is for the sake of an end (198b17–18, 198b10–11, 199a7,
199b32–33), not that nature is an efficient cause for the sake of an end.
 Man from man but not bed from bed 99
imitates the circular motion of the heavenly bodies by moving rectilinearly
on the heels of air’s upward movement in the summer (thereby forming the
generative cycle of water and air), while animals and plants do so by
generating another of their kind (GC ii.10, 336b27–337a8).23 Aristotle
argues, “coming-to-be and passing-away will, as we have said, always be
continuous (συνεχὴϛ)” (GC ii.10, 336b25), for God “fulfilled the perfection
of the universe by making coming-to-be uninterrupted; for the greatest
possible coherence would thus be secured to existence, because that
coming-to-be should itself come-to-be perpetually is the closest approx-
imation to eternal being” (GC ii.10, 336b32–337a1). We have now arrived at
a precise articulation of the type of regularity exhibited on the natural
model of teleology: that coming-to-be itself comes-to-be regularly. This
regularity is initiated by the sun – the sun’s circular motion ensures not
only that sublunary elemental transformation and locomotion will come to
be in an imitative pattern, but also that animals generate another individual
of their own kind in an imitative pattern: “since the upper movement is
cyclical, the sun moves in this determinate manner; and since the sun
moves thus, the seasons (ὧραι) in consequence come-to-be in a cycle, i.e.
return upon themselves; and since they come-to-be cyclically, so in their
turn do the things whose coming-to-be the seasons initiate [e.g. plants and
animals]” (GC ii.11, 338b3–5; see also Cael. ii.3, 286a13–286b2).24 And it is
important for my interpretation that Aristotle has this relationship
between the sun and generable things clearly in mind in Physics ii itself
when he says, “man is begotten by man and by the sun as well” (Phys. ii.2,
194b13).
So far I have argued that the same type of regularity is exhibited in both
elemental and animal generation. I now want to show that this regularity
comes in degrees according to an entity’s rank on the scala naturae.
Generation of Animals ii.1 lays out the ranking as follows: Beginning at
the top of the scala naturae, Aristotle places the heavenly bodies (both
living and eternal), then living things (living but not eternal), and finally
the sublunary elements (neither living nor eternal) (731b24–732a1). This
passage understands rank according to goodness, while the texts I consider
next suggest that ranking is alternatively calibrated according to the degree
of regularity exhibited in the entity’s imitative activity. These two descrip-
tions of ranking ultimately come to the same thing: Since the prime mover
23
In fact, On the Soul ii.4 maintains that, for living things, generating another of their own kind is “the
most natural” (φυσικώτατον; 415a26) act to which all other natural activities are subordinate (415b1–2).
24
On my view, the seasons initiate coming-to-be in the sense that animals make use of the seasons for
the sake of growth. See Scharle 2008: 161–165.
100 margaret scharle
is the best thing in the cosmos, the more closely something approximates
its activity, the better it is; and the more closely it approximates the prime
mover, the more regular and uninterrupted its activity.
The introduction to the Meteorology confirms that the sublunary ele-
ments exhibit a “regularity less (ἀτακτοτέραν) than” the heavenly bodies
moving in a circle (Meteor. i.1, 338a20–b4). This contrast in degree of
regularity is also at work in Metaphysics xii.10’s household analogy: Taken
together with the introduction to the Meteorology, the analogy suggests that
the heavenly bodies are to the sublunary elements as the freemen are to the
slaves and beasts. Thus, we should expect that the heavenly bodies “have
least license to act as they chance to, but all or most of what they do is
arranged (τέτακται),” while the sublunary elements “can do a little towards
what is communal, but act mostly as they chance to” (1075a19–22).25
Although the circular motion of the heavenly bodies always imitates the
activity of the prime mover, not all sublunary elemental movements
imitate the activity of the prime mover, but only those movements, such
as winter rain, that take the sun as their efficient cause.
On the Soul confirms that this regularity varies in degree even within the
stratum of living things: “since then no living thing is able to partake in
what is eternal and divine by uninterrupted continuance (for nothing
perishable can forever remain one and the same), it tries to achieve that
end in the only way possible to it, and success is possible by the more and
the less (τὸ μὲν μᾶλλον τὸ δ’ ἧττον)” (ii.4, 415b4–6). Thus, Aristotle seems
to think that the heavenly bodies exhibit the highest degree of the type of
regularity at issue, followed by animals and plants, which exhibit a lower
degree of regularity (and greater and lesser degrees within this stratum),
and, at the lowest level, sublunary elements exhibit the lowest degree of
regularity.
Physics ii offers two anti-materialist arguments that clearly capitalize on
the degrees of regularity Aristotle articulates. First, Physics ii.4 (196a25–b4)
argues against Democritus’ view that animals and plants come to be by
mind or nature and not by chance, but that the heavenly spheres did come
to be by chance. Aristotle argues that the movements of the heavenly
spheres are much more regular than the generation of plants and animals,
so if Democritus agrees that the regularity with which plants and animals
come to be cannot be due to mere chance, a fortiori it is so in the case of the
heavenly spheres.26 Democritus cannot consistently maintain that the

25 26
Trans. Sedley 2000: 328. See also PA i.1, 641b10–23.
 Man from man but not bed from bed 101
generation of animals is not due to chance while arguing that the move-
ments of the heavenly spheres are.
Physics ii.8 itself offers an analogous argument, which I will call the
“olive-headed-vines argument,” found just thirty lines down from the
winter-rain argument and levied against the same Empedoclean opponent:
Again, in plants too we find that for the sake of which, though the degree of
organization (διήρθρωται) is less. Were there then in plants also olive-
headed vine-progeny, like the “man-headed ox-progeny,” or not? An absurd
suggestion; yet there must have been, if there were such things among
animals. (199b10–13)
Once again, Aristotle takes as a premise a claim about the degrees of
regularity exhibited in the generation of things at different levels of the
scala naturae: The generation of animals is more regular than the genera-
tion of plants, so that if the generation of animals is due to chance, as
Empedocles suggests, a fortiori the generation of plants is, in which case
plant generation should exhibit the botanical analogue of man-faced ox-
progeny (e.g. olive-headed vines).27 But Empedocles never mentions any-
thing like olive-headed vines, which shows his inconsistent application of
the notion of chance.

IV Reconsidering the winter-rain argument

I now submit that Physics ii.8’s argument for Premise (2) is yet another
instance of the a fortiori argument pattern exhibited in both Physics ii.4’s
argument against Democritus and the ii.8 olive-headed-vines argument
against the Empedoclean opponent. In fact, Aristotle suggests this in the
introductory passage directly preceding the statement of the problem:
Here Aristotle complains that his predecessors posit the existence of causes
like love or nous that could operate teleologically, but fail to use them in the
proper explanatory contexts. As we have seen in the a fortiori arguments,
they “touch on it” to explain some phenomena, but fail to use it to explain
the phenomena that require it even more (καὶ γὰρ ἐὰν ἄλλην αἰτιαν
εἴπωσιν, ὅσον ἁψάμενοι χαίρειν ἐῶσιν; Phys. ii.8, 198b13–16). The origins
of this complaint can be found in Plato’s Phaedo (97b–99d), where Socrates
expresses disappointment in Anaxagoras’ positing nous, but failing to make
use of it as a teleological cause of phenomena that require one.
Recall the sub-argument at issue in the first argument of ii.8:

27
Simplicius’ commentary on Physics ii agrees with Alexander that this is another a fortiori argument.
102 margaret scharle
A. (Accepted Premise): “It does not seem to be from luck or from
coincidence that it rains often in winter, but if in the dog-days; nor
that there are heatwaves in the dog-days, but in winter.” (There is a
regularity in winter rain, and in summer heat waves, which cannot be
due to coincidence.)
B. (Disputed Premise): “These things [e.g. animals] and all things that
are by nature, come to be in this way either always or for the most
part.” (Animals and things that are “by nature” come to be with the same
type of regularity as winter rain and summer heatwaves.)
C. (Disputed Premise): “Nothing from luck or chance does [i.e. nothing
from luck or chance comes to be in this way either always or for the
most part].” (Nothing that comes to be by chance comes to be with the
same type of regularity as winter rain and summer heatwaves.)
2. (Conclusion): “These things [e.g. animals] are not able to [come to]
be from coincidence or from chance.” (Therefore, animals and things
that happen “by nature” cannot be from chance.) (A, B, C)
While the Physics ii.4 argument against Democritus relied on the claim
that the heavens display a greater degree of regularity than animal- and
plant-generation, and the ii.8 olive-headed-vines argument relied on the
claim that there is a greater degree of regularity exhibited in animal
generation than in plant generation, here in the ii.8 winter-rain argument
Aristotle suggests that animal generations exhibit the same type of regu-
larity as elemental generations in the form of winter rain. And just as the
Physics ii.4 argument showed Democritus mistakenly to attribute to
chance the heavenly motions, even though they are more regular than
the generations he does not attribute to chance, and just as the olive-
headed-vines argument shows the Empedoclean opponent mistakenly to
attribute to chance animal generations, even though they are more regular
than the plant generations, he does not attribute to chance (at least in so far
as he does not posit the existence of anything like olive-headed vines), so
too this winter-rain argument shows the Empedoclean opponent mista-
kenly attributes animal generations to chance even though they exhibit the
same type of regularity whose degree is the same as (if not more than) the
elemental generations exhibited in winter rain, which the opponent does
not attribute to chance. Not only does my interpretation make better sense
of the dialectic between Aristotle and his Empedoclean opponent, it reveals
that Aristotle thought two of his formidable materialist opponents –
Democritus and Empedocles – fell prey to the same kind of error, simply
from different ends of the scala naturae.
 Man from man but not bed from bed 103
So understood, Premise (A) supports (C) directly by showing that the
type of regularity exhibited in winter rain – a kind of continuity of
generation – is the kind of regularity that cannot be chalked up to chance.
And if the regularity in (B) is this same type of regularity that (C) claims
cannot be due to chance, then neither can the regularity by which animals
are generated be chalked up to chance. By accepting (A), then (C), then
(B), the Empedoclean opponent reaches conclusion (2).
But why would the Empedoclean opponent admit that there is any
continuity at all in animal generation such that it displays the same type of
regularity as winter rain? Even though Aristotle thinks the species are
eternal, and thereby disagrees with the Empedoclean opponent’s version
of “natural selection,” they both share the view that animal generation is
itself continuous: Even the Empedoclean opponent thinks that the animals
that generate do so continuously, for he claims that the ones whose parts
were unsuitably arranged die and continue to die (Phys. ii.8, 198b29–33). As
Aristotle highlights in the next set of arguments, Empedocles thought that
animals are generated from seeds: Aristotle even quotes Empedocles’ poem
that “what was ‘undifferentiated first’ was seed (σπέρμα)” (Phys. ii.8,
199b8–9). As they do for Aristotle, Empedocles’ seeds serve as the link
between one generation and the next, and can thereby suggest that he is
committed to a kind of continuity of generation, even if it is not eternal.
Aristotle wants to point out that in so far as the Empedoclean opponent
thinks this, there is, after all, a regularity found in animal generation that
cannot be simply coincidental, for this regularity is the same type of
regularity – a kind of continuity in generation – that the Empedoclean
opponent agreed could not be due to mere coincidence in the case of
winter rain. Water’s falling in winter (when the sun recedes) ensures (for
the most part) its subsequent evaporation (when the sun returns); in so
doing, winter rain metaphorically “sows the seeds” of the next winter’s
rainfall. For although it may rain in the summer, this kind of rain does not
form a generative cycle with air that will lead to another iteration of rain
the following winter. Likewise, although men may give birth to ox-faced
progeny, those are not the ones that will, in turn, generate. And even if
there are plenty of episodes of coincidental rain that randomly come to be
here and there out of season, and even if there are plenty of random ox-
faced monstrosities that come to be, nonetheless the rain that forms part of
the generative cycle with air will (for the most part) come to be again, and the
animals that give birth to animals that survive and generate will (for the most
part) generate again. This is the distinctive regularity found on the natural
model.
104 margaret scharle
At this point one might worry that my interpretation pushes Aristotle’s
question-begging back just one step, in a way that is similar to Code’s
interpretation. It is worth returning to Code’s interpretation in order to
distinguish mine. As I note above, Code’s interpretation requires that “the
opponent must admit that it is always or for the most part the case that if in
human development a front tooth is formed, then it is suitable for biting,”
and more generally, “for any given natural kind K, tooth formation of a K”
regularly results in functional teeth.28 I argued that this assumption would
violate Premise (4), understood as neutral between Aristotle’s own position
in which there are robust kinds above the level of the elements, and the
Empedoclean opponent’s view in which there are not. By contrast, my
interpretation of the argument does not violate Premise (4), interpreted
neutrally. I argue that Aristotle first sets out the example of winter rain as
the natural model of teleology to which the Empedoclean opponent’s own
version of animal generation conforms. The opponent admits that the
animals that generate do so continuously and thus must be explained
teleologically, just as winter rain must be. Once Aristotle has identified
this set of animals – the ones that generate according to the natural model
exemplified in winter rain – he can then say that in those things tooth
formation regularly leads to functional teeth, while remaining neutral as to
whether it is the elemental natures or the natures of the animals that are
responsible for the continuity. That is to say, Aristotle thinks he must first
establish the need for teleological explanation at the level of the generation
of whole animals – for this is the level that exhibits the distinctively natural
pattern of teleology – and then consider the teleology of the formation of
their parts as part of the generative pattern.
In my reading, all the Empedoclean opponent needs to accept is that
winter rain exhibits a regularity that cannot be due to coincidence, and that
animal generation exhibits the same type of regularity. Thus, animal
generation, like winter rain, requires a teleological explanation, and given
that these phenomena are by nature (as opposed to by thought), nature will
be the teleological cause. To accept this, the Empedoclean opponent need
not accept (i) Aristotle’s particular teleological explanation for the regular-
ity, although Metaphysics xii.10 later makes an argument to this end, nor
must the opponent accept (ii) Aristotle’s appeal to natures above the level
of the elements. If the Empedoclean opponent thinks he can show that
elemental natures are robust enough to explain the continuity of animal
generation, so be it (as far as this argument is concerned).29

28 29
Code 1997: 131. Pace Judson 2005: 349.
 Man from man but not bed from bed 105
That said, Aristotle may think that he also has given the Empedoclean
opponent reason to doubt his ability to explain the continuity of animal
generation by appeal solely to the four elemental natures. Although
Aristotle does not spell this out in the winter-rain argument, if he can
convince the Empedoclean opponent that winter rain displays not only the
same type of regularity, but that winter rain is less regular than animal
generation (as Aristotle himself thinks), then he would have given the
Empedoclean opponent reason to doubt that elemental natures could
explain the continuity of animal generation. In other words, if elemental
processes (exhibited in winter rain) are less regular than biological gen-
erative processes, the former could not explain the latter. But, again,
Aristotle need not argue this in order to reach his conclusion.

V Reconsidering Physics ii.8’s place in Physics ii

With my interpretation that winter rain is teleological in its own right, I am
in the unusual position of claiming that Aristotle extends natural teleology
even to the level of the elements. This novel approach allows me to
construct an argument I believe to be more interpretatively satisfying and
more dialectically satisfying than other existing readings. In addition to
this, my interpretation suggests a unity for the set of arguments in Physics
ii.8 and for the whole of Physics ii. Let me conclude with a brief sketch of
this unified picture.
Aristotle’s conception of nature in Physics ii is presented as opposed to
both a Platonic tradition, in which nature is art, and a pre-Socratic one, in
which nature is not properly distinguished from chance. Rejecting the pre-
Socratic tradition, both Plato and Aristotle insist that the world’s order and
structure are explained teleologically and are not due to chance; however,
Aristotle disagrees with Plato, who holds that nature would have to be art
in order to explain such order. Overall, Physics ii insists that art and nature
are analogous in being the source (ἀρχή) of order, yet distinct in that
nature is an internal source of this order, while art is an external one (Phys.
ii.1, 192b9–34). In arguing against his pre-Socratic predecessors’ concep-
tion of the natural world as operating by chance, however, Aristotle’s
nuanced conception of the relationship between art and nature is at once
a help and a hindrance.
Consider Aristotle’s response to the pre-Socratic predecessors’ argument
in Physics ii.1 for the claim that nature is matter (193a13–25). Aristotle
depicts the pre-Socratics as taking lessons about nature from artificial
examples: Just as matter is the nature of the bed, so too, they think, is
106 margaret scharle
matter the nature in the natural cases of bronze, bone, and wood. In
modeling nature after art, the predecessors relegate much of nature to
chance, for if solely matter (earth, air, fire, and water) is nature, then
everything else is “its affections, states, and dispositions” (Phys. ii.1,
193a24–25). But in response to this materialist challenge, Aristotle argues
that form “is more (μᾶλλον) nature” than matter (Phys. ii.1, 193b6) by
sharply distinguishing art from nature: “man is born from man but not bed
from bed” (Phys. ii.1, 193b9–10). This statement reiterates Aristotle’s
introductory claim that nature is an internal principle of production
(ποιήσεωϛ; Phys. ii.1, 192b29).
This sharp distinction between nature and art gets blunted by Physics
ii.4–6’s insistence, against his pre-Socratic predecessors, on nature and
thought’s allied priority over chance (Phys. ii.6, 198a9–10). I have argued
that it is this shared priority of nature and thought that led the
Empedoclean opponent of Physics ii.8, once again, to mistakenly model
nature after the agency we find in art. Here we find the opponent erro-
neously attempts to model a “coincidental” development of animals
(which Aristotle goes on to argue against) on the lack-of-agency coinci-
dence between rain and corn growth. Through the winter-rain argument,
Aristotle restores nature’s priority over chance by returning to Physics ii.1’s
sharp distinction between art and nature: “man is born from man, but not
bed from bed.” The distinctive regularity by which natural things generate
could not be found in the case of Zeus’s raining for the sake of growing
crops, and the example of winter rain is meant to highlight the distinctive
feature of natural generation in terms that are mutually acceptable to both
Aristotle and his opponent. Aristotle is then in a position to move the
opponent, through an a fortiori argument, to see that animals generate in
this distinctively natural way as well.
This unified reading of Physics ii makes sense of Physics ii.8’s second
argument to the conclusion that nature is prior to art (199a9–33) as an
attempt to block, once and for all, the opponent’s tendency to take lessons
about nature from artificial examples, and to be misled, thereby, into
thinking that nature operates by chance. In so doing, Aristotle brings
together the arguments of Physics ii.1 (that form is more nature than
matter) with the first argument of Physics ii.8 (that nature is for the sake
of something) to conclude that “form must be the cause in the sense of that
for the sake of which (ἡ οὗ ἕνεκα)” (Phys. ii.8, 199a33); in other words, that
form is for the sake of form, or “man is born from man.”
 chapter 6

In defense of the craft analogy: Artifacts

and natural teleology
Charlotte Witt

In the Physics Aristotle distinguishes artifacts from natural beings, noting

that the latter have an internal origin of motion and rest, which artifacts
lack. Still, it is striking that artifacts appear frequently in the passages of the
Physics that introduce ideas like matter and form, which suggests that
artifacts are ontologically similar to natural substances in some important
respects.1 Further, artifacts figure centrally in Aristotle’s arguments for
natural teleology, which also suggests that he finds significant similarities
between artifacts and natural beings. Nonetheless, many commentators
argue that artifacts do not even count as substances for Aristotle, which
must mean that artifacts and natural beings are of a different ontological
type and radically dissimilar. Some think that artifacts lack the unity of true
substances, being only accidentally one (Kosman 2013); others think that it
is an artifact’s inability to reproduce that is key (Katayama 1999); and still
others think it is the fact that artifacts are not self-regulating (Shields 2008)
and lack intrinsic ends (Shields 2008 and Katayama 2009). The idea that
artifacts and natural beings are radically dissimilar casts doubt on the role
that artifacts play in Aristotle’s argument for natural teleology, perhaps
leading some scholars to downplay the significance of these arguments or
to omit them altogether.
In this chapter I focus on the idea that artifacts lack intrinsic ends, since
this claim is directly relevant to the legitimacy of Aristotle’s use of artifacts
in his arguments for natural teleology. For, if artifacts lack intrinsic ends it
is difficult to see what role they could play in an argument meant to
establish that natural beings have intrinsic ends. But it turns out that

1
Of course, Aristotle’s use of artifacts in relation to natural beings (and substances) is found in many
texts. For a discussion of texts other than the Physics, see Katayama 1999 and Shields 2008. In this
chapter I am interested in the relatively narrow question of what Aristotle’s argument for natural
teleology suggests about the ontological status of artifacts.

107
108 charlotte witt
Aristotle’s use of artifacts in his argument for natural teleology in
Physics ii.8 casts doubt on the claim that artifacts lack intrinsic ends in
the relevant sense. Indeed, I think that one of Aristotle’s arguments, the
argument from mistakes, presupposes that artifacts have intrinsic ends or
proper functions, a view we find him explicitly endorsing in another text.
But if Aristotle does make a distinction between the intrinsic ends or
proper functions of an artifact and its possible uses, then it is perfectly
appropriate to use an analogy between artifacts and natural beings in
arguing for intrinsic ends in nature, or natural teleology. Moreover,
Aristotle’s legitimate use of artifacts in his argument for natural teleology,
in turn, casts doubt on the claim that artifacts are not substances, and are
either of a completely different ontological type or not beings at all.

I Introduction: The craft analogy

Aristotle’s discussion of nature and natural teleology in the Physics opens
with a distinction between artifacts and natural beings. Things that exist
by nature have an inner origin of motion and rest, which includes not
simply locomotion, but also growth and diminution and alterations. “On
the other hand, a bed and a coat and anything else of that sort, qua
receiving these designations – i.e. in so far as they are products of art –
have no inner origin of motion and rest” (192b16–18). It might be
reasonable to expect that, if this were a basic ontological distinction,
artifacts and art would both recede from the subsequent discussion of
nature and natural teleology. Or, at the very least, we might think that
artifacts would not play a central role in Aristotle’s explanation of, and
argument for, natural teleology. Alternatively, we might expect Aristotle
to warn us of the ontologically inadequate and foreign character of his
favorite examples. But it turns out that Aristotle uses craft examples
frequently, and draws freely on the structural likenesses between nature
and craft in his discussion of nature and in his argument for natural
teleology. Borrowing the term from David Sedley (2010), I will call both
Aristotle’s use of craft examples and his references to specific similarities
between craft and nature Aristotle’s “craft analogy.”
To modern readers the craft analogy might seem mistaken, even mis-
leading. In particular, Aristotle’s use of the craft analogy to argue for
natural teleology might seem confused. After all, absent a Platonic cosmic
craftsman, it just seems wrong to argue from the existence of function and
purpose in artifacts to the existence of function and goal-directedness in
natural beings, or to think that processes of craft creation will tell us
 In defense of the craft analogy 109
anything useful about processes of natural generation.2 Some scholars
think that Aristotle used the craft analogy simply for pedagogical reasons
because crafts and artifacts are more familiar to us (or to his audience) than
nature and natural beings, but he didn’t really think there were important
similarities between the two that could be legitimately used in an argument
by analogy for natural teleology. I discuss the pedagogical interpretation of
Aristotle’s craft analogy and its implications further in section II.
Alternatively, one might take the craft analogy at face value, which is
that Aristotle thought there were important similarities between craft
and nature that he could legitimately exploit in his argument for natural
teleology. If that is right, then Aristotle could not have thought that
artifacts were ontologically in a different category from natural beings, or
not beings at all. In section III I focus on Aristotle’s arguments for
natural teleology in Physics ii.8, and I develop the idea that Aristotle’s
use of artifacts to explain and to justify natural teleology is cogent and
not misguided or mistaken. The basic assumption underlying Aristotle’s
use of artifacts in the argument for natural teleology is the mimetic
relationship between art and nature: the idea that art “imitates” nature.
Two features are particularly important. First, Aristotle points out that
both artistic creation and natural generation consist of a sequence of
stages, and that their ordering and the relationship of the stages to one
another are themselves ordered by their being for the sake of the end. A
second important feature shared by artistic creation and natural
generation is the possibility of mistakes or errors. Aristotle argues that,
like art and artistic creation, nature and natural generation are subject
to defect and error.
If then in art there are cases in which what is rightly produced serves a
purpose, and if where mistakes occur there was a purpose in what was
attempted, only it was not attained, so must it also be in natural products,
and monstrosities will be failures in the purposive effort. (199a31–b4)
But this argument and the notion of a mistake or a monstrosity pre-
suppose that both artistic and natural processes have intrinsic ends. To see
this, it is helpful to distinguish two points: first, whether or not a being
has an intrinsic end, and, second, where the intrinsic end originates or
comes from. Artifacts have intrinsic ends, with external origins in the
artist’s intentions and motions; natural beings have intrinsic ends, with
internal origins in that the end develops and unfolds through internal

2
See Sedley 2010 for a useful discussion of Platonic and Aristotelian teleology.
110 charlotte witt
processes.3 But if artifacts have intrinsic ends, then one reason for holding
that artifacts and natural beings are ontologically different, or that
artifacts are non-beings, is mistaken.
In section IV I address the apparent tension between the idea that
artifacts have intrinsic ends, and Aristotle’s definition of them as having
an external origin of motion and rest. If an artisan creates an artifact (and is
its external origin in that sense), then doesn’t the form or end have an
external origin? In what sense, then, does an artifact have an intrinsic end?

II The pedagogical interpretation of the craft analogy

A preliminary issue is whether the craft analogy is central to Aristotle’s
explanation and justification of teleology, and, if so, in what way?4 It is
useful to distinguish two questions. First, is the craft analogy central to
Aristotle’s explanation and justification of natural teleology? Since exam-
ples drawn from crafts, like building, and arts, like medicine, appear
throughout Physics ii.8 it is clear that Aristotle uses them in his exposition
of natural teleology. So much should be uncontroversial. But, we might
wonder, in what way are crafts important? How does Aristotle use them?
According to the pedagogical interpretation of the craft analogy, Aristotle
uses examples drawn from the crafts to illustrate points about causation
and teleology in nature and natural beings because they are more familiar
(or accessible) to us than natural beings and processes, and not because they
share any important ontological features with natural beings.
Aristotle’s extensive and careful use of the model of arts in his discussion of
the teleological aspects of natural things is due to the fact that technology is
more familiar to humans than other natures, not because he thinks that
technology is fundamental to nature. On the contrary he holds that art
imitates nature. (Johnson 2005: 9)5
In support of the pedagogical interpretation there are a number of texts in
which Aristotle articulates a distinction between principles that are familiar
to us or more knowable to us and principles that are more knowable by

3
It is worth noting that Aristotle often identifies the male parent (or the motions in the semen) as the
moving origin of animal reproduction. So, even in the case of natural beings, the original source of
the generative motion is external to the developing offspring.
4
For differing views on this issue, see Cooper 1982 and Sedley 2010.
5
“Aristotle himself uses the art model. It will become clear that Aristotle does not intend it to explain
anything, but merely to give us a notion of how causality itself works on the basis of facts most readily
available to ourselves (i.e. how intentional agency and the crafts can produce definite results)”
(M. R. Johnson 2005: 126).
 In defense of the craft analogy 111
nature or without qualification.6 Indeed, the Physics opens with this
observation on the method of the inquiry:
The natural way of doing this is to start from the things which are more
knowable and clear to us and to proceed to those which are clearer and more
knowable by nature; for the same things are not knowable to us and
knowable without qualification.
Aristotle illustrates this point using the example of universals (more know-
able to us) and individuals (more knowable by nature). It is worth pointing
out that the examples Aristotle uses to illustrate his distinction, namely
universals and individuals, fits very well with the subsequent discussion of
the principles invoked by the natural philosophers (and Plato) to explain
change. But it is not equally obvious how to apply Aristotle’s epistemic
distinction to the case of artifacts and natural beings, to art and nature,
since the paired items seem to be of the same ontological type. Artifacts and
natural beings are individuals; art and nature are universal principles.
Moreover, it is not intuitively obvious that artifacts are more familiar to
us than natural beings. Why and in what way is a bed, for example, more
familiar than a goat? And in what way is the production of a house or a
temple more understandable (to someone who is not a builder) than the
generation of an animal? Surely, animal reproduction would be better
understood by a farmer than how to design and build a temple. Finally,
the pedagogical interpretation holds that Aristotle’s argument for natural
teleology moves from the allegedly clear and easy example (artifacts) to the
confusing and difficult case (natural beings). But we don’t actually find this
direction of argument in the text. Instead, Aristotle’s exposition and
argument weave back and forth between artifacts and natural beings in a
free-flowing manner.
So it just isn’t clear how to apply Aristotle’s methodological principle –
of moving from what is more knowable to us to what is more knowable by
nature – to the case of artifacts and natural beings; nor is it clear that
Aristotle intends us to do so. This gives us reason to doubt the relevance of
Aristotle’s methodological principle to his use of the craft analogy in his
argument for teleology. Moreover, even if we are meant to read this text in
the light of Aristotle’s distinction between what is more knowable to us and
what is more knowable by nature, the pedagogical interpretation leaves
open the question of whether Aristotle thinks – in addition – that there are
important ontological similarities between artifacts and natural beings and

6
See e.g. Phys. i.1, 184a19–20; APr. ii.23, 68b35–37; Metaph. vii.3, 1029b3–12; and EN i.3, 1095b1.
112 charlotte witt
significant structural similarities between goal-directed processes in craft
and in nature to which he can appeal in his argument for natural teleology.
In other words, the pedagogical interpretation is compatible with the claim
that the craft analogy plays a central and legitimate role in Aristotle’s
explanation and justification of natural teleology. The pedagogical inter-
pretation, in itself, does not determine whether the craft analogy is used
simply for purposes of illustration and teaching or whether it is meant as a
piece of argument that points to, and presupposes, genuine and important
similarities between craft and nature. A close look at Aristotle’s argument
for natural teleology reveals that there are good reasons to endorse the
second alternative.

III Art imitates nature

In Physics ii.8 Aristotle makes several arguments for natural teleology. The
first (and most extensive) argument does not make use of the craft analogy
(198b1–199a8).7 But the two following arguments, plus an important dis-
claimer, do employ the craft analogy. Looking just at the number of
arguments in this chapter, it is fair to say that the craft analogy is central
to Aristotle’s case for natural teleology. First, Aristotle uses the craft
analogy to make what I will call the structural argument: the idea that in a
natural process of development, each stage in the sequence of stages is for
the sake of the end or goal of the process (199a9–19). Next, Aristotle uses
the craft analogy to make what I call the argument from mistakes to show
that there are intrinsic ends in natural beings and natural processes
(199a34–b8). In order for the argument from mistakes to work, however,
Aristotle must think that there are intrinsic ends in art and artistic processes
as well, since he employs the craft analogy to make his point. Finally,
Aristotle clarifies the connection between deliberation and goal-
directedness, again employing the craft analogy. While the clarification is
not part of a direct argument for natural teleology, it does support the
legitimacy of the craft analogy in the justification of natural teleology by
defusing an obvious criticism, namely that goal-directedness requires
conscious thought or deliberation (199b26–33).
These uses of the craft analogy in Aristotle’s argument for natural
teleology establish that he thought there were important similarities

7
For an interpretation of this argument, see Cooper 1982 and Leunissen 2010: 16–18. Cooper down-
plays the role of the craft analogy in Aristotle’s argument for natural teleology, and Leunissen is
skeptical about its significance in Aristotle’s argument.
 In defense of the craft analogy 113
between craft and nature that could be legitimately exploited to establish
the existence of ends in nature. Before entering into the details of the
arguments, it is useful to explore the conceptual basis for these similarities.
In the Physics, Aristotle says that art “imitates” (and sometimes
“completes”) nature (Phys. ii.2 194a21–22 and ii.8, 199a15–17). Initially it
might seem that the term “imitation” has connotations of Platonic
metaphysics, and these connotations might seem to reinforce the idea
that art and nature are radically, ontologically dissimilar, like Platonic
forms and particulars. It would then seem misguided for Aristotle to use
the craft analogy in his argument for natural teleology. However, it is
unlikely that Aristotle is referring to Platonic imitation, since Aristotle
claims not to understand Plato’s relationship between forms and particu-
lars, and, in particular, Aristotle seems puzzled by the metaphoric language
Plato uses in describing the relationship (Metaph. i.6, 987b13–14).
There are two contexts outside of the relationship between art and
nature in which Aristotle makes fairly extensive use of the notion of
imitation: in his theory of tragedy and the “mimetic arts,” and in his
account of elemental transformation. Aristotle says that elemental
transformation imitates the eternal movement of the celestial circles.
Johnson (2005: 147) is helpful in pointing out that we need not interpret
Aristotle as personifying the elements; rather, Aristotle holds that
elemental transformation resembles eternal celestial motion in that it is
continuous. In the Poetics Aristotle uses the term “mimesis” to express
what tragedy is: “the imitation of an action that is serious,” and he refers
to painting, tragedy, epic, and the like as “the imitative arts” (8, 1451a31).
“Tragedy is essentially an imitation not of persons but of action and life”
(8, 1450a16). However, it is likely that the term has an importantly
different connotation in the Poetics from that in the Physics. Halliwell
(2002: 26) distinguishes:
Aristotle’s specific conception of the mimetic arts (mimetic qua forms of
intentional representation) from his quite separate and general principle of
“the imitation of nature” by human productive craft (technê), the latter
counting as “mimetic” not in terms of representation but by virtue of
analogy or analogousness to natural processes of production.
Technê does not imitate nature in the sense of represent nature; rather, craft
imitates nature because it is like nature or resembles nature in certain
respects. These similarities or resemblances are what the craft analogy
depends upon, and in order to see what similarities are important for
Aristotle’s defense of natural teleology, we need to return to the Physics.
114 charlotte witt
Let’s begin with an argument from early in the Physics that uses the craft
analogy and the notion of mimesis in order to make an epistemological
point. This will allow us to appreciate that Aristotle uses the craft analogy
in contexts other than his argument for natural teleology, which underlines
the breadth of the analogy. In a discussion of whether the natural scientist
should study both matter and form, Aristotle draws a comparison with
what a doctor or a house-builder should know. This comparison begins
with the claim that “art imitates nature”:
But if, on the other hand, art imitates nature, and it is the part of the same
discipline to know the form and the matter up to a point (e.g. the doctor has
a knowledge of health and also of bile and phlegm, in which health is
realized, and the builder both of the form of the house and the matter,
namely that it is bricks and beams, and so forth); if this is so, then it would
be part of natural science also to know nature in both its senses [as matter
and as form]. (Phys. ii.2, 194a22–27)
Notice first that in this text the notion that art imitates nature has a
metaphysical meaning; both natural beings and artifacts are composites
of matter and form. Then notice that the shared hylomorphic structure of
artifacts and natural beings is taken to have an epistemological significance.
Aristotle uses the craft analogy and in particular the relationship of imita-
tion to further an argument and to establish a conclusion. After all, if the
imitation relationship did not hold between artifacts and natural beings,
then there would be no way to infer from some property that is true of
artifacts to the corresponding property in natural beings. We will see that
the craft analogy in Aristotle’s argument for natural teleology follows the
same pattern. First, there is a claim about a feature of crafts or craft
production, then the imitation relationship between craft and nature is
invoked (or clearly presupposed), and finally a conclusion is drawn about
natural beings or their development.
The structural argument (Phys. ii.8, 199a8–32) opens with the claim that
where an end exists, all the preceding steps are for the sake of that end.
Aristotle concludes the structural argument: “if, therefore, artificial pro-
ducts are for the sake of an end, so clearly also are natural products”
(199a18–19). Notice the direction of argument – it runs from a claim
about artifacts to a claim about natural beings. Aristotle assumes that we
will agree about the teleological structure of artistic processes, and, since art
imitates nature (explicitly stated at 199a15–17), the same structure will also
be found in natural processes: “The relation between the later to the earlier
items is the same in both artificial products and natural products” (199a19).
 In defense of the craft analogy 115
Aristotle devises a thought experiment about a house to express the
structural similarity between craft and nature: If a house were a natural
product, then it would be produced in the very same way by nature
(meaning via the same teleological sequence) as it actually is by art. He
adds that things made by nature, if they were made by art as well, would
come to be in the same way. Notice how strong Aristotle’s point is here; the
teleological structure of artistic productions and natural generations are the
same, not just similar in some respects.
Aristotle does not mention the obvious difference between the two
teleological sequences, namely that in an artistic creation the artisan is the
external origin of the motions that constitute the teleological sequence,
whereas in a natural generation after the initial motion (from the male
parent) the origin of the development is internal. This omission is all the
more noteworthy since Aristotle differentiates artifacts and natural beings
precisely on this point. It seems that, for Aristotle, the similarity in the
teleological structure of artistic creation and natural development is
independent of the causal origin of the respective processes. It seems
therefore that the teleological character of the artistic process with which
Aristotle compares the natural process is not simply determined by
the intentions of the artist alone. This is a very important point, to
which I return in section IV.
The argument from mistakes (Phys. ii.8, 199a32–b7) turns on another
similarity between artifacts and natural beings, namely that, like artistic
creations, natural generations sometimes go awry and the result is a mistake
(hamartia) or a monstrosity (terrata). The terminology here is interesting.
Like imitation (mimêsis), the notion of a mistake (hamartia) appears in
Aristotle’s theory of tragedy in the Poetics to refer to the kind of error made
by a tragic character like Oedipus, who unknowingly killed his father and
married his mother. The vocabulary of monstrosity (terrata), in contrast,
finds its home in Aristotle’s biological writings and refers primarily to the
deformed offspring that result from botched processes of reproduction.
Aristotle reasons from the presence of mistakes in artistic processes and
products to botched natural developmental processes and offspring. It
sometimes happens that “the literate man makes a mistake in writing
and the doctor pours out the wrong dose” (199a32–35). So, too, in nature
we find defective seeds or processes of development that result in
monstrosities like the “ox-progeny.” Hence in nature (as in art) there are
goal-directed processes that do not reach their goals or ends. We tacitly
recognize the existence of goals or ends in nature when we recognize
deformed or monstrous outcomes. The notion of a mistake or monstrosity
116 charlotte witt
presupposes that another outcome ought to have happened; it presupposes
that there is an end or form the product ought to have realized. But notice
that these presuppositions also hold in the case of artifacts. The notion of a
mistake also presupposes that there is a form or function the artifact (or
outcome of the artistic process) ought to have or ought to serve, even if in
fact the artifact is defective and cannot serve that function. The function of
the artifact has a normative component; it is not simply what it is doing or
what it can do. Consider the defective house that has lost its roof in a
hurricane; it no longer actually provides shelter for humans and animals,
and it no longer can serve that function. Nonetheless it is a house – albeit a
defective one. It – so to speak – ought to be able to provide shelter for
humans and animals. The argument from mistakes presupposes that
artifacts have proper functions, which are sometimes not realized owing
to a defective process of creation. Similarly, Aristotle reasons, natural
processes can also produce deformed animals and plants.8
In order for the argument from mistakes to be a cogent argument for
natural teleology, Aristotle must think that mistakes in art and in nature
presuppose that there are proper ends to be realized. And these proper
ends are not simply whatever the product can be used for or is being used
for.9 That is, in order that the notion of a mistake do the work that
Aristotle assigns it in this argument, he ought to distinguish the proper
end of an artifact, which it realizes through the appropriate process of
creation and is internal to it, from the many uses an artifact might be put
to. But if this is correct, then Aristotle ought to recognize that artifacts,
like natural beings, have intrinsic ends, or, in other terminology, proper
functions.
And, indeed, we find that Aristotle does recognize that artifacts have
proper functions in a text from the Politics. Aristotle contrasts kinds that
are by convention alone, like money, that have no intrinsic value or use-
value, with artifacts like a shoe that have a proper function. Money is by
convention alone; it has no intrinsic value or use-value.10
Others maintain that coined money is a mere sham, a thing not natural but
conventional only, because, if the users substitute another commodity for it,
it is worthless, and because it is not useful for any of the necessities of life,
and, indeed, he who is rich in coin may often be in want of necessary food.
(Pol. i.9, 1257b10–14)

8
For a discussion of deformed animals in Aristotle, see Witt 2012.
9
For a further discussion of the proper ends of artifacts, see Witt 2015.
10
Aristotle understands money to have exchange-value rather than use-value, although there are
tensions in his theory; see Meikle 1994.
 In defense of the craft analogy 117
The context for this comment is Aristotle’s discussion of wealth acquisition,
which has a proper use in the household, just as a shoe has a proper function
or use in relation to protecting the foot. Although shoes can be used to
acquire wealth or to barter, this is not their proper use (Pol. i.9, 1257a5–10).
Money, in contrast, has only an exchange value, and – unlike the shoe –
money has no proper function (oikeian kresin) at all. Aside from the
conventions governing money and its use, it has no intrinsic value or proper
use. So, not only is the distinction between what an artifact can be used
for and its proper function presupposed by the argument from mistakes,
but we also find Aristotle explicitly drawing that distinction in another text.
The idea that artifacts have intrinsic ends or proper functions appears to
be inconsistent both with the idea that artisans create them (and so their
ends have an external origin in the artist’s intentions) and with the idea that
artifacts serve the ends of their creators (and these ends are extrinsic to the
artifact). I address the first inconsistency in the next section. The second
inconsistency might mean one of two things. The idea might be that in any
given case an artifact serves the purpose of its creator (or user), or the idea
might be that all artifacts, taken collectively, serve the purposes of human
beings (or other tool-using animals), and in that sense they do not have
intrinsic ends. The first claim blurs the distinction drawn above between
the proper function of an artifact (in relation to which the language of
mistakes is appropriate) and what it is used to do or can be used to do. We
can use an ashtray as a hammer, but it is still an ashtray after all. Similarly,
we can use a horse to pull a plow, but that is not its natural function. This is
the sense in which artifacts, like natural beings, have intrinsic ends or
proper functions. The second claim, that artifacts taken together serve the
collective purposes of human beings (or other tool-using animals) is
correct, but it is also compatible with the idea that artifacts have intrinsic
ends in the sense specified above, i.e. proper functions. And this is the sense
required for the argument from mistakes for natural teleology to be cogent.

IV Art does not deliberate

Recall that Aristotle differentiates between artifacts and natural beings
because artifacts have an external origin of motion and rest in contrast to
the internal principle that natural beings have. Animal reproduction and
artifact production seem to differ on this point; the developing animal
embryo has an internal source of change whereas each stage in the produc-
tion of the artifact is shaped by the activities of the artisan under the
guidance of her intentions. This seems to be a crucial point of difference
118 charlotte witt
between artifacts and natural beings, and one that is directly relevant to the
issue of intrinsic ends. The artisan’s individual psychological state (her
intention) is the external origin of the movements that produce the artifact
with its proper function, and this seems to be a significant and relevant
difference between artistic production and natural generation, namely that
the intrinsic end of an artifact has an external origin in the artisan’s
intentions. The craft analogy falters due to this difference, which seems
to create a fatal ambiguity in the notion of an intrinsic end. That an artifact
has a particular proper function is causally dependent upon the individual
psychology of the creator. In contrast, that the natural being has a parti-
cular proper function or intrinsic end is not causally dependent upon an
external source in the same way.11
This is a reasonable criticism of Aristotle’s use of the craft analogy, since
it points to another apparent ambiguity in the notion of an intrinsic end.
And it is a criticism that resonates with those contemporary readers of
Aristotle who think that Aristotle’s craft analogy is confused, since the idea
of creating an object with purpose or function makes no sense independent
of the individual psychology and intentional motions of an artisan.
However, in a few difficult comments, Aristotle eliminates the apparent
ambiguity by explicitly bracketing the process of artistic deliberation and
individual psychological intention from the craft analogy.
It is absurd to suppose that purpose is not present because we do not observe
the agent deliberating. Art does not deliberate. If the ship-building art were
in the wood, it would produce the same results by nature. If, therefore,
purpose is present in art, it is present also in nature. (Phys. ii.8, 199b27–30)
One way to interpret this comment is that Aristotle means that a true
artisan does not need to deliberate about how to create an artifact; she is so
adept and practiced that she simply acts without deliberation or reflection.
Indeed, Aristotle remarks earlier in the chapter that non-human animals
make things without art, inquiry, or deliberation (199a20–21).
However, it is not likely that this is what Aristotle means, because in
other contexts he emphasizes the process of means–ends reasoning used by
human artisans. It is more likely that Aristotle is directing our attention to
the central role of craft knowledge as what is responsible for the existence of
the artifact, an entity with a proper function or intrinsic end. That “art

11
It is worth remarking again that in one sense a natural being does have an external origin of motion,
namely the male parent. But the developing natural being is not causally dependent on the male
parent throughout the process of development as the artifact is dependent upon the artificer during
the entire process of creation.
 In defense of the craft analogy 119
does not deliberate” means that craft is the knowledge of what is to be
made and how to make it, which is to be distinguished from an artisan’s
individual psychological process of reasoning out how to achieve a parti-
cular end-result and the subsequent series of motions of which she is the
origin. So Aristotle distinguishes between the individual psychological
process of the artisan, who deliberates, and is the external origin of the
product and the craft itself, which is (as it were) a stationary body of
knowledge. Hence, Aristotle’s use of the craft analogy in his argument for
natural teleology does not use an ambiguous notion of intrinsic end or
proper function, because the terms of the comparison do not include the
origin of the motion or the individual psychological process that an artisan
might employ in determining what to do.
Where does this leave us? On the one hand, Aristotle distinguishes
beings that are products of art from natural beings in terms of the locations
of their moving causes or principles of origin. In this context, the indivi-
dual psychological processes of the artisan and the motion of the male
parent (or sometimes the motions in the semen) are the relevant items.
However, when Aristotle deploys the craft analogy in his argument for
natural teleology, he explicitly brackets the issue of the origin of artistic
creations and natural generations, which makes sense, since this is a point
of disanalogy between art and nature. With the issue of the moving cause
off the table, the worry about a possible ambiguity in the notion of an
intrinsic end or proper function is defused, since the origin of the intrinsic
end is not relevant to Aristotle’s craft analogy and, hence, not relevant to
his argument for natural teleology.

V Conclusion
Readers of Physics ii must be struck by Aristotle’s copious use of the craft
analogy in his explanation of what nature is, and in particular its central
role in his arguments for natural teleology. Yet scholars argue that artifacts
are not substances, and even not beings at all, which would undermine the
value of the craft analogy especially in Aristotle’s arguments for natural
teleology. In this chapter, I have challenged one important thesis central to
the ontological demotion of artifacts, namely the claim that they lack
intrinsic ends and proper functions and in this way are importantly
different from natural beings.12 The idea that artifacts have intrinsic ends
12
Of course, the question of whether or not artifacts have intrinsic ends is only one issue that is
germane to the topic of their ontological status. For a discussion of other important topics, see
Katayama 1999, Kosman 2013, and Shields 2008.
120 charlotte witt
and proper functions is of central importance for the cogency of using the
craft analogy to argue for natural teleology. Since my interpretation grants
intrinsic ends and proper functions to artifacts it supports the cogency of
Aristotle’s arguments for natural teleology that make use of the craft
analogy. It emerged that when Aristotle uses the craft analogy to argue
for natural teleology, he does so by bracketing the causal question of the
principle or origin of the creation or generation, and in this way eliminates
a difference between craft and nature that might seem to undermine his use
of the craft analogy. Hence, my interpretation provides a partial response
to those who might think that the craft analogy is mistaken or misleading
on the grounds that it occludes a crucial difference between art and nature,
namely the role of the artisan or creator.
 chapter 7

The origins of Aristotle’s natural teleology

in Physics ii
Robert Bolton

I The background
Aristotle discovered natural science. Or so he believed, at any rate, as we
can see most clearly from the opening book of his Metaphysics (Α.3–7).
There, in presenting the historical background to what he offers in the
Physics, he argues at some length that none of his predecessors, not even
those who by tradition were known as students of nature (οἱ φυσιολόγοι)
or naturalists (οἱ φυσικοί), had any adequate or workable conception of the
subject. The earliest of these physikoi, says Aristotle, were interested chiefly
in the nature of matter or in the ultimate material constituents of things,
those constituents from which other things come to be and into which they
perish while they themselves, in their substance, persist unchanged (A.3,
983b6–18). They offered no proper or adequate account of why (dia ti) the
ultimate material constituents are modified in the ways that they are, or of
why and how they manifest regularity and order in their patterns of
alteration (984a19–25, b11–22). Some, such as Parmenides, offered no
account of the patterns of change in nature in so far as they denied the
reality of change altogether (984a29–b4). Others, such as Plato, believed
that natural change was too irregular and fluctuating to be a suitable
subject for genuine science or scientific explanation (A.6, 987a32–b1; cf.
Metaph. M.4, 1078b12–17). Still others, such as Empedocles and the
Atomists, offered, at best, only partial or inconsistent explanations for
the regular patterns of natural change (A.4, 985a10–29, b19–20).
In addition, Aristotle claims, none of the earlier thinkers made any clear
or proper use of the essence or form of natural objects, in definition or in
explanation concerning them. Plato only made gestures in this direction, as

Earlier versions of this chapter were presented at Toronto and at USC. I am grateful to those present,
and especially also to Mariska Leunissen, for very helpful comments and suggestions.

121
122 robert bolton
did the Atomists in their use of the shapes and positions of the atoms to
explain their patterns of interaction (A.7, 988a34–b1; A.4, 985b10–19 with
H.2, 1042b11–15, 1043a5–7, 14–21). Finally, and most importantly for our
purposes here, Aristotle claims that no one at all before him had any real
understanding of natural teleology, or of those goals for the sake of which
things come about in nature (A.7, 988b6–8). The main lesson that Aristotle
wants us to draw from his historical survey is, of course, that genuine
natural science must bring into play each of his so-called “four causes” of
what it studies – material, efficient, formal, and final. As he presents it in
Metaphysics A, this is the heart of a proper conception of natural science, a
conception that he supposes he was the first to properly grasp (cf. Physics
ii.1–3, Metaphysics H.4).
As one might expect, the accuracy of Aristotle’s survey of the details and
of the deficiencies of the approaches to causality of his predecessors has
often been questioned, most especially perhaps his claim that none of his
predecessors had identified the final cause. This claim has struck many
recent commentators as surprising and even, as some have said, “startling,”
particularly in view of the extensive use that Plato makes of teleology in the
Timaeus.1 Plato’s specific accounts, in the Timaeus, of the functions and
goals of various natural entities are often alluded to and sometimes
endorsed by Aristotle himself in his own biological works.2 So why does
Aristotle think that even Plato, like his other predecessors, had no grip at all
on the final cause, that is, on the way in which goals figure as genuine
causes (aitiai) in nature? This is a difficult problem which was recognized
already in antiquity, for instance by Alexander, whose own brief attempts
at a solution, however, take inadequate account of the special difficulty
posed by the Timaeus.3 The problem clearly requires resolution if we are to
properly understand what is distinctive about Aristotle’s own teleology in
his Physics and how it differs from approaches offered by his predecessors.
Let us begin, then, by considering how it is that Aristotle defends his
judgment that his predecessors totally neglected the final cause in nature.
This defense is laid out in Metaphysics A.7, in the following passage:
That for the sake of which actions and transitions and changes take place
they [the earlier thinkers] do speak of as a cause in a certain fashion but not
in this way, that is, not in the very way in which this is in nature a cause. For

1
See already Ross 1924: 179 and 1936: 38; Cherniss 1944: 454.
2
Cf. e.g. Tim. 91e–92a with PA iii.10, 686a25–31, and Tim. 76c–d with PA ii.14, 658b2–10. See also
Philebus 53–54, and Sedley 2010.
3
Alexander, In Metaph. 59.28ff. See now also M. R. Johnson 2005.
 Aristotle’s natural teleology in Physics ii 123
those who speak, for instance, of intelligence (νοῦς) or of friendship (φιλία)
do posit these causes as something good. But they do not at all speak as
though anything is or comes to be for the sake of these things. Rather, they
speak as though changes [to good results] originate from these things. In the
same way, those who speak of the one (τὸ ἕν) or of being (τὸ ὄν) as an entity
of this kind [as a cause which is good] say that it is a cause of the substance
(οὐσίας) [of certain things] but not that it is for the sake of this [good] that
anything is or comes to be. (988b6–14)
In this passage, Aristotle first concedes that some of his predecessors did
indeed identify certain things that functioned as causes for them as things
that are good, and responsible for good in some way. The latter point is
emphasized also in an earlier related passage in A.3, 984b11–22. In Α.3, as in
Physics ii.3, Aristotle himself says that the goal or final cause of a thing is the
good of that thing in some way (983a31–32, 195a23–26.) But this, Aristotle
contends, is not sufficient for any of these presumed causes to count as final
causes or as that for the sake of which something is or comes about in nature.
A cause may be good and of something good, Aristotle indicates, but still be
only a cause from which change originates – that is only an efficient cause, as
the operations of intelligence and friendship were for those who offered
them as causes. Aristotle has in mind here especially Anaxagoras and
Empedocles as those who, respectively, made use of the operations of nous
and philia as efficient causes. This is clear from many other passages, for
instance in Metaphysics Λ.10, where he develops further this complaint
against them (1075b1–16, cf. PA i.1, 640b5ff.). Earlier, in A.3, Aristotle claims
that Anaxagoras in particular identified his cause of what is good, or well, in
nature with an efficient cause (984b20–22). But, as he also indicates in A.3,
and as we shall shortly see in some detail, he does not have only these two
figures in mind in A.7 as those who erred in the way they did (984b18–22).
One main thing, then, that is very clear from the first part of our passage
in A.7 is that, in Aristotle’s view, a final cause in nature, properly under-
stood, cannot be an efficient cause in the sense that it cannot be something
that serves somehow to literally originate change, even regular change to
some good result. Any conception of a goal according to which it does serve
somehow to literally originate change, even to a good result, will fail to
capture the way in which goals serve distinctively to explain things in
nature. This is a point that Aristotle also emphasizes elsewhere, for instance
in another important background passage for our purposes in On
Generation and Corruption i.7, where he says:
The productive agent [i.e. the efficient cause] is a cause in the sense of that
from whence change originates. But that [goal] for the sake of which
124 robert bolton
something obtains is not a productive agent, which is why health [the
doctor’s goal in healing] is not a productive agent [of the healing] except
metaphorically (κατὰ μεταφοράν). For when the productive agent is present
[e.g. when the doctor is healing] the patient is coming to be in respect of
something [e.g. health]. But when the [resulting] stable states are there [e.g.
health], the patient is no longer coming to be but is [e.g. healthy, so that the
efficient cause is no longer present], and the forms [e.g. of health, produced
by the doctor in the patient] and ends (τέλη) are such [resulting] stable
states. (324b13–18)
Here Aristotle emphasizes again that final causes are not efficient or
productive causes. He allows that a final cause, in some cases, may in a
metaphorical or non-literal way be called a productive cause, as health or
the form of health is when it figures in the desire and aim of a doctor to
bring about health.4 But the health that figures in the efficient cause in this
case, as only an intended object, is not a true literal final cause or goal of the
doctor’s healing efforts since, Aristotle says, the actual goal is the stable
resulting state of health which is not present until the efficient causal
process is finished. For Aristotle, then, the doctor’s desire in healing is
not for the actual concrete resulting state, since no such particular state is
yet actual, and, we may add, none may ever be actual. His desire is rather,
we should say, that a certain type of state or form – say a certain balance of
hot and cold – should be realized in a certain actual body, and there is no
reference in this description of his desire for the particular final state which,
Aristotle says, is the true end. If the doctor’s desire is never satisfied, that
does not alter the content of his desire, which is the same whether the desire
is satisfied or not. So reference to that actual final state and true end which
exists only when the desire is satisfied cannot be required for, or included
in, the correct specification of the content of the desire itself, or thus of the
efficient cause of healing.
In Poetics 21 Aristotle helps us to understand better what he means when
he says in On Generation and Corruption i.7 that health is an efficient cause
of the doctor’s healing only metaphorically. He says there: “Metaphor
consists in giving a name to something that belongs [strictly] to something
else, where the transference is from the more inclusive (genos) to the specific
(eidos), or from the specific to the more inclusive, or from specific to
specific, or by analogy” (1457b6–9).
It is useful here to see, first, that the terms genos and eidos in this passage
do not mean genus and species. In one of Aristotle’s examples in the text,

4
See also Metaph. Ζ.7, 1032b6–10, Λ.10, 1075b8–10, and further below.
 Aristotle’s natural teleology in Physics ii 125
the entity large number stands as genos to the number 10,000 (τὸ μυρίον)
as eidos, so that one can use the term “10,000” as the Greeks did, non-
literally, to mean “a large number” (1457b11). We use the term “a million”
in this same non-literal way today when we say, for instance: “There are a
million things wrong here.” We mean literally: “There is a large number
of things wrong here.” But the number 10,000, or a million, is not a
species of the kind (genos) large number, in Aristotle’s normal sense of his
term for species. These are, rather, particular paradigm instances of large
numbers. So the terms genos and eidos are used in Poetics 21 to refer in a
very general way to the more and the less inclusive. In the case in question
in On Generation and Corruption i.7, the desired goal, health, is a
productive or efficient cause metaphorically, in that it is a less inclusive
part or feature of the strict, more inclusive, efficient cause, namely the
doctor’s deliberated desire for health, and it stands for this whole efficient
cause in any proper reference to health itself as an agent. The metaphor or
figure of speech in question here is an example of what we now call
synecdoche – the use of a part to stand for the whole. The actual final
cause or goal itself, then, is not ever literally an efficient cause or
productive agent, although the desire to achieve a goal of a certain type
may well be. In Physics ii.7 Aristotle does say that “often” efficient and
final causes “go towards one,” but only “in kind” (198a24–27). His
example is that of a human being, since a human being is the goal of
human generation, and a human being – another human being the same
in kind – is the efficient cause of this. Here the efficient and final causes
are not literally identical. Indeed, in Metaphysics A.3 Aristotle says plainly
that the final cause is “opposed to” the efficient cause (983a31).
In the second part of our passage in A.7, Aristotle turns more explicitly
to Plato, though without mentioning him by name (988b11–14). He
introduces certain features of Plato’s so-called late ontology or “unwritten
doctrines,” features that he has earlier described in A.6 and 7, in particular
the doctrine that “the one” is the cause of the substance of certain other
things. In Aristotle’s earlier report, “the one” is in fact for Plato the cause of
the substance of his forms (987b18–21, 988b4–6). However, although “the
one” may be good, as it is in Plato’s account, and may serve as a source of
goodness for the things that share in it, “the one” is not the goal or that for
the sake of which any of its participants are or come to be. So for this reason
there is no final causality in this case at all but only at most, as Aristotle says
earlier in A.7, a gesture at a type of formal causality (988a34–b6). For Plato,
“the one” is doubtless something by reference to which someone can
understand what it is for something to be or to come to be good, by virtue
126 robert bolton
of its unity, but that is not sufficient to make it a final cause. This can be
true of a purely formal cause.5

II Aristotle on Plato’s teleology in the Timaeus

In this attempt – Aristotle’s only attempt in Metaphysics A – to show that
no earlier thinker, including Plato, had introduced genuine final causes in
nature there is no direct reference to the Timaeus. But how surprising, or
“startling,” is this? Although there is no explicit objection to the adequacy
of the teleology of the Timaeus as such or by name in A.7, there is,
arguably, a clear indication there of what Aristotle’s main objection to it
would be. As we have seen, there is a direct objection in A.7 to the
attempt at natural teleology of Anaxagoras, and, as commentators have
frequently noted, the introduction and use of intelligence or nous by
Anaxagoras, in order to account for the goodness and order of things in
nature, are revived by Plato himself in the Timaeus. In a section of the
Phaedo well known by Aristotle, where Plato first mentions Anaxagoras’
approach, it is rather quickly put aside, purportedly because of
Anaxagoras’ failure to adequately exploit it, but with the wish expressed
that someone might make proper use of it (97b–99d; cf. Aristotle,
Metaph. A.4, 985a18–21, Phys. ii.8, 198b14–16). As, again, commentators
have often observed, Plato fulfills his own wish in the Timaeus where the
things in nature that happen or are present for the sake of some good are
all the result of the operation of intelligence or nous, either the divine nous
of the Demiurge or that of the world soul. In On the Soul i.2–3, Aristotle
himself uses similar language to that he uses in A.7 both to describe
Anaxagoras’ conception of nous as practical intelligence (phronêsis),
which is “the cause of what is well and rightly ordered in nature,” and
to describe the work of the Demiurge in the Timaeus, who also designs
things in nature practically in accord with his calculation of what is
“better” (404b1–6, 407b9–11).
Given this close connection, we can see from Aristotle’s critique of
Anaxagoras in Metaphysics A.7 exactly what his main objection to Plato’s
approach to teleology in the Timaeus is going to be. In Plato’s approach
there our eye structures, for instance, do enable us to regularly achieve
certain goals in our standard environment (45b–e). But those eye
structures, in Plato’s account, are not literally in place for the sake of
those goals. Rather, they are in place because a sufficiently powerful

5
Contrast Frede 1980.
 Aristotle’s natural teleology in Physics ii 127
intelligent designer desired that goals of this type should be achieved by us
on account of their perceived goodness and chose these eye structures for
us to best fulfill that desire. So it is not the actual goal-state itself, in
Plato’s account, or the fact that the means to it secure it, that explains the
presence of the means. It is, rather, the goal as an object of perceived
goodness and of desire, or, in plain language, it is the desire to realize a
goal of a certain type, perceived as good by a suitable rational agent, that
explains the fact that the means that secure it are there. In the language of
On Generation and Corruption i.7, the goal in question is the cause of the
means to it, in Plato’s account, only metaphorically, as a part or feature of
the true, more inclusive efficient cause, namely the desire to realize a goal
of that type. The intended goal only explains the means in this way, as an
aspect of their efficient cause. As such, the goal in question is not the
actual final resulting stable state, which is for Aristotle the only true final
cause of the means and the change to it. Aristotle can himself speak of a
desired goal in such cases as an unmoved mover, as he does in GC i.7,
324a30–b4 and in DA iii.10, 433b15–17. But this does not belie the fact
that it is only a mover at all metaphorically. So, as Aristotle says in
Metaphysics A.7, in this approach, the approach of Anaxagoras and also,
as it turns out, of Plato in the Timaeus, goals as objects of intelligence and
desire explain only as parts or features of efficient causes, not strictly as
genuine final causes. Thus, the sorts of goals that do figure as causes in the
Timaeus are not, for Aristotle, final causes at all, at least not of the sort
that are operative for him in nature.
It is worth noting further that Aristotle’s view here fits closely with his
standard account of practical action (praxis) and of animal self-
locomotion, where the intended goal, as the object or content of
awareness and desire, explains strictly as a feature of the efficient cause,
not as the actual resulting end of an action. As Aristotle says in On the
Movement of Animals 6:
It is the object of desire and of understanding that first initiates movement –
not every object of understanding but only [one which is] a goal of action.
Thus, it is [practical] goods of this sort [i.e. goods desired as goals] that
originate movement, not everything fine . . . For the animal is moved and
advances from desire [for] or choice [of a means to a goal] when it is affected
by perception or an appearance [of that type of goal as good] . . . So it is in
this way that animals are impelled to move and act, the ultimate [efficient]
cause of their movement being desire; and this arises either from perception
or from the appearance and understanding [of a type of good as goal].
(700b24–701a36, cf. DA iii.10, 433b11–12, EN vi.3, 1139a31–33)
128 robert bolton
So in Metaphysics A.7 Aristotle does not in fact simply “strangely ignore”
the approach to natural teleology found in Plato’s Timaeus.6 Rather, he
indicates clearly to us what he thinks is misguided about that type of
approach and why he, therefore, does not count Plato as a genuine
proponent of natural teleology. This point deserves special emphasis in
any attempt to understand Aristotle’s natural teleology in the Physics and
its originality by comparison with Plato’s, since it is often not properly
appreciated in the literature. It is sometimes supposed, for instance, that in
the Timaeus, and in the anticipations of its approach that we find in the
Phaedo, Plato does hold, in agreement with Aristotle, that when in nature
certain means are regularly fit for securing certain goals it is the goal itself
that is the cause of the means to it as that because of which the means are
present.7 At one point in the Phaedo Socrates does seem to speak in this
way when he says that where things are ordered by intelligence (nous) the
cause (aitia) of this order is that things are best that way (98a–b). But
Socrates quickly clarifies this with the remark that when something is done
“with intelligence” it is done, strictly, because of “the choice of what is best”
on the part of the intelligent agent (99a–b). This, of course, is also the
official view of the Timaeus, where the means are for the sake of an end in
Plato’s teleology only because they are such in the mind and plan and
deliberate activity of an intelligent agent (30a).8

III Implications for Physics ii

Some have argued, however, against this, that Aristotle’s natural teleology
as portrayed in Physics ii is less different from Plato’s than we might
suppose from our results to this point, on the following grounds.
Although for Aristotle, unlike Plato, nature is not the product of deliberate
craft, divine or otherwise, nevertheless in Physics ii, it is claimed, Aristotle
views the operations of nature and of craft as so closely parallel that he
virtually erases the difference between them, just as Plato does in the
Timaeus. We can see this most clearly, it is argued, in Physics ii.8, to
which we should now turn. There Aristotle most fully exploits the analogy
between nature and craft. He says, strikingly:
If a house, for instance, were one of the things that come to be by nature,
it would come to be in just the way as it now does by art (technê). And if
the things that come to be by nature were not only to come to be by

6 7 8
Ross 1936: 38. See e.g. Lennox 2001a. Cf. Johansen 2004: 106–110.
 Aristotle’s natural teleology in Physics ii 129
nature, but also by art, they would come to be [by art] just as they
naturally do. (199a12–15)
In context, we need to remember, the immediate point that Aristotle wants
to make and to support here, in emphasizing the similarity of natural and
of artistic or craft production, is just that in both “all of the earlier stages are
for the sake of the end” (199a8–9). Aristotle’s argument is that, given this
similarity, if it is admitted that there are goals for the sake of which such
stages occur in the case of art, this should be admitted, by analogical
argument, also in the case of nature. This still leaves room for drawing a
clear distinction between nature and art, as in On Generation and
Corruption i.7, on the grounds that in the art the stages are strictly only
for the sake of an end in the plan and deliberate activity of a craftsman, so
that, in this case, goals explain only as features of efficient causes, while in
nature this is not so. But many have wanted to argue that the similarity
between art and nature for Aristotle runs much deeper than this.9 This
comes out most clearly, it is suggested, at the end of Physics ii.8, where
Aristotle echoes his point from the earlier passage that we have just
considered, with a special addition. He says:
It is strange to maintain that there is no coming to be for the sake of
anything [in nature] unless we see that the productive agent [there] has
deliberated. Yet art too does not deliberate. If the art of shipbuilding were in
the wood, it [the art] would make the ship in like manner as something is
made in nature. (199b26–30)
Here Aristotle attempts to discount an alleged difference between
production in art and in nature that, as he indicates, some had used to
argue that there is no final causality in nature. The alleged difference
depends on the assumption that in art the productive agent is a conscious
deliberator, and the argument is that it is only as the intended object of a
deliberator that we can make any sense of goals at all. So, in this objection,
there is no genuine final causality in nature since there is no deliberator
there. This is a line taken, of course, by many philosophers of biology
today, and even by some interpreters of Aristotle who hold that for him, as
for Kant, there is only as if teleology in nature because it is as if there were a
deliberator there.10 Aristotle’s own response to the objection, however, is
that conscious deliberation is not essential even in art. Elsewhere, in
Metaphysics A.1, Aristotle makes a similar point. He says that in the case
of production in the crafts: “Artisans (χειροτέχνας) make things just like

9 10
See e.g. Sedley 2007, 2010. See e.g. Nussbaum 1978, Wieland 1975.
130 robert bolton
some inanimate things do, for instance as fire burns. But while the
inanimate things do each of these things by some nature [that they possess]
the artisans act by [acquired] habit” (981b2–5, cf. MA 10, 703a29–36).
So even in the crafts, for Aristotle, production is not necessarily based on
explicit deliberation. It can be due simply to the exercise of an acquired
habit that can operate in the same way as fire burns. In Metaphysics A.1
Aristotle does reserve the term art (technê) for the special skill of director-
craftsmen (architektones) who can explain why skilled artisans correctly
make things as they know how to do. But more commonly he ascribes
technê, in its more standard sense of skill in making, to all types of skilled
producers necessary in society, which would include shoemakers, builders,
and even architektones, who clearly do not and practically speaking could
not consciously deliberate about every single thing that they do in exercis-
ing their art (Pol. iv.4, 1291a1ff., viii.2, 1337b8ff.).11
Here in A.1 Aristotle closely aligns production by art, in certain respects,
with the activities of inanimate natural objects such as fire. In so doing he
certainly distances himself from the Timaeus, where Plato strongly opposes
this and argues that the work of art, which is for him necessary for the non-
accidental securing of goals in nature, always essentially involves conscious
intelligent deliberation (46d–48b).
But still, in Physics ii.8, as we have just seen, Aristotle does say in
emphasizing the similarity of art and of nature that if the art of shipbuild-
ing were in the wood from which a ship is built, the art itself would make
the ship. This initially somewhat mystifying remark may be compared with
an earlier passage, in Physics ii.3, where Aristotle says:
It is always necessary to seek the cause that is topmost (τὸ ἀκρότατον),
just as in other matters. For example, a [particular] man builds because he
is a builder, but the builder builds by virtue of [possessing] the art of
building. This cause [the art of building], then, is prior, and so is it in all
cases. (195b21–25)

11
Alternatively, as we shall see in further detail below, Sedley 2010 connects Aristotle’s claim that “art
too [like nature] does not deliberate” with a doctrine in Metaphysics Z.7 that “the art of [for instance]
medicine . . . is the form of health” (1032b13). Aristotle’s reasoning, Sedley claims, is that since forms
cannot deliberate, only people, then if an art is identical with the form of its product, art does not,
indeed cannot, deliberate. But, in Aristotle’s official account of an art, “an art (technê) is essentially
the same thing as a state of capacity to make in accord with a correct course of reasoning [i.e. of
deliberation]” (EN vi.4, 1140a9–10). The form of health by itself is obviously not identical with any
such capacity to make; although of course, as we have already noted, suitable knowledge of this form
and of how to instantiate it is required for medical art for Aristotle, as he shows us in detail in Z.7,
1032a25–b30. This is doubtless why he indicates there that the form is the art only “in a way”
(1032b11). See also Broadie 1990 and Leunissen 2010 for further discussion of this issue.
 Aristotle’s natural teleology in Physics ii 131
Here Aristotle claims that the way, in one particular sense, to specify the
cause of a building is to say that this cause is the art of building, just because
it is by virtue of possessing this art that the skilled builder builds. This seems
to fit closely with his later claim that if the art of shipbuilding were in the
wood, the art itself would be the cause of a ship. But this connection does not
yet fully show us how to demystify that later claim. For that, some argue, as
noted earlier, we need to turn to still another related passage, in Metaphysics
Z.7, where Aristotle also compares production in nature and in art, but
this time with more detail and in a more fully articulate way. With respect to
the products of nature, which he considers first in Z.7, Aristotle says:
That by which (ὑφ’ οὗ) they [natural things] are produced, the so-called
nature in the sense of the form (τὸ εἶδος), is their nature. This [nature] is
formally the same (ἡ ὁμοειδής) [as the nature in the product], although this
[productive nature] is in another individual, since it is a[nother] human
being that generates a human being. (1032a24–25)
Here Aristotle ascribes a certain productivity, or causality, in natural
production, to form. He describes the form in the parent as that “by
which,” in some sense, natural human offspring are produced. He then
proceeds, in the text, to urge that there is a close parallel in production by
art to this type of causality in nature.
[In the same way as in nature a human being comes from a human being
so, in production by art,] in a certain fashion (τρόπον τινά) health comes
from health and house from house. That is, that [health or house] which
has matter comes from that [health or house] without matter (τῆς ἄνευ
ὕλης). For the art of healing and the art of building [from which, respec-
tively, health and a house come] are the form of health and of the house.
When I speak of the substance without the matter [i.e. the form] I mean
the essence. (1032b11–14)
This passage specifies for us now, so it is argued, the respect in which an art
is the cause of its products as that “from which” the product comes. An art
is the cause of its products in this sense – as that “from which” – in that the
form of the product is the cause of the product. So, as far as its causality
goes, the art is just the form of the product (cf. GA ii.1 735a2–3). So, in fact,
in art just as in nature the form of the product is the true agent or efficient
cause of the product. When Aristotle says, then, in Physics ii.8 that if the art
of shipbuilding were in the wood from which the ship is built, it would, as
efficient cause, make the ship, he means that, just as in nature, if the form of
the ship were per impossibile in the wood from which the ship is built, it
would, as efficient cause, make the ship.
132 robert bolton
This now gives us, it is claimed, the means by which to get a clear picture
of Aristotle’s natural teleology that shows us how close in fact it is to
Plato’s. We start from Plato’s approach in the Timaeus, where nature is the
product of divine craft. There, form (Platonic form, of course) serves, in
the mind of the divine craftsman, as the intended goal or object of desire
that he aims to imitate and instantiate as far as possible in constructing the
world. If we now subtract from this picture the causal work done by the
desires, calculations, and choices of the divine craftsman made with a view
to the realization of Platonic form, and we assign this causal work rather in
each case to the form alone, but now the immanent Aristotelian form of the
natural object, which itself works through various instruments to achieve
the goal, then what we are left with is Aristotle’s teleology. Here we see,
then, what it is that grounds Aristotle’s teleology, and we also see how close
Aristotle is to Plato in spite of their differences. Or so it is argued.12

IV Questions for this account

Here then is one conception, compelling in various ways, and influential,
of the main core of Aristotle’s teleology in Physics ii with an indication of
its historical origin in Plato and of the textual basis for it. But how well does
this reading square with other things that Aristotle says? Let us first ask how
it fits with what we find in Metaphysics A.7. There, as we have seen,
Aristotle argues in effect that even Plato in the Timaeus has no proper
conception of a final cause. The reason for this is that in Plato’s account the
goal of some natural process or structure – as an object of desire – explains
not as the actual final cause of the process or structure but only as a part or
feature of the efficient cause. That is, the process or structure is not there in
Plato’s story strictly because it serves to realize the actual goal, but rather
because it serves to realize the desire for a certain type of unrealized goal on
the part of a sufficiently powerful divine craftsman, and is chosen and put
in place as such.
Suppose, then, that we eliminate the divine craftsman from this account
and substitute for it Aristotelian form, understood as a non-conscious,
non-deliberative productive agent which prompts and directs via certain
instruments the development of, for instance, the eye structures in the

12
Some of the main elements of this line of interpretation, which is now fully developed by Sedley
(2007, 2010), are laid out already by Cherniss (1944), who was himself following LeBlond. Cherniss
used it as a part of his attempt to convict Aristotle of bad faith in claiming in Metaphysics A.7 that
Plato had no final cause, since Aristotle’s natural teleology, according to Cherniss, very closely copies
that of Plato. See Cherniss 1944: 454ff., esp. at n.411.
 Aristotle’s natural teleology in Physics ii 133
human embryo so that they are fit for achieving certain goals. Here too the
goal, understood as the unrealized object of this impetus or direction from
form, explains as a feature of the efficient cause of our eye structures, not as
their true final cause. That is, our eye structures are not present in this type
of account because they secure the actual final goal-state itself; they are
there because they satisfy the sufficiently powerful impetus from form to
secure this type of not yet present goal. Thus, the goal, so understood, as
the unrealized object of this impetus or direction from form, is as much an
efficient cause, or a part or feature of an efficient cause, as it is in
Anaxagoras or in the Timaeus. This type of account, then, is as much
ruled out by Aristotle’s strictures in Metaphysics A.7 as is that of Anaxagoras
or Plato.
It is a further problem for this approach that in Aristotle’s own account
of animal generation and development our human form is not actually
present to direct anything when our eye structures, for instance, are first
being developed in the human embryo. Only the potential for this form is
present (GA ii.3, 736a27–737a18). It is argued by some that human form
and function is present even in the earliest embryo, in a very low degree of
actuality.13 But that the locomotive, perceptual, and rational capacities that
are necessarily included in human form are actually present to any degree at
all in the earliest embryo seems a very difficult idea, and one ruled out by
Generation of Animals ii.3.14
Matters are not improved, moreover, if we say, as some do, not that our
form is the agent of the development of our eye structures in the embryo,
but rather that the “irreducible potential” for this form is the directing
agent.15 For here too the eye structures do not strictly come to be because of
their actual goal or end; rather, they come to be because they satisfy the
non-conscious impetus for the unrealized end or the “primitive directive-
ness upon the end,” as some put it, which the irreducible potential for form
supplies. So once again, the unrealized end, as the object of this non-
conscious impetus or “directiveness,” explains not as a final cause but
rather as a feature of an efficient cause.
This same difficulty also faces those accounts in which the agent of our
development for Aristotle is some so-called “program” in the embryo for a
certain type of form and end, a program that causally directs development
and is, in what it directs, sensitive to changes in the environment of the
embryo.16 Here too the structures developed are not present for the sake of

13
Code 1987. 14 Cf. Lewis 1988. 15 Gotthelf 2012.
16
See e.g. Bradie and Miller 1999 and M. R. Johnson 2005.
134 robert bolton
their actual end, but rather because they are brought about by the context-
sensitive “program” for this as yet unrealized type of end. As Metaphysics
A.7 and On Generation and Corruption i.7 show us, this sort of account
of what grounds Aristotle’s natural teleology does not explicate this
teleology – it eliminates it. It is argued by some that the understanding
of final causes as features of efficient causes is the approach that Aristotle
should have taken in Physics ii, though not the approach that he actually
did take. They see that, in the case of human action, for Aristotle the goal
explains as a feature of the efficient cause and argue that Aristotle should
have carried this over to his natural teleology.17
Another serious difficulty for this approach arises from the fact that, in
this type of account, a distinct, intrinsic, efficient causal role, in embryolo-
gical development for instance, is assigned by Aristotle to form or,
alternatively, to the “directiveness” supplied by the irreducible potential
for form. A form itself, as we have seen Aristotle explicitly say in
Metaphysics Z.7, is not a full substantial thing, nor is the so-called irreducible
potential for form given that it, like form or soul, is only a capacity of such a
thing and thus not a full substantial thing itself, as Aristotle points out in DA
ii.1, 412a16–22. The same holds of the so-called program for form which
some introduce. However, to saddle Aristotle with entities such as human
form as intrinsic efficient causal agents is to treat a form as a full substantial
thing and to turn him into a soul–body dualist, a stance that he clearly rejects
in On the Soul. As we know from DA ii.1, the form of a living natural entity
for Aristotle is its soul (412a19–21). But Aristotle refuses in On the Soul to
ascribe intrinsic efficient causal agency to the soul itself. As he says in i.4:
We say that the soul [i.e. the human form] is pained and rejoices, is bold
and afraid, and further that it is angry and perceives and reasons . . . Yet to
say the soul is angry is just as if one should say that the soul weaves or
builds. Doubtless it is better not to say that the soul pities or learns or
reasons [or weaves or builds] but that a human being does this by virtue of
[possessing] its soul [e.g. by virtue of being able to desire and to choose to
pursue goals of various types] . . . The intellect (nous) might seem [to be an
exception to this and] to come to be in us as a [self-subsistent] substance
and so to be imperishable [since the intellect itself reasons] . . . But reason-
ing, and loving and hating, are not attributes of the intellect but only of the
one who has it insofar as he has it . . . These were not ever attributes of the

17
See Irwin 1988: ch. 5. Irwin blames Aristotle’s failure on the fact that he uses “pure dialectic” as his
method of inquiry in the Physics and not the “strong dialectic” that Irwin finds in the Metaphysics.
But, as we have seen, Aristotle maintains the same firm distinction between efficient and final causes
that Irwin finds in the Physics also in the Metaphysics, in A.7. He does the same in A.3, 983a30–32, and
in Δ.2 and in Η.4.
 Aristotle’s natural teleology in Physics ii 135
intellect [itself] but rather of the common thing [i.e. the ensouled or living
body]. (408b1–29, cf. i.1, 403a3ff.)
Here Aristotle denies that the soul, that is, the human form, builds or
weaves or reasons or is in any way on its own an agent, or a patient, of craft
activity or other change. It is only “the common thing,” the ensouled body
or person, that has such attributes. No entity such as the soul or human
form, or indeed the art or form of health, has any such attributes. Soul or
form is only a power to do things, not an agent that does anything.
Aristotle uses this point here to undermine an argument for soul–body
dualism. The argument is that the soul, at least the intellectual soul, is a
substance because it can do things, such as reasoning, on its own. Aristotle
rejects this argument on the grounds that its premise – that the soul, the
human form, is on its own an agent – is false. If human form or soul were
on its own an agent, Aristotle supposes, dualism would follow.
There are passages, for instance in DA ii.4, where Aristotle does say that
the soul is in one way an origin of change (ἀρχὴ κινήσεως), or a mover, using
language like that which he standardly uses to describe an efficient cause
(415b9–12, 22ff.; cf. PA i.1, 647a27–28). But he has already warned us in
i.4 that while he, or we, may say this sort of thing, it is not to be taken
literally. The soul, or human form, is never literally an agent or efficient
cause. In DA ii.4 itself, to show us that and in what way the soul is a “source
of change,” Aristotle cites the fact that the soul is a source of qualitative
change (alloiôsis), since the soul is a source of perception: “For perception is
held to be a type of qualitative change and nothing which lacks a soul has
perception” (415b24–25). So the soul is a source of change not because the
soul is the efficient cause of perception. For Aristotle, as we know well, it is
the external object of perception, not the soul itself, which is his efficient
cause of perception (DA ii.5 417b20–21). The soul is a cause or source of
perception and thus of change, as he says, only in that something must be an
ensouled thing of a certain sort in order to have perception.
Our passage in DA i.4 is important also for a proper understanding of
the text we considered earlier, in Physics ii.8, where Aristotle says that if the
art of shipbuilding were in the wood, it, the art, would make the ship. As
the reference to building in DA i.4 shows, for the art of shipbuilding to be
in the wood some shipbuilder and his body would have to be in the
wood, and all of his tools as well, and the only shipbuilding that could
be done would be done by the exercise by this shipbuilder of his art
(cf. i.3 407b15–26). This shipbuilding might be done non-deliberatively,
to a greater or lesser extent, by habit, but that would not affect this point.
136 robert bolton
So when Aristotle says in Physics ii.3 that the causality of the art is prior to
that of the craftsman, he cannot mean to imply that there is any literal
efficient causal agency of the art itself that is responsible for the product.
He must mean rather that it is the possession of the art, achieved through
habituation and other means, which explains why the active craftsman is
the true agent and not vice versa. That is, the priority in question here, of
the art over the craftsman, is not ontological priority or, thus, any priority
of the art as an actual agent or efficient cause. In Physics ii.3 itself, Aristotle
points out that when we are specifying causes, efficient or otherwise, we
must consider whether we are talking about the particular cause of some-
thing or about the cause understood in some general way, and we must
specify also whether we are talking about the actual cause or only some
potential cause. In the case of a building, Aristotle says, the actual particular
efficient cause on any given occasion is “the builder engaging in building”
(195b6). So the art of building or the form of a building is never the actual
particular efficient cause in any case of building.
Given this, we cannot use the fact that in Metaphysics Z.7 Aristotle says
that the art of housebuilding is the form of the house to argue that the
actual particular agent in housebuilding is ever the form of the house. For
the art of building is not the cause of a house in the sense that it is the actual
particular agent or efficient cause. Thus, we cannot infer from Aristotle’s
remark that the art is the form, taken together with the parallel between art
and nature, that in nature the form, e.g. of a human being, is the actual
particular agent in human generation and in embryological development.
As we have seen, the only actual particular agent of such development is the
bodily thing that does have the form or soul and does exercise the capacity
which that confers. That is, the only actual particular originating efficient
cause of human generation for Aristotle is the father fathering, through
appropriate instruments, of course. In Parts of Animals i.1 Aristotle says:
The productive agent [i.e. the efficient cause, of human generation] is prior
[to the product] not only in explanation (logos) but also in time. For a
human being [is the productive agent that] generates a human being, so that
it is because that human being [the father] is such that the generation of this
human being is so. (640a24–26)

V Approaches to an alternative conception of Aristotle’s teleology

This conception, then, that we have been considering of Aristotle’s
natural teleology, in which the agency of the immaterial Demiurge in
 Aristotle’s natural teleology in Physics ii 137
the Timaeus in securing goals in nature is replaced by the agency of
Aristotelian form (or by the agency of an equally non-substantial irredu-
cible potential for form or program for form), not only deprives Aristotle
of final causes by his own lights by turning them into efficient causes
or aspects of efficient causes: It is also inconsistent with Aristotle’s
hylomorphism. We need then to try to find an alternative conception
which avoids these pitfalls. In a word, this alternative conception of
Aristotle’s natural teleology must not be inconsistent with Aristotle’s
view of the soul–body relation, and, most importantly, it must be a
conception in which the means to actual goals in nature literally are,
without qualification, there for the sake of those actual goals and
not because they were desired or chosen, or non-consciously urged or
directed to be present by some productive agent in order to achieve
some as yet unrealized type of goal. It is not sufficiently well appreciated,
I believe, how difficult it is to articulate such a conception.
To help bring out further the difficulty I see here let us turn now briefly
to the contemporary scene and consider perhaps the most prominent
account in the recent literature of purpose or function, namely the so-
called etiological account, one that has been used by many to interpret
Aristotle’s teleology.18 In this etiological account, our heart, for instance,
has the function or goal of circulating our blood if, and only if, two
conditions are satisfied. The first condition is that our heart does effectively
circulate the blood, and the second is that our heart is present in us because
it circulates the blood. Here the notion of cause, or of “because,” covers,
and was intended to cover paradigmatically, causes introduced in evolu-
tionary explanation. In such an explanation, very roughly, we have hearts
because they circulate blood, in the sense that the presence of efficient blood
circulation by the heart conferred a survival advantage on our ancestors
who thereby lived to produce us with a genetic makeup that gives us hearts.
One might complain that what is explained here is not so much the fact
that we have hearts as that we exist at all with hearts. One could argue that
the real explanation of why we as humans have the type of heart we do has
to do with our human genetic makeup.19 But, however that may be, the
supposed cause of our having hearts in this case is not an Aristotelian final
cause. It is, rather, a complex efficient cause, or efficient/material cause,
having to do with our evolutionary history, one in which the fact that we

18
See e.g. Godfrey-Smith 1994, 2014; Millikan 1989; and Wright 1973, 1976. For applications to
Aristotle, see Depew 1997; Gotthelf 2012; and Lennox 2001a.
19
Cf. Nagel 1977.
138 robert bolton
have hearts that circulate blood is given a certain historical cum genetic
explanation.
To see this more clearly, let us spell out the causal sequence in question
somewhat more fully. There is, first, genetic variation in a population
which gives rise to a certain type of heart in certain members of the
population. Then, the action of this heart in these members of the
population gives rise to a certain pattern of blood circulation for them by
comparison with other members of the population. This pattern of blood
circulation then gives rise to an improved survival and reproduction rate
for these members of the population, and for their offspring, including us,
to whom they transmit such hearts. We, then, have hearts because of this.
Notice, however, that in this story it is not even mentioned that the
function or goal or final cause of the heart is blood circulation, but rather
only that we have hearts that do circulate blood in a certain manner,
together with a certain efficient/material causal story as to why that fact
came to obtain. It is this that reveals that our heart has the function or goal
of blood circulation in the etiological account.
To see the import of this point for our present discussion, consider more
abstractly now the contemporary etiological analysis of function when
adapted as an analysis of Aristotelian final causality. In such an analysis:
B is the final cause of A if and only if:
(1) A accomplishes B and
(2) A is present because it accomplishes B.
The first question to ask about such an analysis from the Aristotelian
point of view is: What kind of causality is introduced by the “because” in
clause (2) of the analysans? For Aristotle there are four kinds of causality, so
we need to determine which kind or kinds are intended or permitted by
(2). It is clear enough that the one mode of causality that cannot be
intended or permitted in (2) is final causality, since that would make the
analysis trivially circular. If the causality in (2) were final causality, then
clause (2) would in effect say or be satisfied by: A is present for the sake of B,
or, in other words, B is the final cause of A. But final causality is just the
item of which we are trying to give an analysis. So the causality covered in
clause (2) must be other than final causality. (Matters are not improved if
we write clause (2) as: A came to be present because of B, where that means
or allows: A came to be present for the sake of B. Here again we have a
circular analysis.) This shows, from the Aristotelian perspective, that in this
type of analysis we are in effect explicating or defining final causality
 Aristotle’s natural teleology in Physics ii 139
reductively, in terms of other types of causality. That, of course, is just the
sort of thing one might expect from a contemporary philosophical analysis
where the aim, often, is to define some presumed problematic notion –
here final causality – in terms of less problematic notions. This analysis
permits us, in effect, to understand the presumed unclear notion of final
causality in terms of the clearer notion of efficient or efficient/material
causality. By this means we sanitize, or naturalize as some would say, the
notion of final causality. Many philosophers of biology today would argue
this is the best that we can hope for if we want to find respectable
teleological explanation in real biology.20 But this aim is not something
with which Aristotle would agree. For him, as we have seen, final causality
is not, and is not an aspect of, efficient or efficient/material causality.
So how do we locate a conception of final causality in Aristotle that
avoids this reductive character found in the contemporary analysis and
also, I have been arguing, in the views of Anaxagoras and of Plato? We do
best, I think, to start in a simple way, from Aristotle’s claim in Physics ii.8
that nature brings things about for the sake of something, and that in
nature the final cause or end of a thing is due to its nature or is natural to it
(198b10ff., 199a30–32). Among other things, to take one of Aristotle’s own
examples from ii.8 (198b23ff.), this means that the fact that our broad back
teeth are fit for breaking up food is itself something that holds by nature
(cf. Phys. ii.1, 192b35–193a2). That, indeed, is the main thing that Aristotle
emphasizes in his critique in ii.8 of Empedocles, who claims, according to
Aristotle, that the fact that our broad back teeth are fit for breaking up food
only holds by chance (198b27). To say, against this, that this fact holds by
nature implies, given Aristotle’s view of nature and of natural science, that
this fact is either a primitive principle of natural science which explains
other things or that it is a derivative theorem or conclusion explained
ultimately by such primitive principles.
In Physics ii.8 itself, that our broad back teeth are fit for, and do
effectively perform, the beneficial function of breaking up food by nature
is presumed to be confirmed not by reference to any analysis or definition
of the notion of function or of final cause, but, in the first instance, by
simple empirical means. Aristotle takes it to be clear from our ordinary
experience that in humans, “always or usually,” broad back teeth are fit for
breaking up food. (He says “or usually” to take account of the fact that
there are defective humans (199a33–b7).) This, he supposes, is an adequate
ground for taking this fact to be a natural regularity (198b34–199a8). Some

20
See e.g. Godfrey-Smith 1994, 2014.
140 robert bolton
critics of Aristotle’s stance here have worried that he allows himself too easy
a victory over Empedocles in supposing that in humans broad back teeth
are fit for breaking up food. But, as Aristotle presents Empedocles’ view,
Empedocles agrees that as things stand our broad back teeth are fit for
breaking up food. What they disagree about is the status of this fact. Here is
how Aristotle presents Empedocles’ own view in Physics ii.8: “Our teeth
come up of necessity, our sharp front teeth fitted for tearing, our back teeth
broad and useful for breaking up food . . . [But] things are so organized in a
fitting way (ἐπιτηδείως) by chance (ἀπὸ τοῦ αὐτομάτου)” (198b24–31).
Contrary to what is sometimes assumed in the literature, the main point
of difference between Aristotle and Empedocles here does not concern how
to explain the fact that we have broad back teeth. In particular, Aristotle
and Empedocles do not disagree here over whether there is only a mechan-
istic explanation, or only a teleological explanation, or both, or neither, of
the fact that we have broad back teeth. Rather, their disagreement here
concerns the status of the very different fact that our broad back teeth are fit
for breaking up food. According to this passage, Empedocles agrees with
Aristotle that our broad back teeth are fit for breaking up food – that is, he
agrees that as things stand that is the good and beneficial result that they
efficiently secure. Empedocles even argues further, according to Aristotle,
that the fact that our broad back teeth do efficiently accomplish this
beneficial result confers a crucial survival advantage on us so that, in that
sense, our broad back teeth are present in us because they perform this task,
given appropriate attendant circumstances (198b27–32). This means that
Empedocles actually accepts teleology in nature by our contemporary
etiological standard. That is, in Empedocles’ account our broad back
teeth do regularly break up food, and they are present in us – in the
sense that we survive with them – because they do this. This shows us
again that the contemporary etiological account cannot be Aristotle’s, since
Aristotle clearly does not count Empedocles as a proponent of natural
teleology. But, in any case, Aristotle does not allow himself any easy victory
over Empedocles on the point that, as observation shows, our broad back
teeth are fit for and do efficiently perform the beneficial function of
breaking up food, since Empedocles agrees with this.
Where Empedocles differs from Aristotle here is only in his contention
that the fact of this easily observed fitness holds by chance. This means, as
we have seen, that while our broad back teeth are fit for and do efficiently
function to break up food as things stand, and we survive with such teeth
because of this, the fact that they are fit for this is not a natural regularity or
what we would call a natural law. That is, again in more Aristotelian terms,
 Aristotle’s natural teleology in Physics ii 141
the fact that our back teeth are fit for this neither is one of the fundamental
necessary principles that govern and explain what happens in nature, nor is
it a regularity explained by these principles. Aristotle, by contrast, holds
that the fact that this functional regularity obtains is guaranteed by the
fundamental principles of nature in the sense that it either is one of these
principles or it is in a suitable way deducible from and explained by them.
(Aristotle’s critique of Empedocles leaves each of these two options open.)
That is the key in Physics ii.8 to the teleological outlook that Aristotle
espouses there by contrast with the outlook of Empedocles, and, we may
add, of his contemporary followers today.
Some would discern two contrasted theses of Empedocles at 198b24–30:
(1) We have teeth of necessity, but (2) we have teeth fit for eating food by
chance.21 However, there is nothing in the text indicating a contrast
between “our teeth come up of necessity” and “our sharp front teeth
[come up] fitted for tearing” (198b24–25). The latter phrase is grammati-
cally in apposition to the former and, thus, is explicative of it. So, as
Aristotle indicates, for Empedocles, our sharp front teeth come up fitted
for tearing by necessity (198b24–25). For Empedocles and the others whom
Aristotle is considering here, everything happens by necessity, so if for
Empedocles it is a fact that our sharp front teeth come up fitted for tearing,
then that fact holds for him of necessity (198b12). But this is not incompa-
tible with the claim that this fact holds by chance and not by nature as
Aristotle understands this distinction. If this fact is neither a primitive
explanatory principle of natural science, nor a conclusion explained by
such principles, then it holds not by nature but by chance, even if it holds –
for as long as it does hold – by necessity. Not everything determined or
necessitated by the fundamental principles needs conform to a fixed
natural law or be explained by such laws. In Metaphysics Z.7 what happens
by chance is opposed by Aristotle to what happens by nature or by art, not
to what happens by necessity (1032a12ff).22
So we may say that for Aristotle, according to Physics ii.8:
B is the final cause of A if and only if:
(1) A is fit for B and
(2) that A is fit for B is a natural regularity.

21
See e.g. Cooper 1982.
22
For a related account of the difference between Aristotle and Empedocles, from a different but at
least partly complementary perspective, see Sauvé Meyer 1992. For further discussion of the issues,
with additional references, see Leunissen 2010.
142 robert bolton
Aristotle’s own notion of final cause, then, is based on these two other
notions – that of fitness and that of natural regularity. The notion of fitness
is, then, a thinner one than that of final cause, although it is still, of course,
an evaluative notion, and ineliminably so for Aristotle.23 As we have seen,
moreover, in Aristotle’s approach in Physics ii.8 the fact that it is a natural
regularity that our broad back teeth are, evaluatively, fit for the breaking up
of food is taken as something easy to directly confirm, at least in an initial
way. It is confirmed, Aristotle supposes, by the easily observed fact that in
humans broad back teeth are fit for breaking up food “always or usually”
(198b35–36). Of course, Aristotle would need to show that this presumed
natural regularity does fit into the whole body of natural scientific regula-
rities in an appropriate way – either as principle or as derived theorem – to
finally confirm this. Otherwise, his victory over Empedocles is perhaps too
easy. But this is all that he needs to do.
To help see this more clearly consider, finally, how different the main
problem is that motivates Aristotle’s stance on teleology in Physics ii.8 from
a problem that has motivated contemporary analysts. In recent discussion,
one main problem has been to show how to distinguish the goal or final
cause of some structure or process, when it has one, from a mere regular
consequence of that structure or process. For instance, we today want to be
able to say why the circulation of the blood is the goal or function of the
heart while the production of heartbeat noises is not. But this is not
Aristotle’s problem. He does understand this distinction. He notes in
Physics ii.2, as something of a joke, that death is the end of human life in
a different way from that in which our best condition is the end of human
life (194a28–33). But he regards it as very easy, on the whole, to distinguish
(evaluatively good) goals from other regular consequences. In Physics ii.3,
when he introduces final causality, Aristotle does so simply by means of
what he takes to be uncontroversial examples. He does the same in
Metaphysics Δ.2 and H.4, and elsewhere. He never evinces the need for
an analysis of function or of final cause to enable him to distinguish goals
from other regular consequences. His problem, rather, as we have seen, is
to show that in many cases where things obviously are fit to achieve certain
goals and beneficial results, that they do achieve these results in a fitting
way is something that holds by nature and not by chance.
So if we want to understand Aristotle’s own special conception of
natural teleology in Physics ii we need to see that it has these two crucial
components: (1) Final causality is not reducible to, or analyzable or

23
On the issue of value in teleology, see Nagel 2012 and references there.
 Aristotle’s natural teleology in Physics ii 143
explicable in terms of, any other kind of causality, such as efficient
causality; and (2) in the natural world, that things have the sort of fitness
they clearly now do is typically something that holds by nature, not by
chance. This shows us how very different Aristotle’s approach to natural
teleology is not only from the alternative of Empedocles but also from that
of Plato.24

24
See Bolton 1997, 2011 for further discussion and for a defense of the view that the natural fitness of
means to goals is not primitive for Aristotle but is explained by primitive principles.
 chapter 8

Substantial generation in Physics i.5–7

Devin Henry

The focus of this chapter is on the conceptual machinery of Aristotle’s

theory of generation in Physics i.5–7. According to the standard interpreta-
tion, Physics i.7 imposes two requirements on all processes of change. First,
all change requires an underlying subject from which the change proceeds.
During the course of the change this subject comes to be F from being not-
F. Second, the subject must remain through the change to become a
constituent of the finished product. For example, Socrates survives the
healing process and simply exchanges sickness for health, and bronze
survives the production of the statue and simply changes from being
unshaped to being shaped. The central question of this chapter is whether
or not Aristotle thinks that both of these conditions hold in cases of
substantial generation. Does he think that when a new substance comes
into being there is a pre-existing subject that survives that change and
becomes a constituent of what comes to be? The orthodox view is that he
does, and that his argument for this is to be found in Physics i.7. This, it is
held, forms part of Aristotle’s argument for hylomorphism, the claim that
every particular thing is a composite of matter and form. My aim in this
chapter is to re-evaluate this entrenched view of the Physics. I shall argue
that Physics i.7 is silent on the question of whether or not substantial
generation requires a persistent subject, while evidence external to the
Physics strongly suggests it does not. In this reading, while Aristotle thinks
substantial generation requires a subject from which the change proceeds,
he does not claim that this subject survives as a constituent of the substance
that comes into being.
To begin it will be useful to call attention to the distinction Aristotle
makes in the works on natural science between qualified coming to be
(γίγνεσθαι τι) and coming to be in the unqualified sense (γίγνεσθαι
ἁπλῶς).1 All change for Aristotle is a form of coming to be (or passing

1
See also Jones 1974: 478–480.

144
 Substantial generation in Physics i.5–7 145
away) in one of these two senses.2 As a first pass at this distinction we can
say that when “coming to be” (γίνεται, γίγνεσθαι) is used in the qualified
sense, the verb is always transitive, in the sense that it remains incomplete
without specifying what the subject comes to be. For example, we say that a
human comes to be healthy, or that he comes to be bigger. When used in
the unqualified sense, the verb is intransitive, in the sense that it is complete
all on its own. In this case we say that a human comes to be (full stop), not
that he comes to be so-and-so. This linguistic distinction has a metaphy-
sical basis. For the difference also turns on which category of being the
change concerns.3 When a thing comes to be in the qualified sense, the
change always occurs with respect to one of the non-substantial categories.
Alteration (ἀλλοίωσις) is coming to be in the category of quality (e.g.
Socrates comes to be healthy), growth is coming to be in the category of
quantity (e.g. Socrates comes to be bigger), and locomotion is coming to be
in the category of place (e.g. Socrates comes to be in the market). All of
these count as instances of coming to be in the qualified sense because the
substance (Socrates) already exists and merely becomes qualified in some
particular way. Coming to be in the unqualified sense, by contrast, is a
change in the category of substance; it is substantial generation. In this
case, a new substance comes into being that was not there before. In what
follows I shall reserve the term “generation” for coming to be in the
unqualified (ἁπλῶς) and strict (κυρίως) sense, and use the more general
term “change” (μεταβολή, κίνησις) for the broader category that includes
coming to be in the qualified sense (alteration, growth, locomotion).
At the outset of the Physics Aristotle tells us that when the objects of any
inquiry have principles, causes, or elements (ἀρχαὶ ἢ αἴτια ἢ στοιχεῖα),
knowledge and understanding are acquired by grasping these: “For we do
not think we know a thing until we are familiar with its primary causes or
first principles and have carried out our analysis as far as its elements.
Clearly, then, also in the science of nature our first task will be to determine
its principles” (184a10–16). Aristotle’s positive inquiry into these principles
is set out over Physics i.5–7.4 According to the results of that inquiry, every
change involves a subject, S, that undergoes the change (τὸ ὑποκείμενον)
and a pair of contraries, F and G, that represent the two extremes of the

2
For the main distinction, see Phys. i.7, 190a31–33; On Generation and Corruption i.3, 318a34–35. My
focus in this chapter is mainly on coming to be, although much of what I have to say applies equally
to passing away.
3
With the following, see GC i.3, 317a27–28; i.4, 319a11–16; and Metaph. xii.2, 1069b9–14.
4
Compare GC ii.9, and Metaphysics vii.7 and xii.4. Of these four places only Metaphysics xii.4 offers
the complete list of principles of coming to be (form, privation, subject/matter, primary moving cause).
146 devin henry
change. Within this tripartite schema, S is always a body of some kind (e.g.
human, wood, seed), while the contraries are identified with the form (the
possessed state, ἕξις: e.g. musical) and its privation (the absence of that
state, στέρησις: e.g. unmusical). When the change proceeds from the
privation to the form, it is coming to be. When it proceeds in the opposite
direction, from the form to the privation, it is passing away. Together
Physics i.5–7 argue that these three items (form, privation, and subject)
constitute the principles of coming to be for all natural things: Physics i.5
argues that the contraries are principles; Physics i.6 shows that there must
be some third thing besides these; Physics i.7 argues that this third thing is
the subject that underlies the change.5
Aristotle’s main position is developed at length in Physics i.7. The core of
that theory can be extracted from two key passages: 190a13–21 and 190b10–
23. For our purposes we can focus on the celebrated passage at 190a13–21:
Once these distinctions have been made one can grasp the following from all
the cases of coming to be, if one looks at them as we suggested (ἐάν τις
ἐπιβλέψῃ ὥσπερ λέγομεν), namely, that there must always be an underlying
subject (which is what comes to be), and that this, though always numeri-
cally one, is not one in form. By “in form” I mean the same as “in account”;
for what it is to be a human is not the same as what it is to be musical. And
one survives, while the other does not: for what is not an opposite survives
(for the human survives), but not-musical or unmusical does not survive,
nor does the compound of the two, namely, the unmusical human.
Aristotle makes two main claims here. First, coming to be requires an
underlying subject, which is the thing that comes to be (τὸ γιγνόμενον).6
This subject is numerically one, but not one in form; rather, it changes
from F to G as a result of coming to be. Second, in the process of change
one part survives (the subject), while the other does not (the privation or
the subject-cum-privation). I shall call these the Subject Requirement and
the Survival Requirement, respectively:
Subject Requirement. Coming to be requires an underlying subject from
which the change proceeds.
Survival Requirement. When something comes to be, the subject
endures and remains as a constituent of the finished product.

5
For a general discussion of these chapters, see Bostock 2006: ch. 1 and Kelsey 2008.
6
Aristotle uses τὸ γιγνόμενον to refer both to the subject from which a change proceeds (the terminus a
quo) and to the thing that results from that change (the terminus ad quem). In the present context
Aristotle means the former, while at 190b10–23 (quoted below) τὸ γιγνόμενον picks out the latter. In
Generation of Animals Aristotle uses τὸ γιγνόμενον almost exclusively for the embryo that comes to
be an F.
 Substantial generation in Physics i.5–7 147
Most scholars take Physics i.7 to be developing the concept of underlying
subject in close connection with the idea of being a continuant of change,
so that the Subject Requirement and the Survival Requirement come out as
equivalent. Bostock (2006: 9), for example, argues that the expression τὸ
ὑποκείμενον is being made to do “double duty” in Physics i.7, both for the
starting-point of the change (the terminus a quo) and for what persists
through the change. On this reading, to say coming to be requires a subject
that underlies the change just means there must be something that persists
through the change and survives as a constituent of the finished product.7
In the reading I shall defend, the concept of subject and continuant come
apart: The concept of subject (τὸ ὑποκείμενον) is playing only the one role
of starting-point of change; it is a separate claim to say that this subject also
persists through the change.8
It is obvious that cases of non-substantial change meet both require-
ments. For example, when Socrates comes to be healthy from being sick
there is an underlying subject from which the change begins (Socrates). And
that subject persists through the change and remains as a constituent of the
finished product (healthy-Socrates). However, while Physics i.7 argues that
substantial generation must also satisfy the Subject Requirement (190b1–5),
there is a question about whether or not Aristotle means to commit himself
to the view that in cases of substantial generation this subject survives as an
element of the substance that results from that change. Before turning to
that I want to start by clarifying the Subject Requirement.
The Subject Requirement says that in (all) coming to be there must be a
subject from which the change begins. Like the contraries, the concept of
an underlying subject of change was not Aristotle’s invention. At 189b5–6
Aristotle suggests that some of his predecessors already invoked the idea of
a ὑποκείμενον as something different from the contraries, although they
disagreed over what that underlying subject was. But how does Aristotle
understand the concept? Physics i.7 offers plenty of examples of subjects,
including a human (190a17, b14, b20), a seed or embryo (190b5; cf. GA i.18,
724b14–18; i.20, 728b34–35; and i.23, 731a2–4), bronze, gold, stone (190a25,
190b16–17), and wood (191a7). We can shed some light on this motley crew
of subjects by turning to Metaph. vii.13, 1038b2–7:

7
See also Broadie 1982: 47; Gill 1989: 6, 90, 106; and Witt 1989: 66.
8
Compare Charlton 1970: 77 and Bostock 2006: 7–8. Charlton denies that substantial generation
involves a persistent subject. While I shall eventually side with this reading, here I only mean to make
the point that the Subject Requirement and the Survival Requirement are independent of one
another. That point does not require taking a stand on the question of whether or not Aristotle is
committed to the Survival Requirement in cases of substantial generation.
148 devin henry
Just as the underlying subject, the essence, and the composite of these are
called substance, so too is the universal. We have already spoken about two
of these, namely the essence and the underlying subject, and the latter we
said underlies in two senses: either as a determinate individual, which is the
way that an animal underlies its attributes, or as the matter underlies the
complete actuality.9
Here Aristotle contrasts matter with the determinate individual (e.g. the
individual human or horse) as two kinds of underlying subject. The latter is
the more familiar sense of subject from the Categories. These primary
subjects are independently existing “thises” (τόδε τι) that remain numeri-
cally one and the same throughout their existence, and stand as ultimate
subjects of predication (Cat. 5). The other sense of subject is the proximate
matter that underlies the form of the composite individual. In the case of
living things, for example, the material subject is the living body whose
organs have been specially adapted to realize the capacities of soul that
make up its form (DA ii.1, 412a17–19).10
What I want to suggest is that these two senses of underlying subject –
subject as individual determinate substance and subject as matter – track
Aristotle’s distinction between qualified and unqualified coming to be,
respectively. The underlying subject of qualified coming to be is always
some determinate individual; for example, human is the subject in the
change from an unmusical human to a musical human. By contrast, the
subject of unqualified coming to be is always matter of some kind; for
example, bronze is the subject in the production of a bronze statue. While
Aristotle does not avail himself of this distinction in Physics i.7, it is clearly
implicit in his account. For example, when he is enumerating the princi-
ples of naturally generated substances at the end of Physics i.7 he attempts
to elucidate the concept of “the underlying nature” (ἡ ὑποκειμένη φύσις)
by setting up a contrast between matter, on the one hand, and the
individual substance, on the other (191a7–12), and stresses that the former
underlies the change, but not as a unified “this” (οὐχ οὕτω μία οὖσα οὐδὲ
οὕτως ὂν ὡς τὸ τόδε τι; 191a12–13). Again at 190b13–14 (see below) he
distinguishes bronze or stone from the human being as two kinds of

9
See also Metaph. ix.7, 1049a34–b1.
10
Gill 1989: 31 argues that the aim of Metaphysics vii.3 is to refine the conditions on subjecthood so
that matter actually turns out not to be a subject. In her reading, the refined concept only picks out
the individual, i.e. what is separate and a “this” (τόδε τι). But that is not what Metaphysics vii.3
concludes. Aristotle says that being separate and a “this” are characteristics of substance, not subject,
and that for this reason matter fails to be a substance. Aristotle never retracts the original claim that
matter is an underlying subject. And Metaphysics vii.13 presents that claim as something already
established.
 Substantial generation in Physics i.5–7 149
underlying subject. So there is at least some recognition in Physics i.7 that
when τὸ ὑποκείμενον is used for the material subject (e.g. bronze, gold,
stone, wood) it is to be contrasted with those subjects that correspond to
individual substances (e.g. human, horse).11
According to the Subject Requirement, then, all coming to be requires
an underlying subject from which the change proceeds. When a thing
comes to be in the qualified sense this will be a determinate individual (e.g.
human). In cases of substantial generation the subject from which the
change proceeds is always matter of some kind. The question is whether in
Physics i.7 Aristotle thinks the subject must persist in all cases of change,
including substantial generation, or whether this is true only for changes in
the non-substantial categories.
Before turning to that question let us complete the basic model of
coming to be by introducing the other core passage from Physics i.7:
From what has been said, it is clear that everything that comes into being (τὸ
γιγνόμενον) is complex.12 There is, on the one hand, what comes into being
(τι γιγνόμενον) and, on the other, something that comes to be that (τι ὃ
τοῦτο γίγνεται). And the latter in two senses: either the subject or the
opposite. By “the opposite” I mean the unmusical and by “the subject” I
mean the human. Likewise I call the absence of shape, form, or organization
“the opposite” and the bronze or stone or gold “the subject.” Clearly, then, if
there are causes and principles of naturally existing things from which (ἐξ ὧν)
they primarily are and have come to be (I mean what each is said to be
according to its essence and not incidentally), then everything comes to be
from both the subject and the form (ἔκ τε τοῦ ὑποκειμένου καὶ τῆς μορφῆς).
For in a certain way the musical human is composed of human and musical,
for you can analyze it into the account of each. It is clear, then, that things that
come to be (τὰ γιγνόμενα) have come into being from these. (190b10–23)
Consider the generation of an animal. The thing that comes into being as a
result of the change is the fully developed animal that can be analyzed into
a body, which is the underlying subject, and a soul, which is predicated of

11
Aristotle’s remarks at the beginning of GC i.5 also suggest that these two senses of underlying subject
mark the distinction between qualified and unqualified coming to be. There he says that matter in its
“most proper and strict sense” is the underlying subject of generation and destruction, although “in a
certain way” it is also the subject of other changes (320a2–5). What he means, I take it, is that when
the musical human comes to be musical, “man” can be treated as the matter in so far as it stands in an
analogous relation to the (un)musical human as bronze stands to the statue or wood to the bed (cf.
Phys. i.7, 190b20–23). Thus, in a way, matter is the subject even of non-substantial changes
(including alteration, increase, and locomotion). But matter in the “proper and strict sense” is the
subject only of unqualified coming to be and passing away.
12
I take τὸ γιγνόμενον here to refer to what results from the change (the terminus ad quem). For an
alternative reading, see Gill 1989: 102–108.
150 devin henry
the body as its subject (DA ii.1, 412a16–19). The soul is the form picked out
by the definition specifying what comes to be (τι γιγνόμενον),13 while the
thing that comes to be that (τι ὃ τοῦτο γίγνεται) is either the embryo
(Physics i.7, 190b1–5, 9–10) or the privation of the form (Physics i.5, 188b20;
cf. Plato Phaedo 70b10–72e2).
The remainder of the passage is meant to answer the central question of
Physics i: What are the principles from which (ἐξ οὗ) naturally existing
things primarily are and come to be? Aristotle’s answer can be understood
by drawing on a distinction he makes in Metaphysics xii.3 this time
between two kinds of principle: “Since not only the elements that are
present in a thing (τὰ ἐνυπάρχοντα) are its causes but also something
external (ἐκτὸς) to it, viz. the moving cause, clearly while ‘principle’ (ἀρχὴ)
and ‘element’ (στοιχεῖον) are different both are causes, and ‘principle’ is
divided into these two kinds” (1070b21–25). The two kinds of principle
here are (1) those that are present in a thing as its constituent elements and
(2) those that are not constituents of the product but are external to it. Of
the three principles listed in Physics i.7, the privation counts as a principle
only in the sense that natural generation originates from it. But it is not a
principle in the sense of an element of the thing that comes to be. For
it does not remain in the finished product as a constituent (οὐκ
ἐνυπάρχοντος γίγνεταί τι; Phys. i.8, 191b16), but is destroyed in the
change. Only subject and form are principles in the sense of being con-
stituent elements of naturally generated substances. This is what Aristotle
means when he says that everything that comes into being is composed
from subject and form (γίγνεται πᾶν ἔκ τε τοῦ ὑποκειμένου καὶ τῆς
μορφῆς; 190b20).14
Traditionally scholars have taken this as an argument for hylomorph-
ism. But that is not quite what Aristotle says here. While matter is certainly
13
Compare Metaph. vii.7, 1032a13–25 and xii.3, 1069b36–1070a2. In both places Aristotle uses the
expression τι γιγνόμενον to pick out the form that tells us what it is that comes into being. While it is
true that the Metaphysics vii.7 passage offers “human” and “plant” as its examples of τι γιγνόμενον,
which look like composite substances, Aristotle insists that the same term can designate both the
form and the composite (Metaph. viii.3, 1043a29–37). That he has the form of the product in mind
here is confirmed by GC ii.9, where he explicitly identifies the three principles of coming to be with
the matter, form, and primary efficient cause. Of the first two he says: “This [i.e. the capacity to be
and not be] is a cause in the sense of matter for those things that are such as to come to be, while the
cause in the sense of their end is their shape or form, which is also the account stating the substance
of each of them” (335b4–7). Once again contrast Gill 1989: 121.
14
Physics i.7 does not list the primary moving cause among the principles of naturally generated
substances. However, in Physics ii we learn that form and matter are, in fact, both inner principles of
change for naturally generated substance. The external efficient cause mentioned in the Metaphysics
passage is the (form of the) generating parent (Phys. ii.7, 198a25–28; Metaph. vii.7, 1032a20–25; GA
ii.1, 735a2–4; and PA i.1, 640a19–26).
 Substantial generation in Physics i.5–7 151
one kind of constitutive subject, we have seen that Physics i conceives of τὸ
ὑποκείμενον in the broader sense that includes both matter (e.g. bronze)
and the determinate individual (e.g. human). This is explicit even in the
current passage, for “man” is there identified as the subject in the compo-
site “the musical human.” So if there is an argument for hylomorphism in
Physics i.7, it is not the explicit aim of this passage.
As we have seen, most commentators assume that Aristotle’s concept of
underlying subject already involves the idea of being a continuant of
change, so that the Subject Requirement and the Survival Requirement
collapse into a single condition imposed on all coming to be.15 In this
section I re-examine that orthodoxy. We can carve up interpretative space
into two broad readings. According to the orthodox view, Aristotle holds
that all change requires a subject that persists through the change and
remains as a constituent in the finished product, including substantial
generation. We can call this the Inclusive Reading, since it takes the
Survival Requirement to be inclusive of all forms of coming to be.16 By
contrast the Exclusive Reading takes the Survival Requirement to apply
exclusively to coming to be in the qualified sense (change in the non-
substantial categories).17 In this reading, while Aristotle thinks substantial
generation requires a subject from which the change proceeds, he does not
claim that this subject must survive as a constituent of the substance that
comes into being. What I want to argue is that the Exclusive Reading is
more attractive as an interpretation of Physics i.7 than the more traditional

15
I suspect this reading gains a lot of its naturalness from the fact that most gloss τὸ ὑποκείμενον as
“that which underlies the change.” To avoid begging the question I have adopted the more neutral
expression “underlying subject,” without any assumptions about what this subject underlies (the
form/privation or the change itself).
16
This reading is by far the most common. It is defended in some form or another by Broadie 1982:
46–47; Gill 1989: 6, 98–108; Code 1995: 415–417; Bostock 2006: 8–9; and Kelsey 2010: 112. Broadie
(2004) seems to have changed her position on the matter. Broadie 1982: 47 endorses the Inclusive
Reading and takes the concept of underlying subject to entail the idea of persistence. However,
Broadie 2004: 129–130 takes the weaker view that Physics i.7 is actually silent on the issue, while GC
i.3 positively rejects it in favour of the Exclusive Reading. (In light of this I take her remarks in her
2004 essay in favor of the Inclusive Reading to be offering reasons why one might be tempted by it
without actually endorsing it.) Bostock defends the Inclusive Reading, but he is careful to restrict his
defense to Physics i (2006: 8). He does not speculate as to whether or not Aristotle endorsed the
Survival Requirement outside Physics i, specifically, whether or not GC i.1–5 actually rejects it.
Finally, Kelsey parses Phys. i .7, 190a13–21 slightly differently, so that the Survival Requirement is not
actually at issue. In his reading, Aristotle argues (1) that coming to be always requires an underlying
subject, and (2) that this subject must be “two in form.” He then takes the claim that one part
survives and one part does not survive as evidence for (2). However, Kelsey counts himself among
those who accept the Inclusive Reading (personal communication).
17
Charlton 1970: 77, 135; Jones 1974; and Broadie 2004.
152 devin henry
Inclusive Reading. There are two reasons for this. First, although most
commentators prefer the Inclusive Reading, the evidence from Physics i.7 is
not conclusive enough to force that interpretation on us. As we shall see, it
is possible to read that chapter in a way that is compatible with the
Exclusive Reading. Second, there is evidence external to Physics i that
strongly suggests Aristotle does not think substantial generation involves
an enduring subject. This creates a strong presumption in favor of the
Exclusive Reading as an interpretation of Aristotle’s more general views
about coming to be in the unqualified sense. In light of this I shall propose
that Physics i.7 is best read as remaining silent on the question of whether
or not substantial generation must meet the Survival Requirement. This
conservative reading has the obvious virtue of avoiding a conflict among
Aristotle’s different discussions of change.
Traditionally, Physics i.7 has been seen as the locus classicus for the
Inclusive Reading. The most explicit piece of evidence for this reading is
190a13–21 (translated above), where the Survival Requirement is first men-
tioned. Aristotle introduces that passage by saying: “one can grasp the
following from all the cases of coming to be” (ἁπάντων τῶν γιγνομένων;
cf. 189b30–31).18 Most take this as definitive evidence that coming to be in
the unqualified sense is subject to the Survival Requirement. Bostock
argues that 190b10–23 also implies that some part of the subject remains
even in cases of substantial generation:
The only ground Aristotle could have for saying that whatever comes into
being is composite (sunthetos) is that we can distinguish in it two “elements,”
one the persisting element (what underlies) and the other the acquired
element (the form). If the element said to underlie did not persist in the
end product, there would be no ground whatever for saying that the end
product was composite, and Aristotle explicitly claims that all products of
becoming are composite. (Bostock 2006: 8)
Finally, scholars typically point to the example of a statue coming to be
from bronze at 190a23–26 as a case of substantial generation. And there
Aristotle definitely says that the bronze survives.19
The trouble starts for the Inclusive Reading when we turn to On
Generation and Corruption. The question of whether or not the subject
persists through the change lies at the heart of GC ’s distinction between
qualified and unqualified coming to be. In GC i.1 Aristotle tells us that
the monists, who constructed all things out of a single element, were

18
Bostock 2006: 8; cf. Broadie 2004: 129–130.
19
For what is in my view an unsuccessful argument against the latter point, see Jones 1974: 483–488.
 Substantial generation in Physics i.5–7 153
forced to reduce substantial generation to alteration precisely because they
held that all change requires a persisting subject: “For they must affirm that
the underlying subject (τὸ ὑποκείμενον) always remains one and the same:
and we call this sort of thing ‘being altered’” (314b3–4). On Aristotle’s view,
when a thing undergoes alteration (or growth or locomotion), the subject
survives the change and simply exchanges one property for another. In
substantial generation the subject does not survive the change, but is
completely transformed in the process (cf. 332a8–10). There are at least
three passages where Aristotle makes his position explicit:
T1. Coming to be and passing away in the unqualified sense are not effected
by combination and disassociation but when a thing changes from this into
that as a whole (μεταβάλλῃ ἐκ τοῦδε εἰς τόδε ὅλον). But they suppose that all
such change is alteration, whereas in fact there is a difference. For one
component of the underlying subject corresponds to the form and the other
the matter. There is coming to be and passing away when there is a change in
these, but when the change is in its incidental properties it will be alteration.
(GC i.2, 317a20–26)
T2. Seeing that we distinguish between the underlying subject and the
property whose nature it is to be predicated of that subject, and that
change occurs in each of these, it is alteration when the perceptible subject
remains but changes in its own properties (the properties in question
being either opposites or intermediates); for example, the body, although
enduring as the same body, is now healthy and now ill, and the bronze is
now spherical and now angular, and yet remains the same bronze. But
when the thing is transformed as a whole (ὅλον μεταβάλλῃ) and nothing
perceptible remains as its subject (e.g. when an embryo as a whole is
changed into blood20 or water into air or air as a whole into water) such an
event is the coming to be of the one sort of thing and the passing away of
the other, especially when the change proceeds from an imperceptible
something to something perceptible (either to touch or to all the senses),
as when water comes to be out of or passes away into air (for air is pretty
imperceptible). (GC i.4, 319b8–21)
T3. Our account of growth must preserve the features of the subjects that
are growing and diminishing (τὰ ὑπάρχοντα τῷ αὐξανομένῳ καὶ φθίνοντι).
And these are three: First, any and every part of the growing magnitude is
made bigger (e.g. if flesh grows, every particle of flesh gets bigger); second, it
gets bigger by the accession of something; and third, it does so in such a way
that the growing thing is preserved (σωζομένου) and endures (ὑπομένοντος)

20
The example is peculiar and does not match anything in Aristotle’s own account of animal
generation. It would make better sense if “embryo” and “blood” were transposed, or if “blood”
was replaced by (e.g.) “animal.” As far as I know, there is no manuscript evidence for that.
Nevertheless, Aristotle’s point is clear.
154 devin henry
the change. For, whereas a thing does not endure (οὐχ ὑπομένει) in the
process of unqualified coming to be and passing away, in both growth and
alteration the growing subject itself and the thing that undergoes alteration
does endure (ὑπομένει), although in alteration the quality (τὸ πάθος) and in
growth the size (τὸ μέγεθος) does not remain the same. (GC i.5, 321a17–26)
In each of these passages Aristotle uses the Survival Requirement as
a means for distinguishing growth and alteration from generation.
When the underlying subject persists through the change and simply
exchanges one accident for another, it is growth/alteration. Generation
occurs when the subject from which the change proceeds does not
endure but is wholly transformed in the process of becoming a new
substance, which Aristotle glosses in T1 as changing in both matter and
form together.21
In T2 Aristotle again distinguishes between a subject and its accidents
(πάθος here corresponds to the “incidental affections” mentioned at the
end of T1, 317a26). And again alteration is said to occur when the subject
remains one and the same and changes in one of its accidents, while
generation occurs when there is complete transformation (ὅλον
μεταβάλλῃ) so that nothing remains of the original subject.22 T3 is even
more explicit about this. Here Aristotle is discussing growth (coming to be
in the category of quantity) rather than alteration (coming to be in the
category of quality). He mentions three characteristics of genuine cases of
growing. The last of these is that the subject is preserved (σωζομένου) and
endures (ὑπομένοντος) the change just as in cases of alteration, “whereas it
does not endure (οὐχ ὑπομένει) in the process of unqualified coming to be
and passing away.”
Some commentators have tried to read the GC doctrine in a way that
brings it in line with the Inclusive Reading. In T1 Aristotle tells us
that the subject from which substantial generation proceeds is itself a
complex of matter and form. For example, the subject from which
animal generation proceeds is the embryo, which is a compound of
21
For a similar reading of GC i, see Broadie 2004. See also Metaph. vii.7, 1033a20–23. There Aristotle
says that the matter from which a thing comes to be does not persist (οὐχ ὑπομένοντος) but is
transformed (μεταβάλλοντος) in the process of coming to be.
22
Defenders of the Inclusive Reading who take the continuant of substantial generation to be prime
matter emphasize Aristotle’s reference to the destruction of a perceptible subject. To say that
nothing perceptible remains (this reading holds) leaves it open that something imperceptible
survives the change, namely, prime matter. In my view the focus on a distinction between a
perceptible and an imperceptible subject is a red herring. If Aristotle meant that some imperceptible
subject always endures, we would expect him to make that crucial point here. And yet nowhere in
GC does he ever mention an imperceptible continuant. For more discussion of this point, see Gill
1989: 42–53 and Broadie 2004: 126ff.
 Substantial generation in Physics i.5–7 155
menstrual blood from the female and form from the male. According to
Gill, when Aristotle says that the underlying subject “does not remain”
(μὴ ὑπομένοντος), he means it does not remain what it was at the start of
the change. For it loses the form that determined its original identity in
the process of being transformed into something new.23 But the subject
does remain qua matter; the matter of the underlying subject does not
undergo a change. While this way of understanding the GC account
would save the Inclusive Reading, that is not what Aristotle says in the
above passages. T2 and T3 say nothing like that. In T1 Aristotle says that
one component of the subject corresponds to form and the other matter,
and that generation occurs “when there is a change in these” (not “in one
of these”). The most natural antecedent for the plural τούτοις here is
the form and the matter (not the form alone). That, I take it, is what
Aristotle means by saying that in generation and destruction the subject
changes “as a whole.” If Aristotle had meant to say that the subject of
generation and destruction changes with respect to its form but not with
respect to its matter, we should have expected him to make that crucial
distinction here – especially since he has made it a point of emphasis that
those who posit an enduring subject are committed to the view that
unqualified coming to be and passing way are just forms of alteration
(GC i.1, 314b3–4).
According to GC i.1–5, then, what distinguishes substantial generation
from changes in the non-substantial categories is precisely the fact that in
the latter case the underlying subject persists and changes only in its
incidental properties, whereas in the former case the subject does not
remain but is completely transformed in the process of becoming a new
substance.
Another place to look for a defense of the Inclusive Reading is
Aristotle’s mature theory of animal generation in Generation of Animals.
If Aristotle had strong reasons for insisting that substantial generation
must meet the Survival Requirement, then we would expect this to
form an important part of that theory, especially since he treats animal
generation as substantial generation par excellence. Unfortunately,
there is no evidence that Aristotle was concerned in GA to show that
the subject from which the process begins persists through the change to
become a constituent of the mature animal. By contrast, much of the
inquiry in GA is arguably motivated by a concern with the Subject

23
Gill develops this reading over the course of chapters 2 and 3 of her book. For a precise statement, see
Gill 1989: 62.
156 devin henry
Requirement. For example, GA i.18–23 is about the nature of sperma,
which is characterized as “that sort of thing from which naturally generated
organisms first come into being” (724a18–20). Again, in GA ii.1 Aristotle
investigates how the parts of the animal are formed. That inquiry takes as
its starting-point the principle that “everything that comes to be by nature
or by craft comes into being by the agency of something existing in
actuality from what is already potentially that sort of thing (ἐκ τοῦ
δυνάμει τοιούτου)” (GA ii.1, 734b20–22; see also 734b34–36). This reflects
Aristotle’s general solution to the Eleatic challenge, defended in Physics i.8
and GC i.3, that coming to be in the unqualified sense proceeds from what
exists in potentiality, and that this potential F is the proper referent of τὸ
μὴ ὄν ἁπλῶς: “In one way things come to be from what is not in the
unqualified sense, yet in another way they always come to be from what is.
For there must be something existing beforehand (προυπάρχειν) that is
potentially but not actually an F. And this is spoken of both as being and as
not-being” (317b14–18). Aristotle repeats this a few lines later: “If some-
thing comes to be <in the unqualified sense>, then clearly there will be
some substance that exists potentially but not in complete actuality
(ἐντελεχείᾳ) from which coming to be proceeds and into which the thing
that is being destroyed necessarily changes” (GC i.3, 317b23–25).
Freeland (1987) draws attention to Aristotle’s claim that blood serves as
both the matter for the parts of the mature animal and the matter out of
which the embryo is originally constructed (PA ii.3, 650a32–b13; GA i.18,
725a21–28; i.19, 726b1–15; and ii.4, 740b34–35). She takes this as evidence
that blood is the persisting substratum of biological generation, which
satisfies the Physics i.7 requirement that all change requires an enduring
subject. If Freeland is right, then Aristotle must have abandoned the
GC distinction between growth and generation in the biological works.
For GC is categorical that in the case of growth the subject endures
through the process, but when a thing comes to be in the unqualified
sense the subject does not endure but it is wholly transformed in the
process (see T3 above).
Yet Aristotle did not develop his theory about blood as the matter of
the parts with the Survival Requirement in mind, and so by saying that
“it is the same matter by which the <existing animal> grows (αὐξάνεται)
and out of which it is first constructed” (GA ii.4, 740b35–36), he does not
mean to commit himself to any views about a persistent substratum. So we
can’t take this claim as a defense of the Survival Requirement. More
importantly, this claim does not entail that blood is a persistent substra-
tum. Freeland simply helps herself to this by assuming that the material
 Substantial generation in Physics i.5–7 157
from which generation begins must persist to become the proximate matter
out of which the adult animal is constituted.24 But that does not follow.
When Aristotle claims that the matter out of which the animal grows and
the matter out of which it is first constituted are the same (GA ii.4,
740b35–36), he need only be saying that growth and generation begin
from the same material, i.e. that blood plays the role of (material) starting-
point in each case. This does not commit him to the further claim that
blood from which generation begins persists through the change to become
the proximate matter out of which the adult animal is constituted. And
there are good reasons to think that Aristotle wants to resist this conclu-
sion. First, as we have seen, Aristotle treats growth and generation as two
distinct kinds of change, and insists that only the former involves a
persistent subject: “For, whereas a thing does not endure (οὐχ ὑπομένει)
in the process of unqualified coming to be and passing away, in both
growth and alteration the growing subject itself and the thing that under-
goes alteration does endure (ὑπομένει), though in alteration the quality (τὸ
πάθος) and in growth the size (τὸ μέγεθος) does not remain the same.”
Second, according to Aristotle’s mature theory of animal generation, blood
could not satisfy the Survival Requirement. To meet the Survival
Requirement, the blood from which generation begins must remain
numerically the same throughout the change (“this, though always
numerically one, is not one in form”). Yet Aristotle holds that the blood
that is used to form the embryo is not even specifically the same blood that
is later used to grow and maintain the parts of the completed offspring (GA
ii.6, 744b32–745a4).25

24
See Freeland 1987: 406.
25
Much of Freeland’s argument depends on the claim that menstrual blood is the proper subject of
animal generation in the Physics sense, which I think is incorrect. Aristotle associates the subject
that serves as the starting-point of substantial generation with the potential F (e.g. Phys. i.8,
191b27–29 and GC i.3, 317b23–25). And according to Metaphysics ix.7, a thing counts as the
potential F when it is in a state where nothing further must be added, removed, or changed before
it can straightaway be transformed into an F (1049a10–11) and, in cases of natural generation,
possesses the generative principle within itself so that it develops into an F through itself, if
nothing external impedes it (1049a13–18). While Aristotle thinks that menstrual blood is the
matter out of which the embryo is formed, and while he does say in GA that it is already
potentially an animal of the same kind as the mother (GA ii.3, 737a22–25; ii.4, 738b3–4 and
740b18–20; ii.5, 741b7–8; cf. 729a32–33), there is good reason to think that it is not the subject (τὸ
ὑποκείμενον) of generation in the Physics’ sense. It fails the conditions for being the starting-point
of the change laid down in Metaphysics ix.7. For it is not yet in a state where it will straightaway
change into a new organism “through its own principle” (διὰ τῆς αὑτοῦ ἀρχῆς). It still needs to
undergo a further change before it is in that state, namely, it needs to be fertilised by the semen
and then transformed into an embryo (cf. GA i.20, 728a26–30). When we apply the Metaphysics
ix.7 test to the GA, it turns out that it is only when the embryo has been formed and has developed
to the point where it has a rudimentary heart that can process nutriment that it is in a state where
158 devin henry
Defenders of the Exclusive Reading can use the evidence from GC i.1–5
and Aristotle’s mature theory of animal generation as a reason to reconsider
the standard account of Physics i.7. Consider, first, the passage at 190a13–21.
It is possible to read the opening sentence of that passage as restricting its
scope to those cases of coming to be that Aristotle has just been discussing
(ἐάν τις ἐπιβλέψῃ ὥσπερ λέγομεν), namely those where a thing comes to be
in the qualified sense (189b32–190a12).26 (Remember that in GC Aristotle
says in at least three places that when a thing comes to be in the unqualified
sense the subject does not survive.) In support of this is the fact that Physics
i.7 does not present any argument for the (surely controversial and not
obvious) claim that even in substantial generation something persists
through the change. Aristotle does argue that when a thing comes to be
in the unqualified sense there must be some underlying subject from which
the change proceeds (190b1–5). But he does not also say that this subject
must persist through the change to remain as a constituent of the finished
product.
Bostock claims that 190b10–23 implies that the subject persists
through the change even in cases of substantial generation. But this is
an interpretative move on his part. At 190b10–23 Aristotle says that it is
clear “from what has been said” that what comes into being is always
complex. Now, what has just been said is that all cases of coming to be
involve a subject from which the change proceeds (190b9–10). Bostock
simply assumes that the concept of “underlying subject” already includes
the idea of persistence. Thus, he reads 190b10 as a straightforward
inference from the claim that the subject persists to the claim that
what comes into being is therefore complex. Moreover, he claims to
be unable to see any other reason for drawing such a conclusion: “If
the element said to underlie did not persist in the end product there
would be no ground whatever for saying that the end product was
composite, and Aristotle explicitly claims that all products of becoming
are composite.” 27 But this is not the only way to read the passage.
Aristotle had already made the point that what comes to be is complex
back at 190a3–4, prior to introducing the idea that something might
survive the change (190a17–21). So that point could hardly be said to

it will immediately develop into a new animal when nothing external impedes it (GA ii.1, 735a12–
26 and GA ii.4, 735b33–740a13). This makes the primitive embryo the proper subject of animal
generation. (This is confirmed by the fact that throughout GA Aristotle refers to the embryo, and
not the menstrual blood, as τὸ γιγνόμενον, which in the Physics is synonymous with the under-
lying subject of change.) For a similar reading, see Gill 1989: 228–233.
26
This is the sense of the Oxford translation. Cf. Bostock 2006: 8. 27 Bostock 2006: 9.
 Substantial generation in Physics i.5–7 159
depend on that idea. Instead, Aristotle’s grounds for claiming that what
comes to be is complex seem to be linguistic: It is implicit in the way we
speak, for example, “when we say the non-musical human becomes a
musical human” (190a3–4). The bulk of the chapter is an attempt to
unpack this implication with more precision, using his own distinction
between subject and contrary. In light of this, “from what has been said”
should be taken to refer to everything that precedes 190b10 (rather than
just the immediately preceding lines, as Bostock assumes), so that the
whole passage can simply be read as a reiteration of the chapter’s central
point.
Finally, while Aristotle does say at 190a23–26 that, in cases where a
statue comes to be from bronze, the bronze survives the change,
defenders of the Inclusive Reading take it for granted that he is
using the example to make a point about substantial generation.28
But there are good reasons for thinking he is not. First, Aristotle
does not extend the argument of the chapter to substantial generation
until 190b1. Prior to that, he seems to be speaking only about changes
in the non-substantial categories. Second, and more importantly,
there is no evidence that Aristotle actually thinks of this as a case
of genuine substantial generation. The only clear examples of sub-
stantial generation in Physics i.7 are cases of biological generation (e.g.
when an animal or plant comes to be from sperma, 190b4). And in T2
(GC i.4, 319b8–21) the case of a statue coming to be from bronze
is explicitly treated as an alteration of the bronze, which Aristotle
contrasts with coming to be in the unqualified sense.29 At the very
least Physics 190a23–26 should be read with caution. For Aristotle
elsewhere says that we should not speak of a statue coming to be
from bronze without qualification, because that from which the
change begins does not persist (οὐχ ὑπομένοντος) but gets trans-
formed (μεταβάλλοντος) in the process of becoming a statue
(Metaph. vii.7, 1033a20–23).30
The evidence from Physics i.7 is thus not strong enough to force the
Inclusive Reading on us. In light of this I think Broadie (2004: 129–130)

28
For instance, Code 1976.
29
Code 1976: 357–358, 365 claims that in GC Aristotle thinks we can describe the same change as either
an alteration or substantial generation. This cannot be right. The central aim of GC i.1–5 is to
establish that substantial generation exists. And the way Aristotle does this is by showing that it is
distinct from and not reducible to changes in the non-substantial categories.
30
For an interpretation of this passage that is consistent with the Inclusive Reading, see Code 1976:
363ff.
160 devin henry
has it right when she says that the model of substantial generation on
offer in Physics i.7 is at best undeveloped with respect to the Survival
Requirement. While Aristotle is concerned to show that substantial
generation must proceed from an existing subject (190b1–5) – a claim
that forms the basis of his response to the Eleatic challenge in Physics
i.8 (191a30–31) – he is not focused on the further issue of whether or not
this subject also persists through the change. If this is right, then Physics
i.7 is not incompatible with GC, as it would be on the Inclusive
Reading. For Aristotle did not mean to commit himself in the Physics
to any particular view concerning the persistence of the subject of
generation. When it came to substantial generation Aristotle’s primary
concern was simply to establish that substantial generation exists as a
genuine type of change (cf. GC i.1, 314a4–6; i.2, 315a26–28). This only
required defending the Subject Requirement (Phys. i.7, 190b1–4). On
the one hand, Aristotle must show that there are genuine cases of
substantial generation where something new comes into being without
committing himself to the possibility of generation ex nihilo.31 He
meets this challenge by arguing that even in the case of substantial
generation the change proceeds from some pre-existing subject
(Phys. i.7, 190b1–4), and that this subject is the potential F (GC i.3,
317b14–25). This solution to Parmenides’ challenge does not depend on
the further claim that this subject also persists through the change to
become a constituent of the finished product.32 On the other hand,
Aristotle needs to show that when a new substance comes into being it
is not reducible to a change in one of the non-substantial categories
(GC i.1, 314a4–6). And this, we have seen, required him to reject the
Survival Requirement.

31
Aristotle agrees with Parmenides that generation ex nihilo is impossible.
32
Some argue that Aristotle must take this subject to be persisting through the change in order to
avoid the idea that substantial generation involves sheer replacement (Gill 1989: 7; cf. Code 1995).
If there were no persisting subject, the argument goes, then substantial generation would be a
case where the pre-existing subject perishes into nothing and is replaced by a product that
emerges out of nothing, which violates Eleatic principles that Aristotle accepts. I find this
argument for the Survival Requirement unmotivated by the text. Even if there is a philosophical
problem here (to me this sounds like a pseudo-problem), there is no suggestion that Aristotle was
concerned with it in the works on natural science, let alone that he developed the Survival
Requirement in response to it. Nor does positing a persistent subject solve that problem
(cf. Ebrey unpublished). It is also worth noting that in GC i.4 Aristotle actually characterizes
substantial generation in a way that sounds like sheer replacement. In T2 he says that with
substantial generation the thing “is transformed as a whole” so that “nothing perceptible remains
as its subject.” “Such an event,” he says, “is the coming to be of one sort of thing and the passing
away of the other.”
 Substantial generation in Physics i.5–7 161
If the arguments of this chapter are reasonable, then we should be
cautious about attributing to Aristotle in Physics i.5–7 the view that sub-
stantial generation involves a persistent subject. Not only is there a shortage
of conclusive evidence in those chapters for the Inclusive Reading, the bulk
of the evidence outside Physics i shows that Aristotle had little reason to
endorse it and every reason to reject it.33

33
Phys. i.9, 192a13–14 remains problematic for this reading. There Aristotle uses “what persists” as
shorthand for matter. But this may just be a simplification of the Physics i.7 doctrine that ignores the
difference between qualified and unqualified generation. As we have seen, when he is careful to
distinguish the two, he does so by reference to the survival requirement (cf. Metaph. i.3, 983b7–19).
 chapter 9

A dynamic ontology: On how Aristotle arrived

at the conclusion that eternal change
accomplishes ousia
Diana Quarantotto

Nature is difficult. As Aristotle reminds us at the opening of Physics i, the

principles of nature are obscure to us, at least at the beginning, although
they are clearer in their own right. So, Aristotle’s preferred methodology in
approaching the study of nature is to start with what is more familiar to us
and to move from there to what makes more sense in its own right.
Aristotle’s use of the craft analogy is a well-known instance of this meth-
odology. He begins with a miscellany of craft-related examples to elucidate
the distinction between the four causes before moving to its application to
nature, where the four causes are much trickier to disentangle.
This chapter sets out to show that Aristotle applies this methodology
also to a fundamental aspect of the distinction between being and change.
Broadly, Aristotle starts his physical inquiry by relying on a clear-cut
distinction between being and change and by giving priority to being
over change. Going on, approaching more and more nature’s inner func-
tioning and deep structure, and considering things from a different, more
fundamental and more revealing perspective, Aristotle arrives at the con-
clusion that the substantial being (ousia) of natural entities is accomplished
through a particular kind of change (eternal change), and hence that being
in its primary sense is itself dynamic.1 This is a major innovation by
Aristotle within the Greek tradition of natural philosophy and of its
treatment of the relation between being and change. Aristotle’s Physics
provides a long and important segment of the investigatory path that
brought Aristotle to this result.

1
While reconstructing Aristotle’s research on this issue, I shall clarify and define progressively what I
mean by ‘ousia accomplished through eternal change’, and what the interpretation proposed here is
an alternative to.

162
 A dynamic ontology 163
The topic proposed here is not easy to tackle or even to frame. As we
shall see, the thesis that being is accomplished through change is formu-
lated along a research path and from a point of view that, albeit funda-
mental, are both rare and not fully articulated by Aristotle. However, there
is enough evidence in Aristotle’s writings to go on, and the matter is too
important because of its implications just to be left aside. Basically, the
issue is intimately connected to how the prime mover can be both the first
principle of being and the first principle of eternal change for natural
entities. Of course, my remarks here do not amount to a full treatment
of the topic. Take them as early steps along a road that has been little
travelled,2 but which may make a big difference.

I The initial evidence: The road less travelled by

Our starting-point will be GC ii.10, 336b26–337a7. This is one of the rare
passages where Aristotle offers an overview of the natural world, looking at
it, as he says elsewhere, ‘from above’ (GA ii.1, 731b23). This means looking
at it from the point from which ‘nature and the heavens are suspended’
(Metaph. xii.7, 1072b14), i.e. from the perspective of the prime mover
(PM), and to follow the path that proceeds from this principle towards the
different levels of the hierarchy of the natural world.3
[The continuity of generation and corruption] happens with a good reason.
For we say that nature, in all things, always desires what is better, and that
being is better than not-being (it has been said elsewhere how many senses
there are in which we say ‘being’), but it is impossible for this [i.e. being] to
belong to all things, since they are far removed from the principle. Therefore
god filled up the whole in the way that was left, by making generation
unceasing. For this was the way to connect being together as much as possible,
since coming-to-be eternally and generation are the nearest things there are to
ousia. The cause of this, as has frequently been said, is circular locomotion,
since this alone is continuous. That is why also the other things that change
into each other in respect of their affections and powers, as do the simple
bodies, imitate circular locomotion. For when air is generated from water and
fire from air, and in turn water from fire, we say that the generation has come
round in a circle, because it has returned again to the beginning. So also
rectilinear locomotion, by imitating circular locomotion, is continuous.

2
To my knowledge, the only scholar who has addressed this area is Michael Frede (2000: 12–17). He
claims that the relation between being a sensible substance and change is much more intimate than
one may think, i.e. that the ousia of natural entities is identical with a certain pattern of change.
3
On the distinction between the path towards the principle and the path from the principle, see
NE i.2, 1095a30–b4.
164 diana quarantotto
The immediate goal of the passage is to provide a teleological explanation
of the continuity and eternity of natural generation. To do this Aristotle
sets natural generation in the wider net of the relations that link it to
celestial bodies and to God. Aristotle describes the world as a hierarchy
ordered according to the teleological principle that being is better than not-
being and according to the degree with which different kinds of natural
entities manage to accomplish being. Being is conceived as eternal being
and as ousia.
Aristotle claims that, although every natural entity strives for being, this
cannot belong to every one of them. Among natural entities, being belongs
to celestial bodies, which accomplish it through their circular locomotion,
whereas it does not belong to sublunary entities. Yet sublunary natural
entities, due to their striving for being, accomplish eternal generation,
which is described as the nearest thing to ousia. This idea of closeness
and of approximation to being and ousia is also expressed by means of the
concept of mimêsis: eternal generation is a mimêsis of the process whereby
celestial bodies accomplish being/ousia. Lastly, a divine principle is
described as what first causes both celestial locomotion and sublunary
generation to be eternal. Arguably, the likeliest candidate for this role is
the first unmoved mover,4 who is indeed both the first cause of being and
the first cause of the eternal movements of natural bodies.5
The topic of this passage is also tackled in GA ii.1, 731b18–732a3 and DA
ii.4, 415a23–b7. However, in this On Generation and Corruption passage
there are at least two particular points that offer an unusual perspective on
the matter.
First, Aristotle employs a single notion of ‘being’ (eternal being con-
ceived as ousia) and provides a gradualist perspective on it. He adopts a
point of view whereby, strictly speaking, only God and eternal natural
entities have being/ousia, whereas perishable natural ones just manage to
get as close as they can to this goal. Aristotle’s purpose is to explain the way
that being stems from a single divine source and reaches towards all natural
substances.
Secondly, eternal change is presented as the means whereby eternal
natural entities achieve being/ousia and perishable ones approximate to
it. Here eternal change is considered not simply as what brings about ousia,

4
Joachim 1922: ad loc.; Kahn 1985: 185; Williams 1982: ad loc.
5
In this passage, Aristotle uses locutions that credit God with a providential action and that are, I
think, rightly regarded as figurative (Kahn 1985: 185; Sedley 2007: 168). However, I shall claim that,
this metaphorical language notwithstanding, the passage provides a true Aristotelian view on God’s
causal role.
 A dynamic ontology 165
viewed as a mere result of change, but, more intrinsically, as what
constitutes it. Indeed, while being is conceived rather literally as ‘full
being’ and applied primarily to God, coming-to-be eternally and genera-
tion are said to be ‘the nearest things there are to ousia’, what ‘connects
being together’ (suneiroito to einai) and the means whereby God ‘filled up
the whole’ (suneplêrôse to holon). In other words, sublunary generation is
described as a process that blends parts of being and fills up gaps in order to
accomplish full being.

II Some objections from the road more travelled by

The perspective offered by On Generation and Corruption ii.10 clashes with
some common ideas both about the relationship between being and
change and about the relationship between the being of the prime mover
and that of natural entities.6 A possible formulation of these common ideas
might be the following:7
The ‘being-est’ of beings is the PM, which is eternally unchanging. The
PM is the first cause of being for everything. Moreover, it is also the first
cause of the eternal change whereby natural entities imitate its perfect
being. However, the being of celestial bodies lies not in their eternally
changing but in their eternality. Similarly, the elements and natural corrup-
tible substances ‘aspire’ to perfect being by constituting an eternal-process,
not an eternal-process.
This interpretation ignores both the two points of the On Generation
and Corruption ii.10 passage that I have just highlighted: the gradualist
perspective on being/ousia8 and the idea that the ousia of natural entities is
accomplished through eternal change. According to this interpretation, the
reason natural entities are involved in an eternal change is to emulate God’s
perfect being, and not to accomplish their own being: having an ousia and
striving for eternal being are independent or loosely connected matters.
Secondly, this interpretation ascribes to Aristotle a universal application of
the clear-cut distinction between being and change: it conceives of the

6
Note, however, that the proposed interpretation of what Aristotle says in On Generation and
Corruption ii.10 is compatible with its other parallel passages (i.e. GA ii.1, 731b18–732a3 and
DA ii.4, 415a23–b7). The only difference between them is in focus and emphasis.
7
I picked up on these ideas reading the comments by the editor of this volume and one of the
anonymous readers, whom I thank for this. The unanimity of these comments encourages me to
think that this interpretation is de facto standard, although I have not found it explicitly formulated
in the literature.
8
Instead this interpretation recognizes a gradualist perspective on divine/perfect being.
166 diana quarantotto
eternity of natural entities as a mode of their being, which is distinct from
the eternal change that accomplishes it.
Now, I shall suggest that this interpretation, despite its sounding
Aristotelian, is simplistic and that the root cause of this is that it takes
concepts that properly belong to the starting phase of Aristotle’s physical
inquiry (the path towards the principle) and tries to apply them to a later
and more advanced phase, to which they do not pertain (the path from the
principle).
Let us consider the first point. A gradualist perspective on being/ousia is
rare in Aristotle’s writings. However, this is not sufficient reason to dismiss
it. Not only is it clearly documented by the On Generation and Corruption
ii.10 passage, but it is also not unique to it, as it is attested in other passages
from Aristotle’s writings,9 and, as we shall see, it was endorsed also by some
of Aristotle’s predecessors. Moreover, there are reasons that account both
for its rarity and for its importance within Aristotle’s thought.
A gradualist perspective on being/ousia is rare along the research path
that predominates in Aristotle’s writings. This is the path that proceeds
from what is better known to us and that corresponds to the starting phase
of Aristotle’s inquiry, which is focused on a local and domain-specific level
of explanation. Indeed, while proceeding along this path, and for a long
segment of it, we do not even know that there is a single, changeless, and
immaterial principle ‘from which nature and the heavens are suspended’.
Therefore, this initial inquiry might give the impression that the immanent
and local causality Aristotle focuses on can be made fully intelligible with-
out any appeal to a universal cause, or that the PM’s causal action concerns
some extra aspects of the being of natural entities (for instance, their
attempt to imitate the PM’s perfect being, as the standard interpretation
mentioned above claims). But this impression is misleading. For while
reaching the first principle and while starting the path that proceeds from
it, one gets a different view on the natural world. One learns that the PM’s
influence on the natural world is omnipresent. Moreover, along the path
from the principle, a gradualist perspective on being/ousia is perfectly
legitimate, and perhaps even necessary. Indeed, this path aims at showing
both that and how all beings stem from one single being, i.e. both that and
how being, in its primary sense, stems from a single source towards the
natural world. The PM is the source of being for all natural entities, and

9
Examples of gradualist perspectives in Aristotle: Cael. i.9, 279a23–30; ii.12, 292b1–293a15; Meteor.
iv.12, 389b29–390a9, esp. 390a4–7; GC i.3, 318b35–319a1; Cat. 5, 2b22–28; Metaph. ii.1, 993b24–31.
On the gradualist aspects of Aristotle’s ontology, see Morrison 1985 and Rashed 2005: lxiii–lxxxv.
 A dynamic ontology 167
not only of their approaching its perfect being. Despite the rarity of his
treatment of this issue, Aristotle does not give us reasons to think that the
causal action of the PM on the natural world is twofold, or that it is limited
to natural entities’ approaching its perfection and therefore does not
concern their own being – as the standard interpretation claims.10 In
other words, Aristotle does not give us reasons to think that the passages
where he uses a gradualist perspective on being/ousia and describes natural
entities’ striving for eternal being concern something different from and
further than their own being. Indeed, these are the rare passages where we
find clues as to how the PM’s causation of being is exerted.11
Similar points can also be made about the second aspect of the standard
interpretation, i.e. the clear-cut distinction between being and change. As we
shall see, a clear-cut distinction between being and change is proper to the
path that proceeds from what is more familiar to us towards what makes
more sense in its own right. But things change when one looks at the natural
world from above. Indeed, the PM is the first cause of being for all natural
entities and it is such insofar as it is the cause of their eternal change. Again,
Aristotle does not give us reasons to think that the PM’s causal action is
twofold, i.e. that there is a distinction between the PM’s causation of being
and the PM’s causation of eternal change.12 There is simply no other way
whereby the PM exerts its causal action on the natural world than by
determining an eternal change. This is why the first principle is presented
as a mover and indeed called the ‘first mover’. So, insofar as the PM is
concerned, the causation of being is causation of change. I suggest that the
PM’s causation of being is causation of change because natural entities
accomplish being/ousia through eternal change.13
This idea may sound less awkward if one recalls that the being of the PM
is not a state but a process, i.e. a living activity. So, change – or better, a

10
Of course, this means neither that the PM is the only cause of the being of natural entities nor that
their being is identical with God’s being.
11
It is known how scarce Aristotle’s statements on this issue are. Even Metaphysics xii does not offer
much detail. There, the divine substances are called ‘primary substances’ (1074b9). However, as
Frede observes (2000: 50), Aristotle does not tackle the issue of the sense in which they are so and
hence, more generally, the question of how these primary substances are causes of being for other
substances.
12
E.g. Metaph. xii.8, 1073a23–25, 1074b35–36.
13
An alternative hypothesis may be the following: the PM causes eternal change directly and the being
of natural entities indirectly. I do not see clear signs of this two-step causality in Aristotle’s writings
and believe that this hypothesis is undermined by all the passages where the PM is presented as the
first being and the first source of being in a sense that seems stronger than the abovementioned one
(besides On Generation and Corruption ii.10, see, for instance, Metaph. xii.8, 1074b35–36 and
Cael. i.9, 279a28–30).
168 diana quarantotto
particular kind of change (eternal change) – can be viewed as the means
whereby natural entities get as close as they can to that particular kind of
dynamic being that the PM both is and is the source of.14 Indeed, this may
be why in Cael. i.9 279a28–30 Aristotle presents the PM as the source from
which ‘is suspended the being and life for other things, for some in more
distinct form, for others in a dim way’, and why at the beginning of
Physics viii Aristotle describes eternal change as a sort of life for all things
composed by nature.15
Naturally, it is always possible to make distinctions and to claim that
there is a difference between an eternal-process and an eternal-process.
However, I wonder whether this distinction is so important in Aristotle’s
work, especially where the path from the principle is concerned. I suggest,
by contrast, that much can be learnt by emphasizing the conceptual shifts
that the view from above engenders, among which the idea that being is
accomplished through eternal change.
Of course, by claiming that the ousia of natural entities is accomplished
through eternal change, I do not mean to deny that Aristotle also speaks of
ousia in other senses. Rather, my proposal is that these different senses, or at
least some of them, stem from and belong to different and progressive
paths of Aristotle’s investigation. Moreover, my proposal is that of empha-
sizing that only once the path from the principle is followed is Aristotle’s
inquiry on nature really completed. Indeed, this is, at least in part, the
methodological and theoretical meaning of Aristotle’s introduction of first
philosophy and its distinction from physics, i.e. of Aristotle’s conviction
that first philosophy is the ultimate fulfilment of the human desire to
know, and that the understanding of God, conceived as the principle of all
things and as a non-physical entity, is necessary to fully understand the
natural world.16 This is especially important to recall, since each of these
two paths provide a different view on the natural world. Indeed, grasping
the cause of all things shows fundamental connections and relations that
are invisible from the bottom up, and hence requires a modification of our
understanding of the things that are caused by the first principle. Thereby

14
It is true that the PM is of itself entirely changeless. However, this does not imply that a similar
distinction between being and change, or even a similar extraneousness of being from change, holds
also for natural entities (I tackle this issue in Quarantotto 2014). Note, moreover, that there is
another reason the particular kind of relationship, suggested here, between the immutable eternity
of the PM and the mutable one of natural entities may sound less awkward: it has a precedent in the
Platonic description of time as a ‘movable image of eternity’ (Tim. 37c).
15
According to DA ii.4, 415b13, the being of living things is life: it is not a state but a particular kind of
process. I tackled this issue in Quarantotto 2010.
16
Frede 2000 and 2004.
 A dynamic ontology 169
the view from above determines a revision of (at least some of) the
conceptual tools and frameworks that originated from, and were used in,
the previous, more localised, perspective. The idea that ousia is accom-
plished through and realised in eternal change is among these revisions.17
Now, in what follows, I shall provide some independent evidence that
Aristotle has taken the point of view from above seriously. I shall suggest
that, by adopting this point of view and by claiming, as a result, that the
ousia of natural entities is accomplished through eternal change, Aristotle
inserts a novelty into the age-old debate on the relationship between being
and change, and on the status of the being of natural entities. Moreover,
this further evidence will help us understand the meaning of the On
Generation and Corruption passage and reconstruct the research path that
Aristotle followed to reach the result presented there.

III Some independent evidence on ousia, eternity,

and eternal change
Within Aristotle’s Physics, the topic of natural entities’ ousia is dealt with
for the first time in Book ii, chapter 1. Book i provides the first steps of an
investigation into the principles of natural science and only mentions the
problem of ousia (Phys. i.7, 191a19–20). Instead, in Physics ii.1, the issue of
ousia is explicitly addressed (193a9–11, 16, 20, 25). Aristotle considers two
possibilities: the ousia is either matter or form. The first hypothesis is
attributed to some of Aristotle’s predecessors and is argued for with their
own arguments. According to these thinkers, only matter can be the ousia
because only matter meets a certain requirement: to be eternal.
Some identify the nature or substance of a natural object with that immediate
constituent of it which taken by itself is without arrangement . . . As an
indication of this Antiphon points out that if you planted a bed and the
rotting wood acquired the power of sending up a shoot, it would not be a bed
that would come up, but wood which shows that the arrangement in accordance
with the rules of the art is merely an accidental attribute, whereas the substance is
the other, which, further, persists continuously through the process. But if the
material of each of these objects has itself the same relation to something

17
I suspect that one of the causes of this conceptual revision is the shift, which results from taking up
the view from above, from a perspective that focuses on finite entities and processes to a perspective
that focuses both on eternal entities and processes and on the way in which finite entities and
processes are part of something infinite. This is also suggested by a closely related topic, i.e. by the
difficulties arising from the application of the same principles to both perishable and eternal sensible
substances and by the enormous strain under which the notions of these principles are put (cf. Frede
2000: 17). I shall tackle this last issue in section VI.
170 diana quarantotto
else . . . that, they say, would be their nature and substance . . . For
whatever any one of them supposed to have this character . . . this or
these he declared to be the whole of substance, all else being its affections,
states, or dispositions. Every such thing they held to be eternal (for it could not
pass into anything else), but other things to come into being and cease to be
times without number). (193a9–28)18
This passage shows that the idea that ousia is something eternal is an
ancient and traditional one. Eternity is considered by the pre-Socratics
both a distinguishing trait of ousia and a fundamental requirement some-
thing must fulfil to be an ousia. These thinkers are presented by Aristotle as
holding that only what is eternal can be an ousia, and that therefore what
changes, what is generable, and perishable, cannot be an ousia. This is why
they ruled out the possibility of attributing the role of ousia to form.
Aristotle emphasises this point also in other passages. For instance, in
Metaph. i.3, 983b6–18, he claims:19
Of the first philosophers, most thought the principles which were of the
nature of matter were the only principles of all things; that of which all
things that are consist, and from which they first come to be, and into which
they are finally resolved (the substance remaining, but changing in its mod-
ifications), this they say is the element and the principle of things, and
therefore they think nothing is either generated or destroyed, since this sort of
entity is always conserved . . . So they say nothing else comes to be or ceases to be;
for there must be some entity – either one or more than one – from which all
other things come to be, it being conserved. (trans. Ross 1928)
The idea that ousia is something eternal, together with the associated
contrast between being and becoming, had a lasting career: it was also
endorsed by Plato, who in certain contexts conceives of being as what
always is and never comes to be and contrasts it with what always comes to
be and never is.20
First then, in my judgment, we must make a distinction and ask: what is that
which always is and has no becoming; and what is that which is always
becoming and never is? That which is apprehended by intelligence and reason
is always in the same state; but that which is conceived by opinion with the
help of sensation and without reason is always in a process of becoming and
perishing and never really is. (Tim. 27d–28a, trans. Jowett 1875)

18
Translations are by Hardie and Gaye 1930, sometimes with small changes. Emphasis is always mine.
19
Cf. also Metaph. i.3, 984a9–16 (esp. 983b10).
20
Significantly, this is also the point of view adopted by Aristotle in On Generation and Corruption
ii.10: he identifies being with eternal being and claims that being cannot belong to all natural
entities. However – breaking with this tradition – here Aristotle also claims that eternal generation is
the nearest thing to ousia.
 A dynamic ontology 171
Aristotle accepts this position, but only to a certain extent. In Phys. ii.1,
193a9–28 and 193b8–11 he accepts the criterion for ousia lying behind
Antiphon’s argument, but uses it to prove what Antiphon and other pre-
Socratics denied: i.e. that form is (in a relevant sense) eternal, and so can be
an ousia.21
Again man is born from man but not bed from bed. That is why people say
that the shape is not the nature of a bed, but the wood is – if the bed
sprouted, not a bed but wood would come up. But even if the shape is art,
then on the same principle the shape of man is his nature. For man is born
from man. (193b8–12)
This shows that, according to Aristotle, form is an ousia (also) because it is
eternal and that its eternity is accomplished by means of circular generation
(‘man is begotten by man’).
In the first chapters of Physics ii Aristotle uses the idea that ‘man is
begotten by man’ also while dealing with the issue of the kind of separation
that is characteristic of natural forms (194b12–13): ‘[The student of nature]
is concerned only with things whose forms are separable indeed, but do not
exist apart from matter, for man is begotten by man and by the sun as well.’
In this connection, he criticises other kinds of separation, especially the one
Plato attributed to forms:
Thus on the second account of nature, it would be the shape or form (not
separable except in account) of things which have in themselves a principle of
motion. (193b3–5)
The holders of the theory of forms do the same, though they are not aware of
it; for they separate the objects of natural science, which are less separable than
those of mathematics. (193b35–194a1)
From these passages and remarks it may be deduced that the point of
Aristotle’s criticism is to show that forms need not be separate in order to
be eternal, since ‘man is begotten by man’.22 More generally, drawing the
right consequences from the fact that ‘man is begotten by man’ is sufficient
to prove – against the pre-Socratics – that form too is eternal, and – against
Plato – that to ensure the eternity of forms it is not necessary to conceive of
them as separate from matter and change.

21
This move is prepared by Aristotle’s insistence in Physics i on the requirement that any principle, as
such, must be permanent and eternal (188a27–28, 192a25–34; Metaph. iii.4, 1000b24–28; vi.1,
1026a16–17). This makes impossible the interpretation that Aristotle’s acceptance of the pre-
Socratics’ criterion for ousia in Physics ii.1 is simply dialectical.
22
Metaph. i.6, 987a 29–b18; iii.2, 997b6–12; vii.8, 1034a2–5.
172 diana quarantotto
In sum, Physics ii.1 records that Aristotle’s predecessors consider eternity
a necessary requirement for something to be an ousia, that Aristotle himself
shares this idea and that he nevertheless believes this requirement can be
met without also endorsing (at least not completely) the contrast between
being and becoming; in fact, Aristotle believes that a certain kind of
becoming (the one picked out by the ‘man is begotten by man’ formula)
is what enables forms to be eternal and hence to be ousia.
Now, this is very close to what Aristotle claims in On Generation and
Corruption ii.10: eternal generation is the means whereby perishable
natural entities accomplish, as far as they can, eternal being and ousia.
Hence, Physics ii provides independent evidence to support the
proposed interpretation of On Generation and Corruption ii.10.
Moreover, Physics ii shows that the idea expressed in On Generation and
Corruption ii.10 – that the ousia of natural entities is accomplished through
eternal change – is already germinating in the starting phases of Aristotle’s
investigation into nature.23 However, this idea still needs to be developed
and clarified. We still need to know what change is and what kind of
change is able to accomplish ousia. And, especially, we still need to under-
stand how and in what sense change (even a special kind of change)
accomplishes ousia.

IV The first steps of Aristotle’s inquiry into change: Things

more familiar to us
The idea that change accomplishes ousia is an enormous innovation by
Aristotle in the traditional debate about the relation between being and
becoming. There is even something paradoxical about it, since tradition-
ally – e.g. according to the Heraclitean flux theory – what was contrasted
with being and ousia was an eternal change.24 Naturally, in the Sophist,
Plato rejects the contrast between being and becoming, judging as absurd
the hypothesis that to pantelôs on is without movement or life and putting
forward an ample ontology in which being includes ‘all changeless things
and all changing things’ (248e–249d). But Aristotle goes a step further than
this. He claims that being/ousia not only includes change and all changing

23
Note, moreover, that in Phys. i.7, 190b17–20 Aristotle speaks of the principles of change and of being
without making any clear-cut distinction.
24
On the flux theory, cf. Plato Theaet. 152e2–8, 160d6–8; Crat. 402b1–5. Of course, the eternal change
that, according to Aristotle, accomplishes ousia is different from the eternal change of the flux theory
(Phys. viii.3, 253b6–254a1, 8, 265a2–12).
 A dynamic ontology 173
things, but also that the being/ousia of natural entities is accomplished
through eternal change.
Aristotle introduces and develops this innovation gradually. Here I shall
highlight the main steps of the path that, within the Physics, brought him
from the traditional idea that change is something highly problematic and
close to not being to the idea that it is the means whereby being, in its
primary sense, is accomplished. I shall claim that, in the starting phases of
his investigation (especially in Physics i and iii), Aristotle employs the
ordinary perspective – also adopted by his predecessors – which distin-
guishes being from change, and gives priority to being over change. This
perspective highlights the problematic nature of change and makes it
appear close to not-being. Later, in a more advanced phase of his inquiry,
Aristotle introduces a different perspective. In Physics viii, the concept of
‘change’ is extended and put under strain: Aristotle presents a kind of
change whose distinguishing features seem antithetical with those whereby
he defined change in Physics iii.1–3: e.g. the former is said to be something
complete (teleion),25 whereas the latter is described as incomplete (ateles).26
The change introduced in Physics viii is circular locomotion, which is one
of the changes that, according to On Generation and Corruption ii.10,
accomplishes ousia and which is introduced, tellingly, at the beginning of
Physics viii as ‘a sort of life, as it were, to all naturally constituted things’.
Let’s start with Physics i. In this book, Aristotle deals with the first
principles of natural entities and addresses some of his predecessors’
theories, including the most radical and influential one: Eleatic monism
and immobilism. He criticises both the Eleatic theory that being is one and
changeless, and one of its main assumptions: the idea that being has only
one meaning (186a22–25). In Aristotle’s view, this assumption was the
main cause that brought both the Eleatics to their conclusions and other
philosophers to unsatisfactory accounts of change (185b31–32, 186b35–
187a10).
In Physics i.2–3, Aristotle challenges the Eleatic theory by setting it
against the thesis that ‘being’ and ‘one’ have many different senses and
hence that being is not one in the way they think it is (185b25–186a3).
Thereby the issue of change is dealt with on the basis of a theory of being
and of its meanings. This becomes particularly clear in chapter 8, where the
problem of change is directly addressed. Here Aristotle presents the Eleatic
dilemma used to deny the possibility and intelligibility of change. Change

25
Phys. viii.8, 264b27–28; viii.9, 265a13–27 (esp. 17, 23).
26
Phys. iii.2, 201b32; viii.5, 257b8–9; Metaph. ix.6, 1048b28.
174 diana quarantotto
is denied because a changing thing should meet two requirements that
seem to be incompatible: what changes must come from something that
both underlies (i.e. from being) and is not already present (i.e. from non-
being) (191a27–31). So, a changing thing appears to be impossible, because
it must both be and not be, or it must come to be neither from being nor
from not-being. The main tool Aristotle uses to solve this dilemma is his
innovative idea that being has many senses: what changes, changes both
from what is, but not insofar as it is, and from what is not, but not insofar
as it is not (191b13–15, 17–18). So a changing thing can both be and not be
without there being any contradiction and without violating the principle
of excluded middle.
The inquiry into change continues in Physics ii, where Aristotle focuses
on teleological processes and, as we have seen, hints at the idea that eternal
change is the means whereby natural entities accomplish ousia. However, it
is only in Book iii that Aristotle finally provides a definition of change. In
Physics iii.1 Aristotle defines change as the ‘actuality [entelecheia] of that
which is potentially as such’ (201a10–11) and then rephrases this definition
in various ways. These different formulations record the difficulty of
defining change and of grasping its way of being. This is particularly
clear with ‘energeia ateles’ (201b31–33), which – especially when energeia is
replaced by entelecheia (257b8–9) – is a genuine oxymoron, at least on the
level of the expression: it says that change is an incomplete completeness,
i.e. that it both has and does not have the telos (entelecheia ateles).
In Aristotle’s view, the problematic nature of change stems from the fact
that ‘it can be classed neither in the dunamis of things that are nor in the
energeia’ (201b28–29): again, also from this point of view, change seems to
violate the principle of excluded middle. Because of this, change is
something indeterminate (aoriston). This is why some of Aristotle’s pre-
decessors defined it in privative terms as difference, inequality, or not being
(201b20–26). Aristotle’s solution is to characterise change neither as
potentiality (dunamis) nor as actuality simpliciter (energeia haplê), but as a
sort of actuality (energeia tis), i.e. as an incomplete actuality (energeia ateles)
(201b31–33).27
From a terminological point of view, Aristotle’s account of change is
built around the notion of goal (telos/ergon): change is defined according to
its relation with the telos. Change is something insofar as it is an entelecheia/
energeia; change lacks something insofar as it is ateles. What brought
Aristotle’s predecessors to define change in terms of privation (i.e. the

27
Cf. also Metaph. ix.6, 1048b28; EN x.3, 1174b2–5.
 A dynamic ontology 175
indeterminate nature of change) is conceived by Aristotle in terms of
incompleteness: the lack of the telos.
Change is incomplete – says Aristotle – because the potential being it is
the entelecheia of is itself incomplete (201b31–33). The potential being is
incomplete because it lacks the telos towards which change is directed, and
change inherits this privative nature. Hence, change is characterised by a
lack, i.e. by the fact that it does not have the telos towards which it is
directed. It is a being that comprises not-being within itself.
This is difficult to grasp, but not impossible to be (202a2–3). So Aristotle
comments on the result reached in this phase of his investigation into
change and its relationship to being.
Now, what is the scope of the account of change provided in
Physics iii.1–3? This is a difficult question, but it is necessary to tackle it,
especially if we want to understand how eternal change is able to
accomplish ousia. Scholars tend to see it as a sign of correctness (or, at
least, as a strength) for an interpretation of the definition of change
provided in Physics iii.1–2 if it picks out both finite and infinite changes.
Instead, here I shall explore another possibility. In my view, there are
reasons to think that the definition presented in Physics iii.1–2 is modelled
primarily on finite changes,28 since they are what is more familiar to us and
are therefore the first objects of Aristotle’s scientific inquiry into change.29
This is suggested, first of all, by the way the investigation into change is
set up from its very beginning in Physics i. Here Aristotle describes change
as something that occurs between opposites (188a19, 188b21–26). This
description suits only finite processes. For, as we learn from
Physics viii,30 changes that occur between opposites cannot be continuous
and eternal.
The same point of view is taken up in Physics iii.1–3. This inquiry is
introduced both by a reference to the oppositions characteristic of
change (201a3–9) and by a distinction between what is only in actuality
and what is both in potentiality and in actuality (200b26–27). What is
both in potentiality and in actuality is then identified with the kineton
(201a19–25). This suggests that the change Aristotle is here focusing on is

28
Cf. Metaph. iii.4, 999b9–10: ‘no change is infinite, but every change has an end’.
29
This, however, does not imply that the definition given in Physics iii.1–2 cannot be adapted
subsequently to eternal motions. Indeed, it can be adapted since eternal motions can be considered
also as combinations of finite parts (cf. section VI). The question whether and in what sense eternal
motion is a kinêsis is closely connected with the question of whether it is or it is not an energeia (in
the sense of Metaphysics ix.6). On this latter issue, see my (Quarantotto 2014).
30
Phys. viii.8, 261a28–b3, 2, 252b7–12, 252b28–253a2, 7, 261a29–b3.
176 diana quarantotto
temporally limited: a kineton is conceived as something that does not
always have a certain attribute (i.e. it is not only and always in actuality a
certain thing), but, if it does not have it at t1, it has the potentiality to
acquire it and to have it at t2.31
The same conclusion can also be drawn from the fact that, in Physics iii,
change is characterised as incomplete, i.e. in terms of its lacking the telos/
end-point towards which it proceeds.32 Indeed, in Phys. viii.9, 265a20–21
Aristotle explicitly links the notion of finiteness to that of incompleteness,
by claiming that a finite change is incomplete.
Note, moreover, that the notion of ‘infinite’ (and the questions whether
it is something or not, and what it is) is introduced and discussed only after
the conclusion of the inquiry into change provided in Physics iii.1–3, i.e. in
chapters 4–8 of this book. Also, the very possibility and existence of eternal
change are demonstrated not before Book viii. By contrast, in Book viii,
finite change is considered evident and used as a starting-point of the
investigation. At the end of this inquiry eternal change is then described
with properties that are antithetical to those ascribed both to intermittent
change and to change as defined in iii.1–3: eternal change does not occur
between opposites and is complete (264b27–28).
Another clue in this direction is given by the way in which the definition
of kinêsis provided in Book iii is used in Book viii. In Phys. viii.1,
251a8–16, from this definition Aristotle infers that ‘there must be some-
thing burnable before it is burnt and something able to burn things before it
burns them’, i.e. that the subject of change that this definition picks out is
something that is not always in motion. Now, in Phys. viii.7, 261a31ff.
Aristotle makes the following claims: 1) this kind of subject of change
pertains to change insofar as it occurs between opposites; 2) not all changes
occur between opposites; 3) eternal change does not occur between
opposites and this is why it is eternal. In other words, in this passage
Aristotle claims that the properties of change and of what changes he

31
Cf. also Phys. viii.1, 251a8–16; iii.1, 201b5–11.
32
Phys. iii.2, 201b31–32. Scholars generally maintain that the interpretation that allows the definition
of kinêsis given in Physics iii.1–3 to pick out both finite and infinite changes is the one according to
which change is the actuality of a potentiality for changing. Instead, I believe that this interpretation
either does not explain why change is incomplete or, if it does, it does not allow the definition of
kinêsis to pick out infinite changes. Arguably, the only way for this interpretation to account for the
incompleteness of change is to claim that change does not accomplish its final goal (e.g. a house =
goal1), but only the one that is the actuality of a potential for changing (e.g. the construction of a
house = goal2). Now, if this is why change is incomplete, then the definition does not pick out
infinite processes. For, on Aristotle’s view, only circular processes are infinite, and circular processes
are said to be complete in the sense that they always reach their goal in each of their points, i.e. they
always reach the goal that, in this interpretation, kinêsis does not reach (goal1).
 A dynamic ontology 177
deduced in Phys. viii.1, 251a8–16 from the definition of change provided in
Book iii do not belong to eternal change or to what is in eternal motion.
If this is true – if the definition of change given in Physics iii.1–2 is
provided to account, at least primarily, for finite processes – then change
thus defined is not a good candidate for being the means whereby ousia is
accomplished. Because of its relationship with what is not, it cannot be
eternal and complete. Put more carefully, because of its occurring between
opposites, change must come to a stop once it has reached its telos.33 Using
the language of On Generation and Corruption ii.10, we may say that
change thus defined cannot accomplish ousia because it does not ‘connect
being together’. However, this is not surprising, given Aristotle’s progres-
sive research. Indeed, in Physics viii Aristotle not only – as just noted – uses
the concept of change introduced in Physics i and iii to tackle some
arguments against the possibility of an eternal motion, but he also sets
up to demonstrate the existence of an eternal motion. Moreover, this
eternal motion is identified with the circular motion of celestial bodies
and, unlike the kinêsis dealt with in Physics iii.1–3, it is described as
complete (teleion).34 So it seems that Aristotle’s inquiry proceeds by first
focusing on finite changes (which are more familiar to us) and then going
on to consider infinite changes (which are less familiar to us, but arguably
make more sense in their own right). Put differently, Aristotle moves from
one kind of change, which cannot accomplish ousia, to another kind of
change, which has exactly the features needed to accomplish ousia: eternity
and completeness.35
Let us see then what circular locomotion is and whether Aristotle’s
account of it in Physics viii gives us some clues to understanding how
and in what sense eternal change accomplishes ousia.

V Eternal changes accomplishing ousia (circular

locomotion): Things less familiar to us
In Physics viii the existence of an eternal change is the object of a
demonstration. Aristotle’s inquiry begins with the question ei esti (is
there eternal change?), and then, once it is established that there is eternal
change, it continues with the question ti esti (what is eternal change?).

33
Cael. i.8, 277a14–23; Phys. viii.7, 261a31–b15, 8, 261b31–36.
34
Phys. viii.8, 264b27–28, 9, 265a17; Cael i.2, 269a18–21. Phys. viii.9, 265a17 suggests that also
rectilinear motion can have a degree of completeness, which it possesses insofar as it doubles back
(Phys. viii.9, 265a20–22).
35
Eternal motion is also complete: Phys. viii.8, 264b9–28, 9, 265a13–22; Metaph. xii.6, 1071b10–11.
178 diana quarantotto
Unlike intermittent change (i.e. change of things that are sometimes in
motion and sometimes at rest, and which occurs between opposites),
eternal change (i.e. the change of one single thing that is always in motion)
is not considered evident, i.e. something we can observe in ordinary
experience and take for granted. Moreover, both the existence and struc-
ture of eternal change are studied starting respectively from the existence
and from the structure of the intermittent change: here, too, Aristotle
proceeds from what is more familiar to us towards what makes more sense
in its own right.
Insofar as the structure of eternal change is concerned, Aristotle defines
it by first showing why motion between opposites cannot be eternal and
what ‘modification’ it must undergo to be so. He claims that motion
between opposites cannot be continuous and eternal but only consecutive.
This is because what moves from opposite to opposite with a rectilinear
motion, once it has reached the terminus ad quem, if it is to go on, must
move on a straight path at the same time from one point towards the same
point. So this kind of change cannot be continuous, but there will be a time
interval between changes. In sum, ‘motion that doubles back along a finite
line is a composite motion consisting of two parts, while if it does not
double back, it is incomplete and perishable’ (265a20–22).
To be continuous and everlasting, a motion must be such that what
moves makes its movement at the same time from one point towards the
same point (264b10–11).36 This is not possible for rectilinear motion, but it
is possible for circular motion, i.e. it is possible if the beginning and end,
which in rectilinear motion are distinct and opposite, come to coincide and
become the same point, like bending a straight line so that its two ends
meet: the continuity and eternity of circular motion are conceived in terms
of the overlapping and coincidence of what is instead distinct and opposite
within straight and finite motion (264b27).
This is also why circular motion, unlike rectilinear motion, is complete
(264b27–28). Circular motion is complete since it has an end (telos), but no
limit (peras), i.e. since, unlike straight motion, each of its points is both a
beginning (archê) and an end (telos).37 Each point is an end, since each is
where the movement is completed; and each point is a beginning, since
each is where the movement ‘starts’ again and again. So, to have an end
without being limited means reaching completion without coming to a
stop. Whereas rectilinear motion is incomplete, since it does not have the

36
The possibility of an infinite rectilinear motion is ruled out in Phys. viii.9, 265a17–20.
37
Phys. viii.8, 264b9–28, 9, 265a27–265b1; Cael. i.9, 279b2–3.
 A dynamic ontology 179
telos towards which it proceeds, circular motion is complete, since it always
reaches this telos in each of its points.

VI The problem: How can eternal change accomplish ousia?

Do the structure and completeness of eternal motion enable it to accom-
plish ousia? How, and in what sense? What does it mean for a process to
accomplish ousia? Note that Physics viii does not explicitly aim at answer-
ing these questions. The book deals more generally with the existence and
structure of eternal change and with its cause. However, given that
Aristotle has never fully articulated the view from above, this book –
together with some passages from Cael. – is our only source of information
on this topic. Let us then attempt to provide an answer. We shall see that
this inquiry will give us clues to understand better how and in what sense
eternal change ‘connects being together’ and how it is the means whereby
God ‘filled up the whole’, as the On Generation and Corruption ii.10
passage claims.
Remember that it is mainly the relationship of change to not-being that
determined the contrast between being and becoming as formulated by
Aristotle’s predecessors.38 Moreover, in Aristotle’s account, this very same
relationship prevents kinêsis, as defined in Physics iii.1–2, from being
complete and everlasting, i.e. from having the features whereby the pro-
cesses that accomplish being/ousia are described in On Generation and
Corruption ii.10. Now, insofar as kinêsis (in the sense of incomplete and
finite change) is concerned, not-being is given by the opposition between
beginning and end: one of these points is the privation of the other. So it
seems that, in order to be eternal and complete, a process must ‘neutralise’
the not-being that is present in finite changes: again, borrowing the
language from the On Generation and Corruption ii.10 passage, it must
‘connect being together’ and ‘fill up the whole’. And, indeed, this is what
happens with circular motion: in circular motion the opposition between
the starting and the end points are neutralised because each point is both a
beginning and an end. This, however, does not mean that there is no
opposition at all between these points. If that were so, there would be a
state, not a process. So we probably need to distinguish between the
neutralization and the complete nullification of the opposition. With
this distinction in mind we may say that insofar as circular motion is

38
Cf. also Phys. i.9, 192a15, where Aristotle speaks of the ‘destructive power’ inherent in change.
180 diana quarantotto
concerned, opposition is in a sense absent and in another sense present, i.e.
it is not completely nullified but disabled by the circularity of the motion.
So here is my proposal: for a process to accomplish ousia it needs to
neutralise (without nullifying) the not-being, which an incomplete and
finite change instead includes; and this is what the coincidence of begin-
ning and end, characteristic of circular motion, amounts to; moreover, this
is what enables a process to go on eternally.
This proposal has several advantages. As we shall see below, the idea that,
in a circular process, the opposition between the starting point and the end
point (and hence the not-being) is neutralised is suggested by some crucial
passages of Aristotle’s description of circular motion. In particular, this
idea offers a way to explain some problematic and, apparently, conflicting
aspects of this description. Moreover, the proposed interpretation is a good
fit with the picture of the world’s structure offered in On Generation and
Corruption ii.10: it enables us to deepen our understanding of the way
whereby being/ousia stems from its primary divine source to the different
levels of the natural world’s hierarchy, as it gives celestial motion a
mediating role in this causal chain.
Let us start with the textual evidence. The kind of neutralization I am
suggesting is particularly clear in Aristotle’s description of circular motion
in Phys. viii.9, 265a27–b1. Here he claims that, whereas in rectilinear
motion the beginning, middle, and end points are defined (horistai), in
circular motion they are undefined (aorista), since each is at once a begin-
ning, a middle, and an end. This shows that there is a difference between
these points, but this difference is neutralised by their ability to interchange
their roles within the movement.
The hypothesis just sketched may explain why Aristotle gives apparently
conflicting descriptions of circular motion, saying sometimes that it occurs
between opposites and sometimes that it does not. As we have just seen,
Aristotle claims that, insofar as circular locomotion is concerned, what
moves from point A moves at the same time and in virtue of the same
direction towards point A, without undergoing contrary movements. By
contrast, in other passages, circular motion is said to occur between points
that are either opposites to or, at least, different from each other. In
Metaph. viii.8, 1050b20–22, for instance, we are told that what moves
circularly is endowed with a potentiality of being and not-being in respect
of place. Similarly, in Metaph. xii.7, 1072b3–10, Aristotle claims that what
moves circularly, insofar as it moves, can be otherwise with respect to place,
unlike the first unmoved mover, which cannot be otherwise simpliciter. In
Cael. i.8, 277a24–26, Aristotle says that circular locomotion, just like
 A dynamic ontology 181
rectilinear locomotion, has in a way the opposites. Lastly, the matter that
celestial bodies are made out of is called pothen poi hulê (‘from one place to
another’), since it is potential being with respect to place,39 and this
potentiality is called by Aristotle ‘potentiality for opposites’.40
One way to account for this discrepancy is to attribute it to Aristotle’s
use of two points of view on circular locomotion. This hypothesis is
suggested especially by Phys. viii.8, 264b24–28 and Cael. i.8, 277a23–26.
Here the double nature of circular locomotion seems to stem from the
possibility of viewing it both as an infinite whole and as a combination of
finite parts. If one considers just a segment of the process, a body that
moves circularly moves between points that are opposites to each other.
Instead, if one considers the whole process, i.e. if one connects the end with
the beginning (Phys. viii.8, 264b27–28), the opposition disappears: the
circularity of the process neutralises, as it were, the opposition between
being and not being with respect to place.41
The explanation for why Aristotle considers circular locomotion from
these two perspectives may be found in the difference between the
completeness of circular motion and that of a state or that of the PM’s
activity.
The completeness of circular motion is different from the completeness
of a state (e.g. being pale in actuality).42 The completeness of a state is – let
us say – simple, whereas the completeness of a process is complex, as it
involves two elements: according to Aristotle, every change is from one
point to another point.43 If the opposition between these points is kept, the
process is incomplete and finite. Instead, if they are neutralised without
being nullified, the process is complete and everlasting without turning
into a state. Indeed, if the opposition were nullified, there would not be any
difference between a complete process and a complete state.
Similarly, the completeness of circular motion is different from the
completeness of the PM’s activity. Whatever it is, this activity is not a
state, but shares the simplicity of a state: the PM is described as a point-like

39
Metaph. ix.8, 1050b20–22; xii.2, 1069b24–26. 40 Metaph. ix.8, 1050b8–9.
41
This point is connected with the 10th aporia of Metaphysics iii: whether perishable and eternal
substances have the same principles or not. In the light of the interpretation advanced so far, I
suggest that eternal substances have principles that, from a certain point of view, are the same as
those of perishable substances (i.e. three principles: two contraries and matter), whereas, from
another point of view, they are different (i.e. one principle: since the contrariety is neutralised, the
three principles become one). On the hypothesis that celestial bodies, unlike perishable ones, have
just one principle, cf. Fazzo 2013.
42
Phys. iii.1, 201b5–13. 43 Phys. viii.2, 252b10; Cael. ii.12, 292b6–7.
182 diana quarantotto
entity that lacks any kind of duality (Cael. ii.12, 292b4–7).44 Now, unlike
the PM’s activity, circular motion entails a duality, but this duality is
neutralised (without being nullified) by the circularity of the process, and
this is what turns it into something complete.
So, circular motion seems to have a double nature. Aristotle says that
what moves circularly ‘both moves and is at rest’ (265b1–2). What moves
circularly moves, since, while rotating, its parts occupy opposite points.
What moves circularly is at rest, since the opposition between these points
is neutralized by the circularity of the process.
The fact that circular motion has a double nature may stem from its
being the intermediate element between the PM and the sublunary world,
i.e. from its having a mediating role in the ‘transmission’ of being/ousia
from the first immaterial point-like unmoved mover to the material com-
plexity of sublunary natural entities. Indeed, circular motion has this
intermediate position both spatially and functionally: it is the first natural
member of the causal chain that stems from the PM, and so it is both what
is moved and what moves (Metaph. xii.7, 1072a24). Moreover, it is what
ultimately brings, and accounts for, the opposition present in the sublun-
ary world (Phys. viii.6, 260a1–5). But how does it do so? I suggest that it
does so because, unlike the PM, circular motion is a change, i.e. something
that entails a duality or an opposition and so partakes in the nature of
change; however, unlike sublunary changes, circular motion neutralizes
this duality and opposition and so partakes in the simplicity of the PM.
This is why Aristotle calls it the ‘first motion’.45

VI Eternal changes accomplishing ousia (circular

generation): Things less familiar to us
According to On Generation and Corruption ii.10, sublunary natural
entities accomplish being/ousia through circular generation. Actually,
through circular generation sublunary natural entities manage to get as
close as they can to being/ousia: they do not return upon themselves so that
the same individual comes to be a second time, but what recurs is the same
only in species (GC ii.11, 338b6–19).
Circular generation achieves this result because it is an imitation
(mimêsis) of circular locomotion, i.e. it is similar to but not identical

44
On the similarities and differences between the PM’s activity and that of natural entities, see my
(Quarantotto 2014).
45
Phys. viii.9, 265a13–27, b10; Metaph. xii.7, 1072b8–9.
 A dynamic ontology 183
with circular locomotion.46 Circular locomotion and circular generation
are similar insofar as both are circular, everlasting, and complete processes.
Their differences are due to the different matter that celestial and sub-
lunary bodies are respectively composed of.47 Celestial bodies accomplish
being/ousia through circular locomotion, since they are composed of a
matter that is non-generable and imperishable, and whose nature is to
move circularly and eternally.48 Instead, sublunary matter is not one single
element endowed with a circular and eternal locomotion. Rather, it is a set
of four different elements, each of which has its own movement that is not
circular and eternal, but rectilinear and finite (Cael. i.2, 269a25–27).
Therefore, insofar as they are endowed with rectilinear and finite move-
ments, sublunary elements are numerically generable and corruptible, and
are such that every compound of them shares their same destiny of
numerical generability and corruptibility. This means that, differently
from celestial locomotion, sublunary generation is a process that, albeit
unitary, comprises a multitude of rectilinear and finite movements as its
parts.49
This suggests a further implication: sublunary generation is prior to all
other sublunary changes,50 as celestial locomotion is prior to sublunary
generation.51 Indeed, according to the On Generation and Corruption ii.10
passage, circular generation is the sublunary process that is directly
involved in the top-down causal chain of the world stemming directly
from the PM. Because of this, circular generation is prior to all other
sublunary changes not only from an ontological point of view, since it is
the process whereby the ousia of sublunary natural entities is accomplished,
but also structurally, since it is a complete process, i.e. it is the whole of
which all other sublunary changes are parts, and hence the whole without
which either of them could occur. Indeed, every one of the partial changes
comprised by circular generation is a rectilinear, finite, and hence
incomplete process (i.e. a process that occurs between opposite: being

46
GC ii.10, 337a1–7.
47
An interesting passage about hierarchical systems where entities belonging to different levels reach
or approach a certain goal through different means is Cael. ii.12, 292a22–b19; on this, see my
(Quarantotto 2014).
48
Metaph. xii.2, 1069b24–26; Cael. i.2, 269a2–9.
49
This does not imply that celestial motions have no internal complexity (cf. Cael. ii.12, 292a22–b19).
50
Spontaneous generation would be an exception.
51
This is why Aristotle calls eternal locomotion the ‘first kinêsis’ (cf. note 45 above) and celestial matter
the ‘first element’, the ‘first body’, and the ‘first substance’ (Cael. iii.1, 298b6; Meteor. 338b2,
339b16–17, 340b11; i.3, 270b3, 10; ii.12, 291b32). Note, moreover, that eternal being, which is
accomplished through eternal change, is prior to the other kinds of being (Metaph. ix.8, 1050b19;
Int. 13, 23a21–26; GA ii.1, 731b23–31).
184 diana quarantotto
and not-being in respect of place or quality or size or ousia). Therefore, for
sublunary generation, to be complete means turning a multitude of
rectilinear, finite, and incomplete changes into a circular, infinite, and
complete one. And this is what Aristotle may mean when, in On
Generation and Corruption ii.10, he says that ‘also rectilinear locomotion,
by imitating circular locomotion, is continuous’.
Insofar as the generation of organisms is concerned, this is why Aristotle
describes generation and the partaking in the eternal and divine as the
goal of the first soul – which accordingly he calls gennêtikê (DA ii.4,
416b23–25) – as the most natural of the activities of organisms
(415a26–b1) and as that for the sake of which they do whatever they do
by nature (415b1–2). Moreover, the view from above enables us to
understand these claims in a way that is different from how they may be
understood in the ordinary perspective: generation can be seen as the
complete whole of which nutrition and the self-maintenance of each single
organism are incomplete parts,52 i.e. as a unitary process that goes on
through the metamorphosis of matter (i.e. through the life-cycles of a
series of individual organisms).
Something similar holds for sublunary simple bodies. In this case, too,
circular generation is prior to the rectilinear motions performed by each
single body. Indeed, the reciprocal transformation of sublunary elements
is one of the main reasons the whole sublunary world has not broken
down into its parts (GC ii.10, 337a7–15). In this sense, sublunary elements
are not independent from each other: their rectilinear motions are just a
phase within a broader and complete process that is their reciprocal
transformation.53
So, the sublunary world must be kept in continuous change to avoid its
own destruction. Continuous/circular change ensures the being of the
whole and of its parts through the neutralisation of not-being (GC i.3,
318a13–26). This may be one of the reasons why, in On Generation and
Corruption ii.10, Aristotle describes eternal generation as the process
whereby God ‘filled up the whole’ and ‘connected being together’. These
expressions suggest that the being of the whole is something that has gaps
of non-being within itself that must be filled up, or holes that must be
neutralized by means of connections between ‘parts’ of being.

52
The living activity of individual sublunary organisms is a complete process (cf. Metaph. ix.6,
1048b18–36), but it is so in a weaker sense than circular generation, since it has temporal limits
and is therefore, from this point of view, rectilinear (GC ii.11, 338b6–11).
53
Cf. Bodnár 1997: 81–117.
 A dynamic ontology 185

Conclusion: The view ‘from above’

Following the bottom-up path, you will first meet a clear-cut distinction
between being and change. Moreover, you will first meet only rectilinear
and finite changes. From the bottom-up, you will see parts without
knowing that they are parts of something, what they are parts of, and
how. But when you come to the top and have a view from above, you will
see the entire picture, the building blocks, and the main relationships
between them.
You will see that the world is a hierarchy comprising three main levels,
each of which causes the being of the lower one without being caused by it.
These three main levels are similar to each other, since each is something
everlasting, complete, and capable of accomplishing ousia: the complete
being of the first unmoved mover, the complete locomotion of celestial
bodies, and the complete generation of sublunary natural ones. These are
the building blocks on which the being and continuity of the world
depend. All the other kinds of being and of change, those occurring in
the sublunary world, are incomplete parts of sublunary generation.
Moreover, from above you will see that, insofar as natural entities’ ousia
is accomplished through change, Aristotle’s ontology is radically dynamic
and that the world keeps being insofar as it keeps moving and changing. As
already noted, this may sound less awkward if one recalls that the being
that is the primary source of being for all other things is not a state but a
living activity. Hence, perhaps it is not just because of rhetorical emphasis
that, at the beginning of Physics viii.1, Aristotle describes eternal change as
a zoê tis ousa for all things composed by nature (250b14). But what does it
mean? Is it a sort of life or a life of a certain sort (Cael. i.9, 279a23–30)? Let’s
keep investigating.
 chapter 10

Aristotle’s processes
David Charles

I Introduction
Aristotle describes *changes1 (kinêseis) as continuous (Phys. v.4, 229a1,
viii.10, 267a20), as quick or slow (v.4, 228b29), as lasting from one time
to another (vi.8, 239b1ff.), as going on for ever (viii.7, 261a30ff.), as even
or uneven (v.4, 228b18 and 228b27ff.), incomplete or complete (EN x.4,
1174a28ff.), interrupted or uninterrupted (Phys. viii.4, 255b7, ii.8, 199b16),
accelerating or slowing down (vi.7, 238a6ff.), and as forced or in line with
nature (iv.8, 215a1ff.), and also talks of different *changes proceeding in
different directions (v.6, 229b27ff.).
As these descriptions show, particular *changes have different properties
at different times: some are initially slow and then fast. *Changes are (non-
derivatively) the subjects of these properties at different times. They last
through time and move through space. Some are interrupted, failing to
reach their goal: these could have continued longer.
In this chapter, I shall consider some aspects of Aristotle’s account of
*change, beginning with his attempt in the first three chapters of Physics iii
(hereafter Physics iii.1–3) to say what *change is. There, he suggests that it is
‘the actuality (entelecheia) of that which is potentially as such’ (Phys. iii.1,
201a11–12), illustrating this suggestion by examples such as ‘when the

I am indebted to Tom Ainsworth, Elena Cagnoli Fiecconi, Jason Carter, Ana Laura Edelman, Paolo
Fait, Verity Harte, Mariska Leunissen, Fred Miller, Takashi Ochi, Michael Peramatzis, Diana
Quarantotto, Stefan Sienkiewicz, and Breno Zuppolini for their comments on the penultimate draft
of this chapter. I also gained from earlier discussion with Lucas Angioni, David Bronstein, Laura
Castelli, Kei Chiba, Timothy Clarke, and Ursula Coope. The present chapter engages with, and re-
assesses, issues already discussed in Charles 1984.
1
I shall use *change (with an asterisk) to stand for whatever is denoted by the Greek term kinêsis. It is a
substantial thesis that Aristotle’s *changes are changes. I use this notation so as not to assume from the
outset that kinêseis are, for example, events that occur at or between times, rather than processes that
continue through time.

186
 Aristotle’s processes 187
buildable, in so far as we describe it as such, is in actuality, there is building’
(201a16–18). In sections II and III, I shall consider differing interpretations
of these telegrammatic remarks, placing them in the context of Aristotle’s
discussion of *changes elsewhere in the Physics and Metaphysics Theta. I
shall next examine his remarks on the individuation of *changes in Physics
iii.3 (section IV), and finally locate his account of *change in the context of
his discussion of substances and activities (energeiai) in Metaphysics Theta
(section V).

II The definition of *change: ‘a *change is the actuality of that

which is potentially . . . as such’ (Phys. iii.1, 201a11–12)
Several interpretations have been offered of Aristotle’s initial account of
what it is to be a *change.2 They differ in their understanding of its two key
terms: ‘actuality’ and ‘what is potentially as such’.
‘Actuality’ is sometimes understood as an actualisation. However, this
seems problematic if ‘actualisation’ means the process of transitioning to
an end-state:3 for this appears to make Aristotle’s account immediately and
unacceptably circular. Many, as a consequence, translate ‘actuality’ as
actuality, taking this to be different from the process of actualisation.
But what is an actuality? For some, it is the state achieved when the
*change is completed. They take this to be the only alternative to the
(immediately circular) account in terms of actualisation. So understood,
Aristotle’s subject of change has the potentiality to be in a given state: the one
achieved at the end of the change. However, as it stands, their account is
flawed. It fails to capture the idea of the subject proceeding to, and having
the potential to proceed to, the relevant end-state. But it is precisely because
it has this latter potential that the subject comes to be at the end-state.
Two lines of interpretation address this difficulty. One, favoured by
David Ross, Robert Heinaman, and my earlier self, understands the phrase
‘what is potentially as such’ to refer to something qua having potentiality to
*change or be *changed.4 This potentiality is realised when there is a
*change. However, this interpretation (which I shall call Interpretation
A) is, as it stands, incomplete. What is needed is an appropriate way,
suggested in Aristotle’s own writings, of understanding the term ‘actuality’.
I shall seek to fill this gap in section V.

2
Some are contrasted by Coope 2009.
3
Although this view is defended by Kostman 1987, I shall not consider it further here.
4
Ross 1936, Charles 1986, and Heinaman 1994. See also Beere 2009.
188 david charles
A second interpretation, initially proposed by Aryeh Kosman and sub-
sequently developed by Sarah Broadie and (with reservations) Ursula
Coope, has recently gained currency. It takes Aristotle’s phrase ‘as such’
to modify not the potentiality of the subject to be in a given state but the
type of actuality involved.5 The relevant actuality is not the end-state
achieved but the actuality-as-being-potentially-in-a-given-end-state of
what has the potential to be in that end-state. Housebuilding, so under-
stood, is a distinctive kind of actuality of the potential to be a house: the
actuality-as-what-is-potentially-a-house of that which has the potential to
be a house. When building occurs there is an actuality-as-the-potential-to-
be-a-house of the bricks’ potential to be a house. Although this proposal
(which I shall call Interpretation B) is ingenious, it is, I shall argue, subject
to serious objections.6
Two considerations are taken by proponents of Interpretation B to estab-
lish that, whatever its difficulties, it must be preferable to Interpretation A.
The first runs as follows. In Phys. viii.5, 257b6–9, Aristotle describes a
*change ‘as the incomplete actuality of what is *changed’. However, since
all potentialities are incomplete, he cannot be referring to a special kind of
incomplete potentiality, but rather to a special kind of incomplete actu-
ality. Interpretation B takes the potentiality at issue to be the one which (in
the case of housebuilding) bricks possess to be a house. Thus, when
Aristotle distinguishes complete from incomplete actualities, he is identi-
fying the former with actually being a house, the latter with actually being
potentially a house (or, as it is sometimes said, ‘being active as a potential
house’).7 This passage, in their view, points to a special kind of incomplete
actuality: the actuality-as-potentially-a-house.
It is important to note that Aristotle, in 257b6–9, does not explicitly
mention the potentiality to be, for example, a house. He refers only to the
subject of *change (e.g. the bricks). Nor does he distinguish, as he does
elsewhere, the actuality of the bricks as what is buildable (namely the
building) from the actuality as a completed house (compare Phys. iii.1,
201b10–12). This latter distinction, however, is significant. Once the house
is completed, the *change is over. There is no longer an actuality of the
subject qua what is buildable. The subject is no longer correctly described

5
Kosman 1969, Waterlow/Broadie 1982, Coope 2009, and Kosman 2013. An interesting variant of this
view is suggested by Hussey 1983.
6
I shall refer to proponents of these interpretations as A- or as B-interpreters. My interest is to see how
far each line of interpretation can be developed and defended, not merely to criticise or support what
has so far been said in support of either.
7
Kosman 2013: 68.
 Aristotle’s processes 189
in this way. By contrast, the actuality characteristic of a *change has as its
subject something not yet itself complete (as Aristotle notes in 201b11–12:
see also DA iii.7, 431a6–7): something that is not yet a house. With this
distinction in place, Interpretation A need not take 257b6–9 to refer to a
special kind of incomplete potentiality. The relevant potentiality can, on
the A-reading, be that of a special kind of subject: something, not yet a
house, with the potential to become a house. The bricks in question might
have two distinct potentials: one to be a house, the other to *change into a
house. Qua possessing the latter, they are – according to A-interpreters – a
special kind of subject of change.
But why does Aristotle refer to the *change as ‘an incomplete *actuality’
in 257b6–9? *Changes, as he often notes, are, in some important respect,
incomplete (see Phys. iii.2, 201b32; Metaph. ix.6, 1048b28: ‘every *change
is incomplete’). This is because, in his view, housebuilding is only complete
when the house is built. While the building is under way, the *change is
incomplete. This is why *changes are hard to pin down in terms of
actuality and potentiality (Phys. iii.2, 201b28–202a2): unlike other *actu-
alities, they are essentially incomplete even though they differ from mere
potentialities (which are present before the process has begun). These
remarks suggest an appreciation of what is now called the ‘imperfective
paradox’: *Changes, while they are going on, exist but are not yet complete.
However, the incompleteness of the actuality, so understood, is a general
truth about *changes, which both Interpretations A and B can easily
accommodate. The phrase ‘incomplete actuality’ in 257b6–9 does not
support either view. It is simply a way of distinguishing the *change
(which is an incomplete actuality) from the state (of being a house)
which is, by contrast, a complete actuality. Aristotle’s brief remark in
257b6–9 is consistent with both Interpretations.
A second consideration, emphasised by B-interpreters, is that Aristotle
could not have accepted as the definition of *change the actuality of what is
potentially to *change or be *changed (as proposed by Interpretation A).
This definition is, they say, ‘circular.’ Charity requires us to see Aristotle as
proposing the non-circular definition suggested by Interpretation B.
It is important to spell out this argument carefully. There is no circu-
larity in defining a *change as the actuality of the potential to *change or be
*changed, assuming that the notion of actuality is not itself that of a
*change. For the potential to *change is not the same as the *change itself.
(In this, the present suggestion differs from the immediately circular
definition of *change in terms of actualisation.) The problem arises only
when one combines this definition with one of the potential to *change in
190 david charles
terms of *change (its specific actuality) – a step Aristotle appears to take in
Metaphysics Theta when emphasising the definitional priority of actuality
over potentiality.8 Once these accounts are combined, a *change is defined
as the actuality of the potential to *change (or be *changed) and the latter
defined in terms of *change. Even if Aristotle’s definition of *change is not
immediately circular, it is one of a pair of definitions, each of which
depends on the other. Aristotle does not succeed in breaking out of a
small circle of interdefined terms (potentiality to *change and *change).
Since, for B-interpreters, this failure violates Aristotle’s basic requirements
on acceptable definitions, the A-interpretation of his definition of *change
must be incorrect. Further, as they point out, there is no explicit reference
in Physics iii.1 to the capacity for *change (postulated by A-interpreters).
While Aristotle speaks of what is buildable, he does not specify what it is
for something to be buildable. So, B-interpreters ask, why saddle him with
Interpretation A’s ‘circular’ account when there is no textual need to do so?
It is, of course, true that the two accounts – that of *change as the
actuality of what is capable of *change, and that of what is capable of
*change as what *changes – when combined, depend on each other. They
do not, taken together, offer a reductive definition of *change in terms
independent of the notion of *change. If Aristotle was aiming to break out
of this circle of terms, he failed. However, there is reason to believe that this
was not his goal. Or so I shall argue.
In Metaphysics Theta 6 Aristotle writes that one should not seek a
definition of everything (1048a36ff.), noting that one may not be able to
offer definitions of such basic terms as activity (energeia) or potentiality
(capacity).9 The basic term ‘activity’ is applied elsewhere to actualities
(1047a30, 1050a21–22). In Theta 6 Aristotle simply gives examples of
what is capable: what is capable of building and seeing, and matter capable
of being a ‘worked-up’ object, such as a statue (1048a31ff.). The basic
notions of potentiality and activity (and, by implication, actuality) are
not defined independently of each other. Indeed, Aristotle seems to think
that it cannot be done. When discussing his examples, he describes them as
analogous to each other. There is no attempt to analyse one group in terms
of the other. He simply says: ‘as capacity stands to activity, so matter stands
to form’ (1048b6–8).

8
See Metaph. Theta 8, 1050a4ff. For an illuminating discussion of the relevant types of priority, see
Peramatzis 2011.
9
Indeed, an activity may be understood as the way a thing is when it is not in the way in which we
would describe it as capable of being or of becoming so and so.
 Aristotle’s processes 191
Aristotle is offering a conceptual map on which to locate varying kinds
of potentiality and activity (including actualities). On one side, there are
capacities to be, the kind of capacities matter possesses. On the other, there
are capacities for *changes: the capacities to build or to be built with. So far
from attempting to give a definition of the latter in terms of the former,
Aristotle simply presents them as analogous. He is not attempting the
ambitious project required by the B-Interpreter: of defining *change in
terms independent of *change.
In Theta 6 Aristotle develops his account by means of examples: as that
which builds to that which is ready to build, that which is awake to that
which is asleep, so that which is seeing is to that which has sight (1048a37–
b4). In the first case, he contrasts that which is actually building with that
which is potentially building. ‘Actually building’ and ‘potentially building’
are two different ways of building, as ‘actually seeing’ and ‘potentially
seeing’ are different ways of seeing. These different ways of seeing or
building are compared with actualities and potentialities in the case of
substances. There is no attempt at an independent definition of any of
them. Nor does he do so in Metaph. Delta 7, 1017a35 when listing cases,
some involving *changes (e.g. building), some activities (e.g. seeing), still
others substances (like being a statue). While all can be thought of in terms
of being potentially and actually something, there is no attempt to reduce
one set of cases to the other.
Aristotle takes a further step in Theta 7. He suggests that when some-
thing is potentially a statue, it has the capacity to be a statue and distin-
guishes two relevant capacities: the capacity to be a house and the capacity
to become a house. He writes (1049a10ff.):
Something is potentially a house, if nothing in the matter prevents it from
becoming a house, and there is nothing which needs to be added or taken
away or changed for it to become a house, it is potentially a house. The same
account applies in all the cases where the starting-point of the process of
coming to be is external to the matter.
The matter has two distinct capacities: one to be a house, first acquired
when it has the potential to become a house, the other to become a house.10
While the latter requires that the matter in question has the potentiality to
be a house, the two potentialities are not to be identified. Indeed, when the
house is built, the object no longer has the potential to become a house. So
understood, a *change is the realisation of the specific capacity to become a

10
I discussed this passage in more detail in Charles 2010c.
192 david charles
house which the matter possesses before it has become a house and loses
once it has become one.
In sum: Interpretation B’s second consideration, with its emphasis on
the dangers of ‘circularity,’ seems ill-founded. Aristotle is not seeking in
Theta or elsewhere a reductive definition of *change in terms indepen-
dently of *change. Instead, he is trying to show that various differing kinds
of potentiality and activity (and actuality) share a similar structure. The
presence of this circle of interdefined terms is, from his point of view, no
objection to its philosophical validity. So, even if B-Interpreters were to
succeed in devising, using Aristotelian machinery, a reductive definition of
*change, their ingenuity would be misplaced. There is no reason to think
that Aristotle himself was engaged in any such reductive project. (I shall
return to this issue in sections iii and V.)
This said, Interpretation B is correct to emphasise that there is no
explicit reference in Physics iii.1 to the capacity, for example, to be
*changed into a house. While Aristotle talks of the buildable, he does not
specify this in terms of the capacity to *change (or be *changed) into a
house. But what does this show – given Aristotle’s willingness, in
Metaphysics Theta 7, to distinguish the capacity to become a house and
the capacity to be a house? Perhaps Physics iii.1 contains a preliminary
discussion of *changes, aiming only to specify the extension of the term
‘*change’, not to give a complete account of what a *change is. So under-
stood, its ‘definitions’ are not designed to give a full account of what it is to
be a *change.11 Instead, Aristotle is introducing an account to be refined
later in Physics iii and elsewhere. Indeed, subsequent passages in Physics iii
seem designed to elucidate key phrases such as ‘the buildable’ (as used in
Physics iii.1). Let us examine this suggestion further.
In Physics iii.3, Aristotle gives a ‘more knowledgeable’ account of what
*change is (202b26ff.): the actuality of what is capable of doing and of
suffering as such. A more knowledgeable account is presumably closer to
the first principles of an account of what it is to be a *change. The actuality
of a capacity is a *change if the relevant capacities are ones for acting or
suffering, ones whose actuality is acting or being acted upon. However,
Physics iii.3 does not suggest the approach favoured by B-interpreters.
Aristotle focuses on making and suffering (poiein and pathein), which he
separates from the result: the thing made or what is suffered (pathos or
11
One might regard Physics iii.1 as giving an account of what the term ‘*change’ signifies to be
contrasted with subsequent accounts of what it is to be a *change. For this distinction, see APo. ii.9,
93b29–94a10 (discussed in Charles 2000: 23–56). A similar distinction may be suggested at Phys. ii.7,
198a26ff.
 Aristotle’s processes 193
poiêma) (202a25). He speaks throughout of *changing, making and suffer-
ing, not the result of the *change (the completed house). In this more
advanced definition, Aristotle refers to the capacity to act or to be acted
upon, not to the potential to become a result. As Physics iii progresses, he
elucidates the kind of potentialities involved: they are ones required for
doing and suffering, not those required for the presence of the relevant
result (being a house).12
B-interpreters, however, understand the capacities to make or to suffer
differently, taking the capacity of the builder not to be to build but rather
as one for the house (the result) to be.13 According to this suggestion, the
builder has the capacity for the relevant matter (such as bricks and stones)
to be in a new state. Aristotle, so interpreted, focuses on the result and
understands the builder’s skill as his capacity for something else to be in a
new state.
This ingenious suggestion has its own disadvantages. Skill is usually (and
plausibly) characterised by Aristotle as the agent’s ability to make some-
thing (as for example, in EN vi.4, 1140a7ff.). The builder does not have the
bare capacity for something else to be in a new state, rather the ability to
bring about this new state (by building). The former capacity (if it were to
exist) would not be unique to builders: their employers would also have it
and exercise it by providing builders with the resources to do the building.
However, the employers lack the capacity to bring about the new state
themselves by building. Indeed, this is why they need a builder.
Aristotle’s account of the relevant kind of capacity in Physics iii.3 makes
explicit his earlier thought: the potential of what is *changeable as such is to
be understood in terms of what has the potential to make or suffer as such.
He does not seek to define *change in terms independent of capacities for
*change. Instead, in this more developed account, Aristotle specifies the
capacity of the *changeable in terms (familiar from Categories 8, 9a14ff.) of
the capacities of the agent to make something (poiein) and the patient to
suffer (paschein). This approach is to be expected if Aristotle is following
the strategy favoured by Interpretation A.
To conclude: first, while Interpretation B offers a possible reading of
some key phrases in Physics iii.1 and viii (taken in isolation), it is less
convincing as an account of the unfolding argument in Physics iii.1–3.
Second, its central motivation – that of presenting Aristotle as seeking a
way to break out of the relevant circle of terms – is not one which he
shared. Indeed, he may well have thought it misguided.

12 13
See Heinaman 1994. See Coope 2009.
194 david charles
III The definition of *change (2): Further problems
for Interpretation B
Interpretation B fares no better as Physics iii progresses further. Aristotle
talks later of infinite *changes that go on forever (Phys. iii.8, 208a16–21; see
also iv.1, 208a20 and viii.7, 261a30ff.). Since these have no end-point, they
cannot be the actuality of the ability to be at an end-point as that ability. In
his account, the infinite is ‘that of which there is always something outside’
(Phys. iii.6, 207a1–2). In the case of geometrical construction, ‘it is always
possible to think [of a bigger number] in the direction of the more, since
bisections of a magnitude are unlimited’ (Phys. iii.7, 207b10–11). To grasp
that the procedure is infinite is to grasp that it will not actually terminate.
In a similar way, to grasp that a given *change is infinite is to grasp that it
too will not terminate (208a16–21).14 Nor can these remarks be dismissed as
irrelevant to Aristotle’s definition of a *change. Physics iii, which begins
with that definition, ends with an account of infinitely long *changes. Such
*changes (including those involved in geometrical construction) play an
essential part in his account of the ‘potential infinite’. Aristotle could not
have consistently developed that account had his definition of *change
been as the actuality of what is potentially in such-and-such a given end-
state. There is simply no such state in the case of infinite *changes.15 Since
the B-interpretation attributes to Aristotle a major inconsistency in Physics
iii as a whole, it seems preferable to understand his account of *changes in
the way recommended by Interpretation A: as the actuality of the capacity
to *change.16 *Changes, so understood, can go on for ever.17
Does Interpretation B succeed in attributing, in the case of finite
*changes, a plausible non-circular definition of *change to Aristotle?
Does the full actuality of the ability to be a house, as the ability to be a
house, capture all and only cases of housebuilding?18

14
Contrast the idealist reading rejected in Phys. iii.8, 208a18–19: ‘something is outside not because
someone thinks it but because it is so’.
15
White 2009: 260–263.
16
For further discussion of Aristotle’s idea of the potential infinite, see Coope 2012.
17
Eternal circular *changes are said to be complete (teleion: Phys. viii.8, 264b28 and 9, 265a21) because
they are continuous and indestructible. However, they are still ‘incomplete’ in a different way: it will
never be true to say of an object, ‘this is moving in a circular way and has already completed that
individual circular movement’ (compare Metaph. Theta 6, 1048b25–31). I am indebted at this point
to discussion with Diana Quarantotto.
18
A. Anagnostopoulos (2010) offers several important criticisms of the translation of key phrases
adopted by defenders of Interpretation B. I seek to address his concerns about Interpretation A
below.
 Aristotle’s processes 195
It is not clear that it does. Many think that bricks can manifest the ability
to be a house as that ability without actually being built with: for example,
when lying about before being used. According to Interpretation B, how-
ever, they do not (at this stage) manifest an ability to be a house (as that
ability), only an ability to have the ability to be a house.19
Did Aristotle himself think in this way? In Metaph. Theta 7, 1049a10ff.,
he notes that bricks first have the potential to be a house when nothing
needs to be added or taken away from them for them to be acted on by the
builder to make a house. They have this ability when builders can take
them to use in building. Nor should this surprise us. Their ability to be a
house is used (as that ability) by builders who choose to use them for
building or retain them for later such use. Further, brick-makers use the
fact that bricks (when finished) actually have the ability to be a house when
they stop working on them once they have reached the point at which they
can be used (without further addition etc.) in house construction.
Aristotle, it seems, took the (commonsense) view that a collection of bricks,
when fully made, have (and manifest) the ability to be a house, not merely
the ability to have the ability to be a house.20
The case of bricks standing in a half-built house abandoned by the
builder raises similar problems. They too seem to manifest the ability to be
a house as that ability. So too a sapling, considered as a half-grown tree: it
appears to exercise its ability to be an oak (as that ability) even though
being a sapling (the state) is not itself a *change. Interpretation B, it seems,
does not (as stated) succeed in picking out all and only cases of *changes. Its
proposed non-circular definition does not yield an extensionally correct
account of what a *change is.
Some B-interpreters reply as follows: the ability to be a house as that
ability is only constitutively realised when building is under way. But why,
one may ask, is building the only ‘constitutive realisation’ of the ability to
be a house? Why aren’t the bricks, before being built with and when in a
half-built house also ‘constitutively realising’ (albeit in a different way) the
relevant ability? Unless the notion of ‘constitutive realisation’ is itself
defined in terms of *change itself, it will apply to many irrelevant cases.
Some B-interpreters, at this point, fall back on the idea of activity
(energeia), suggesting that there is only an activity of the relevant ability
when the house is actually being built. However, this proposal (without
further development) seems vulnerable to a similar pattern of objection:
either the notion of the activity of what is potentially a house as potentially

19 20
See Kosman 2013: 61–63. Similar remarks apply to an acorn before it has been planted.
196 david charles
a house is itself to be understood on the basis of that of *change or it applies
to states such as being a sapling which are not *changes at all.
The difficulty is a general one: B-interpreters seek to avoid the impasse
noted in section I by introducing a further actuality – the actuality of the
bricks as potentially a house. To avoid immediate circularity, they insist
that this is not an actualisation. However, if this new actuality is under-
stood in such a way that states or objects can also manifest it, it will be
present in phenomena other than *changes. Merely adding this further
complex actuality does not yield the non-circular definition of *change that
B-interpreters seek. They seem not to have escaped from the impasse noted
in section II.
The (charitable) conclusion to be drawn is, I suggest, that Aristotle was
not proceeding in the way suggested by Interpretation B. That said,
Interpretation A still faces a serious challenge: to locate an Aristotelian
notion of the actuality of the capacity to *change (or be *changed) which
avoids immediate circularity. In the final section of this chapter, I shall
address this issue. However, before doing so, there is a further aspect of
Aristotle’s discussion of *changes to consider: his account of their
individuation.

IV *Changes and their individuation: Another look at Physics iii.3

Physics iii.3 is a much-discussed chapter.21 In this section, I shall argue that
it is less committal, in certain crucial respects, than it is often taken to be.
Its goal is to uncover just enough structure for Aristotle’s immediate
purposes, not to give the complete, worked-out, account of *change
individuation that recent authors have assumed.
Although *changes are characterised in terms of capacity and actuality
(entelecheia) in Physics iii.1–2, several issues remain unresolved. In Phys.
iii.3, 202a20–21 Aristotle notes that, in the case of acting and suffering,
there is one activity (energeia) of what is moved and of what is capable of
moving but not one account of it (logos). How is his claim to be under-
stood? I shall begin with a brief overview of his subsequent discussion.22
Aristotle presents a general difficulty (202a21–22) and gives two argu-
ments against the view that there are two distinct *changes (such as
teaching and learning) which occur in the patient (202a28–36). Teaching

21
See Gill 1980; Charles 1984; Coope 2004; and Marmodoro 2007. My present view differs from the
one developed in 1984.
22
For more detail, see Charles 1984 and Coope 2004.
 Aristotle’s processes 197
cannot be in the learner. It is not possible for there to be two *changes
which end at the same point in the same object. (Let us call the first
argument I, the second argument II.) He then offers two arguments against
the idea of there being one *change in the cases under review. The first is
that there cannot be one activity of two capacities that differ in form
(argument iii). The next argument (argument IV) is that if the two
*changes were identical, all predicates true of one would be true of the
other; but this is not the case. The teacher in teaching does not learn what
he teaches (202a36–b4).
Aristotle considers these arguments in turn. He rejects argument I:
teaching, he says, can occur in the learner (202b6–8). He next rejects
argument iii: there can, he claims, be one activity of two capacities
(202b8–9). Finally he rejects IV, saying that, even if there were one
*change, not all predicates true of one would be true of the other
(202b12–16). But what is his attitude towards argument II: the contention
that there cannot be two *changes that end at the same point in the same
object? Does he accept it, holding that there is only one *change? Or does
he accept it, in a quite different way, holding there are two *changes? While
he is clear that there is in some way only one *change, how is this claim to
be understood? Is there one *change solely in virtue of the fact that doing
and suffering happen at the same time in the same place? (For this use of
‘being the same,’ see Metaph. Gamma 2, 1003b22.) So understood, there
will be one *change in nature because acting and suffering always (or
necessarily) co-occur. Or, alternatively, did he think that there is only
one *change because there is, in reality, only one nature present with one
essential feature?
Aristotle’s remarks seem inconclusive. In 202b8ff., he talks of one
activity (energeia) which is not the same in being (tôi einai), terminology
used previously to describe things which are (in some way) the same but
differ in the time of their occurrence or some other aspect (201a20). The
cases under review in Physics iii.3 are not the same in the account of what it
is to be each thing (to ti ên einai). They are to be contrasted with his favoured
example of cloak and garment (which are the same in the relevant type of
account: 202b12–13).23 If things differ with regard to the account of what it
is to be each of them, they will differ in definition and essence. But if they

23
It is important to distinguish between difference ‘in being’ (frequently used when there are different
accidental descriptions of the same thing) and difference in ‘in being what it is to be the thing’
(where the description is an essential one). I know of no case (in Aristotle) where there is something
one in number but with two (non-equivalent) essential descriptions. (I take ‘being a cloak’ and
‘being a garment’ to be, in this sense, equivalent descriptions.)
198 david charles
differ in these ways, they must be numerically distinct. (How can there be
one thing with two distinct essences?) Therefore, if the definitional
accounts of the two *changes (doing and suffering: 202b11) are distinct,
these *changes must be numerically distinct. (There is, it should be noted,
no further *change mentioned in the immediate context of which these are
accidental properties.)
Aristotle, if this is correct, holds that there are two *changes which
necessarily accompany each other but are not strictly identical. They will
be the same in the less demanding manner already mentioned: they
necessarily co-occur but are not the same in definitional account. There
will be a distinct individual *change for each separate capacity realised: the
capacity to do (teach) and the distinct capacity to learn (suffer). The two
*changes will further differ in having different subjects and different goals.
(The goal of learning will be, let us say, the acquisition of knowledge (the
pathos: 202a24), that of teaching to ensure that someone else acquires
knowledge (the poiêma: 202a24).) In having these different goals, the two
*changes resemble the road from Athens that ends at Thebes and the road
from Thebes that ends in Athens (202b12–14). Further, the two *changes,
although they co-occur, may have different properties (as the road to
Thebes is uphill, that to Athens downhill). They will be one *change solely
in virtue of their necessarily co-occurring.
While this proposal constitutes an adequate answer to Aristotle’s puzzle,
he immediately considers another:
However (‘not but what’, 202b16f.), not even if teaching and learning were
one, would it be the case that to teach is to learn . . .24
Since this suggestion is contrasted with the previous one, the *change now
under consideration must be one in account. While before Aristotle had
previously been thinking of two *changes, different in definition, which
necessarily accompany each other, he now considers an alternative: the
*change in question is one in definition (one in being what it is). However,
even so, he notes, to teach will not be the same as to learn. There is one
*change (one in definition) describable as a teaching and as a learning,
where these descriptions do not mark out distinct essences of different
*changes. In a similar fashion, one might say that there is one and only one
distance that can be described either as the distance from A to B or B to
A. So understood, the descriptions ‘the distance from A to B’ and ‘the
distance from B to A’ are not essential descriptions of different distances.

24
For more on this locution, see Denniston 1934.
 Aristotle’s processes 199
Instead, they are different ways of describing one and the same distance,
which (in Aristotle’s terminology) is one in definition but distinguishable
in thought (and accidental properties).
Aristotle’s next sentence introduces a relevant generalisation:
But to speak generally, the teaching and the learning are not the same, nor is
the acting and the suffering in the strict sense; but that to which both
belong, that to which the teaching and the acting belong, the process, will be
the same in the strict sense [as something unspecified]. For the activity of
this in that and of that under the influence of this differ in account.
(202b19–22)
This passage adds two points. In the case of acting and suffering:
(i) the suffering (the pathêsis) is not strictly identical (kuriôs) with the
acting (the poiêsis);
(ii) that to which the suffering belongs, the *change, is the same in the
strict sense [as something unspecified].
(i) is clearly true in the second account just sketched. Although acting
and suffering are the same in definition, ‘suffering’ and ‘acting’ are
two different ways of describing the same *change. Suffering and
acting are not one in the strictest sense because they are distinguish-
able in thought. However, (i) is also true in Aristotle’s earlier
account since, so understood, suffering and acting differ in defini-
tion as well in thought. The generalisation holds true in both the
accounts Aristotle has so discussed.
(ii) is more difficult to interpret because the final clause is incomplete.
Some take its full expansion to be:
(ii)* ‘the *change is the same in the strict sense as the suffering . . . ’
In their view, the suffering (of A under the influence of B) is strictly
identical (inseparable in thought and in definition) with the *change.
Others take this clause to mean:
(ii)** ‘the *change is the same in the strict sense as itself . . . ’
where the *change is said to be strictly identical with itself, not with either
the suffering or the acting.
The first reading seems preferable. The final clause is naturally read as
understanding the whole phrase ‘is strictly the same as the learning/suffer-
ing’ rather than just a sub-phrase ‘is strictly the same’. Second, it makes
a substantial point, recalling Aristotle’s earlier remark that the *change is
the actuality of what can be moved under the influence of the mover
200 david charles
(202a15–16): the *change is strictly identical with the suffering in the
patient. This point is reinforced in the final lines (202b21–22), where a
difference in account (logos) between suffering and acting allows the former
but not the latter to be strictly identical with the *change. By contrast,
(ii)** makes only the obvious point that the *change in question is identical
with itself, without specifying what that *change is. But why add this self-
evident point here?25
It is important to note that (ii), understood as (ii)*, will be true in both
the proposals he is considering. In the first, the acting differs from the
suffering in definitional account; in the latter, in thought alone. In both,
the *change (understood as the actuality of what can be moved) may be
taken as strictly identical with the suffering, spelling out the remark in
202a15ff.; they will be inseparable in both definition and thought. So
understood, Aristotle’s generalising remark (‘to speak generally’) applies
to both his first and second proposal. In both, there are two accounts (two
logoi) which describe what occurs, which is (in some way) one. In the first,
these two accounts define two distinct activities that necessarily co-occur.
In the second, there is one *change, one in definition, even though it can be
thought of in two ways. Both spell out ways in which there can be one
activity that is two in account (202a18–21).26
Did Aristotle decide between these alternatives in the present passage?
Or was he content to point to a feature common to both: that there is (in
some way) one activity which can be seen (in both these the two ways
indicated) as falling under two different accounts? There is no need for him
to go further at this point. On either view, what occurs is in some way one
and the same activity of the capacity of the sufferer and of the maker, even
though the two models he has canvassed differ in their ways of making this
thought determinate.
Aristotle did not need to go further if his aim in the chapter is to defend
the following claims:
(1) There is one activity (energeia) of both capacities (202a18).

25
Bishop Butler, it should be noted, only made his famous comment ‘everything is what it is and not
another thing’ when rejecting an alternative theory. It was not put forward as part of a positive
definition of any particular type of thing. Aristotle mentions a similar point briefly in Metaph. Zeta
17, 1041a19–20, only to lay it quickly aside.
26
For cases where Aristotle notes several ways of addressing a problem without finally deciding
between them, see GC i.3, 319a29–b5. Two more committed strategies have been suggested:
(a) take 202b19–21 as resuming from 202b16 with 202b16–19 as an aside (prioritising the first
proposal): see Charles 1984: 14f.; (b) take 202b19–21 as applying only to the proposal made in
202b16–19 (prioritising the second proposal). See Coope 2009.
 Aristotle’s processes 201
(2) The activity occurs in the patient and is not different from the
activity of the agent (202a15–16).
(3) The *change is the activity of what is capable of *changing under the
causal influence of the *changer (202a19–22).
All three claims are true in both models, even though they offer different
accounts of the key notions of ‘one activity’ in (1) and ‘not different from’
in (2). Both can agree that (3) is strictly true of the *change which they
distinguish either in thought or in definition from the activity of the agent.
Both develop (albeit in different ways) the idea that the *change is, strictly
speaking, identical with the activity of the patient brought about by the
agent. In both, housebuilding can be understood as the actuality of what is
capable of acting and of suffering, as such (202b26–28), although they
disagree as to the precise form of sameness involved.
In recent discussions, some have taken Aristotle to favour the first
model, others the second (in one of several forms).27 My present suggestion
is that Aristotle did not, in fact, decide between these alternatives in Physics
iii.3. He could leave this issue open since he has set out sufficient con-
ceptual resources to defend in some way or other his three initial claims
(noted above). Since he aims to address a problem, an aporia, by outlining
possible solutions, he does not need to determine here the exact kind of
sameness in question. His present project is a modest one, with the more
minimal goal of finding at least one way out of the aporia.
There are, of course, several ways in which Aristotle’s remarks could have
been developed further. However, he does not do so here. In Metaphysics
Gamma 2, he notes two ways in which things can be the same: either by
being necessarily coextensive or by having the same definition. In Physics
iii.3 either suggestion is sufficient for his purposes. (Nor is it clear that he
decided between them elsewhere. But the resolution of that question awaits a
further study of many difficult texts throughout his writings.)

V. The nature and role of actualities: interpretation A and Aristotle’s

ontological project in Metaphysics Theta
There remains a major problem: can Interpretation A offer an Aristotelian
understanding of actuality (entelecheia) in a definition of *change

27
I took the first approach in Charles 1984, while Ross 1936 preferred the second, individuating
*changes in terms of the route followed from beginning to end. Others who prefer the second
approach individuate the relevant *change as the joint realisation of two capacities (see Coope
2009). These issues are discussed in Marmodoro 2007.
202 david charles
acceptable by Aristotle’s own standards? Such a definition must apply in a
non-immediately circular fashion to all and only *changes. It should also fit
into Aristotle’s ontological project in Metaphysics Theta.
Aristotle, as we saw, was happy to take potentialities and actualities (or
activities) as basic features of his ontology. In Metaphysics Theta 6 to 8, he
seems to see the relevant potentialities for *change, together with *change
itself, as part of a circle of interdefined terms. But how did Aristotle
understand actuality?
Some remarks in Metaphysics Theta suggest an ontologically less com-
mitted way of understanding ‘actuality’. In Theta 8, 1050a15ff. Aristotle
writes:
The result (ergon) in the case of action is a goal, and the activity (energeia) is
the result. This is why that term (activity) is predicated on the basis of what
is the result in the case of actions and extends also to the actuality
(entelecheia).
The goal-like status of actions and of actuality allows him to refer to both as
‘activities’. In the case of flute-playing discussed in the previous lines, the
goal of the teacher is that the pupil plays the flute. While the pupil’s actions
exhibit the fact that the teacher has taught him how to play, what the
learner aims at (and what the teacher teaches) is playing the flute. This
is what has the goal within it. The pupil displays that he has achieved
this goal in playing the flute. In other cases the relevant goal is seeing or
thinking: something present in my or your thinking and seeing
(1050a35ff.). In others, what counts as success is something beyond the
action. The latter is exemplified by the house produced by house-building
(1050a26ff.): the result of the action. However, both housebuilding and the
resultant house are described as ‘more a goal’ than the capacity for house-
building (1050a27ff.).
Aristotle’s remarks suggest the following account. An actuality is some-
thing which when present is the realisation of the relevant potentiality.
One might (with due circumspection) describe the actuality as the success
condition for the realisation of that potentiality. In the case of teaching
someone to play the flute, the success condition is playing the flute. This is
what teachers aim to pass on and what is achieved when the pupils’
potential to play the flute is realised (1050a21–23). When the potentiality
is to see, this will be realised when there is seeing (1050a23–25). If the
potentiality is to be a house, this will be realised when there is a house.
Similarly, if the potentiality is that to *change or be *changed into a house,
this will be present when there is a *change of this type (1050a25–29).
 Aristotle’s processes 203
Success conditions need not be states. They are simply whatever has to be
present for the relevant potentiality to be realised. In some cases, they will
be activities such as seeing or thinking, in others states (such as being a
house), in others a *change. Since the notion of a success condition applies
widely in this way, it can be used without circularity in the definition of a
*change. Of course, the success condition for the specific potentiality to
*change is the *change itself: but this is precisely what is required. There is
no danger of immediate circularity, since the notion of actuality is not
analysed in terms of *change.
The notion of ‘success condition’ can be developed further. In the case
of substances, the goal (as in the case of activities such as seeing) is present
in the activity itself (1048b21). By contrast, in the case of *changes, the goal
is reached only at the end of the *change. Although the builder is engaged
in housebuilding, the *change is not completed until the house is built.
(This is the imperfective aspect of *change noted in section II.) The
relevant success condition for the realisation of the capacity to build does
not require that builder complete the house. Housebuilding can be going
on even in cases where the builder does not live to finish his project.
Although the house may be the final goal, some goal is achieved in the
very activity of housebuilding itself (1050a25–28). The success condition
relevant to capacities for *change or being *changed is the achieving this
intermediate goal. *Changes can do this without reaching their final goal.
Actualities in Theta 8 (so interpreted) are one part of Aristotle’s basic
ontology. In the case of substances and *changes, the actuality, understood
as the relevant success condition, is what has to be achieved for the capacity
to be realised. When the capacity is one for being a substance, the relevant
success condition is being that substance. When the capacity is one for
*change, the success condition will be the *changing. In the first case, the
success condition (being an axe or being a human being: the form; see
1050a15ff., 1050b2) will be met when there is a particular substance (such as
an axe). In the second, when the relevant success condition is met, one has
a particular *change (such as a particular housebuilding). The notion of
actuality in these cases is simply what has to be achieved for the capacity to
be realised. This general notion can be used without circularity in the
definition of *change.
Aristotle is setting out a map of the relevant ontology. On it he can
locate various types of continuant (including *changes and substances), all
which can have different properties at different times. (A human can be
short at some time, tall at another, as a particular *change can be slow at
one time, fast at another.) From this perspective, substances, *changes and
204 david charles
activities such as seeing are differing types of continuant, each with its own
distinctive features.28 This is why Aristotle presents *changes and sub-
stances as having analogous but irreducibly distinct roles (see 1048b9ff.).
His aim, in talking of potentiality and actuality, is to compare and contrast
these differing types of continuant.
Aristotle’s strategy throughout Metaphysics Theta is designed to generate
this overview. When he says at the outset ‘Let us get clear about the basic
case of capacity, an actuality which is not most useful for our purposes’
(1046b36f.) his aim is to get clear about the basic case (of *changes) and
then to examine the use of ‘actuality’ and ‘potentiality’ beyond it. There
can be genuine cases of capacities (or potentialities) beyond those for
*change introduced in 1046a4–19 and, equally, beyond capacities for
activities like seeing. There are, as is clear in Theta 6 and 7, capacities to
be a given object distinct from those for undergoing a *change or activity.29
While the most useful sense of ‘potentiality’ is that in which something is
said to be something potentially, this is not a capacity for a *change (or for
an activity like seeing) but rather a capacity to be. Aristotle carries through
the project, outlined in the introductory chapter (1045b35ff.), in the
remainder of Theta, pointing to different kinds of continuant: *changes,
substances, and activities such as seeing, with differing features.30 His
ontology, as becomes clear as Theta develops, is made up of these varying
types of continuant. Interpretation A captures this ambitious but stead-
fastly non-reductive account of *changes.
What are *changes in Aristotle’s account? As a distinctive type of
continuant, they are, it seems, unfoldings or processes (changings), with
different properties at different times. They can, as already noted, move
through space and end before they should. So understood, his *changes are
not changes, understood as events. Events do not move through space.31 If
they unfold, they do so by one (sub-)event happening after another. There
is no one event that accelerates and decelerates. Events are not the kind of
thing to speed up or slow down.32 There is no one event that is non-

28
For the general notion of a continuant, see W. E. Johnson 1921.
29
For this approach, see A. Anagnostopoulos 2010 and Johansen 2012a. It was developed in some
detail in Friday morning discussions on Theta 1 in Oxford in 2008. For a contrasting view, see Frede
1989 and Beere 2009.
30
Examples of differences: (a) matter, conceived as the capacity to be (e.g.) a human, plays an essential
role in the case of substances (1050a15–16) but not in that of *changes; (b) in the case of substances,
the relevant success condition is a form, such as being an axe or being disposed to cut, not the
activity of cutting itself. The latter is the success condition for the potentiality to cut.
31
This point is made by Dretske 1980.
32
For further discussion, see Stout 1996, and Galton and Mizoguchi 2009.
 Aristotle’s processes 205
derivatively fast at one point and slow at another. Events are occurrents,
not continuants: they happen at times, or occur between times, but do not
themselves move through time. Although events unfold, they are distinct
from the process of their unfolding.
To conclude: Aristotle’s *changes (kinêseis), as continuants, are pro-
cesses, not events.33 A proper appreciation of this point has important
implications for our understanding of his accounts of action, time, and
causation. The end of this chapter is not the end of the still unfolding story
of Aristotle’s processes.

33
Most recent commentators talk of *changes as events: see, for example, Penner 1970 and Coope
2007, some (it seems) misled by translating *changes (kinēseis) as changes at the outset. For some
relevant conceptual distinctions, see Fine 1999 and Haslanger 2003.
 chapter 11

Physics v–vi versus viii: Unity of change

and disunity in the Physics
Jacob Rosen

Introduction
You toss an apple straight up into the air and let it fall back into your hand.
The apple moves up, and then the apple moves down. There is an upward
motion of the apple and then a downward motion of the apple. Do these
two motions compose a single whole motion? Is there such a thing as the
up-and-down motion of the apple?
Another question. You read Physics v–vi, with its general theory about
changes and continua. Then you read Physics viii, with its cosmological
arguments about the eternity of motion and the existence of a first
unmoved mover. Have you just read a single text, something we could
call a ‘continuous treatise on movement’ (Ross 1936: 3)?
The first question matters to Aristotle’s cosmology. Aristotle believes he
can show, in Physics viii.8, that the answer is ‘No’. He purports to prove
that when something moves back and forth along a straight line, its
successive motions in opposite directions do not compose a single motion.
This result, together with other theorems of his physics, entails that there
cannot be an eternal motion along a straight line. From this he infers that
the only possible eternal motion is circular motion. Since, independently
of this, he thinks he has proven that there needs to be an eternal motion, he
can assemble a demonstration for the existence of eternal circular motion.
His candidate for such a motion is the motion of the heavenly sphere in
which the stars are fixed. Ultimately, then, his argument in Physics viii.8
offers a measure of confirmation for certain astronomical theories of his
day, and a sort of explanation for the existence and rotation of the outer-
most heavenly sphere which is posited by these theories.
The second question matters to the history and interpretation of
Aristotle’s writings. Its answer admits of degrees. It will affect, among
other things, the extent to which we use passages in one book to elucidate
206
 Physics v–vi versus viii 207
passages in another, and the extent to which we may combine doctrines
from the different books and then offer up the resulting whole as a view
held by Aristotle.
In this chapter I will develop some thoughts about the second question,
the one about the text, by way of considering how Aristotle reaches his
answer to the first question, the one about the apple. Aristotle offers several
arguments in Physics viii.8 for his thesis that, when something moves
back and forth, it does not undergo a single motion. These arguments
occur against the background of a sophisticated theory, expounded in
Physics v–vi, of the basic structure of motions and of other continuous
entities such as times and magnitudes. The arguments in Physics viii.8
stand in a surprisingly complex relation to that theory. On the one hand,
Aristotle evidently relies on the theory in a number of crucial steps.1 Yet in
other steps he seems to contradict or misapply the theory. This situation
offers the occasion to examine Aristotle’s views about some fundamentals
in the metaphysics of motion, while also raising questions about the unity
of the text which has come down to us as the Physics.
Let me signal in advance one of the questions we will encounter. In one
of the arguments of Physics viii.8, Aristotle introduces a thesis about
continua that we may call the Potentiality Doctrine. According to this
doctrine, a continuous entity has no actually existing proper parts and no
actually existing middle-points. Rather, it has parts and middle-points only
potentially or in capacity. The Potentiality Doctrine is not affirmed in
Physics v–vi. To the contrary, Aristotle often refers in these books to parts
and middle-points without ever suggesting that they are only potentially
there. It is natural to wonder whether we may attribute to Aristotle a single
overall theory of continua in which the Potentiality Doctrine is combined
with the theory of Books v–vi. The answer will depend on many con-
siderations, but an important one is this: when we come to Aristotle’s
argument for the Potentiality Doctrine in Physics viii, we will see that it
rests on an assumption that is contradicted by a theorem in Physics vi.5.
That is reason to doubt whether the Potentiality Doctrine is going to
combine successfully with the theory of Physics v–vi. At the least, it speaks

1
For example, Aristotle’s rejection of eternal rectilinear motion in Physics viii.8 tacitly relies on the
following two theorems from Physics vi: (1) no motion traverses an infinite straight line (Phys. vi.10,
241a26–b10); and (2) no motion takes an infinite time to traverse a finite distance (Phys. vi.2,
233a31–34, and vi.7, 237b24–25). These theorems imply that if something moves forever along a
straight line, it must sometimes turn around; Physics viii builds on this by arguing that what turns
around does not undergo a single motion.
208 jacob rosen
against regarding these three books of the Physics as a single, continuous
exposition of such a combined theory.
In all, there are many points of agreement and more than one point of
tension between Books v–vi and Book viii. I will begin by presenting a
few fundamental points of agreement. Then I will argue that there are (at
least) two tensions. First, in Physics viii.8, Aristotle assumes that any single
motion must be homogeneous (in a sense to be explained), whereas
according to Physics v–vi a motion is not homogeneous. Second, in
Physics viii.8, Aristotle assumes a beginning of change (in a sense to be
explained), whereas in Physics vi.5 he proves that there is no such thing.
These tensions both undermine Aristotle’s justification of the Potentiality
Doctrine and affect his strongest arguments against the existence of eternal
rectilinear motion. We will thus need to ask where these tensions leave us
in assessing the unity and success of Aristotle’s project in the second half of
his Physics.

I Common ground
Let us begin with three doctrines about change that are endorsed by
Aristotle both in Physics v–vi and in Physics viii. The first is the thesis
that there are changes. The second concerns the way in which changes
occupy time. The third concerns the conditions under which a given
change is ‘one’ (full stop), or ‘one with’ a given change.
The most fundamental commonality between the different texts is their
commitment to the existence of such objects as changes and motions.2 This
is a substantial theoretical commitment, going beyond the more modest
claim that things move and change, or that there are moving and changing
things. (Davidson posed the question: ‘Things change; but are there such
things as changes?’3 Aristotle’s answer, like Davidson’s, is ‘Yes.’) Aristotle
writes:
It is necessary that if a motion is present, then something moves, and that if
something moves, then a motion is present.4 (Phys. vi.1, 231b25–26)
I should mention that this biconditional is preceded in the text with an ‘if
indeed’ and followed by a ‘then . . . ’. It is the antecedent of a conditional

2
In my translations of Aristotle, I will use the word ‘motion’ and its cognates to translate κίνησις and
its cognates, and use ‘change’ and its cognates to translate μεταβολή and its cognates. Every motion is
a change, but not vice versa. The distinction between motion and change is not important for this
chapter, and I will not be careful about it.
3
Davidson 1970: 25. 4 Translations are my own.
 Physics v–vi versus viii 209
claim, not something that Aristotle asserts on its own. Still, it is beyond
doubt that he endorses it. A sign of this is his easy way of switching back
and forth between verbs predicating of things that they move or change,
and nouns denoting motions or changes. He does so both in Physics v–vi
and in Physics viii.5
As it stands, Aristotle’s biconditional is fairly weak. It merely says that
something moves if and only if there exists a motion. It does not explicitly
say whether there is a motion of each thing that moves; or whether a motion
is present at every time when something moves. Nor does it specify whether
motions have particular kinds and characters, corresponding to how things
move. (For example, is it the case that something moves fast if and only if
there is a fast motion? Or that something moves to Venice if and only if
there is a motion to Venice?) We must return to the last question later, but
the first two have natural, obvious answers, and it seems clear that Aristotle
accepts these obvious answers. He assumes that a subject S moves if and
only if there is a motion of S, and that something moves during time T if
and only if there is a motion of it occupying T. So I will take it that in
Physics v–vi and viii, Aristotle endorses the following principle:
Noun–Verb Translation Principle: S moves during T iff there is a
motion of S that occupies T.
A second point of agreement between Physics v–vi and viii is their
understanding of the way in which changes are temporally extended.
Aristotle in both texts conceives of changes as spread out in time like
sails. By this I mean that, where a change occupies a given time, it has
different parts occupying different parts of the time.6 (Physics vi contains
an analogous claim involving spatial extension, namely that, where a
change belongs to a given body, it has different parts belonging to different
parts of the body.7) We should appreciate that this is not the only possible
philosophical view Aristotle could have taken. In principle he might have
conceived of a change as a continuant: something that endures through
time in much the way that substances are naturally thought to do. Thus, he
might have thought that a change is wholly present in every part of the time
for which it exists. I have heard it said that Aristotle sometimes regards

5
See, for example, Phys. viii.8, 265a7–12.
6
See, for example, Phys. vi.4, 235a18–24; viii.8, 263a27–29, 264a24–26. As already mentioned,
according to viii.8 a change has these parts only potentially, not actually (whatever exactly that
means).
7
Phys. vi.4, 234b21–24 and following. This claim is neither affirmed nor contradicted in Physics viii,
to my knowledge.
210 jacob rosen
changes as continuants in this way. I do not know whether that is true, but
in any case he does not so regard them in the texts I am discussing.
The third point of agreement between the texts is a shared framework
for addressing issues of unity for changes. Recall that Aristotle’s question in
Physics viii.8 is whether there can be a single change of a certain sort
(namely, a single change back and forth along a straight line). In order to
answer it, he builds upon a discussion of criteria for oneness of change that
was presented in Physics v.4. (He refers back to this discussion at viii.8,
262a1.) Aristotle there discussed three main ways of being one: being one in
genus (genei), being one in kind (eidei), and being one ‘simply’ or ‘without
qualification’ (haplôs).
The last, and strongest, way of being one is the topic of Aristotle’s
concern in Physics viii.8. On the surface, at least, he adopts the same
criteria for it as were laid down in Physics v.4. According to both texts, a
change’s being one without qualification depends upon three factors: what
changes (i.e. the subject of change), when it changes (i.e. the time of
change), and that ‘in which’ it changes (i.e. the path of change, if I
understand rightly).8 A change is one without qualification if and only if
(i) its subject is one, (ii) its time is one and without gaps, and (iii) what it is
in is ‘one and indivisible’ (v.4), or ‘undifferentiated in kind’ (viii.8).9 The
third criterion appears to be the same as the criterion for a change’s being
one in kind.10 This explains Aristotle’s statement that ‘necessarily, a
motion that is one (namely, without qualification) is also one in kind,
although it is not necessary for a motion that is one in kind to be one
without qualification’ (Phys. v.4, 228b9–10).
In the next section I will discuss a problem in Aristotle’s treatment of
oneness in kind in Physics viii.8. But before I can do that, there is a point
that requires clarification.

Interlude: One-place oneness and two-place oneness

Aristotle’s discussion is subject to a complication that will be crucial in the
next section of this chapter. The complication is that he shifts between
what we may call a ‘two-place’ use and a ‘one-place’ use of the notion of
oneness.
A two-place statement of oneness has the form ‘change A is (or is not)
one with change B’. For example, Aristotle tells us that every locomotion is

8 9
Phys. v.4, 227b23–26; viii.8, 262a1–4. Phys. v.4, 227b29–228a3; viii.8, 262a1–2.
10
Phys. v.4, 227b6–7, 19, 27–28.
 Physics v–vi versus viii 211
one in genus with every other locomotion (Phys. v.4, 227b5). He tells us that
every whitening is one in kind with every other whitening (227b11). And he
tells us that one man’s restoration to health is not simply one with another
man’s restoration to health (228a1–3). He gives reasonably clear explana-
tions of what he takes such claims to mean.
A one-place statement of oneness has the form ‘A is (or is not) one’. It is
sometimes, but not always, clear what Aristotle means to say with a
sentence having this form. A fairly clear case is when the term substituted
for A signifies a type of change. In this case the sentence can be understood
as equivalent to a certain two-place oneness claim: it means that each
change of the type is one with each other change of the type. For example,
Aristotle says that learning is, to a degree, one in kind (227b13), and this
seems to mean that every learning is one in kind with every other learning.
Another sort of case is trickier, although the basic intention behind it is
still recognisable. This is the case where the term substituted for A purports
to refer to an individual change, and the sentence either affirms or denies
that the term’s referent is one without qualification. Consider an example
(in this passage it is clear from context that ‘one’ means ‘one without
qualification’):
The motion is not one but many, if there is rest between them.
Consequently, if a motion is separated by stationariness, it is not one or
continuous. (Phys. v.4, 228b4–6)
What makes these sentences difficult is that the term ‘the motion’ or ‘a
motion’ seems as though it must refer to a single thing, if it refers at all. And
yet the term is used to say such things as ‘the motion is not one’ or ‘the
motion is many’. Are such claims capable of being true? Aristotle seems to
be struggling somewhat to express himself properly; indeed he mixes
grammatically singular and plural forms in a way that tests the rules of
syntax. For example, in his phrase ‘the motion is not one but many’ there is
a predicate (‘many’, pollai), which does not agree in number with its
subject (‘the motion’, hê kinêsis).
Nevertheless, Aristotle’s basic intention seems reasonably clear. The idea
is that there are some changes (plural) that we are interested in, and we are
interested in whether or not these changes compose a single change. In
order for them to do so, there are some conditions which they jointly must
satisfy: for example, they must jointly occupy a time that has no gaps in it.
A term such as ‘the motion’, although grammatically singular, can be used
to say things plurally about the changes of interest. Thus, for example, ‘the
motion is separated’ is made true not by the fact that any single individual
212 jacob rosen
is separated, but by the fact that the motions in question (plural) are
separated. (The plural predication here is like what we use when we
say of some people, ‘they are crowded together’, or ‘they are arranged
in a square’. Such sentences do not say something about each person
individually – they do not say that any individual person is crowded
together or arranged in a square – nor do they posit a whole item,
composed out of the people, which is a crowded or square item. They
simply attribute a predicate to some people plurally.)

II Oneness in kind and homogeneity

As we saw above, Aristotle holds that oneness without qualification implies
oneness in kind. Oneness in kind is among the criteria for unqualified
oneness.
This point is fairly clear when we deal with two-place statements of
oneness. Suppose that my face simultaneously heats up and turns red, and
we want to know whether the heating is simply one with the reddening.
Well, the subject of the heating is one with the subject of the reddening,
and the time of the heating is one with the time of the reddening. Two of
Aristotle’s three criteria are met. However, his third criterion is not met:
the heating is not one in kind with the reddening. Hence the heating is not
unqualifiedly one with the reddening, because it differs from it in kind.
When we turn to one-place statements of oneness, we encounter a
difficulty. Suppose we now ask whether my face’s heating and its reddening
are, i.e. compose, a single change (we could put this by asking whether my
face’s growing flushed is unqualifiedly one). Or, returning to our old
example of the apple, suppose we ask whether the apple’s motion from
the apex of its arc down to your hand is simply one. (Do its motion from
apex to the level of your eyes and its motion from there to your hand come
together to compose a single change?) To answer such questions we will
need to apply Aristotle’s three criteria in new ways. We will need to know
what it means to say that a change’s subject or its time ‘is one’: not one with
something, but just plain one. And we will need to know what it means to
say that a change ‘is one in kind’: again, not one in kind with something,
but just plain one in kind.
Above I suggested that we can reconstrue such claims in a plural form.
We can look at the components of a putative single change and ask of
them, plurally, whether they belong to one subject, whether they occupy one
time, and whether they are one in kind. But understanding these plural-
form questions is not straightforward. We should notice, to begin with,
 Physics v–vi versus viii 213
what the sentence ‘they occupy one time’ does not mean. It does not mean
that each of the changes in question occupies the same time as each other
change. For example, assuming that the apple’s motion from apex to eyes
and its motion from eyes to hand compose a single motion from apex to
hand, we will need to say that they occupy one time. But the motion from
apex to eyes does not occupy the same time as the motion from eyes to
hand. One occurs after the other. The thought behind Aristotle’s criterion,
presumably, is that the times occupied by these two motions ‘add up’ to a
single, continuous time.
I want now to focus on the criterion involving oneness in kind. In Physics
viii.8, I will suggest, Aristotle interprets this criterion precisely along the
lines that we have just rejected for the criterion involving time. That is, he
appears to hold the view that if a change is unqualifiedly one, then each of
its parts must be one in kind with each other part. Then I will argue that
this view is inconsistent with the theory of Physics v–vi.

II.1 Homogeneity in Physics viii.8

Let us say that a change is homogeneous if and only if each part of the change
is one in kind with each other part of the change. In Physics viii.8, Aristotle
apparently holds that a change is one without qualification only if it is
homogeneous.
We find this view at work in Aristotle’s first argument for the thesis that,
if something moves back and forth along a straight line, it does not
undergo a single motion. The argument begins:
For it turns back, and that which turns back on a straight line undergoes
contrary motions (261b32–34)
And he concludes:
but contraries differ in kind, and are not one. (262a5)
In this argument, Aristotle notes that a supposed single back-and-forth
motion would have two contrary motions as parts; he observes that these
contrary motions differ in kind from each other; and he infers that there is
no single back-and-forth motion. This appears to be an application of the
view that a single motion must be homogeneous.
Aristotle gives further indication of the homogeneity of changes later on
in the chapter. The context is the following claim: if something undergoes
a motion which it was not always undergoing, then the motion is preceded
by the sort of rest which is opposed to it. (Rest at A is opposed to motion
214 jacob rosen
from A to B.) Thus, for example, for some period before you walk from
your office to the beach, you are at rest in your office. Now, Aristotle
recognises that his claim must be restricted. For motions have smaller
motions as parts, and Aristotle does not want to say that all these smaller
motions are preceded by periods of rest. On your way from the office to the
beach you pass a lemonade stand and an ice-cream booth, and there is a
part of your walk by which you traverse the distance between them.
Aristotle does not think that this part of your walk must be preceded by
a period of rest in front of the lemonade stand. So, he restricts his claim in
such a way that it does not apply to motions that are parts of larger
motions. He states his restriction in the following way:
I mean those motions that are different in kind, and not if it is some part of
the whole motion. (264a25–26)
Aristotle here contrasts motions that are parts of motions with motions
that are different in kind. This suggests that the parts of a motion are one in
kind with the wholes of which they are part, and with the other partial
motions that precede and succeed them.

II.2 Heterogeneity in Physics v–vi

According to the theory of Physics v–vi, by contrast, an unqualifiedly
single change will not be homogeneous. The reason is this. Every change
is a change from something and to something;11 let us call these its starting-
point and its end-point. A change is one in kind with another change only
if it has the same end-point as it. (There are further conditions, too: it must
also have the same starting-point, and indeed the same path from starting-
point to end-point, as the other change.)12 But the different temporal parts
of a single change have different end-points. For example, the apple’s
whole downward motion, from its apex to your hand, is unqualifiedly
one change. But one of its parts is a motion from the apex to a place before
your eyes, and another of its parts is a motion from there to the place in
your hand. These parts have different end-points, so they are not one in
kind with each other. Hence the apple’s whole downward motion is not
homogeneous.
A thought along these lines finds explicit expression in Nicomachean
Ethics x.4 (complete with an apparent cross-reference to Physics v–vi).13

11
Phys. v.4, 225a1; vi.4, 234b11 and 5, 235b6. 12 Phys. v.4, 227b14–20.
13
NE x.4, 1174a31–b5. Aristotle uses this line of thought in order to argue that pleasure is not motion.
 Physics v–vi versus viii 215
What matters to us here, and what I would now like to show, is that the
line of thought brings out a genuine consequence of the doctrines of
Physics v–vi.
According to Books v–vi it is clear that, where a motion occupies a
given time, the motion has a part in each part of the time. These parts are
themselves motions.14 And in the lesser times occupied by these lesser
motions, the moving thing moves (kineitai) or traverses (dierchetai) lesser
magnitudes. Thus, we find statements such as the following:
It is evident that the magnitude is continuous if the time is, since in half the
time it traverses half the distance, and in general a lesser distance in a lesser
time. (Phys. vi.2, 233a13–16)15
Suppose, for example, that there is a motion from A to C, and that B lies
between. Then the motion from A to C has, as parts, a motion during
which the thing traverses magnitude AB and a motion during which the
thing traverses magnitude BC. It is reasonable to think that these latter
motions are, respectively, a motion from A to B and a motion from B to
C. Aristotle himself does not have occasion to describe them in precisely
this terminology, but arguments can be given to show that this is what
they are.
The main thing needed in order to show this is a supplementation to the
Noun–Verb Translation Principle stated earlier. We need to add a correla-
tion between a characteristic of a moving thing, namely that it moves from A
to B, and a characteristic of a motion, namely that it is a motion from A to B.
If we are justified in doing this, then we may attribute to Aristotle the
following expanded principle:
Expanded Noun–Verb Translation Principle: S moves from A
to B during T iff there is a motion of S from A to B that occupies T.
There is good evidence for such an expanded principle in Physics v–vi.
Here, for example, are two passages from v.1:
Every motion is from something and to something, for that which primarily
moves is different from that to which it moves and that from which it
moves. (Phys. v.1, 224b1–2)
(a) Since every change is from something to something . . . (b) that which
changes could change in four ways: either from a subject to a subject, or

14
See for example Phys. vi.4, 235a18–24. At Phys. vi.1, 232a8, an assumption for reductio has the
consequence that ‘the motion would not be composed out of motions’, and this is treated as an
unattractive consequence.
15
See also Phys. vi.2, 232b7–8, a34–b2, 6, 236b34–237a3, and 7, 237b23–24.
216 jacob rosen
from a subject into not a subject, or not from a subject into a subject, or not
from a subject into not a subject . . . (c) Consequently, it is necessary from
what has been said that there are three changes: that from a subject to a
subject, that from a subject to not a subject, and that from not a subject to a
subject. (Phys. v.1, 224b35–225a10)
It seems clear in the first of these passages that Aristotle is equating what a
motion is from and to with what a thing moves from and to. The equivalence
is especially clear in the second passage. The details of interpretation of this
passage, for example the question what exactly Aristotle means by ‘subject’,
need not concern us. What is noteworthy is Aristotle’s switch back and
forth between saying that a change is from and to something, in points
(a) and (c), and saying that a thing changes from and to something, in point
(b). Aristotle’s inference from (b) to (c) shows that he regards the two ways
of speaking as equivalent. In other passages in Books v and vi as well,
Aristotle switches fairly casually between speaking of what a change is from
and to, and of what a thing changes from and to.16
Given the expanded translation principle, it only remains to convince
ourselves that, during the different parts of a motion from A to C, a thing
moves from A to B and moves from B to C. Here we must pause to note
an ambiguity in the verbal form ‘moves’, or rather in the corresponding
Greek present tense form. On the one hand, the form can be understood
as having perfective aspect. So understood, the statement that S moves
from A to B during a time implies that, at the end of the time, S has moved
from A to B. (This in turn implies that S is at B at the end of the time,
Phys. vi.5, 235b7–8.) But, on the other hand, the form could be under-
stood as having imperfective aspect, equivalent to ‘is moving’. The
statement that S is moving from A to B during a time does not imply
that S has moved to B by the end of the time. For example, if it takes you a
whole day to walk to Thebes, then it is not true that you walk (perfective)
to Thebes in the morning, but it is true that you are walking to Thebes in
the morning.
The question is which verbal aspect figures in the translation principle?
Is a motion from A to B a motion during which something moves
(perfective) from A to B, or a motion during which something is moving
(imperfective) from A to B? If it turned out to be the latter, then perhaps
motions would be homogeneous after all. For it is plausible that a thing is
moving from A to B during every part of a motion from A to B.

16
See, for example, Phys. v.1, 224b7–10, 12–15; vi.4, 234b10–13 and 10, 241a26–b11 (esp. b9); see also
vi.5, 235b13–14, 236b2–4.
 Physics v–vi versus viii 217
In my view, it is more natural on the whole to adopt the perfective
reading of Aristotle’s present tense ‘moves’. This reading is also supported
by certain considerations of detail.17 If you accept my view of the matter,
then you will join me in drawing inferences from claims made by Aristotle
in the perfect tense, such as the following:
Let AB have moved from B to C primarily . . . If BC is divisible, there will be
something before C to which AB has changed, and another in turn before
that. (Phys. vi.5, 236b11–14)
Here Aristotle says that, before having changed from one point to another,
a thing has changed to an intermediate point. I infer that the thing changes
to an intermediate point, and, applying the translation principle, conclude
that it undergoes a change to an intermediate point.
Fortunately, there is one passage in which Aristotle himself uses the
present tense, so that we may apply the translation principle directly,
without settling questions of verbal aspect or the relation between present
and perfect tense forms. In this passage, Aristotle supposes that something
has changed from C to D. He argues that CD is not indivisible. Then he
proceeds:
Necessarily, what is in between is a magnitude and is infinitely divisible.
Consequently, it changes to those beforehand. (Phys. vi.6, 237a33–34)
Here Aristotle says that, before having changed from C to D, the thing
changes to the various points between C and D (i.e. the points at which
CD can be divided). He says this using the present tense. Applying the
translation principle, it follows that the thing undergoes a change to each of
the intermediate points. Each of these changes is, to review, different in
kind from the others, and they are all parts of the change from C to
D. Consequently, the change from C to D is not homogeneous.
To conclude, some of Aristotle’s arguments in Physics viii.8 against the
possibility of eternal rectilinear motion are based on the principle that a
change must be one in kind in order to be unqualifiedly one. This principle
is common to Books v and viii. But Aristotle interprets the principle to
mean that if a change is unqualifiedly one, then each of its parts must be

17
Here are two considerations. (1) At Phys. v.6, 230a4–5, Aristotle says that if something stands still at
A, then a motion to A is or coincides with a coming-to-rest. His claim seems plausible only on the
assumption that at the end of a motion to A the thing is at A (so that it will rest immediately after the
motion). This assumption is validated by the perfective, but not by the imperfective, reading of the
translation principle. (2) At Phys. vi.5, 235b6–8, Aristotle juxtaposes present and perfect tense claims,
in a way that suggests (even if not, I grant, strictly entailing) that what a thing changes to and what it
has changed to are the same for any given change.
218 jacob rosen
one in kind with each of its other parts. This contradicts the theory of
Physics v–vi, according to which a single change has, as parts, changes to
different endpoints, with the result that its parts differ in kind from each
other.

III The potentiality doctrine and the beginning of change

Aristotle has other arguments in Physics viii.8 that are based on another of
his criteria for unity of change. In these other arguments, he appeals to the
principle that a change, in order to be one, must occupy a single time
without gaps. He thinks he can show that this criterion is failed by back-
and-forth motion: when something moves back and forth along a straight
line, it must rest for some time at the point where it turns around. (Indeed,
he thinks you can see that it rests there, Phys. viii.8, 262a18.)
In the course of developing these arguments, Aristotle introduces a new
general doctrine about continua, which I will refer to as the Potentiality
Doctrine. This is the claim that a single continuous thing, such as a motion,
line, or time, has parts and middle-points only potentially or in capacity, not
in actuality. On the other hand, he argues, when something moves back
and forth, the point where it turns around is in actuality a middle-point of
its motion. Since its motion has an actual middle-point, it follows that its
motion is not one. Along the way, he also uses the Potentiality Doctrine to
answer one of Zeno’s paradoxes of motion (263a4–b9).
Aristotle seems conscious in Physics viii.8 that the Potentiality Doctrine
is a new thesis relative to the theory presented in Books v–vi. At least, he
thinks that the doctrine provides a new and better response to Zeno.18
There were gestures in the direction of this doctrine in earlier books,
during Aristotle’s discussions of infinity and of time (Phys. iii.6,
206a14ff., iv.11, 220a10ff.). But it is nowhere endorsed in Books v–vi.
This is somewhat surprising, given that many of Aristotle’s proofs in
Physics v–vi appeal to the existence of parts and of intermediate points.19
One would expect it to make a difference to these proofs whether or not the
parts and middle-points actually exist, as opposed to being mere possibilia,
and yet he says nothing about the question.
The Potentiality Doctrine is rather difficult to understand, but it is
clearly important to Aristotle. It is closely connected with some claims

18
He claims to improve on the solution given ‘in the first discussion of motion’ (ἐν τοῖς πρώτοις
λόγοις τοῖς περὶ κινήσεως, Phys. viii.8, 263a11), apparently referring to Phys. vi.2, 233a21ff.
19
See, for example, Phys. vi.2, 232a32ff., 232b27ff., 4, 234b23ff., 6, 236b32ff., and 10, 241a8ff.
 Physics v–vi versus viii 219
about potentiality found in the Metaphysics, in particular the doctrine that
no substance is composed out of actually existing substances (Metaphysics
Z, 13 and 16). (The doctrines are related but not the same, since motions,
lines, and times are not substances.) Aristotle seems to need it in order to
combine his belief in the existence of continua with his denial of the
simultaneous actual existence of infinitely many things.20 Commentators
often regard it as one of his central doctrines about continuity.21
What, then, should we make of the Potentiality Doctrine’s absence
from Physics v–vi? Here are two opposed and extreme views. On a
unitarian view, we might simply infer that Books v–vi are not intended
as a complete, self-sufficient treatment of continua. Part of their job is
to build up to Book viii, and they were written with the intention of
their being supplemented by further refinements such as the Potentiality
Doctrine. On an opposite view, we might infer that Books v–vi were
written in isolation, independently from the concerns of the rest of the
Physics. Issues about causation, potentiality and actuality, and infinity
are simply not on their agenda. They are a more or less mathematical
text (relating perhaps to a part of mathematics in the way that optics
relates to geometry),22 plonked into the middle of a more natural-
philosophical one.
Presumably the truth lies somewhere between these extremes, and I will
not try to determine where precisely. But I would like to point out two
issues that are relevant to the question. The first issue is whether Aristotle’s
argumentation in Physics v–vi is even compatible with the Potentiality
Doctrine. Do his proofs go through when we add the premise that none of
the parts and middle-points appealed to therein have actual existence?
There is no space here to settle this issue, since it depends on both a
detailed interpretation of the Potentiality Doctrine and a detailed analysis
of Aristotle’s proofs. I am optimistic that the central proofs of Books v–vi
can be recast so as to be compatible with the Potentiality Doctrine.
The second issue concerns the argument that Aristotle gives in support
of the Potentiality Doctrine in Physics viii.8. The argument is unsound.
Worse, the flaw in Aristotle’s argument for the Potentiality Doctrine

20
See Coope 2005: 10. Aristotle’s main target in his discussion of infinity (Physics iii.4–8) is the
question of infinitely extended magnitudes. But he also speaks of number, and he seems to deny any
simultaneous, actually infinite number of things at iii.7, 207b11–15.
21
See, for example, Ross 1936: 68.
22
The relation in question I have in mind is described by Aristotle as ‘being under’. See APo. i.13,
78b32ff. for this relation both within mathematics and across the boundary between mathematics
and natural science.
220 jacob rosen
appears to rest precisely on a failure to appreciate a significant theorem
from Physics vi.5.

III.1 The argument of Physics viii.8

When you threw the apple into the air and let it fall back into your hand,
there was a point where the apple turned around. That point was an actual
middle-point of the apple’s movement up and down. It served as the end-
point of the apple’s motion up and the starting-point of the apple’s motion
down. The apple arrived there from your hand and departed from there
back to your hand. For Aristotle in Physics viii.8, these claims imply that
the apple rested for some time at the point, and hence that it did not
undergo a single motion up and back.
The point where the apple turned around is different from a point
before your eyes which the apple passed through on its way down. At the
point before your eyes, the apple did not rest for any time. It did not arrive
there or depart from there. The point did not serve as the end-point of a
motion or as the starting-point of a motion. It was not an actual middle-
point of the apple’s downward motion. A single motion has no actual
middle-points.
Aristotle can parlay his result about middle-points of motion into a
general doctrine about parts and middle-points of continua. For, first,
Aristotle presumably thinks that a motion has an actual middle-point if
and only if it has actual parts: if there is a middle-point, then there are
parts that meet at the middle-point, and if there is a part, then there is a
middle-point where the part meets its remainder. This explains why,
having argued that a continuous motion has no actual middle-points, he
later asserts that it has no actual halves (viii.8, 263a28–29). Second,
Aristotle seems to think that a motion has actual parts if and only if the
time occupied by the motion has actual parts, and if and only if the
magnitude traversed by the motion has actual parts (compare Phys. vi.4,
235a15–17ff.). This explains why he is prepared to assert that continuous
lines and times, in addition to motions, have only potential parts
(viii.8, 263b3–9).
Let us follow Aristotle through his argument about middle-points of
motion. The central passage is Physics viii.8, 262a19–b8, with the argu-
ment then being carried on at 262b17–263a3. Aristotle’s first premise is this:
if a point is (actually) a middle-point of a motion, then it is (actually) the
end-point of a motion (namely, of the part of the whole motion of which it
is the final boundary) and the starting-point of a motion (namely, of the
 Physics v–vi versus viii 221
part of the whole motion of which it is the initial boundary). He expresses
this as follows:
There being three things, beginning, middle, and end, the middle is both in
relation to each. (Phys. viii.8, 262a19–20)
Aristotle’s next two premises are these. If a point is actually the starting-
point of a motion, then at some time the moving thing has departed
(apogegone, apelêluthe) from the point. And, if a point is actually the end-
point of a motion, then at some time the moving thing has come to be
(gegone) at the point. Thus he writes, of something which has undergone a
motion from A to C:
It has departed from point A, i.e. the beginning, and has come to be at C,
when it finishes and stops. (Phys. viii.8, 262b7–8)
From these premises it follows that, if a point is actually a middle-point of
something’s motion, then at some time the thing has come to be at the
point, and at some time the thing has departed from the point. So far,
Aristotle’s reasoning is tenable. What comes next, however, is not so good:
It is impossible for A simultaneously to have come to be at B and to have
departed. Hence it does the one in one point of time, and the other in
another. Hence there will be a time in the middle, and consequently A will
rest at B. (Phys. viii.8, 262a32–b2)
Aristotle adds a further detail to his argument a little later (by which time
point B has been replaced by D):
It is not the case that it simultaneously has come to be at D and has departed
from D: for it would simultaneously be there and not be there in the same
instant. (Phys. viii.8, 262b26–28)
In these passages, Aristotle assumes that if A has come to be at a point and
has departed from the point, then there is such a thing as the instant, or
point of time, when A has come to be at the point, and such a thing as the
instant or point of time when A has departed from it. At the instant when A
has come to be at the point, it is at the point, and at the instant when A has
departed from the point, it is not at the point. Given the Principle of
Non-Contradiction, it follows that the instant when A has come to be at
the point is different from the instant when A has departed from the point.
Between any two instants there is a period of time. So, there is a period of
time between these two instants, during which A has come to be at the
point and has not yet departed from the point. Throughout this time, A is
at the point. Hence, A rests at the point.
222 jacob rosen
From the perspective of Physics vi.5, many of the steps in this argument
are acceptable.23 But there is a crucial error, namely, Aristotle’s assumption
that there is such a thing as the instant when A has departed from the point.
It is, of course, sometimes true to say of A that it has departed from the
starting-point of its motion. But for every instant at which it is true to say
this, there is an earlier instant at which it is also true to say this. There is no
first, primary time or instant in which A has changed some, in which A has
departed from the starting-point of its motion. This is argued at some
length in Physics vi.5 (236a13–27), and corollaries are derived from it in the
remainder of that chapter (236a27–b18).
When we hold this lesson in mind from Physics vi.5, we can see that the
argument of viii.8 does not go through. True, there is such a thing as the
instant when A has come to be at the middle-point, and yes, at this instant
A is at the point. Every instant at which A can be truly said to have departed
from the point is indeed different from, not simultaneous with, the instant
at which it has come to be there. So, yes, for any given instant at which A
has departed from the point, there is a period of time separating it from the
instant at which A had come to be there. But no, such a period is not a
period during which A has come to be at the point and not yet departed
from the point. For, prior to any instant you choose at which it is true that
A has departed, there is always an earlier instant at which it is already true
that it has departed. (There is a last instant at which it has not departed,
and every instant thereafter is one at which it has departed.)
As far as Physics v–vi is concerned, nothing seems to rule out the
possibility that the apple you threw and caught never rested anywhere
between leaving your hand and arriving back in your hand. What is more,
as far as I can see, nothing in those books rules out the possibility that it
underwent a single motion up and back. Related to this, Books v–vi seem
to be consistent with the proposition that a single motion has actual
middle-points. Now, Physics v–vi do not purport to deliver the whole
truth about motion. Perhaps Aristotle could derive the results he wants by
introducing some further independent principle – something consistent
with, though not entailed by, the theory of Physics v–vi. But that is not
what he does in Physics viii.8. His argument in Physics viii.8 rests on a
tacit premise that is inconsistent with Physics vi.5. Thus, the argument that
Aristotle actually gives for the Potentiality Doctrine, along with the

23
For example, the connections between having departed and not being there, and between having
come to be and being there, appear in vi.5 (235b14–16, 235b7ff.). Similarly, the claim that there is a
time between any two instants is ubiquitous in v–vi.
 Physics v–vi versus viii 223
associated argument against the existence of eternal rectilinear motion, is
incompatible with the theory of Physics v–vi.

IV Conclusion
Let me close with three questions.
The first question is, did Aristotle have a single basic understanding
of continua, or did he rather have multiple overlapping fragments of
theories? In particular, where does the Potentiality Doctrine belong in
his understanding of continua? I have argued that his attempt to
establish this doctrine is defeated by Physics vi.5. Given this, we should
not, without further ado, read the Potentiality Doctrine into the theory
presented in Physics v–vi. Further research is called for here. A minimal
condition on attributing a combined theory to Aristotle would be to
show that the Potentiality Doctrine (setting aside Aristotle’s attempted
proof of it) is consistent with the doctrines and proofs given in
Physics v–vi.
Second, what should we think about the unity or disunity of the second
half of the Physics? The tensions we have seen go beyond, I think, what we
normally find as a result of mistakes or oversights within a single philoso-
phical work. They indicate that Books v, vi, and viii were not all written
in one sitting. This impression is reinforced by differences in technical
terminology between the books.24 On the other hand, the texts do not
seem wholly independent of each other, and it is plausible that Aristotle
himself assembled them together. It would be useful to have a more
complete map of the interconnections and any additional tensions between
the texts, so as to make an informed hypothesis about the manner of their
composition.
Finally, do the tensions we have seen amount to a major problem for
Aristotle’s physics and cosmology? I have argued that some of Aristotle’s
arguments against the possibility of eternal rectilinear motion in Physics
viii.8 are inconsistent with v–vi. Are any of his arguments consistent with
the earlier books? I count six arguments in all. Two seem to rely on the

24
In most books of the Physics, including v–vi, a point is called a στιγμή (Phys. v.3, 227a27, 28, 31, 4,
227b16; vi.1, 231a25, 26, 30, 231b7, 9, 13, 10, 241a3, 7, 10, 12, 13, 19; there is an exception at vi.9,
240b3). In Book viii a point is called a σημεῖον (viii.8, 262a23, 29, 262b2, 4, 7, 12, 25, 263a24, 31,
b10, 12, 264a3). Furthermore, in Physics v–vi the starting-point and end-point of a change are
normally referred to using the prepositions ἐκ and εἰς, respectively (e.g. v.1, 225a1), while in Book
viii Aristotle uses ἀπό for the one and ἐπί or πρός for the other (e.g. viii.8, 262a7, 9–10, b10–12, 19,
264a29–31).
224 jacob rosen
mistake I described about the beginning of change (262a19–263a3, 264b1–
6). Two seem to rely, more or less directly, on the ‘homogeneity’ inter-
pretation of oneness in kind (261b32–262a6, 264a21–b1), which I have
argued is incompatible with Physics v–vi. One is brief, cryptic, and
prima facie irrelevant to the desired conclusion (264b6–9). This leaves
only one argument which, I think, could be made to work (264a7–21).
And this argument seems more suggestive than demonstrative, resting on
somewhat uncertain linguistic intuitions rather than clear scientific prin-
ciples. (Roughly, the intuition behind the argument seems to be that there
is something wrong with saying of a thing, ‘it is moving back and forth
between A and B’. Instead, you should say, at some times, ‘it is moving to
A’, and, at other times, ‘it is moving to B’.)
In the first book of On the Heavens, Aristotle treats it as a previously
established fact that circular motion, and only circular motion, can be
eternal (Cael. i.2, 269b6–9). We have not questioned Aristotle’s positive
arguments in the Physics for the possibility of eternal circular motion. But it
is troubling that his most powerful-seeming arguments on the negative
side, against the possibility of eternal rectilinear motion, are all defeated by
Physics v–vi. When we take away these arguments we weaken his case for
the priority of circular over rectilinear motion (cf. Phys. viii.9, 265a24–26),
and we hamper his effort to make the stars’ circular motion intelligible
(why is there an eternally rotating sphere rather than, say, an eternally
swinging pendulum?). Still, we should not exaggerate the trouble. Aristotle
has other arguments for the priority of circular motion (Phys. viii.9,
265a16–17, b11–12; Cael. i.2, 269a18–21). And he has other resources for
denying the existence of eternal rectilinear motion. For example, he could
mobilise his doctrines about natural and unnatural motions in On the
Heavens i. According to these doctrines, any simple body that moved back
and forth in a straight line would sometimes be moving unnaturally, and
no body would do that sort of thing forever.
 chapter 12

Perfection and the physiology of habituation

according to Physics vii.3
Mariska Leunissen

Introduction: Habituation as a factor in moral development

In his ethical treatises, Aristotle singles out three factors that play a role in
the moral development of men, namely nature, habit, and reason.1
Although a complete understanding of Aristotle’s moral theory will thus
involve an analysis of all three factors, this chapter focuses mostly on
Aristotle’s discussions of habit and the process of habituation.
Specifically, I aim to offer an interpretation of Aristotle’s “physical” dis-
cussion of the acquisition of character virtue in Physics vii.3 by drawing
from his physiological treatment of habituation in the ethical treatises.
In the Nicomachean Ethics, Aristotle argues – famously – that character
virtues do not come to be by nature (for instance, in the way that humans
realize their perceptive capacities),2 but rather that they are dispositions
(ἕξεις) that we acquire “through habits” (EN ii.1, 1103a17: ἐξ ἔθους)3 and “by
first activating them” (EN ii.1, 1103a31: τὰς δ’ ἀρετὰς λαμβάνομεν
ἐνεργήσαντες πρότερον). Virtues thus come to be as the result of the
repeated performance of virtuous actions, whereby the quality of the
action is said to determine the quality of the disposition that comes to be
(EN ii.1, 1103b6–25). In the Politics, Aristotle claims that also practices such
as bodily conditioning and exercise, imitations in play, and musical
instruction during childhood are productive of the character virtues, and
are therefore forms of habituation (ἐθισμός).4 Together, these practices aim
to change the appetitive soul-capacities of men in such a way that they
become correctly disposed towards undergoing the appropriate

1
EN x.9, 1179b20–21; Pol. vii.13, 1332a38–40 and 15, 1334b6–28; and EE i.1, 1214a14–25.
2
See especially EN ii.1, 1103a17–b2 and 6, 1106a9–10: “we have the capacities by nature, but we do not
become good or bad by nature.”
3
See also EE ii.2, 1220a39–b1 and MM i.6.2.
4
Pol. vii.13, 1332b10–11; vii.15, 1334b8–9; vii.17, 1336a18–19; and viii.4, 1339a7–10.

225
226 mariska leunissen
affections5 and obedient to the rational part of the soul.6 Because of this,
Aristotle singles out habituation as the most important factor in the moral
development of men, as it can override nature and prepares the ground for
moral education.7 However, even though Aristotle is fairly explicit about
the scope, aims, and importance of habituation as a factor in moral
development, the ethical treatises – perhaps because they are practical in
nature – do not provide much information about what exactly happens
physiologically speaking in the bodies and souls of men during habituation.8
This chapter aims to close this gap in our understanding of Aristotle’s
concept of habituation by examining the scant physiological remarks in the
ethical treatises from the perspective of Aristotle’s only extended discussion
of the changes involved in virtue-acquisition in Physics vii.3 and by
offering an explanation of how these physiological changes ultimately
result in the presence of character virtue as a distinct and unified psycho-
logical disposition. Let me note from the outset, however, that the
discrepancies between Aristotle’s ethical and physical accounts of virtue-
acquisition are significant, and that it is not my aim to harmonize the two
accounts in all respects. Rather, I will use Aristotle’s discussion in Physics
vii.3 as a starting-point for what I take to be a plausible, “Aristotelian”
account of the physiology of habituation, and draw from the ethical
treatises in order to flesh out this picture from the Physics and to further
clarify it.
In section I, I briefly examine three claims Aristotle makes regarding
virtue-acquisition in Physics vii.3, namely (1) that their acquisition is a case
of “perfection” (and not of generation or alteration), (2) that character
virtues belong to the category of “relation” (and not to that of quality), and
(3) that their acquisition “always occurs with” the alteration of the percep-
tive part of the soul. In section II, I offer suggestions for how this “natural
scientific” account of perfection can be applied to Aristotle’s characteriza-
tion of habituation in the ethical treatises and present a reconstruction of

5
EN ii.3, 1104b11–13 and x.9, 1179b24–26; Pol. viii.5, 1339a20–25 and 1339a41–b4. See Moss 2011 and
2012 on character virtue as a non-rational (but cognitive) state and on habituation as a non-rational
process of practical induction sufficient for the production of virtue.
6
EN i.13, 1102b25–31; EE ii.1, 1220a10–12 and 2, 1220b6–8.
7
EN x.9, 1179b4–31 and Pol. vii.13, 1332a38–b8.
8
Cf. Tracy 1969: 23: “[Aristotle] presents the physiological and psychological aspect of moral activity,
therefore, only in so far as that is necessary to the practical moralist”; Reeve 1992: 61: “What we
cannot do is explain why it is that changing desires, our own or other people’s, typically involves a lot
of repetition and really hard work . . . There is an explanation of these facts, no doubt, but ‘we must
hear it from the natural scientists’ (1147b8–9), not from philosophers”; and Hursthouse (2006: 112):
“But how do we get from the early tuning of toddlers’ passions to the enjoyment of fine actions? . . .
‘Habituation,’ Aristotle tells us, but, as everyone notes, he tells us little about what this involves.”
 Perfection and the physiology of habituation 227
the various types of change involved in habituation mentioned there. And
finally, in section III, I will use this evidence from the ethical treatises to
make sense of Aristotle’s second and third claim in Physics vii.3 and
provide an account of the changes that occur in the perceptive soul of
men during habituation.

I The acquisition of character virtue in Physics vii.3

In Physics vii.3 Aristotle provides one of the lengthiest discussions of the
physiology of virtue-acquisition in the Corpus.9 However, the main pur-
pose of the chapter is to defend a rather restricted notion of alteration,
which Aristotle needs in order to save his earlier argument against kinetic
infinity. This argument, which is developed in Physics vii.1–2, builds on
the premise that every change requires contact between the origin of
change and the entity undergoing that change, but in Physics vii.3
Aristotle brings up two examples of change – both within the category of
alteration – that appear to violate this requirement of contact: namely, the
change that items undergo when taking on or casting off figures or shapes,
and the change that items undergo when taking on or casting off disposi-
tions (hexeis). Aristotle’s strategy in dealing with these two alleged counter-
examples is, first, to redefine all alteration as change in perceptual qualities
only (245b3–5), thereby restricting alteration to changes in the “affective
qualities and affections, such as hotness and coldness” (which constitute
the third type of quality as defined in Cat. 8, 9a14–10a10);10 and, next, to
show that the two alleged counterexamples are not really cases of alteration
at all. The bulk of the chapter deals with the second example, i.e., the
acquisition and loss of dispositions of both body and soul (see 246a10–
248a9; the example of figures and shapes is discussed in 245b6–246a9), with
dispositions being immediately identified with virtues and vices (246a11–
12). The subsequent discussion then focuses on the acquisition and loss of
these virtues and vices (first those of body, next those of soul), and it is in
this context that Aristotle discusses the acquisition of character virtues. In
what follows, I discuss Aristotle’s account of the acquisition of virtues of

9
In writing this section I have benefitted much from Maso, Natali, and Seel 2012. I have also relied on
their edition of the Greek text of Physics vii.3, which has been transmitted in two different versions –
version α, which is the version printed by Ross 1950, and version β, which is the version known by
Simplicius, Philoponus, and Themistius. My discussion focuses on version α, but also draws from
version β where relevant.
10
Categories 8 discusses three more types of quality: (1) states and conditions; (2) natural (in)capacities;
and (3) shapes and forms.
228 mariska leunissen
the body (in 246b4–20) and that of the acquisition of character virtue (in
246b21–247a19) together, as the two accounts are very similar.11
As stated above, Aristotle begins his demonstration of the proposition
that the acquisition and loss of dispositions are not cases of alteration by
identifying these dispositions with virtues and vices, perhaps because, in
the case of living beings, every disposition makes its possessor either well or
ill disposed (Metaph. v.20, 1022b10–12).12 Based on this identification,
Aristotle argues that, since virtues and vices are not alterations but perfec-
tions, dispositions of the body and soul – and their acquisitions – are not
alterations (246a12–b3). Subsequently, he argues that virtues and vices
belong to the category of relation, thereby reaffirming again that their
acquisition or loss cannot be a change in the category of quality (246b3–
4). Aristotle then works out these two arguments with regard to virtues and
vices of body and of character, while – also in both cases – adding a third
argument about how, although their acquisition is not itself an alteration,
virtue-acquisition always occurs “with” an alteration of something else. Let
me discuss these three arguments in turn.
The first argument identifies virtues and their acquisition with “perfec-
tions,” while identifying vices and their acquisition with “departures”
(246a13–17):
But rather virtue is a kind of perfection (τελείωσίς τις) – for when it acquires
its own virtue (τὴν αὑτοῦ ἀρετήν), then we say that each thing is perfect, for
then it is most in accordance with [its] nature (τότε γὰρ ἔστι μάλιστα [τὸ]
κατὰ φύσιν), just as a circle is perfect, when it has become a circle in the
highest degree and when it is best – and vice is a destruction of this and a
departure [from this].
According to this account, which applies equally to virtues and vices of the
soul (246b21–247a3), acquiring a virtue is reaching a form of perfection
according to which its possessor now exhibits its own nature to the highest
degree or in the best way, while acquiring a vice is degenerating from one’s
nature. Aristotle uses the perfection of houses as an example (246a18–b1):
Evidently it would be “absurd” to call the coping and tiling of houses
alterations rather than perfections, and so the acquisition of virtues (or

11
The section concerning the intellectual virtues (247b1–248a6) falls outside the scope of this chapter;
on this, see De Haas 2012: 99–108 and Natali 2012: 109–117.
12
For the sake of this argument, Aristotle must assume that all dispositions – and not just those of body
and soul – are either virtues or vices. However, elsewhere he also refers to virtues as conditions
(diatheseis – a term that is perhaps exchangeable with hexis in discussions of virtue; see Bowin 2011:
147n.21), and possibly to dispositions of liveless things that are neither virtues nor vices (see e.g. DA
ii.7, 418b18–20 and iii.8, 432a5–6). On this, see further Coope 2012: 59–61.
 Perfection and the physiology of habituation 229
vices) by living beings should also be called perfections (or departures) and
not alterations.
What exactly is entailed by such a process of perfection is unclear, but
the house-example seems to associate perfection with a teleological pro-
cess of generation. Thus, perfection may either refer to (1) the final stage
of such a process, as suggested by “coping”: for, strictly speaking, houses
do not exist without the last (row of) brick(s) being in place yet, so that
coping could not be a case of alteration simply because the subject of
alteration does not yet exist. Or (2), as suggested by the comparison with
“tiling” (and by the example of the circle), perfection may refer to a
change of something that has already reached its final stage of generation
(e.g., the house and circle already exist, but perhaps imperfectly so), but
that is now made to express its own nature to the highest degree. The latter is
not a case of alteration, since the subject does not come to possess any
new attributes. This second reading, which I think is the more plausible
one, also fits best with Aristotle’s characterization of perfection in version
β (246b27–29):
For virtue is a kind of perfection – for each thing is then most perfect
(μάλιστα τέλειόν), when it reaches its proper virtue (τῆς οἰκείας ἀρετῆς) and
is most in accordance with nature, just as a circle is then most in accordance
with nature, when it is most a circle.
In this version, it is clearly not the generation of the circle that is at stake,
but rather the degree to which it exhibits its own nature: by being a circle
most, it is said to be most perfect and to reach its “proper virtue.” Similarly,
in Aristotle’s definition of perfection in Metaphysics v (16, 1021b14–1022a3),
a being – such as a physician or a fluteplayer – is called perfect when “with
respect to its kind it lacks nothing of its proper virtue” (b16–17: ὅταν κατὰ
τὸ εἶδος τῆς οἰκείας ἀρετῆς μηθὲν ἐλλείπωσιν) and when it “lacks nothing
with regard to goodness and cannot be excelled and has nothing that is
found outside of it” (b31–32: τῷ κατὰ τὸ εὖ μηδὲν ἐλλείπειν μηδ’ ἔχειν
ὑπερβολὴν μηδὲ ἔξω τι λαβεῖν). Again, what is at stake in this definition is
not whether something has already realized its own nature or not (the
examples refer to people who are already practitioners of their crafts, not
students in training), but rather whether it has realized it well or badly,
with perfection picking out the condition of that which could not have
realized its nature better – it lacks nothing of its proper virtue and cannot be
excelled. Hence, something is perfect when it does not only have every-
thing that is necessary for it to be what it is (that is, for it to be a normal
functioning member of its class), but when it also possesses those things in
230 mariska leunissen
the best possible condition, such that their interplay and combination
make it unsurpassable in excellence.13
Aristotle’s second argument places the virtues and vices of body and
character in the category of relation (246b3–4 and 247a1–4; and not in the
category of quality, as is suggested in Cat. 8, 8b27–29),14 as they all satisfy
the following two criteria (246b8–9):15 first, “each . . . is in [the category] of
relation (ἑκάστη . . . ἐστι τῷ πρός τι),” and second, “[each] puts what has it
in a good or bad condition (εὖ ἢ κακῶς διατίθησι) regarding its proper
affections,” where “proper” is understood as picking out those affections
that naturally produce or destroy that which is being affected. And, as there
are no changes of relations properly speaking (instead, the items that are
part of the relation change, and thereby change the relation between them,
but this is not a change of the relation itself), the acquisition or loss of these
virtues or vices are not alterations (246b10–14).
Unfortunately, Aristotle does not specify what components are consti-
tutive of the kind of relation that qualifies as character virtue. Perhaps the
analogy with the relational nature of health and fitness (246b4–6) is helpful
here: For, according to Aristotle, the latter are placed “in the mixture and
due proportion of hot and cold [elements] (ἐν κράσει καὶ συμμετρίᾳ
θερμῶν καὶ ψυχρῶν),” which exist either in relation to themselves, intern-
ally in the body, or externally in relation to the mixture of the environ-
ment.16 So, apparently, bodily virtues and vices are either nothing more
than elements that are in some way related to something (as is suggested in
version β, 246b21–22: “just as health is a kind of symmetry of hot and cold
[elements], either with what is internal or in relation to the environment”),
or they are ontologically dependent on those relations without being iden-
tical to them.17 In either case, it seems that, if character virtues are relational
in a similar or analogous way to the way bodily virtues are relational (as is
indicated by ὁμοίως in 246b20), then perhaps character virtues, too, should

13
Cf. Coope 2012: 69–71 and Strohl 2011: 284–285.
14
The Categories lists the genus of dispositions and virtues in the category of relation, but the particular
dispositions or virtues as qualities: see Cat. 7, 6b15–16 and 8, 11a20–36.
15
The Greek leaves open the possibility that only one of the two criteria needs to be satisfied in order to
classify as a relation, but other evidence suggests that health at least requires both: see Freudenthal
1995: 13.
16
See also Phys. iv.3, 210b22–27:
What Zeno is puzzled about, that if there is a place, it must be in something, is not difficult to
solve. For nothing prevents the first place from being in something else, however not as in that
place, but as health is in the hot elements as a state (ὥσπερ ἡ μὲν ὑγίεια ἐν τοῖς θερμοῖς ὡς ἕξις),
and [as] heat [is] in the body as an affection.
17
Bodnár 2012: 75–77.
 Perfection and the physiology of habituation 231
be “placed in” or even identified with a mixture and due proportion – not
of bodily elements, but rather of “elements” of the soul.
The third argument for why the acquisition and loss of virtues or vices
are not alterations claims that this miscategorization is due to the fact that
their generation and perishing necessarily always involve the alteration of
something else. For bodily virtues and vices are said to come to be when the
bodily elements in which these states are placed are altered (246b14–20).
The account concerning character virtues is more complex, so let me quote
it in full (247a6–19):
But it is necessary that they come to be when the perceptive part [of the soul]
is altered (ἀλλοιουμένου τοῦ αἰσθητικοῦ μέρους). And it will be altered by
perceptibles. For every character virtue concerns pleasures and pains of the
body, and these [occur] in acting or in remembering or in expecting. The
ones that occur in action are in accordance with perception, such that they
are being moved by something perceptible; the ones that occur in memory
and in expectation [come] from this: for either the ones who remember what
they have experienced or the ones who expect what they will [experience]
feel pleasure. Therefore it is necessary that every such pleasure comes to be
through perceptibles. Since when pleasure and pain come to be [in the soul]
also vice and virtue come to be (for they [i.e., vice and virtue] concern them
[i.e., pleasure and pain]), and pleasures and pains [are] alterations of the
perceptive part, it is clear that when one undergoes some alteration one
necessarily also loses or acquires those [vices and virtues]. Therefore there is
a coming to be of them with alteration, but they themselves are not
alterations.
In short, Aristotle argues that when the perceptive part of the soul – in
which, presumably, the character virtues are present as relations – is
affected by pleasures and pains (which are themselves produced by sense
perception) and is thereby altered,18 character virtues or vices are acquired
or lost.
As this overview brings out, Aristotle’s remarks about the nature of
virtue-acquistion in Physics vii.3 are very condensed and leave it unclear
as to what extent they apply to his treatment of habituation in the ethical
treatises, where – at least prima facie – virtue-acquisition appears to be a
kind of qualitative change and the language of perfection does not play any
substantial role. In what follows, I will analyze Aristotle’s treatment of

18
This characterization of pleasures as alterations of the perceptive soul seems to clash with Aristotle’s
account of pleasures as perfections of activities in EN x.4, 174b31–33: on this, see Wardy 1990: 224–
225, who suggests that the former is a simplification of a more complex theory of pleasure that is not
relevant in Physics vii.3.
232 mariska leunissen
perfection and the physiology of habituation in the ethical treatises, and
show that, for the most part, it is quite congenial to what has been deemed
to be an “excessively physiological”19 account in Physics vii.3, and that
these materials together can be used to offer a richer picture of what
happens physiologically when men habituate themselves.

II Perfection and the physiology of habituation

in the ethical treatises
In the ethical treatises, the view that the acquisition of character virtues
through habituation is a kind of perfection that results in men who are not
only biologically complete and functional, but who also have reached their
proper virtue and are therefore most in accordance with their human
nature, is never developed in much detail.20 However, as I will argue, it
is nonetheless present and fitting with the account in Physics vii.3, and,
furthermore, its conceptualization in the ethical treatises provides addi-
tional insights into the physiological process of habituation. Since the
Nicomachean Ethics presupposes the student of ethics to have had the
kind of upbringing that is outlined in the Politics, I start my account
with the latter.
In the Politics, the view that habituation is a form of perfecting men
surfaces in Aristotle’s discussion of the tasks of lawgivers in the ideal city.
One of the most important of these tasks is the proper habituation and
education of children who would otherwise remain imperfect both in their
deliberative capacity and in the development of their character virtues (Pol.
i.13, 1260a13–14 and 1260a31–33; a31: ὁ παῖς ἀτελής) and would not be able
to contribute to the virtuousness of the city. Accordingly, Aristotle claims
that the educational program should follow the natural division of the
development of boys rather than the traditional division into seven-year
periods, for “it is the purpose of every craft and [every] education to
complete what is lacking in nature” (Pol. vii.17, 1336b40–1337a3; a2–3: τὸ
προσλεῖπον . . . τῆς φύσεως ἀναπληροῦν).21 This characterization of edu-
cation suggests that Aristotle sees the role of political science as a form of
helping men realize their human nature (“completing,” I take it, is a form

19
See Ross 1936: 676; Viano 2012: 89–93; and Wardy 1990: 223–227.
20
The conceptualization of habituation as a form of perfection is, however, worked out in detail in
Aquinas, who adopted it from Aristotle: see Irwin 2006: 332–335.
21
Cf. EE vii.2, 1237a2–3 and Aristotle’s eugenic program in Pol. vii.16 (especially at 1335a11–17 and
b29–31), which is designed to help avoid the natural occurrence of imperfections in offspring.
 Perfection and the physiology of habituation 233
of “perfecting”: see Protrepticus, Iambl. ix 49.28–50.12),22 since this cannot
be achieved by nature alone.23 The aim of habituation, then, is to complete
nature by instilling the proper virtues in men and thereby producing a kind
of “second nature” in them that replaces the imperfect and unstable one
they have from birth (see, e.g., Rh. i.11, 1370a5–9, Mem. 2, 452a27–30, and
EN vii.10, 1152a29–33). Aristotle does not specify the physiological changes
produced in men during this process of perfection brought about by
habituation, but his discussion of the educational program of boys in the
ideal city includes several components that imply some kind of change:
conditioning, imitations through play, and musical education. Let me
discuss these in turn.
First, since habituation should track the order of the natural develop-
ment of the bodies and soul-parts of boys (see Pol. vii.15, 1334b10–29),
Aristotle stresses the importance of bodily exercise and conditioning
straight from birth. By exposing newborns up until the age of 5 to
particular environments, diets, and motions, lawgivers primarily condition
their bodies for health and (military) strength, and prevent laziness or
deformity (Pol. vii.17, 1336a25–39 with Pol. viii.2, 1337b5–15), thereby
instilling the bodily virtues – a process that is also emphasized during the
ages of 7 to 17. However, these practices also change the physiology of the
blood to promote natural courage (consuming milk, for instance, produces
“a military disposition”: see Pol. vii.15, 1336a3–8), and possibly also alter
the appetitive capacities of the soul by conditioning boys to certain
behaviors and affections (e.g., so that they no longer experience cold
temperatures as painful: see Pol. vii.15, 1336a12–18) – a practice that is
also common among some animals.24
A second process involved in the early habituation of boys within the
household is that of imitation through play. Children should not only be
exposed to morally good examples (whether in the form of behavior,
speech, or artistic representation) while being protected from bad ones,
but they also need to be encouraged to imitate those good examples. For
instance, children should perform those actions in play that are imitations

22
For nature does not imitate craft, but it (i.e. craft) nature, and it exists to help also complete those
things nature leaves out (τὰ παραλειπόμενα τῆς φύσεως ἀναπληροῦν). For some things nature itself
seems capable of bringing to perfection (ἐπιτελεῖν) by itself without actually needing any help, but
others it [brings to perfection] with difficulty or is completely unable to do.
Transl. Johnson and Hutchinson 2005: 259; slightly modified.
23
Reeve 2012: 104. Cf. Phys. ii.3, 194a21–27 and 8, 199a15–17.
24
Pol. vii.13, 1332b3–4. On animals habituating their young, see HA vi.12, 567a5–7, ix.5, 611a19–21;
and ix.7, 612b29–31; on humans habituating animals, see HA vi.19, 573b25–27; vi.21, 575a33–b2;
and ix.3, 610b33–611a2.
234 mariska leunissen
(μιμήσεις) of the serious actions they will perform later in life (Pol. vii.17,
1336a33–34). Because imitations are naturally pleasant to humans, imita-
tion is a natural mode of learning for them (Poet. I.4, 1448b5–24; cf. Rh.
i.11, 1371b5–10), and it is especially suited for the habituation of children,
who easily translate what they see and hear into action (Pol. vii.17, 1336b2–
8). For through imitative play children are encouraged to reinvoke their
perceptions of the actual goods presented to them via imagination – a
practice that, perhaps, sharpens their discriminative ability to perceive the
actual good as good25 – and to take pleasure in the performance of those
good actions: Imitation thus shapes the preferences of boys by making
them enjoy – and love (see: Pol. vii.17, 1336b33) – the performance of
virtuous actions.
The third process involved in habituation is (presumably)26 part of the
education of boys from age 7 to 14: In addition to performing gymnastics
and being educated in letters and drawing, students are to receive musical
education. Aristotle argues that education in music prepares boys for the
proper enjoyment of leisure later in life, once they have become morally
perfected men,27 but he also suggests – and later endorses that this is indeed
the case (Pol. viii.5, 1340a1–12) – that music is productive of character
virtues (Pol. viii.5, 1339a20–25):
Or should we rather think that music contributes something to virtue, on
the grounds that – just as gymnastics prepares our body to be of a certain
quality – so too, music is capable of making our character of a certain quality
(τὸ ἦθος ποιόν τι ποιεῖν), by habituating us to be able to enjoy things in the
right way (ἐθίζουσαν δύνασθαι χαίρειν ὀρθῶς).
Apparently, education in music – which involves both being trained as
audience and, up to a point, also as a performer (Pol. viii.6, passim) – is
capable of producing positive alterations in human character traits (cf. Pol.
viii.5, 1339a41–42: εἰ δύναται τὰ ἤθη βελτίω ποιεῖν) and thereby helps
produce character virtue. And, Aristotle claims, music not only makes men
able to enjoy things in the right way, it also makes them able to judge in the
right way, where the objects of enjoyments and judgments are not just
music itself, but also “appropriate characters and good actions” (Pol. viii.5,
1339a41–b4 and 1340a14–18; cf. Pol. viii.6, 1340b38–39). In this way,
habituation through music not only alters the – presumably – natural

25
Perception is a causal factor involved in action (EN vi.2, 1139a17–18), and virtue requires one to be
able to perceive the actual good as being good for oneself (EN iii.4, 1113a29–33 and 5, 1114a31–b20).
On practical perception, see further Moss 2011: 251–254 and Reeve 2012: 40–50.
26
Reeve 2012: 254. 27 See Pol. viii.3, 1337b25–1338a32 and 5, 1339a32–33, with 1339b15–25.
 Perfection and the physiology of habituation 235
character traits in the direction of natural virtue, but also the appetitive and
the (practical) perceptive capacities of the soul.
Aristotle explains these changes produced by music physiologically in
terms of the special representational nature of music: “by hearing
imitations . . . all people become similarly affected (Pol. viii.5, 1340a12–
14: ἀκροώμενοι τῶν μιμήσεων γίγνονται πάντες συμπαθεῖς)” and “change
with respect to their soul (1340a22–23: μεταβάλλομεν . . . τὴν ψυχὴν),”
implying that the affection that is expressed in music – when experienced –
will produce the corresponding affection (or disposition: 1340a41) in the
listener.28 This in turn happens because sounds possess likenesses to
character traits (1340a29–30: ὁμοίωμα τοῖς ἤθεσιν), with melodies repre-
senting different kinds of natural character traits and rhythms representing
the character of action (1340a38–b10), and humans have a natural affinity
to both of these (1340b16–18; cf. Poet. 4, 1448b20–21). Or, as the author of
the Problems explains it, sounds consist of two sets of motions, one moving
the sense organs, the other moving us in accordance with the character they
resemble (Pr. xix.27, 919b26–37; cf. Pr. xix.29, 920a3–7 and xix.48,
922b22–23). Thus, by exposing them to the right kinds of music,29 boys
regularly realize their natural character traits in the best way possible and
are habituated to experience the kind of emotions and pleasures and pains
that one ought to feel when faced with the real objects and situations (Pol.
viii.5, 1340a23–28). And boys can do so without actually – or yet – having
to perform virtuous actions, the qualities of which would have the same
effect, as Aristotle claims in the Nicomachean Ethics.
In the Nicomachean Ethics, Aristotle uses the term “perfection” explicitly
and in the relevant way in his earliest discussion of what the character
virtues are and how they are acquired in Nichomachean Ethics ii.1.30 There,
after rejecting the view that character virtues come to be by nature, he states
that humans are by nature fit to receive the character virtues, but that “[we]
are perfected through habits” (EN ii.1, 1103a23–26; 25–26: τελειουμένοις δὲ

28
Especially in melody and rhythms, Aristotle says, there is “the greatest likeness to the true natures of
anger and gentleness, and in addition of courage and temperance and of all their opposites and of
other character traits as well” (Pol. viii.5, 1340a18–22).
29
Not all music is appropriate for educational purposes: one should only use melodies that “pertain
most to character (ταῖς ἠθικωτάταις)” and do so in a good way (Pol. viii.7, 1341b38–1342a3), such as
the Dorian melody (Pol. viii.7, 1342b13–17).
30
Although Aristotle does not specify what he means by perfection, he clearly understands it as a
perfection of one’s nature (EN vii.12, 1153a11–12: τῶν εἰς τὴν τελέωσιν ἀγομένων τῆς φύσεως), as
acquiring one’s proper virtues (EN ii.5, 1105b29–31, 6, 1106a15–24; and, perhaps, x.5, 1176a27), and as
resulting in unqualifiedly good men (EE vii.2, 1237a29–30).
236 mariska leunissen
διὰ τοῦ ἔθους).31 Aristotle offers three arguments for the claim that perfec-
tion occurs through habit and not through nature, and because they are
informative for his conceptualization of the physiology of habituation, I
discuss each of them in turn.
First, Aristotle claims that nothing that behaves in one way by nature
can be habituated to act contrary to its nature (EN ii.1, 1103a19–23). Stones,
for instance, have a natural tendency to move downwards, and no amount
of being thrown upwards changes this tendency. However, when it comes
to character, habituation can produce contrary dispositions (see EE ii.5,
1222a36–42, Pol. vii.13, 1332a40–b3, and EN ii.9, 1109b1–7): The natural
character traits we have from birth (which are “natural capacities”) might
push us into one direction, but by “pulling ourselves into the opposite
direction” we can “straighten ourselves out,” “change [those traits] towards
the better,” and attain the mean disposition of character that constitutes
virtue. This kind of change is possible because, physiologically speaking,
humans do not simply behave in the way they do by nature, but “by desire
and choice,” and because the character traits they have from birth –
although natural in origin – are “rational capacities,” which (as opposed
to irrational capacities, which can be found in any natural entity but can
only be realized in one direction) belong uniquely to ensouled beings and
can be realized in opposite directions (Metaph. ix.2, 1046a36–b7 and ix.5,
1047b35–1048a5). Specifically, ensouled beings such as humans can them-
selves determine – via desire or choice – whether to realize their rational
capacities in accordance with their own nature (e.g., actualizing medical
science for the sake of producing health, which is the proper realization of
medical science) or to produce their privations (e.g., sickness, the acciden-
tal realization of medical science: Metaph. ix.2, 1046b7–24 and ix.5,
1048a10–11). On this account, the natural virtues (but not yet the character
virtues, since these are not straightforwardly realizations of natural capa-
cities) are the proper realizations of the natural character traits humans
have from birth,32 which have been stabilized in their best condition due to
men frequently producing a second actuality of these capacities in the
direction of virtue. The change from natural character (a dualizing

31
Acquiring character virtue with practical wisdom is “merely” the secondary of two ways in which
humans reach perfection: “the best and most complete virtue” identified with the human good in
Aristotle’s function-argument (EN i.7, 1098a16–18) is theoretical wisdom, achieved through a life of
contemplation (EN x.7–8).
32
Cf. EN vi.13, 1144b4–9 and perhaps EE iii.7, 1234a24–30: Character traits tend by nature towards
natural virtue, which suggests that virtue is the per se realization of the natural capacities for
character.
 Perfection and the physiology of habituation 237
capacity) to natural virtue (a good disposition), then, is a form of “special
alteration,” which involves the preservation (σωτηρία) of an already exist-
ing capacity.33
That character virtue comes to be in a non-natural way is also the
conclusion of Aristotle’s second argument, which points to a difference
in temporal priority between capacities that develop naturally and those
that develop non-naturally (EN ii.1, 1103a26–34). For while capacities that
develop by nature – such as perception – are first possessed as capacities and
are later displayed as activities, virtue is acquired like crafts, where the
capacity for doing something is acquired by first repeatedly doing it.34
Physiologically speaking, this distinction tracks again the difference
between irrational and rational capacities (Metaph. ix.5, 1047b33–34):
Of all the capacities, some are innate, such as the perceptive ones, others are
by habit, such as the [capacity] to play the flute, others are by learning, such
as the [capacity] of the crafts, and for some it is necessary to possess them by
having previously activated them (προενεργήσαντας) – namely as many as
are by habit and by learning – but it is not necessary for those that are not
like this and that imply being acted upon.
Natural capacities, such as perception, are irrational and are acquired via
the substantial change that takes place during embryology (i.e. the capacity
to perceive comes to be at the same time as the perceptive organs). This
means that we possess such capacities as first actuality from birth (DA ii.5,
417b16–18; cf. GA iv.1, 766a5–10), while their use or activity – as “second
actuality” – is a “special” type of alteration. And since these capacities are
irrational, their activity does not depend on desire or choice, but simply on
the presence of objects of perception that act on the sense organs (Metaph.
ix.5, 1047b35–1048a24). However, acquiring character virtue is, as Aristotle
puts it, just like acquiring the rational capacity to play the flute, which
happens by “earlier engaging in their activities” (EN ii.1, 1103a31:
ἐνεργήσαντες πρότερον) and by doing this frequently (Rh. i.10, 1369b6–
7). In other words, the virtuous activity that constitutes the second actu-
ality produces the virtuous disposition that constitutes the first actuality.
The kind of “earlier activation” that is required for the production of
virtues is spelled out more clearly in Aristotle’s third argument against the
naturalness of moral development (EN ii.1, 1103b6–25): Since activities are
the causes of and means for the production and destruction of virtues (as

33
And not of proper alteration, which involves the destruction of one affection while another takes its
place: DA ii.5, 417b2–19. Cf. Bowin 2011: 138–161 and Polansky 2007: 234–240.
34
Cf. Pol. viii.1, 1337a18–21.
238 mariska leunissen
they are also in crafts), teachers are required to ensure that men perform the
activities that are good rather than bad. For if there were no such causal
relation between activity and virtue, and if men were simply born as good
or bad performers of those activities, teaching would make no difference
and men would have virtues or vices from birth. Aristotle summarizes this
causal relationship as follows (1103b21–23):
In one word, dispositions come to be from like activities, wherefore it is
necessary to display a certain quality of activities (τὰς ἐνεργείας ποιὰς). For
the dispositions follow the differences between these.35
In short, it is the quality of the activity that somehow produces a qualita-
tively identical disposition in the soul. In what follows, Aristotle clarifies
what qualities one’s actions should have: Actions should be in accordance
with correct reason (EN ii.2, 1103b31–32) and should hold the mean
between excess and deficiency in the pursuit of pleasure and pain, and
in the experience of emotions (EN ii.2, 1104a11–27; ii.3, 1104b12–16;
1104b25–26; and 1105a6–7).36 Hence, by performing virtuous actions
men experience the qualitative states of feelings and emotions – presum-
ably as mediated by their natural character traits – that are constitutive of
virtue, which, if repeated often enough, will produce a qualitatively
identical disposition. The physiological account of how one becomes
virtuous by listening to the appropriate kinds of music, as suggested in
the Politics, and the account of how one becomes virtuous by performing
virtuous actions are thus very similar. For both cause the natural character
traits, the appetitive capacities, and the perceptive capacities to be repeat-
edly affected and thereby altered in the way that is appropriate for virtue,
such that feeling pleasure, pain, and emotions correctly becomes second
nature to that person, and he becomes prone to responding to perceptual
inputs in the way that is characteristic of the person with a virtuous
disposition.37 Because of this, there need not necessarily be any qualitative
difference in the appetitive and perceptive capacities of the soul in the
performance of a virtuous action by someone who is habituating himself
and by the virtuous person, although the latter – having also acquired

35
Cf. EN ii.2, 1103b29–31: “it is also necessary to investigate the issues concerning actions, namely how
they should be done: for they are determinative – κύριαι – also of the qualities – ποιὰς – of the
dispositions that come to be.”
36
Note that actions with these qualities not only produce virtues but also constitute the proper
activities of the virtues once acquired (EN ii.2, 1104a27–b3).
37
Cf. Cat. 10, 13a23–31 and EE ii.2, 1220a39–b3, where Aristotle characterizes habituation as being
changed frequently in a certain way, such that the soul is eventually capable of acting in that way.
 Perfection and the physiology of habituation 239
practical wisdom38 – always performs virtuous actions knowingly (even if
this knowledge does not count for much: EN ii.4, 1105b1–2), deliberately
chooses to perform them because of themselves, and performs them from
a firm and unchanging disposition (EN ii.4, 1105a31–33).39
This evidence from the ethical treatises concerning habituation is not
exhaustive; however, I believe that it is sufficient to flesh out Aristotle’s
claims about virtue-acquisition in Physics vii.3 and thereby to present an
Aristotelian account of the physiology of habituation.

III A physiological account of habituation built on Physics vii.3

While the ethical treatises do not often characterize habituation in terms of
perfection, they are explicit about the fact that habituation is a process
through which men realize their human nature to the fullest and are
completed in a way that cannot be achieved by nature alone, but that
requires the help of political science. At least in this respect, the treatment
of habituation in the ethical treatises is consistent with the one in Physics
vii.3. However, Aristotle makes two further claims about the acquisition of
character virtue in the Physics that require clarification: namely, that virtue-
acquisition is not a case of alteration, because virtues are relations, not
qualities, and that, although virtue-acquisition is not itself a case of altera-
tion, virtue comes to be “when the perceptive part [of the soul] is altered.”
In the ethical treatises, Aristotle never refers to character virtues as
relations. On the contrary, Aristotle lists virtues as examples of qualities
in EN i.6, 1096a25, and although he discusses the status of virtues as
dispositions at length in EN ii.5, he does not distinguish there between
dispositions as first actualities of natural capacities and dispositions as
relations among certain elements of the soul. However, perhaps the fact
that Aristotle is explicit in the ethical treatises that virtues are not straight-
forwardly realizations of the natural capacities for character that humans
have from birth is relevant here: Although becoming virtuous involves
frequent realizations of these natural qualities in a good direction, such
that they become altered for the better, character virtue itself cannot be
38
Practical wisdom is the proper virtue of practical reason that is acquired via moral education and
indirectly via habituation, as habituation prepares the ground for moral education: For instance,
Aristotle associates habituation with a form of “fine-tuning” and inundation of souls (EN x.9,
1179b16–18 and 1179b24–26; cf. Pr. xxii.11, 931a6–15) that makes men receptive to moral arguments
(cf. EN i.13, 1102b26–1103a1 and x.9, 1180a7–11). Cf. Burnyeat 1980: 86 and Moss 2011: 239.
39
These latter two qualities are essential for what constitutes full virtue (EN ii.4, 1105b3–4) and are also
acquired through habituation (1105b4–5: “[they] arise from often performing just and temperate
[actions]”). See Moss 2011: 210 for arguments against intellectualist readings of EN ii.4, 1105a31–33.
240 mariska leunissen
identified with the resulting, stabilized dispositions, for these amount at
most to multiple, individual natural virtues. Additionally, even though the
ethical treatises never mention the idea that virtues are themselves relations
between (internal) elements of the soul, they do mention how having
virtues puts one in a good relation with regard to affections (see e.g. EN
ii.5, 1105b25–26 and EE ii.1, 1220a33–34), thereby satisfying the second
criterion for relations as specified in the Physics (vii.3, 246b9 and 247a3–4).
Furthermore, although the ethical treatises never define habituation as a
process of perfection that always takes place with the alteration of the
perceptive part of the soul, we saw above that they regularly associate
habituation with alterations of (natural) character40 and, more generally,
with changes or movements of the soul.41 And it is clear that not the whole
soul is being changed: for, in his ethics, Aristotle works with a tripartite
division of the soul, according to which the nutritive part is always set aside
as having no share in human virtue (EN i.13, 1102a32–b12 and EE ii.1,
1219b20–21; b26–39), while the two remaining parts each receive their own,
exclusive virtue. The part that has reason receives the intellectual virtues,
while the part that is receptive to reason receives virtue of character (EN
i.13, 1103a3–10 and EE ii.1, 1220a8–11; cf. Pol. vii.14, 1333a16–19). However,
with one exception, Aristotle never calls this part the perceptive one;
instead he calls it “that which is capable of appetite and, in general, of
desire” (EN i.13, 1102b30; cf. EE ii.4, 1221b27–32, Pol. vii.15, 1334b10–29,
and MM i.5, 1185b1–12) or “the part that concerns character” (as opposed to
the one concerning belief: EN vi.13, 1144b14–16). Only once does Aristotle
contrast the perceptive and appetitive capacities of the soul together with
nutrition (EE ii.1, 1219b21–23), thereby suggesting that both are involved
in character virtue. But this fits with the depiction of habituation in the
Politics, according to which habituation affects and alters the natural
character traits as well as the appetitive and perceptive capacities of boys,
such that they are capable of seeing the actual good as good for them and to
take pleasure in it.
In fact, according to Aristotle’s exposition in On the Soul, the appetitive
and perceptive capacities are both components of the perceptive part of the
soul. For in this treatise (see DA ii.2–4, especially at 413b11–16, and DA
iii.10, 433b1–4), Aristotle divides the human soul into a nutritive part, a
perceptive part, and a thinking part, while characterizing all the remaining

40
See e.g. Pol. vii.13, 1332a40–b3 and viii.5, 1339a23–24.
41
See e.g. Pol. viii.5, 1340a12–14 and 1340a22–23; EE ii.1, 1220a29–30 and 2, 1220a39–b3.
 Perfection and the physiology of habituation 241

Perceptive part
perception and feeling of desire/
imagination pleasure/pain appetite

natural heating/
character/virtue cooling of
blood action
expanding/
Rational part

contracting
practical wish (=outcome of pneuma
reason deliberation)

Figure 12.1 Schematic representation of the rational and perceptive part of the soul
as involved in the production of human action.

capacities of the soul as belonging to one or more of these parts.42 As for the
perceptive part of the soul, it includes (a) the five perceptive capacities (touch,
taste, vision, hearing, and smell), (b) imagination, (c) desire and appetite, and
(d) pleasure and pain.43 Each of these sets of capacities is causally dependent
on the presence of the others, and together they form one functionally
complex part of the soul (see my schematic representation in Figure 12.1).
Now, if we take seriously Aristotle’s claim in the Physics that virtue-
acquisition involves the alteration of the perceptive part of the soul, this
must mean that all or most of the capacities constitutive of this part are
altered via habituation. And, as we saw above, the ethical treatises provide
evidence that this is indeed the case for at least three of these capacities
(there is no evidence for habituation directly changing imagination, but
since imagination in humans is tightly connected to the operation of the
perceptive capacity,44 these capacities are likely altered together).
First, habituation of the perceptive capacities involves a cognitive training
through which one acquires the ability to recognize and judge correctly
what is good and bad regarding moral phenomena,45 but it also involves an
alteration of the perceptive capacities themselves such that they are in a
perfect mean46 and are able to see correctly and actualize the appropriate
feelings of pleasure and pain. For, according to Aristotle, every perception

42
Corcilius and Gregoric 2010. These three soul-parts also figure in the function-argument in EN i.7,
1097b33–1098a7, which supports my suggestion that the political and natural sciences work with the
same triparte division of the soul.
43
See DA ii.2, 413b22–24 and 3, 414b1–16; cf. PA ii.17, 661a3–8.
44
See DA iii.3, 428b10–16, 429a1–2; and iii.11, 433b31–434a9.
45
See Moss 2012: passim on what she calls practical induction. 46 Polansky 2007: 333.
242 mariska leunissen
is accompanied by a feeling of pleasure or pain and, vice versa, every feeling
of pleasure or pain depends causally on the activity of perceptual discrimi-
nation (objects of perception do not need to be perceived as good or bad to
be followed by feelings of pleasure or pain; instead, perceiving an object
simply is also to feel pleasure or pain with regard to that object),47 such that
an appropriate feeling of pleasure and pain depends on the correct activity
of perceptual discrimination.
Second, one’s capacity to feel pleasure and pain is altered by the frequency
and quality of the feelings experienced in response to one’s perceptive
activity: For feeling pleasure and pain in the right way, that is, in a way that
tracks the actual good or bad of what one perceives, is a sign of character
virtue (EN ii.3, 1104b3–9; cf. EE vii.2, 1237a1–7 and EN iii.12, 1119a11–17).
Since the perceptional mean is naturally biased towards pleasure (EN ii.9,
1109b7–9), habituation is needed to “make one feel pleasure and pain by
the things one should” (EN ii.3, 1104b12–13; cf. EN x.9, 1179b24–26 and
Pol. viii.5, 1339a20–25 and 1339a41–b4).
Third, since the kind of perception that involves feelings of pleasure or
pain is causally followed by an appetite or desire to pursue or avoid the
object of perception (DA iii.7, 431a8–14), and since this desire – in
humans – is a co-cause of action, the capacities for appetites or desires need
to be appropriately calibrated as well, so as to respond correctly to feelings
of pleasure or pain. Physiologically speaking, desires to avoid or pursue are
heatings or coolings of the blood: That is, desires involve alterations of the
blood that are themselves the result of some kind of alteration involved in
perception or imagination (MA 6, 701a4–6 and 7, 701b13–23). These
heatings or coolings of blood result in the contraction or expansion of
the pneuma that is present in the blood, and this, then, results in locomo-
tion or action.48 Regarding this capacity, too, Aristotle emphasizes that the
virtuous person must have the appropriate appetites and desires (see e.g.
EN iii.1, 1111a30–31 and iii.11, 1118b15–21), and that their training is one of
the objectives of habituation (NE I.3, 1095a2–13). In the virtuous person,
then, regardless of whether his virtue is habituated or natural (EN vii.8,
1151a15–19), the actual good will appear as what is good for himself, and it
will be pleasant and will produce a desire to pursue it.49

47
See Corcilius 2011 on the relation between perception, non-rational desire, and pleasure and pain.
48
See MA 7, 701b1–32 and 8, 701b33–702a21; and EN vii.3, 1147a34–35: “for appetite leads [him]: for it
can move each of the [bodily] parts.” Note that wish, the rational desire produced by practical
reason as the outcome of deliberation, also causes such heatings and coolings of the blood, and that
wish is by nature the more authoritative of the two desires (DA iii.11, 434a12–15).
49
In men with full virtue, this desire or appetite will be in harmony with wish.
 Perfection and the physiology of habituation 243
In addition to these three sets of capacities belonging to the perceptive
part of the soul, the ethical treatises mention a fourth capacity – not
mentioned in On the Soul – that is altered by habituation: namely, the
natural capacities for character. These, too, I submit, belong to the per-
ceptive part of the soul, and their role in Aristotle’s moral psychology is
that of a mediator between the activity of perception and the feelings of
pleasure and pain: That is, unless one’s natural character traits are them-
selves in a mean and “good,” they either increase or decrease the amount of
pleasure or pain that is experienced concomitant to the perceptual activity,
regardless of the latter’s accuracy (e.g. cowardice increases the amount of
pain, and hence of chilling, experienced at the perception of real or
apparent danger, beyond what is appropriate, and results in excessive flight
or inaction). In this way, character is also responsible for how we perceive
things (cf. EN iii.5, 1114a31–b3 and Insomn. 2, 460a32–b11), and that is why
it should be changed towards the better, and why Aristotle claims that
having “good” natural character traits is advantageous in becoming
virtuous.
We can now see what Aristotle means in Physics vii.3 when he claims
that character virtue comes to be when the perceptive part of the soul is
altered through perceptibles. For, as we saw, boys should be habituated by
being exposed to morally good examples, by imitating those examples in
play, and by listening to character-shaping music, all of which affects and
thereby alters the perceptive capacities and influences the direction in
which the natural character traits become stabilized. It is also clear what
Aristotle has in mind when he says that pleasures and pains are alterations
of the perceptive part of the soul: Perceptions – whether immediate
perceptions or memories or expectations (the latter of which operate
through “the likenesses” to affections: MA 8, 702a5–7) – are pleasant or
painful, and since feeling pleasure and pain involves a heating or cooling of
the blood that produces a desire to pursue or avoid, this is a kind of
affection of the perceptive part of the soul (and, of course, of the sense
organs with which the perceptive soul forms a unity). And finally, we can
explain why Aristotle identifies the coming to be of pleasure and pain with
the coming to be of virtues and vices of character, for it is through the pains
and pleasures experienced via perceptions that the capacities of the percep-
tive part of the soul are altered.
In conclusion, assuming that Aristotle’s account of virtue-acquisition in
Physics vii.3 can indeed be used to (re-)construct a physiological theory of
habituation, then we may infer that when these capacities of the perceptive
part of the soul are all in their best or worst condition, both internally and
244 mariska leunissen
externally, in relation to each other, that that is when virtue or vice is present.
For, if character virtue is the proportionate relation between the capacities
that are constitutive of the perceptive part of the soul, then it will be
present when those capacities have been altered so as to be stably in the best
condition possible and are related to each other in the appropriate way.
Their coming to be is not itself a generation or an alteration, but the
perfection that arises when all the constitutive elements of the perceptive
part of the soul are in the best condition and are proportionately organized
regarding each other, just as health50 is the perfection that arises from such
an organization of bodily elements.51 If this account is correct, then perhaps
it also helps explain what Aristotle means by the unity-thesis of virtue (EN
vi.13, 1144b35–1145a2). For character virtue is only possible when all of the
constitutive capacities of the perceptive soul are in their best condition and
are proportionally organized, and the resulting virtuous disposition will
constitute one, unified, but distinct psychological state.

50
Cf. Cael. i.3, 270a27–29.
51
Cf. Simplicius (In Ar. Phys. 1071.18): “It is clear that also the virtues of the soul belong to the
relations, since they too are due proportions of those things in which they arise – emotions, and
appetites, and such – just as the bodily [virtues are due proportions] of hot and cold [bodily
elements] and such.”
 chapter 13

Self-motion as other-motion in Aristotle’s Physics

Ursula Coope

In Physics viii.5, Aristotle sets out to provide a positive account of self-

motion. He characterises self-motion as a kind of other-motion: a self-
mover must be composed of two parts, one of which moves the other. The
part that produces the movement is a kind of unmoved mover, in that it
produces motion in a way that does not depend on it itself being moved.
The contrast here is with a moved mover, something that produces motion
just because it itself is moved by something else, as one carriage of a train
might move another just because it in its turn is moved by the engine.
τῆς ὅλης ἄρα τὸ μὲν κινήσει ἀκίνητον ὂν τὸ δὲ κινηθήσεται· μόνως γὰρ
οὕτως οἷόν τέ τι αὐτοκίνητον εἶναι.
In the whole thing, there is [a] that which imparts motion without itself
being moved and [b] that which is moved, for only in this way is it possible
for a thing to be self-moved. (258a2–4)
In this chapter, I shall discuss three questions raised by this account of self-
motion. The first relates to the role of this account in Physics viii.
Aristotle’s overarching aim here is to argue that there must be some first
moved thing that is continuously in motion, and that its motion must be
caused by an ultimate primary mover that is absolutely unmoved. Why is
an account of self-motion given such a prominent place in this argument?
After all, in Metaphysics L Aristotle argues for the existence of an unmoved
mover without appealing in the same way to an account of self-motion.1

1
In On the Movement of Animals i–iv, Aristotle argues that a self-mover must have, both within and
outside itself, a kind of unmoved springboard that it pushes against when it moves itself. He goes on
to ask whether the universe also moves itself in this way, by pushing back against something that is
unmoved, and he argues at length that it does not: there is nothing that could serve as such a
springboard for the whole universe. Note, though, that the possibility considered in these chapters of
On the Movement of Animals is rather different from that discussed in Physics viii. In Physics viii,
Aristotle argues that within a self-mover there must be an unmoved mover. In On the Movement of
Animals, neither the internal nor the external unmoved thing is described as a mover. It is, rather, a

245
246 ursula coope
My second question concerns Aristotle’s central argument that a self-
mover must contain within it a kind of unmoved mover. One obvious
opponent here is Plato, who, like Aristotle, thought that a chain of moved
movers must have its origin in a self-mover, but who described self-
movement as a kind of self-moving motion. Does Aristotle have any
argument against this Platonic view that the ultimate origin of a chain of
moved movers must be something that is itself in motion?
In the third section of my chapter, I consider an obvious objection that
might be made to Aristotle’s account of self-motion: can such an account
really distinguish between a genuine self-mover (a unified thing, one part
of which moves another) and two distinct things, one of which moves the
other? For instance, can Aristotle avoid the consequence that when an
animal moves a stick, the animal plus stick counts as a self-mover? Aristotle
needs to have something to say in answer to this question if his account is
really to be an account of self-motion, rather than an argument that there
is, in the end, no such thing as self-motion.
As we shall see, the first of my questions bears on the other two. Once we
understand the purpose for which Aristotle is giving this account of self-
motion, we can see what constraints he is under in giving it, and which of
the possible objections to it he most needs to answer.

I The role of the Physics viii.5 account of self-movement

in Aristotle’s discussion
At this point in Physics viii, Aristotle has already made the following
claims:
(i) There always has been and always will be movement (viii.1–2).
(ii) Every moving thing must be moved by something (viii.4).
(iii) A chain of movers moved by other things cannot go on to infinity: it
must have an origin (viii.5, 256a4–b3).
From (ii) and (iii), he concludes:
(iv) The origin of a chain of moved movers must either be a self-mover or
an unmoved mover (viii.5, 257a26–27).
He then gives his account of self-motion, arguing that:

kind of springboard: something that the animal pushes against in moving itself. (For discussion, see
the chapters by Coope and Morison, in Primavesi and Rapp (in press).)
 Self-motion as other-motion in Aristotle’s Physics 247
(v) A self-mover must be a composite, one part of which moves without
itself being moved, the other part of which is moved (viii.5, 257a31–
258b9).
He goes on to argue that:
(vi) There must be some first, continuous, eternal, and uniform circular
locomotion. (The continued coming to be of animal self-movers
depends on this.) (viii.6, 258b16–259a6, 259a13–20)
(vii) There must be some single cause of this primary motion. (viii.6,
259a6–20)
(viii) The unmoved mover within an animal self-mover is accidentally
moved when the animal moves. An unmoved mover that is acci-
dentally moved by itself in this way could not produce infinite
continuous motion, and hence could not be the unmoved mover
that explains this primary motion. (viii.6, 259b20–260a10)
(ix) In fact, the cause of the primary motion must be something that is
itself absolutely unmoved. (viii.6, 259b22–31)2
My question is about the role, in this argument, of (v), the claim that a self-
mover is composed of an unmoved mover and the thing it moves. An
interesting feature of the argument I have set out above is that this claim
seems somewhat tangential. In (iv), Aristotle sets out two alternatives: the
mover that originates a chain of moved movers must be either a self-mover
or an unmoved mover. However, what he goes on to say (in (viii) and (ix))
seems to rule out the possibility that what causes the primary continuous
locomotion could be a self-mover, for he claims that the cause of this
motion must be a mover that is absolutely unmoved. A mover that is
absolutely unmoved cannot in any sense ‘belong to’ the thing it moves,
since, according to Aristotle, a mover counts as (at least) accidentally
moved if it belongs to a thing that is moved (or contains such a thing as
a part) (254b9–10). This seems to imply that an absolutely unmoved mover
cannot, together with what it moves, constitute a unified whole that counts
as a self-mover. Hence, the origin of the ultimate primary continuous
motion must be an unmoved mover rather than a self-mover.3 Given this,

2
Aristotle allows that certain heavenly bodies can be accidentally moved by something and yet moved
continuously (as he must, if he is to maintain, as he does elsewhere, that the spheres on which the
planets are set are accidentally moved by outer spheres), but he claims that something that
accidentally moves itself cannot cause continuous motion (259b29–31).
3
Further grounds for this conclusion can be found in the features Aristotle ascribes to self-movers in
Physics viii.4, when he is arguing that the simple bodies are not self-movers. He says there that self-
movement is peculiar to living things and that what can move itself can stop itself (255a6–9).
248 ursula coope
Aristotle’s attempt to establish the precise nature of self-movers can seem
something of a digression.
Why, then, does Aristotle devote the second half of Physics viii.5 to
giving this account of self-motion? There are, I think, two main reasons
why this account is important to Aristotle’s overall argument. Both of them
have to do with Aristotle’s response to Plato.
In Laws x, Plato argued that the ultimate cause of motion was a self-
motion (understood as a self-causing movement). Aristotle could have
responded to this simply by claiming that such a cause would not be able
to produce eternal continuous motion, but in fact he has a more radical
objection. He argues that no origin of movement could be a self-causing
movement, and hence that Plato’s account is mistaken even as an account of
self-motion. This argument is important not simply as a way of responding
to Plato, but also because it provides crucial support for one of Aristotle’s
more positive claims. In particular, I shall argue, it is by understanding
what is wrong with Plato’s account of self-motion that we come to see why
it is that every thing in motion must be moved by something (see section II
below). This, then, is the first reason why Aristotle’s positive account of
self-motion is more than a digression: the argument he gives here against
Plato provides further justification for his claim (ii): the claim that every
moving thing must be moved by something.
However, I suspect that Aristotle also has a second reason for discussing
self-motion more fully here. He wants to say that there is, after all, some-
thing right in Plato’s claim about the primacy of self-motion (even though
Plato had the wrong account of self-motion). An animal self-mover is a first
mover: it (or more precisely, the mover within it) is the origin of a chain of
moved movers. Aristotle wants to emphasise what animal self-movers have
in common with the first mover of the primary eternal motion.4 As we have

Moreover, the fact that each of the simple bodies only moves naturally in one direction shows that
such movement cannot be self-movement (255a9–11). For an alternative interpretation, on which the
first eternal movement of Physics viii is a kind of self-movement, see Kosman 1994. Although I think
Kosman is right to stress that the first eternal movement together with its mover constitutes some-
thing that is in many respects like a self-mover, the possibility that it is strictly speaking a self-mover
is, I think, ruled out by Aristotle’s claim that the first mover is not even accidentally moved, together
with his account (in Physics viii.4) of what it is to be accidentally moved.
4
This perhaps explains why, even though he does not think that the eternal first motion is, strictly
speaking, self-motion, he is nevertheless prepared to say that what moves itself is ‘the principle (archê)
of moved movers and the primary thing among things that are moved’ (261a25–26, cf. similar
remarks at Physics viii.6, 259a33–b1 and at MA 1, 698a7–8). What moves itself is primary among
sublunary things that are moved, and the account he has given of the way in which a self-mover moves
itself applies also to the way in which the eternal first movement is caused by its mover. That is why
the fact that self-motion is a kind of locomotion is a reason for thinking that the eternal first motion
will also be locomotion (261a23–26).
 Self-motion as other-motion in Aristotle’s Physics 249
seen, this primary eternal motion cannot be a kind of self-motion, since its
mover is absolutely unmoved. Nevertheless, I shall argue, Aristotle’s
account of self-motion in Physics viii.5 is framed in such a way that it
applies not only to genuine self-movers, but also to the conjunction of the
first eternal moved thing and its unmoved mover. As such, it aims to give
necessary but not sufficient conditions for being a self-mover. In this
account, Aristotle includes only those features of self-movers that are also
features of a particular conjunction of things that does not constitute a self-
mover: the first eternal moved thing together with its mover.
This raises an obvious question. If Aristotle does not in fact think that
the combination of the first eternal moved thing and its mover constitutes a
self-mover, then why does he go to such pains to formulate an account in
Physics viii.5 that applies not only to genuine self-movers but also to this
combination of the first eternal moved thing and its mover?
One possible answer is that providing such an account helps to establish
something about the unity of the science of physics. In Physics ii.7,
Aristotle says that there are three different realms of study: one concerned
with that which cannot undergo movement, a second concerned with
eternal moving things, and a third concerned with perishable moving
things (198a30–31). The study of things that cannot undergo movement
is, he says, no part of physics (198a28). What, though, of the other two
realms of study? Elsewhere, he implies that physics is a science concerned
with both eternal and perishable moving things.5 This raises a question:
what justifies the assumption that there is one science of both eternal and
perishable moving things, and that the same basic principles apply to both
eternal and time-limited movement? In Physics viii, this question arises
with particular urgency, since Aristotle will argue that the mover that
produces the primary eternal motion is itself absolutely unmoved (and
hence, according to his own earlier remark (198a28), falls outside the scope
of physics). Can a single science concern itself both with movement that is
produced in this way and also with the kind of movement that has its origin
in animal self-movers?
To answer this, it is not enough to show that (as Aristotle argues in
Physics viii.6) animal self-movers are causally dependent on things that are
in eternal motion. Aristotle clearly thinks that X can be causally dependent
on Y, even though X and Y do not fall under the same science.6 What is
5
See, in particular, Metaph. L.1, 1069a36–b2. For an illuminating discussion of Aristotle’s views on the
unity of physics, see Falcon 2005.
6
As we have just seen, eternal motion is causally dependent on an absolutely unmoved mover,
although such a mover does not fall within the scope of physics.
250 ursula coope
needed is an abstract account that will bring out the similarities between
self-movers, on the one hand, and the first eternally moved thing together
with its mover, on the other. It is precisely such an account, I argue, that
Aristotle provides in the second half of Physics viii.5. Because it applies
both to animal self-movers and also to the primary eternally moved thing
together with its mover, this account helps to justify the assumption that
physics is a unified science.
I have argued that Aristotle provides the account of self-motion in
Physics viii.5 partly as a reply to Plato and partly as a way of emphasising
what genuine self-movers have in common with the primary eternal moved
thing together with its mover. In what follows, I first ask how Aristotle’s
argument fares as a response to Plato, and then discuss how the content of
his account is affected by his wish to capture just those features of self-
movers that they share with the first moved thing together with its mover.

II The argument that a self-mover must be composed

of two parts: An unmoved mover and what it moves
I have claimed that Aristotle’s account of self-movers here is written partly
in opposition to Plato’s view in the Laws, according to which the ultimate
origin of a chain of moved movers is a self-causing movement. In this
section, I shall ask whether Aristotle has any good argument against this
Platonic view.
The core of Aristotle’s argument comes in a single paragraph at 257a31–
b13. First (257b2–6), he says that it is impossible that what moves itself
should as a whole (and while remaining one and indivisible in form) both
cause and undergo the same movement. This impossibility is illustrated
with two examples: the same thing cannot at the same time and in the same
respect be teaching and being taught; the same thing cannot at the same
time and in the same respect be becoming healthy and producing health.
Second (257b6–13), he says that what undergoes movement must be in
potentiality whereas what causes movement must already be in activity
(ἤδη ἐνεργείᾳ). This again is illustrated with an example: what produces
heat must itself be hot. From this, it follows that what produces heat must
be distinct from what is becoming hot (since what is becoming hot must
not yet be hot).7 Aristotle then generalises from this example: ‘so too in
7
‘It is that which is hot that produces heat, and in general that which produces the form possesses it.
Consequently [if the mover is not distinct from the moved] the same thing in respect of the same
thing will be at the same time both hot and not hot’ (257b9–11: θερμαίνει τὸ θερμὸν καὶ ὅλως γεννᾷ τὸ
ἔχον τὸ εἶδος. ὥσθ’ ἅμα τὸ αὐτὸ κατὰ τὸ αὐτὸ θερμὸν ἔσται καὶ οὐ θερμόν).
 Self-motion as other-motion in Aristotle’s Physics 251
every case where the mover must have a synonymous property’. From this
he concludes, quite generally, that what first produces a movement must be
distinct from what undergoes that movement.
Much of this passage simply relies on examples. As interpreters, we need
to ask how Aristotle means to use these examples to defend his general
conclusion about self-motion. After all, his aim is not merely to show that
in certain cases the primary origin of a movement is something unchanging.
Although this would establish (against Plato) that it is possible for the
ultimate origin of a movement to be something that is not undergoing
movement, it would not establish Aristotle’s conclusion that within a self-
mover there must be a part that is moved and a distinct, unmoved, part that
is the mover.
Aristotle’s final example (that the cause of becoming hot must be some-
thing that is already hot) suggests the following general argument:
(i) The primary cause of something’s changing towards being F must be
something that is itself already F.
(ii) What is undergoing a change towards being F cannot yet be F.
(iii) Nothing can at the same time and in the same respect both be F and
not be F.
So: (iv) The primary cause of something’s changing towards being F must
be distinct from what undergoes the change towards being F.
So: (v) Since the mover within a self-mover is the primary cause (not
merely an instrumental cause) of the self-mover’s movement, the part
of the self-mover that produces movement must be distinct from the
part that undergoes that movement.
However, if this is Aristotle’s argument, it seems to face two insuperable
problems, both related to premise (i), the claim that a change towards
being F must be caused by something that is already F. The first problem is
that the scope of premise (i) is too restricted for it to do the work needed of
it in this argument. Aristotle himself recognizes that (i) only holds for a
certain range of cases, and, because of this, the conclusion he draws is
carefully qualified: ‘So too in every other case in which the mover must
have the synonymous property’ (257b11–12). This simply invites the
question: what about those cases in which the mover doesn’t have to have
the synonymous property? How could Aristotle go from this carefully
restricted conclusion to the general claim that, within a self-mover, what
is moved must be distinct from the mover? These questions are especially
pressing, since it is hard to see how premise (i) could apply to a particularly
252 ursula coope
central kind of movement: locomotion. Locomotion is not naturally
thought of as the acquisition of a form, and there is no obvious sense in
which the agent of a locomotion must already be in the state that the
moving thing is progressing towards. After all, it is clearly not true that
what moves something to place P must itself already be in place P.8 But
Aristotle’s argument needs to apply to locomotion, since he himself goes
on to say that locomotion is the primary motion and that, in the strictest
sense, the movement with which a thing moves itself is locomotion
(253a14–15, 261a23–25).
Of course, there may be a priori grounds for accepting something like
premise (i), in spite of the apparent counterexamples. However, when we
consider what such a priori grounds might be we encounter a second
problem. A plausible suggestion is that premise (i) is based on what
might be called the ‘transmission’ principle of causation: the principle
that what ultimately causes a thing to be a certain way must itself be that
way. If causation is thought of as a kind of transmission of a property from
one thing to another, then it is easy to see how this principle might be
justified: how could an agent transmit a property to something else if it
didn’t itself have that property?9 The problem with this is that one could
accept this general principle without accepting Aristotle’s application of it,
and hence without accepting premise (i). In fact, one could use this same
principle to argue for the very view Aristotle is attacking: the Platonic view
that what causes motion must itself be in motion. The argument would go
as follows. We are trying to explain X’s motion. We have already agreed
that if X is in motion, it must be being moved by something. But how can
the mover transmit movement to X unless it is itself in movement?
Indeed, something like this view seems to lie behind Plato’s reasoning in
the Laws. Plato agrees that every movement has a cause and that chains of
moved movers must have their origin in some primary cause. But he
concludes that the primary cause is a self-moving movement. Plausibly,
the basis for this view is this alternative application of the transmission
principle: he assumes that what causes movement must itself be in move-
ment. So, if we accept the transmission principle (the idea that what

8
Of course, there are various ways in which one could try to fit locomotion into the model. For
instance, one might say that the agent of locomotion has in its mind the goal towards which the
locomotion is proceeding (as the craftsman has in his mind the form of the thing he is producing).
But this would not be enough to save the apparent argument, since it is, at least prima facie,
possible for one and the same thing (at the same time) both to have the goal F in mind and also to
be (at) F.
9
Waterlow 1982: 244–246 calls this the model of ‘giving and receiving’.
 Self-motion as other-motion in Aristotle’s Physics 253
ultimately causes a thing to be a certain way must itself be that way), there
are two rival conclusions we might draw. We might conclude (with Plato)
that what causes a thing to be in movement must itself be in movement
(only then can it transmit movement to the thing it acts upon), or we
might conclude (with Aristotle) that what causes a thing to become F must
itself be F (only then can it transmit F-ness to the thing it acts upon). The
transmission principle by itself cannot justify the Aristotelian, as opposed
to the Platonic, conclusion.
The fact that this argument is vulnerable to these objections should
make one doubt whether it is, in fact, the argument Aristotle had in mind. I
shall argue that, in fact, Aristotle is defending here a more abstract
principle: that what first causes motion must do so in virtue of being a
certain way, not in virtue of itself undergoing motion. When Aristotle says
that, in certain cases, the cause of becoming F is something that already is F
he is providing an illustration of this more abstract principle, but he is not
meaning to imply that all first causes operate by transmitting a form to the
thing upon which they act.10 His argument, thus, does not rely on the
principle of contradiction, in the way that his remarks about the hot thing
might suggest: he is not saying that the first cause must already be F,
whereas what becomes F cannot yet be F. Instead, his point is that the first
cause is a cause in virtue of stably being a certain way, not in virtue of
changing or becoming. The sense in which a self-mover must be composite
is, then, that it must have two aspects: qua mover it is unchanging (since it
produces movement in virtue of stably being a certain way), whereas qua
moved it is undergoing change. If this is Aristotle’s argument, then his
objection to Plato is that, on Plato’s view, the ultimate explanation of a
movement is itself a movement, rather than a state of being.
If this is right, then we need to ask how Aristotle would justify the claim
that the ultimate explanation of movement must be a thing’s stably being a
certain way. I shall argue that Aristotle provides such a justification in a few
difficult lines in our passage. In these lines, he draws a connection between
movement, incompleteness, and potentiality:
ἔτι διώρισται ὅτι κινεῖται τὸ κινητόν· τοῦτο δ’ ἔστὶν δυνάμει κινούμενον,
οὐκ ἐντελεχείᾳ, τὸ δὲ δυνάμει εἰς ἐντελέχειαν βαδίζει, ἔστιν δ’ ἡ κίνησις
ἐντελέχεια κινητοῦ ἀτελής. τὸ δὲ κινοῦν ἤδη ἐνεργείᾳ ἔστιν
Further, it has been established that the moveable is what is moved. But this
is moved in virtue of potentiality, not fulfilment. But what is in virtue of

10
Or at least, the argument he gives does not rest on the truth of this universal claim.
254 ursula coope
potentiality progresses towards fulfilment, and movement is the incomplete
fulfilment of the movable. But the mover is already in activity. (257b6–9)
The difficulty of these lines stems partly from a problem of translation. The
most obvious translation of τοῦτο δ’ ἔστὶν δυνάμει κινούμενον, οὐκ
ἐντελεχείᾳ would be ‘this is moved in potentiality, not in fulfillment’,
but this translation seems to get the wrong sense. When the movable
thing is being moved, it is not merely potentially being moved. I have,
instead, given the translation: ‘it is moved in virtue of potentiality, not
fulfilment’.11 To understand what might be meant by saying this, we shall
need to look at what Aristotle says elsewhere about the relation between
movement and potentiality.
The argument in these lines rests, I think, on two claims: first, the claim
that what is in potentiality is posterior to, and causally dependent on, what
is in fulfilment (‘what is in virtue of potentiality progresses towards fulfil-
ment’); and second, the claim that a moving thing is, as such, not com-
pletely in fulfilment (‘movement is the incomplete fulfilment of the
movable’). Before we consider how Aristotle is putting these two claims
together in this argument, it will be helpful to look at each of them in turn.
The first claim is not directly about movement at all. It is, rather, about
the relation between what is in fulfilment and what is in potentiality.
Aristotle says that what is in fulfilment is prior, in a certain way, to what
is in potentiality. In our passage, he expresses this priority relation by
saying that what is in potentiality ‘progresses towards’ (badizei eis) fulfil-
ment. With this notion of ‘progressing towards’ he means, I think, to
capture a kind of teleological relation. Being potentially is a way of being for
the sake of some fulfilment. Aristotle spells this point out much more fully in
a discussion of potentiality and fulfilment in Metaphysics Theta 8. He says
there that what comes to be progresses towards (badizei epi) a principle and
end (telos) (1050a7–8). The end is an activity (energeia), and the thing that
comes to be acquires its potentiality for the sake of such an activity. Thus,
animals have sight (a potential) for the sake of seeing (an activity), and
humans have the art of housebuilding (a potential) for the sake of building
houses (an activity) (1050a10–12).
Aristotle takes this fact about teleological directedness to ground a
further claim about causation: the primary cause of something that is in
potentiality must be something that is in fulfilment. His thought seems to
be that anything that is, by its very nature, directed towards (or for the

11
In this, I follow Waterlow 1982: 244n.27.
 Self-motion as other-motion in Aristotle’s Physics 255
sake of) some further thing must be a dependent kind of entity. It must
depend on, and in particular derive its end-directedness from, some dis-
tinct cause. Ultimately, the source of such end-directedness must be some-
thing that is not similarly end-directed. The primary cause must be
something that is just what it is, without reference to anything further.
The primary cause must thus be something that is in fulfilment, not in
potentiality. He illustrates this with examples of cases in which being
potentially F is explained by being F in fulfilment. To be an acorn is to be
potentially an oak. An acorn must, then, have as its primary cause some-
thing that is, not merely potentially but in fulfilment, an oak. To be a
fluteplayer is to have the potential to play the flute; the primary cause of
being a fluteplayer must be the activity of playing the flute (one becomes a
fluteplayer by practising).12
This tells us that being F in fulfilment is causally prior to being F in
potentiality, but it does not, by itself, tell us anything about the relation
between causal priority and movement. For this, we need to turn to the
second claim Aristotle makes in this passage: the claim that a moving thing
is, as such, not completely in fulfilment. In making this claim, Aristotle is
clearly referring back to his definition of kinêsis in Physics iii.
He defines movement/change (kinêsis) there as ‘the fulfilment (entele-
cheia) of what is potentially, qua such’ (ἡ τοῦ δυνάμει ὄντος ἐντελέχεια, ᾗ
τοιοῦτον) (201a9). What exactly this means is controversial. Interpreters
have disagreed over what is meant by entelecheia, over what is meant by
‘qua such’, and over how to understand the phrase ‘what is potentially’ (is
this what is potentially in some state or what is potentially changing?).13 For
now, I want to leave these disputes to one side and instead look directly at
something Aristotle presents as an advantage of this definition. He says that
this definition allows us to capture what was right about earlier, confused,
accounts of movement: accounts which represented change/movement as
inequality, difference, or non-being. Such accounts were confused, in that
they failed to provide necessary and sufficient conditions for movement
(something can, for instance, be different without moving). But these
accounts nevertheless correctly recognised something about the nature of
movement: namely, that movement is essentially privative, or as Aristotle
puts it, incomplete.

12
See Metaphy. Theta 8, 1049b24–32. Aristotle draws the conclusion that the mover is already in
activity at 1049b27, using almost exactly the same wording as at Phys. viii.5, 257b9.
13
For some contrasting answers to these questions, see Kosman 1969, Heinaman 1994, Kostman 1987,
and A. Agnastopoulos 2010. I discuss the first three of these interpretations in Coope 2009.
256 ursula coope
We can learn something about what Aristotle means by this if we look at
the reason he gives for thinking that movement is incomplete:
ἥ τε κίνησις ἐνέργεια μὲν εἶναί τις δοκεῖ, ἀτελὴς δέ· αἴτιον δ’ ὅτι ἀτελὲς τὸ
δυνατόν, οὗ ἐστιν ἐνέργεια.
Movement is thought to be activity of a kind, but incomplete. The cause is
that the potential thing of which it is an activity is incomplete. (201b31–33)14
Movement is incomplete, Aristotle says, because the corresponding dunaton
(or potential thing) is incomplete. This explanation is, at first sight, puzzling.
Given Aristotle’s general views about the priority of fulfilment (or actu-
ality) to potentiality, one would expect features of what is capable of
phi-ing to be explained by features of phi-ing, rather than vice versa.
How, then, can Aristotle maintain that the incompleteness of movement
is explained by the incompleteness of the corresponding potential thing?
Aristotle is, I think, claiming that movement is related in a special way to
the corresponding potentiality. This follows from his definition of move-
ment as ‘the fulfilment of what is potentially, qua such’. Movement is the
proper activity of what is potentially, insofar as it is merely in potentiality.
Because of this, movement inherits certain features of the potential thing,
in a way that other activities do not. Thus, Aristotle is not saying that the
potentiality exercised in movement is incomplete and that there is, in
contrast, some complete kind of potentiality exercised in the kind of
activity that is not movement. Rather, he thinks that all potential things
are, as such, incomplete, but that movement is related to potentiality in
such a way that it inherits the incompleteness of the corresponding
potential thing.15 Aristotle is alluding to this special relation of movement
to potentiality when he says, in our passage from Physics viii, that the
changeable thing is δυνάμει κινούμενον, οὐκ ἐντελεχείᾳ (what I translated
as ‘moved in virtue of potentiality, not fulfilment’).
To be incomplete (ateles) is to be without an end. As we saw, a potential
thing is directed at (‘progresses towards’) some fulfilment. The potential
thing is thus ateles, because it is directed at some end (telos) that is other
than it. Aristotle’s claim is that movement, because of its connection to
potentiality, is similarly end-directed.

14
Here the claim is that movement or change is an incomplete activity (energeia), rather than (as in
viii) an incomplete fulfilment (entelecheia), but Aristotle seems to be using the two words inter-
changeably in these passages, so I think we can take him to be saying the same thing.
15
Here I am in agreement with Kosman 2013: 66, who says, commenting on 201b31–33, ‘This cannot
mean that there is some special kind of incomplete ability, the realization of which is incomplete and
is motion. For all mere ability is incomplete or unperfected – atelês.’
 Self-motion as other-motion in Aristotle’s Physics 257
We are now in a position to arrive at an understanding of our passage by
putting together these claims about incompleteness, posteriority, move-
ment, and potentiality. As we saw earlier, Aristotle thinks that because of
the way in which potentiality is directed towards fulfilment, a thing that is
in potentiality cannot be causally primary. Being potentially F is an
essentially other-directed way of being. Aristotle’s point here is that move-
ment also is an other-directed way of being. In this, movement is much like
potential being. Just as the potential to be a musician is for the sake of
actively being a musician, so also the movement becoming a musician is for
the sake of actively being a musician. Just as the seed is for the sake of
actively being the mature animal, so also growing (a change) is for the sake
of actively being a mature animal. There is reason, then, to think that
movement will be posterior to activity, just as we earlier argued that
potential being is posterior to active being.
Aristotle holds that as movement inherits its incompleteness (and hence
posteriority) from potentiality, so also it inherits its lack of causal primacy.
A movement must have as its first cause something that explains its
directedness towards an end, something that determines what would
count as interference with, or completion of, that movement. What
provides this explanation cannot be something that itself stands in need
of a similar explanation. Thus, a movement must have its origin in some-
thing that is not, similarly, incomplete or end-directed. Thus, Aristotle can
conclude that insofar as a thing is a first cause of movement, it must not
only be active (rather than merely potential), but active in a complete way
(and, hence, not undergoing movement).
Aristotle takes this argument to justify his claim that the first cause of a
movement must be completely active, and hence that a movement cannot,
ultimately, be explained by a movement. The challenge for such a view is to
explain how Aristotle can reply to a Platonist, who holds that what
produces a movement must do so in virtue of itself undergoing movement.
The interpretation sketched above shows how Aristotle might respond. He
argues that movement is a dependent kind of entity: it is an incomplete
fulfilment. Because of this, every movement must have an agent that
produces the movement in virtue of itself being completely fulfilled. Of
course, Plato might disagree with the claim that movement is a dependent
entity, a kind of incomplete fulfilment. Alternatively, he might say, with
some plausibility, that the kind of movements he had in mind in the Laws
were not what Aristotle is calling ‘movements’. Plato lists as examples of
such movements ‘wish, reflection, diligence, counsel, opinion true and
false, joy and grief, cheerfulness and fear, love and hate’ (897a).
258 ursula coope
However, Aristotle does, I think, have an answer to this. Plato himself
agrees with Aristotle that every movement requires an agent. That is
why Plato thinks that the first movement must be a self-movement
(rather than simply being an uncaused movement). What is the justifi-
cation for this assumption that every movement must have an agent?
Aristotle’s answer, I want to suggest, is that it is because movement is an
incomplete fulfilment that every movement must have an agent. As we
have seen, in saying that movement is an incomplete activity, he is
saying that it is a dependent entity that cannot itself be causally primary.
In other words, it is because movement is in this sense dependent, that
it requires an agent.
Aristotle does, then, have something to say to Plato. If Plato disagrees
with Aristotle’s characterisation of movement, he needs to find a different
justification for his assumption that every movement has an agent. So
Aristotle’s challenge would be this: the very thing that justifies Plato’s
claim that every movement has an agent (and hence that leads him to say
that the first movement must be a self-movement) also justifies the
further claim that the first cause of a chain of movements must act as a
cause in virtue of being completely in fulfilment, not in virtue of itself
undergoing movement.
Moreover, we can now see a further way in which this discussion of
self-movers contributes to Aristotle’s overall argument. In the earlier
chapters of Physics viii, he argues that every movement has a mover,
and claims that a chain of moved movers cannot go on to infinity. The
second of these claims is (in these earlier chapters of Physics viii) simply
an assertion,16 and the first is based merely on an examination of
different cases: he tries to show that something has a mover whether it
is moved unnaturally or naturally (and, if naturally, whether it is self-
moved or not).17 His account of self-motion in Physics viii.5 provides
some theoretical justification for both these claims, for it provides an
argument that movement, as an incomplete (and hence dependent)
being, must ultimately depend on something that is not similarly
incomplete. This implies both that every movement must have a
cause, and also that a movement’s first cause must be something that
is not itself a movement.

16
As Wardy notes (1990: 89), Aristotle does defend this claim in the second half of Physics vii.1. My
interpretation implies that Wardy is wrong to think that Aristotle needs this Physics vii.1 argument
to justify what he says in Physics viii.5.
17
Physics viii.4.
 Self-motion as other-motion in Aristotle’s Physics 259

III Primary self-movers

I have claimed that the account of self-movers in Physics viii.5 is framed in
such a way as to apply not only to animal self-movers, but also to the first
eternally moved thing together with its mover. In support of this claim, I
turn now to Aristotle’s discussion of what it is for something to be a
primary self-mover. I shall argue that Aristotle is here responding to a
natural objection to his account of self-motion. The manner in which
Aristotle responds to this objection shows, I think, the care he takes in this
account of self-movers to confine himself to features that genuine self-
movers share with the first eternally moved thing together with its mover.18
The objection to which Aristotle is responding arises because of his view
that self-motion is a kind of other-motion: for a thing to move itself is for
one part of it to move another part.19 The objection is that such an account
cannot distinguish between what we would ordinarily call a self-mover and
a conjunction of two things, one of which moves the other. Aristotle needs
to have something to say about this, in part because animal self-motion is a
kind of natural motion. When an animal moves itself from one place to
another, the source of this movement is in the animal itself, and hence the
movement is natural to the animal and the animal has a nature. Aristotle’s
account of nature would be undermined if he were forced to concede also
that when, say, an animal moves a stick from one place to another, the source
of this movement is within the animal+stick, and hence the animal+stick also
has a nature (since it has an internal source of change and rest).
Aristotle needs some account of how a genuine self-mover constitutes a
unity, but the conjunction of such a self-mover with something else that it
moves does not.
There are resources within Aristotle’s philosophy for answering this. An
obvious answer he might give is that the mover, within a self-mover, must
be the form of the self-mover. The animal is a self-mover because its mover

18
Another place where Aristotle is clearly tailoring this account to fit the eternal first motion and its
mover is 257b20–25. Here he is considering the possibility that the two parts of a self-mover might
move each other. One of his objections to this is that then it would not be necessary that each part
move the other. (The point seems to be that this suggestion only explains why A moves B on the
supposition that B moves A, and only explains why B moves A on the supposition that A moves B. It
does not explain why A and B are in motion rather than not.) But if we are simply concerned with
animal self-motion, it is not obvious why it should be desirable to have an account that presents such
motion as necessary. The need to explain the necessity of movement arises when the movement we
are trying to account for is eternal celestial movement.
19
It is important to recognise that a ‘part’ here need not imply a physically separable part. The mover,
for instance, could be the form of the thing.
260 ursula coope
(its soul) is its form; the animal+stick is not a self-mover, because its
unmoved mover (the animal’s soul) is not the form of the animal+stick.
This answer would enable him to capture the idea that a self-mover must
be a kind of unified whole (since it must have a form, which is its mover),
and also to capture the idea that it is really a self-mover, since it implies that
what moves the composite is itself accidentally moved.20 Given that this is
a good Aristotelian reply to our question, it is noteworthy that Aristotle
does not give it here. In fact, in this section of Physics viii.5 he says nothing
about the mover being the soul or being a kind of form. The reason
Aristotle does not give this answer here is, I think, because it would prevent
his account of self-movers applying to the combination of the prime mover
and what it first moves. The prime mover cannot be the form of the thing it
first moves, since then it would itself be moved accidentally, and – as we
have seen – Aristotle claims that the ultimate prime mover is not, even
accidentally, moved.
Aristotle thus faces a difficult task in attempting to answer our question.
On the one hand, he needs to distinguish between a self-moving composite
(such as the soul+body) and a composite of one thing’s moving another
(such as the soul+body+stick). Such a distinction is needed if Aristotle is to
avoid attributing an internal source of change and rest, and hence a nature,
to all sorts of accidental conjunctions of self-movers with the things they
move. On the other hand, Aristotle’s account needs to apply not only to a
self-moving composite, but also to the conjunction of the first eternally
moved thing with its mover. The account of primary self-movers that
Aristotle gives in the last part of viii.5 (258a9–b9) is, I shall argue, designed
to meet both these constraints.
The account is as follows. If A (the unmoved mover in the self–mover)
moves B, and B in its turn moves C, then the whole, ABC, can be regarded
as a self-mover of sorts, but primarily or strictly speaking the self-mover
will just be AB. For instance, if the dog’s soul moves its body, and its body
in turn moves a stick, the primary self-mover will be the combination of
the dog’s body and soul, in other words, the dog. The primary self-mover,
then, will be the unmoved mover together with the first thing that it moves.
An animal is a self-mover because it is composed of an unmoved mover
(presumably, the animal’s soul) together with the first thing moved by that
mover. This account allows for the conditions on being a primary self-
mover to be satisfied also by the absolutely first mover together with the

20
Aristotle himself makes this point about animal self-movers at Phys. viii.6, 259b16–20. For the
claim that the soul is the origin of an animal’s self-movement, see On the Soul iii.9–11.
 Self-motion as other-motion in Aristotle’s Physics 261
thing it first moves, even though (as we have seen) the absolutely first
mover cannot be the form or soul of the thing it first moves, since it is not
accidentally moved by the movement it causes.
For this account to avoid obvious counterexamples, it is important
that the kind of self-movement Aristotle has in mind here is self-
locomotion. He says that this is, properly speaking, the only self-
movement (253a14–15, 261a23–25). So if a hot thing, without itself
undergoing change, causes something else to become hot, the hot
thing and the thing it heats do not, together, satisfy these conditions
on being a self-mover.21 If one assumes (in accordance with what
Aristotle himself asserts) that, at least in the sublunary realm, every
origin of motion is a self-motion and the origin of the self-motion is an
unmoved mover, then we can always trace a chain of moved movers
back to an unmoved mover. The first of these moved movers, together
with the unmoved mover will constitute a self-mover. Aristotle holds
that, in the sublunary realm, this will be an animal.
Of course, one obvious question about this account is what it can say
about the unity of B and of C (the first thing that is moved by the unmoved
mover, and whatever is in its turn moved by that first thing). What makes
B and C two distinct things, thus justifying the conclusion that B, rather
than B plus C, is the first thing moved? Suppose B is itself composed of
parts (as Aristotle has argued it must be, given that it is moved). What
justifies the claim that B, as opposed to one of its parts, is the first thing
moved?
As Aristotle points out, these questions can be pressed in such a way as to
suggest that there is no first thing moved (and hence no primary self-
mover). According to Aristotle, any thing that is moved is infinitely
divisible: it has parts, which in turn have other parts, and so on ad
infinitum.22 Moreover, its movement depends on the movement of its
parts (at least, in the sense that the movement of the parts is a necessary
condition of the movement of the whole). This raises a problem for
Aristotle, since it suggests that there may be no first thing moved: any
putative first thing moved by the unmoved mover will in fact only be

21
It is less obvious how Aristotle can avoid allowing that this Physics viii.5 account applies to the
conjunction of a magnet and the iron it attracts.
22
He argues for this in Physics vi.4, and he reminds us of this claim at Phys. viii.5, 257a33–b1, just
before giving his argument that a self-mover does not move itself as a whole. The relevance of lines
257a33–b1 in their immediate context is not at all obvious. However, the point Aristotle makes here
about infinite divisibility is important for understanding the aporia (at the end of Physics viii.5)
about the primary thing moved.
262 ursula coope
moved because certain other things (its parts) are themselves moved by the
unmoved mover. Thus, any self-mover, AB, will have a more primary self-
mover that is its part: ‘AB will not be what is primarily moved by itself,
since when something is taken away from AB, the remainder will continue
to move itself’ (258a30–32).
In raising this worry, Aristotle is taking up, and responding to, an
argument that he himself gave in Physics vii.1. It is an argument for the
conclusion that every moved thing is moved by something. Although of
course he still agrees with this conclusion in Physics viii, he realises that he
cannot endorse the argument of Physics vii.1.23 That argument starts out
from the claim that anything that is moved is infinitely divisible. This is
taken to imply that its movement depends on its parts, and hence on
something that is distinct from it (since a thing is distinct from its parts).
Since the parts are divisible, their movement will depend on the movement
of their parts, and so on ad infinitum. In Physics vii.1 Aristotle concludes
from this that a moved thing must be moved by something that is distinct
from it.
In Physics viii.5, Aristotle sees that he cannot endorse this argument,
since, if it were successful, it would imply that there was no first thing
moved. Every moved thing would depend for its movement on the other,
prior, moved things that were its parts. This is a conclusion that Aristotle
cannot accept, since it would undermine his argument that a chain of
moved movers originates in a first mover together with a first thing moved
by that mover.
His response, in Physics viii.5, is to claim that the moved thing is
potentially, but not actually, divided: ‘so that if it is divided it will not
continue in possession of the same nature’ (258b2–3). Although anything
that is moved can be divided into parts, this does not imply that a thing’s
movement is causally dependent on the movement of its parts. In fact, its
parts may only move in virtue of being parts of the whole. A sign that this
is so is that many self-movers no longer move themselves when divided.
Hence, the fact that the moved thing is infinitely divisible does not
undermine the claim that there are primary self-movers. A primary self-
mover will be divisible into parts, but these parts will not be things that
would, if separated from the primary self-mover, be self-movers in their
own right.
23
This, of course, raises interesting questions about the role of Physics vii in the Physics as a whole. It
suggests that Aristotle did have Physics vii in mind when he wrote Physics viii, but that he did not
intend the two books to be part of a single work (since he took some of his arguments in viii to
supersede those he had given in vii).
 Self-motion as other-motion in Aristotle’s Physics 263
As an account of animal self-movers, this response is not wholly success-
ful. As Aristotle himself remarks elsewhere, some animals can be cut in two
in such a way that their parts continue to move themselves. Does Aristotle
have to deny that these are primary self-movers?24 Moreover, even in the
case of animals that cannot be cut in half and continue to move, there are
usually parts one can remove that would still allow what remains of the
animal to move itself. Is Aristotle committed to the claim that it is not
strictly speaking the animal that is the self-mover, but the animal minus its
hair and teeth and . . . ? Finally, it is not clear that even this subtracting
procedure will always arrive at a definite first mover. It might turn out that
the animal would remain a self-mover without part A or without part B,
but not without both. In such a case the animal minus part A and the animal
minus part B would both seem to have equal claim to be the primary self-
mover.
Aristotle does not consider any of these problems here. I suspect part of
the reason for this is that in this context he is not, primarily, interested in
giving a full account of animal self-motion. Instead, he defends the claim
that there can be a primary thing moved, because he will need to assume
this claim in the account he goes on to give of the relation between the first
eternal motion and its mover. In giving that account, he emphasises that
there is one unified, continuous first motion. It is because of the unity and
continuity of this first motion that it requires both a single subject and also
a single absolutely first mover: ‘if motion is continuous, it is one, and it is
one only if its mover is one and the thing that undergoes the motion is one’
(259a17–19). Aristotle’s argument thus relies on there being not only a
single continuous motion, but also a single unified subject for such a
motion. If such a subject is to have the right kind of unity, its movement
cannot be causally dependent on the movement of its parts. This is why it is
so important for Aristotle, in Physics viii.5, to explain what was wrong with
his earlier argument in Physics vii.1.
I have claimed that Aristotle’s account of self-movers in Physics viii.5 is
best understood when we consider its role in his more general argument.
He is not attempting here to give a full account of what is involved in an
animal’s moving itself. Instead, he gives a highly abstract argument that
appeals to his earlier discussion of the ontological status of motion. This
account serves several purposes. It presents a challenge to Plato’s alternative

24
Perhaps he would not be worried by this conclusion. At Progression of Animals 707a24–b4, he
implies that an animal that is divisible in this way is like a single body that contains within it many
living beings (see also Juv. 468b9–15).
264 ursula coope
view of the origin of movement. More positively, it provides an argument
that movement must, ultimately, be caused by something that is not itself a
movement. Finally, by giving this account of self-motion, Aristotle shows
that it is possible, at least at a certain level of abstraction, for there to be a
single account that applies both to animal self-movers and also to the first
eternal movement together with its mover. Insofar as the account is
successful, it thus provides some basis for the view that physics, the study
of motion, is a single unified science.
 chapter 14

The argument of Physics viii

Andrea Falcon

I Introduction
Physics viii is often regarded as an argument in two stages from the eternity
of motion to the existence of a first mover which is eternal and absolutely
unmoved. This twofold reading of Physics viii raises an obvious concern
about the boundaries of the science of nature. The first unmoved mover
does not seem to be a proper object of study for the student of nature. Since
this mover does not possess an internal principle of motion and rest, its
study would seem to pertain to first rather than second philosophy. And
yet the traditional reading does not only make the first unmoved mover an
object of study for the student of nature; it also makes this mover the
ultimate aim of the investigation conducted in Physics viii. It is tempting
to dismiss this concern by suggesting that in Physics viii Aristotle does not
enforce a division of labor between first and second philosophy, but that
instead he builds a bridge between these two sciences. I will not accept this
suggestion. I will argue that the investigation conducted in Physics viii is
not blind to the boundaries of natural philosophy. By my lights, Aristotle
remains a disciplined investigator throughout the whole of Physics viii.
Even when Aristotle is concerned with the first unmoved mover, he does
not lose sight of his stated goal, which is the explanation of the eternity of
motion. As I will argue, his treatment of the unmoved mover does not, and
need not, go beyond the boundaries of natural philosophy.

II The explanatory structure of Physics viii in outline

My reading takes the lead from the opening lines of On the Movement of
Animals (MA). Those lines allude to an investigation that clearly

Many thanks to Mariska Leunissen, Zoli Filotas, Emily Perry, and Charles Ives for their comments on a
draft of this chapter.

265
266 andrea falcon
corresponds to Physics viii. We are told that this investigation develops in
two stages, and that both stages are concerned with motion, or more
precisely with eternal motion:
Now, it was determined earlier that the source of all other motions is that
which moves itself, and [that the source] of this [is] that which is not subject
to motion, and that the first mover must be unmoved, when it was deter-
mined concerning eternal motion, whether or not there is such a thing, and if
there is, what it is. (MA 1, 698a8–11)
In this passage, Aristotle seems to have in mind the two stages of scientific
investigation outlined at the beginning of the second book of the Posterior
Analytics. These stages are introduced by two questions: the ei esti (whether
it is) and the ti esti (what it is) question. By answering an ei esti question we
establish whether something is the case (or not). A fact thus established
calls for a causal explanation. It is quite possible that a thing’s cause
becomes available at the same time as the fact is established, but this is
the exception rather than the rule. Typically, we first establish a fact by
answering an ei esti question, and then we look for its causal explanation.
When we ask a ti esti question, we ask for a definition that specifies the
relevant cause. Consider how Aristotle combines the search for a definition
and the search for a causal explanation in the Posterior Analytics: It is the
same, he says, to search what a thing is (ti esti) and why it is (dia ti) (Apo.
ii.2, 90a15). Aristotle illustrates this point with the help of the following
example: The definition of a lunar eclipse is privation of light coming from
the moon due to the interposition of the earth. This definition tells us what
a lunar eclipse is by singling out the cause of the lunar eclipse, namely the
interposition of the earth between the moon and the sun.
With this in mind we can turn to the opening lines of Physics viii. The ei
esti question is introduced right at the start of the book with the following
words:
Did motion ever come into existence not having existed before, and will it
go out of existence again, so that nothing is in motion? Or is it the case that
motion neither comes into existence nor goes out of existence but instead
always existed and always will exist, belonging to things as something
immortal and unceasing, as if it were a kind of life of all beings constituted
by nature? (Phys. viii.1, 250b11–15)
Following what Aristotle himself tells us at the outset of On the Movement
of Animals, I read this passage as introducing an investigation into the
existence of eternal motion. By the end of Physics viii.1, Aristotle is
confident that he has established that motion is eternal. He tells us that
 The argument of Physics viii 267
there was never a time when there was no motion, and there will never be a
time when there will be no motion (252b5–6). Thus, there exists eternal
motion. If this reading is accepted, the end of the first chapter marks the
end of the ei esti investigation. I will discuss the role of Physics viii.2 in due
course.1 For the time being, I am content to say that Aristotle turns to the ti
esti question only at the beginning of Physics viii.3. We have seen that
answering a ti esti question entails looking for a causal explanation, so it is
not really surprising that the chapter begins with a dia ti question: Why in
the world (dia ti pote) are some things at one time in motion and at another
time at rest (253a23–24)? It is in the context of answering this question that
Aristotle introduces, in Physics viii.6, a first principle that is eternal and
causes motion without being itself subject to motion, namely the first
unmoved mover. However, Aristotle cannot be content with the mere
introduction of such a principle but is required, by his own conception of the
scientific enterprise, to try to get to the most causally precise description of
that principle. This description is reached only at the end of Physics viii.10.
It crucially depends on some of the results reached in Physics viii.7–9. But
to understand how these chapters contribute to answering the ti esti
question we have to go back to Physics viii.6, and in particular to the
claim that there must be a first principle that is eternal and causes motion
without being itself subject to motion. This principle can only move
eternally in one and the same invariable way. By reflecting on this appar-
ently innocuous fact, Aristotle is able to infer that there must exist a single
eternal motion. Note that the existence of a single eternal motion cannot
be logically inferred from the claim that there exists eternal motion (Physics
viii.1). It can be secured only by reflecting on the nature of the first
principle (Physics viii.6). But as soon as this result is in place, we need to
know what kind of motion this is and why. This is the task that Aristotle
sets for himself in Physics viii.7–9.
I have offered an outline of Physics viii. On the proposed reading, there
is a single plan running from the beginning to the end of Physics viii. My
goal in the pages to follow is to shed further light on this plan. I will argue
that Aristotle is following a deliberate argumentative strategy in which
eternal motion is progressively illuminated, and eventually explained, by
arriving at the most precise description of its cause. In other words,
Aristotle adopts a strategy of progressive revelation. As a result of this
strategy, it is only at the very end of Physics viii that we (following

1
See below, section IV.
268 andrea falcon
Aristotle) are able to answer the ti esti question, and hence to say what
eternal motion is.
On this reading, eternal motion remains the focus of study throughout
Physics viii. One might object that this kind of focus does not sit well with
Aristotle’s programmatic remarks in Physics viii.1. There, Aristotle tells us
that the investigation he is about to launch contributes “not only to the
study of nature but also to the investigation about the first principle”
(251a5–8), which suggests that the book is also about the first principle.
To block this objection, it is sufficient to introduce the distinction between
primary and secondary objects of study. Since proper knowledge depends
on grasping the relevant cause(s), it is not merely an accident that Aristotle
is concerned with the first unmoved mover as the cause that explains why
motion is eternal. On the contrary, as we have already seen, Aristotle is
required to search not simply for the relevant cause but also for the most
precise description of that cause. But this fact alone does not make the
unmoved mover the primary object of study in Physics viii (thereby
turning its investigation into a metaphysical one). At most, it makes it a
secondary object of study. The primary investigation, which remains
firmly within the realms of natural philosophy, is about motion, or more
precisely about eternal motion.

III The ei esti stage of investigation (Physics viii.1)

Physics viii.1 is an attempt to answer the question whether motion is
eternal, or so I have argued. As soon as the ei esti question is introduced,
Aristotle outlines three possible answers to the question. He links them to
the names of Democritus, Anaxagoras, and Empedocles. Aristotle is not
interested in a full review of what has been said on the topic of the eternity
of motion by his predecessors. Rather, he is content to highlight what his
(our) theoretical options are. They are the following: (1) motion is eternal
(Democritus); (2) motion is not eternal (Anaxagoras); (3) motion and rest
alternate eternally (Empedocles).
The ei esti question is answered by invoking the definition of motion
given at the beginning of Physics iii.2 It is also possible to detect a less
obvious reference to the treatment of time offered in Physics iv.3 Evidently,
Aristotle uses the results established in Physics iii and iv as his scientific
starting-points (archai). He is able to arrive at the conclusions that there
was never a time when there was no motion and that there will never be a

2 3
Phys. viii.1, 251a8–10. Phys. viii.1, 251b10–13.
 The argument of Physics viii 269
time when there will be no motion, while starting from principles devel-
oped within his own science of nature, and while doing so solely with the
help of conceptual tools he developed earlier in the Physics. What his
predecessors have said on the topic of motion and time does not seem to
play any significant role in how the conclusion stated in calce at the end of
Physics viii.1 is reached.
If this reading of Physics viii.1 is correct, then dialectic understood as the
discussion of the so-called reputable opinions (endoxa) does not contribute
much to answering the ei esti question. Its role is a different one, and one
that does not become fully apparent until in the second part of the chapter.
By then, Aristotle has already established that there is always a motion prior
to any given motion, and that there is always a motion subsequent to any
given motion. At that point, Aristotle returns to the positions defended by
Anaxagoras, Empedocles, and Democritus in order to review how they
failed in their attempt to offer an adequate causal explanation for the
positions they have defended. His discussion of their views (and their
failures) helps us see what is expected from an attempt to answer the ti
esti question, if it is to be successful. The ultimate criteria of success that
can be extracted from the discussion of Anaxagoras, Empedocles, and
Democritus are two: Our answer (1) must do justice to the existence of
order and regularity in the natural world (contra Anaxagoras), and (2) must
explain eternal motion in such a way that it leaves nothing unexplained
(contra Empedocles and Democritus).
So let us see, briefly, what Aristotle thinks we can learn from the errors of
his predecessors. Pace Anaxagoras, there is no disorder in nature and in the
things that happen in accordance with nature (252a11–12). By Aristotle’s
lights, this is an empirical fact and a fact that the cosmological account
advanced by Anaxagoras, which posits nous as an agent initiating motion,
contradicts. Unlike Anaxagoras, Empedocles does not violate the empirical
truth that there is order (taxis) in nature. In fact, his cosmological account
may be taken as a vindication of the claim that there is order in nature.
However, Empedocles fails to offer an adequate explanation of the parti-
cular order he envisions. More specifically, it is not sufficient to say that
change and rest alternate, and to say that there are equally long periods of
change and rest. It is necessary to find a cause of that particular arrange-
ment. In particular, Aristotle says that we should never posit anything
without an argument, and that our argument must be either inductive or
deductive (252a23–25). When Aristotle turns to Democritus, he draws
attention to the following principle that the latter invoked in support of
his claim that motion is eternal: “things happened in the past as they
270 andrea falcon
happen now” (252a35). Democritus may have used this principle in his
explanation of the collision of atoms. His explanation may have combined
two claims: (1) that there is always a collision prior to any given collision
(and presumably there will always be a collision subsequent to any given
collision), and (2) that no further cause for this fact needs to be given. In
short, things happened in the past as they happen now. Aristotle objects to
this explanatory strategy with the help of an example which is familiar to
the reader of the Posterior Analytics: It is always true that a triangle’s internal
angles have the property of adding up to 180 degrees, but this fact never-
theless has a cause. Put differently, it is not sufficient to say that things
happened in the past as they happen now: A true causal explanation of why
things happen always in the same way must be supplied.4

IV From the ei esti to the ti esti question (Physics viii.2)

My starting-point has been what Aristotle himself says in the opening lines
of On the Movement of Animals. Those lines invite the application of the
two-stage theory of scientific inquiry outlined in the second book of the
Posterior Analytics to Physics viii. According to this theory, it is possible to
launch into a search for the cause(s) explaining why something is the case
only when that fact has been established. By the end of Physics viii.1
Aristotle has established that there is eternal motion. The search for the
cause of this fact is a long and complex journey. It is only at the end of
Physics viii.10 that Aristotle appears to be confident that he has reached an
adequate description of the cause and is in the position to say what eternal
motion is. In the following sections, I will outline some aspects of his
argumentative strategy.
Before moving on, however, a few more words on the conclusion
reached at the end of Physics viii.1 are in order. That conclusion not
only commits Aristotle to the thesis that there is no beginning and no
end to motion; it also commits him to the claim that there are no gaps in
the infinite chain of motion. This follows from the rejection of the position
that Aristotle ascribes to Empedocles. Aristotle’s view is that there is always
4
Democritus seems to have employed the same explanatory strategy in his account of animal
generation. In Generation of Animals, Aristotle tells us that Democritus invoked the fact that “this
is how things always come to be” as sufficient explanation of the formation of bodily parts (GA ii.6,
742b17–23). In this case, Aristotle gives some additional information concerning how Democritus
may have reasoned: to ask the question why (dia ti) is to seek for a beginning (archê); however, what is
always is indefinite (apeiron), and there is no beginning (archê) of the indefinite (apeiron). Aristotle
objects to this explanatory strategy, as he does in Physics viii, by recalling his favorite geometrical
example of the triangle and its angles, as described above (GA ii.6, 742b24–29).
 The argument of Physics viii 271
motion without interruption.5 At the end of Physics viii.1, nothing pre-
cludes the existence of a single motion that is both continuous and eternal.
And yet, as we have already said, the existence of such a motion cannot be
derived from the existence of an eternal, unbroken chain of motion. In
other words, the existence of a single motion that is both eternal and
continuous is not a given in Physics viii. It is a result to be secured in the
process of answering the ti esti question.6
What I have said so far draws attention to the fact that the conclusion
reached at the end of Physics viii.1 is in need of further articulation and
elaboration. Let us therefore return to the theory of scientific inquiry
outlined in the second book of the Posterior Analytics. There, Aristotle is
not just content to say that the ei esti question is to be answered before
asking the ti esti question; he also tells us that by answering the ei esti
question we gain some initial grasp of what the thing itself is. More
precisely, we acquire non-accidental knowledge that the thing exists. For
Aristotle, this is knowledge that is already directed towards the ti esti of the
thing (93a28–29). What Aristotle may have in mind is a certain epistemic
state that, among other things, gives us a fix on the thing to be explained
and at the same time motivates us to go further in our inquiry.7 I suggest
that this is exactly the state we are in at the end of Physics viii.1.
In Physics viii.2, Aristotle introduces two objections to the conclusion
that has just been established, namely that motion is eternal. It is important
to understand that his discussion of these objections is not meant to be a
confirmation of the conclusion reached in Physics viii.1. This conclusion is
not in need of further confirmation. At this stage of the inquiry, we do not
know yet whether there is a single eternal motion. But we do know that
there is always motion without interruption. Our knowledge of this fact is
as firm as it can possibly be. Therefore, the question for the reader of Physics
viii is this: What role does Physics viii.2 play in the argument, if it does
not serve to confirm the results achieved in Physics viii.1? I submit that the
discussion of the two objections is not looking backward to the conclusion
established in Physics viii.1 but rather forward to the subsequent investiga-
tion. In dealing with these objections, Aristotle does not only give us a
sense of the investigation to come but also orients it in a certain direction.

5
Cf. how the conclusion reached at the end of Physics viii.1 is recalled at the beginning of Phys. viii.6,
258b10–12: “since there is always to be motion without interruption, there must be something eternal
that first imparts motion (whether one or many), and the first mover must be unmoved.”
6
See below, section v.
7
This aspect of Aristotle’s theory of scientific inquiry has been studied by Rob Bolton. See, in
particular, Bolton 1976: 514–544 and 1987: 120–166 (especially at 130–142).
272 andrea falcon
The first objection is that motion is a self-terminating process, since all
motion seems to be from a contrary to a contrary (252b9–12). Although this
general observation does not really prove that motion cannot be eternal, it
can be taken as initial evidence against the conclusion established in Physics
viii.1. If all motion is bounded by contraries, then it is more difficult to
believe that motion is eternal. In his reply to this first objection, Aristotle
anticipates that his argument will show that all motion does not take place
between contraries, for there is one type of motion, namely circular
locomotion, which is not bounded by contraries.
The second objection begins with an observation about the behavior of
inanimate objects: These objects are in motion at one time and at rest at
another; it would seem that they should be either always in motion or
always at rest (252b12–16). Aristotle builds on this objection by adding an
observation about the behavior of animate beings. It would seem that these
beings can initiate their own motion without any prior motion. If this is
the case, these beings are prima facie evidence for the thesis that there can
be motion without a prior motion (259b17–28). Aristotle rejects the claim
that animate beings are self-movers in the sense that they can initiate their
motion without any prior motion. This qualification is important. Aristotle
does not mean to deny that animate beings are self-movers; he only
disputes how self-motion is typically understood. Clearly, he thinks that
it is an empirical fact that some physical motions caused by the environ-
ment take place in the animal. Those physical motions are not only prior to
animal motion, but also causally connected to it. In other words, there is a
sense in which the animal is causally responsible for its bodily displace-
ment. But that bodily displacement causally depends on other, prior
motions taking place in the body of the animal. Aristotle ends this section
by promising that self-motion will become clear at a later stage of his
discussion (253a20–21). His remark confirms that the chapter as a whole is
looking forward to the second stage of the argument offered in Physics viii.

V The ti esti stage of investigation (Physics viii.3–6)

Throughout Physics viii.2, the focus remains on motion. Aristotle pro-
mises to tell us more about circular motion, the motion of inanimate
objects, and the motion of animate beings. Evidently, these are the types
of motion which are to be integrated into a single causal explanation of why
motion is eternal. It is from the motion of the animate and inanimate
things that the subsequent investigation gets started. These things are at
one time in motion and at another time at rest. This is a fact that is clear
 The argument of Physics viii 273
and obvious to everyone.8 It is also a fact that calls for an explanation, as it
seems to be prima facie evidence against the conclusion reached in Physics
viii.1, namely that motion is eternal. This explains why Physics viii.3
opens with the following question: Why in the world are some things at
one time in motion and at another time at rest?
This question is presented as the starting-point (archê) of a new inves-
tigation. While the investigation up to this point was pre-causal (as it was
prompted by an ei esti question), the new investigation is a causal one, as it
is introduced by a why-question (a dia ti question). Aristotle is remarkably
forthcoming as to where this causal investigation will take us. While the
starting-point (archê) of the investigation is why some things partake of
motion and rest, the anticipated conclusion is that there are three types of
thing that correspond also to three modes of being with respect to motion:
for while some things are not subject to motion, other things are always in
motion, and still others partake of motion and rest. This is how Aristotle
puts it:
The starting-point (archê) of our investigation is the same as the starting-
point of the difficulty mentioned above: Why in the world (dia ti pote)
are some things at one time in motion and at another time at rest? Now,
it is necessary either that all things are always at rest, or that all things are
always in motion, or that some things are in motion and some are at rest.
In this last case, again, [it is necessary] either that the things that are in
motion are always moved and the things that are at rest are always at rest
or that all the things are by nature equally at rest and in motion. Or there
remains also a third possibility: for it is possible that some things are not
moved, others are always moved, and still others partake of both [motion
and rest]. This is what we must say, for this solves all the difficulties
[introduced in Physics viii.2] and is the end (telos) of our investigation.
(Phys. viii.3, 253a22–32)
In stating the telos of his investigation, Aristotle does not make use of causal
language. And yet, it is clear that he will have to do more than just establish
that there are three types of thing (or, if you prefer, three modes of being
with respect to motion). If Aristotle does not want to be open to the
criticism he has leveled against his predecessors, he will have to offer a
causal explanation for the arrangement envisioned in the opening lines of
Physics viii.3. More precisely, he will have to explain this arrangement, and
explain it in such a way that there is no need for us to go any further in the

8
Phys. viii.6, 259a27.
274 andrea falcon
search for an explanation. Hence, he will have to find the ultimate cause of
this order.9
This remarkable explanatory feat is achieved at the end of Physics viii.6.
By then, Aristotle has not just established that there are three types of thing
(three ways of being with respect to motion); he has also shown how they
are causally interconnected. At that stage of the argument, Aristotle seems
to be confident that the question introduced in the opening lines of Physics
viii.3 can be answered by identifying the cause which explains why there is
this threefold relation to motion:
It has become clear from what has been said also what puzzled us at the start,
namely why in the world (dia ti pote) is it not the case that all things are either
in motion or at rest or some always in motion and some always at rest, but
some things are [in motion] at one time but not at another. The cause
(aition) of this is now clear: Some things are moved by an eternal thing that is
unmoved, and for this reason they are always in motion, while others [are
moved] by something that is moved and changed so that they too must
change. What is unmoved, as was said, because it remains simple, identical,
and in itself, will move with a motion that is one and simple. (Phys. viii.6,
260a17–19)
The argument leading to the discovery of the eternal mover which moves
by being absolutely unmoved is best understood as an inductive inference
to the only possible causal explanation of a fact that is forced upon us by
experience, namely that some things partake of both motion and rest.10 I
will not engage in a detailed study of the remarkably complex argument
offered in Physics viii.3–6. Here I am content to point out that Aristotle
does not simply reason from effect (motion) to cause (the first eternal
unmoved mover); rather, he reasons from a remote effect (the fact that some
things partake of motion and rest) to its ultimate cause (the first eternal
unmoved mover). The relevance of these qualifications will become clear in
due course.11 For the time being, I only add that the ultimate cause is
introduced in Physics viii.6 to explain why motion is eternal in the sense of
being a continuous and inexhaustible chain of motion. Aristotle argues
that a collection of perishable unmoved movers, no matter how many they
are, does not suffice to explain why there is always motion without

9
Recall what Aristotle says in connection with Democritus: To look for an explanatory principle is to
look for the sort of thing that explains eternal motion in such a way that the investigation can stop
when the principle is found because at that point nothing is left unexplained.
10
Cf. Phys. viii.3, 254a35–b1: “against all these [views] one piece of evidence (pistis) is sufficient: We see
(horômen) that some things are in motion at one time and are at rest at another time.”
11
See below, section vii.
 The argument of Physics viii 275
interruption. In addition to the existence of perishable unmoved movers,
the existence of an eternal unmoved mover is necessary.12
I have already noted that the existence of a single eternal motion cannot
be derived from the conclusion that motion is eternal (Physics viii.1). The
existence of such a motion must be secured in the process of looking for an
answer to the question why motion is eternal, or so I have argued. Now, the
existence of a single eternal motion is inferred towards the end of Physics
viii.6. It is inferred from a reflection on the very nature of the eternal
mover which moves by being absolutely unmoved. Precisely because the
first eternal mover cannot be moved even in an accidental way, it must
move always in one and the same way. At first sight, this may look like a
fairly innocuous point. And yet it is sufficient to establish that, in addition
to the first unmoved mover, there must be some single motion that is
eternal and continuous:
But if indeed there exists something that is always of such a nature as to
move something while being itself unmoved and eternal, then it is necessary
that the first thing moved by this is eternal too. This is also clear from the
fact that there is no other way for generation and corruption and change to
exist in the other things unless there is something moved that moves. For the
unmoved [mover] will move in the same way and with one and the same
motion, since it will not change in relation to what is moved. But what is
moved by something that, though it is in motion, is moved directly by what
is unmoved, since it will be in a different relation to the things [that it
moves], it will not be the cause of the same motion, but because it is in
contrary places or forms it will cause each of the other things to be moved in
contrary ways, and will cause them to be at one time at rest and at another
time in motion. (Phys. viii.6, 259b32–260a10)
At this stage of the argument, we know very little about this motion. We do
not know, for instance, what kind of motion it is. But we do know that it
enters as an indispensable causal factor into the explanation of the eternal
cycle of generation and corruption. More specifically, the first unmoved
mover cannot account directly for motion as experienced on earth. Hence,
an intermediate mover is required. The eternal motion caused by the first
unmoved mover functions as such a mover. Needless to say, there cannot
be a single eternal motion without a thing that undergoes this motion. In
12
In the opening lines of Physics viii.6, the question of how many unmoved movers are needed is
clearly flagged: “since there is always to be motion without interruption, there must be something
eternal that first imparts motion (whether one or many), and the first mover must be unmoved” (Phys.
viii.6, 258b10–12). Note that in the course of Physics viii.6, Aristotle argues for the view that one
eternal motion is sufficient to account for the existence of a single eternal inexhaustible chain of
motion.
276 andrea falcon
Physics viii, however, Aristotle says very little about this being. In fact, he is
content to say that it performs the function of a moved mover in the
explanation of the eternity of motion.
If this reading is correct, what Aristotle says in the above passage cannot
be dismissed as an afterthought. On the contrary, it does some important
explanatory work. It helps us articulate our initial grasp (and confused
idea) of the eternity of motion. The eternal chain of motion which does not
admit of a beginning and an end, and does not admit of gaps either, turns
out to be a fairly complex state of affairs entailing the existence of a single
thing that is eternally moved in addition to the existence of an eternal cycle
of generation and corruption. Perhaps this result can be recast as follows.
By the end of Physics viii.6, we have discovered not only the cause of
eternal motion, which is to say the first mover that is absolutely unmoved,
but also the proper subject that undergoes eternal motion, namely the first
moved thing.
It may be helpful to recall how Aristotle himself refers to the results he
has achieved in Physics viii.6. In the first book of On Generation and
Corruption, Aristotle refers to Physics viii as one of the investigations
concerned with motion. Interestingly enough, he seems to think that his
argument has established two things: (1) that there is a first unmoved mover
that is not subject to motion; and (2), that there is something else that is
eternally moved by it. Note that Aristotle is quite forthcoming in describ-
ing the first unmoved mover as an efficient cause of motion:
We must deal with these [questions] to the extent that it is possible, and give
the cause of the fact that there is always generation – both unqualified and
qualified generation. Since one cause is that from which we say that motion
originates, and another is matter, it is the latter that we have to discuss here.
For we have dealt with the other cause earlier in our writings on motion when
we said that there is something that is not subject to motion through all time, and
something else that is always moved. (GC i.3, 317a1–5)

VI The ti esti stage of the investigation continued (Physics viii.7–9)

The conclusion of Physics viii.1 is that there must be a chain of motion
with no beginning and no end and no gaps in between. The argument
advanced at the end of Physics viii.6 helps us see that such a chain requires
the existence of a thing which is always moved in the same invariable way.
Admittedly, we know very little about the nature of its motion. By reflect-
ing on the nature of the first principle of motion, we can safely conclude
 The argument of Physics viii 277
that this motion must not only be single, but also simple. More specifically,
if the mover must be absolutely unmoved, then it can only impart a simple
motion. In this case, simplicity entails uniformity. Still, we are not yet in
the position to say what kind of motion this is. Hence, our investigation
cannot stop here. Let us put it differently: Even if we have found the proper
subject and the proximate efficient cause of motion, we still do not know the
form of this motion. The task Aristotle sets for himself in Physics viii.7–9 is
to establish that the first moved mover is moved in a circle. Circular
motion is the only form of locomotion that can be both primary and
eternal.
Aristotle is quite forthcoming about his goal in Physics viii.7–9. His
goal is to establish whether there is a motion that is continuous, and, if
there is one, what it is.13 It turns out to be the motion undergone by the first
mover that is eternally moved, and it enters into the causal system as the
primary, i.e. first, motion:
But once we have made another beginning (archê), these matters will
become clearer. It is to be investigated whether it is possible that there
exist one motion that is continuous or not, and if it is possible, what it is,
and what is the first motion: For it is clear that if it is necessary that there is
always motion, this motion is first and continuous because the first mover
moves this motion, which must be one and the same, continuous, and first.
(Phys. viii.7, 260a20–26)
In the stretch of text that begins with chapter 7 and ends with chapter 9,
Aristotle proves that:
1. locomotion is prior to generation, alteration, growth, and corruption
2. circular motion is prior to rectilinear motion
3. hence, circular locomotion is the primary motion.
Part of the argument for the second claim consists in showing that:
4. no motion other than circular locomotion can be eternal and continuous.
For Aristotle, there can only be finite rectilinear motion. But no finite
rectilinear motion can be eternal and continuous. Two reasons are given
for this claim. First, every motion takes place between contraries. Finite
rectilinear motion is no exception to the rule. Let us treat A and B as
contraries. Since the thing that undergoes rectilinear motion from A to B

13
Note that Aristotle adopts the ei esti and the ti esti questions in this stretch of text. In other words,
even the search for the primary, i.e. first, motion, offered in Physics viii.7–9 unfolds according to the
two-stage theory of scientific investigation outlined in the Posterior Analytics.
278 andrea falcon
or from B to A must turn back as soon as it has reached either A or B, we are
confronted with two motions rather than a single motion, namely motion
from A to B and motion from B to A. Second, turning back necessarily
involves coming to a stop, as there must be an interval of time between the
motion from A to B and the motion from B to A. What we obtain is an
eternal back and forth between A and B, rather than a single continuous
eternal motion. Circular locomotion is the only motion that can be
continuous and eternal. Aristotle treats circular locomotion as motion
from one point to the same point, namely motion from A to A. Contrast
this motion with finite rectilinear motion. In the case of finite rectilinear
motion, it is possible for one thing to move from A to A, but only after
having stopped at B. Moreover, motion from A and motion to A are
contrary motions.14 It is possible to have one thing moving with a single
motion from A and to A provided that the motion in question is in a circle.
This is the only motion that can be continuous, as the thing that undergoes
motion need not stop when it has reached A.
It might be helpful to draw attention to a prominent feature of Aristotle’s
argumentative strategy. This feature is especially (but by no means exclu-
sively) evident in Physics viii.7–9. There, Aristotle is content to speak of
circular locomotion in very general (i.e. abstract) terms. It is fairly clear that
the circular locomotion that enters into the explanation of eternal motion is
the eternal, continuous, and indeed uniform, circular locomotion of the
outermost celestial sphere. This is the sphere of the fixed stars, which is
moved always in the same way by the first unmoved mover. Still, it is
remarkable that Aristotle never describes it in this way. Nor does he distin-
guish the motion of the heaven of the fixed stars from the motions of the so-
called planets. To understand why Aristotle never engages in what we might
want to call (for lack of anything better) celestial physics, we need to recall
that the ultimate goal of Physics viii is to explain why there always was and
always will be motion. It turns out that this explanation requires the
introduction of a causal intermediary between the first unmoved mover
and the eternal cycle of generation of corruption. This causal intermediary is
celestial motion. Aristotle is interested in celestial motion insofar as it enters
as an indispensable moved mover into the eternal chain of motion.15 A full

14
Recall that the identity of motion is established by the contraries involved in the process. Motion
from A is shorthand for motion from A to B, and motion to A for motion from B to A.
15
Contrast this eternal chain of motion with the following chain of motion: man–stick–stone. In this
second chain, the stick is a dispensable moved mover. While the stone is moved by the stick, it is clear
that the man could move the stone without employing a stick. The eternal circular motion
envisioned as an intermediary between the first unmoved mover and the eternal cycle of generation
 The argument of Physics viii 279
study of the nature of celestial motion goes emphatically beyond the scope of
Physics viii. Such a study is the task of celestial physics. It is in the context of
celestial physics that Aristotle offers a study of the heaven of the fixed stars,
including an explanation of its distinctive orientation. Furthermore, it is in
the context of celestial physics that the circular motion of the heaven of the
fixed stars is distinguished from the circular motion of the so-called planets.
Finally, it is in this context that Aristotle introduces the highly controversial
claim that the heavens are made of a special simple body, unique to them.

VII Aristotle’s second treatment of the cause (Physics viii.10)

The last words of Physics viii.9 are a careful recapitulation of the argument
of Physics viii up to this point:
We have argued that there was always motion and always will be motion
throughout all time [Physics viii.1], and we have explained what the
principle of this eternal motion is [Physics viii.3–6]. We have also explained
which is the primary motion and which is the only motion that can be
eternal [Physics viii.7–9], and we have pronounced the first mover to be
unmoved [Physics viii.6]. (Phys. viii.9, 266a7–9)
One might be tempted to think that the argument has reached its natural
conclusion at the end of Physics viii.9. Interestingly enough, Aristotle does
not stop there. He immediately goes on to announce a second treatment of
the cause, namely the first unmoved mover:
Now, we have to say that this [i.e. the first unmoved mover] must be
without parts and must have no magnitude. (Phys. viii.10, 266a10–11)
At least at first sight, this return to the cause of eternal motion is surprising.
Recall that Aristotle has already offered an inference to the ultimate cause
of eternal motion. We have seen that the argument that started at Physics
viii.3 resulted not only in the discovery of such a cause, but also in the
discovery of the genuine subject of eternal motion. Arguably, the greatest
triumph of Physics viii.6 is the conclusion that, in addition to the first
mover that moves by being absolutely unmoved, there must be something
else that is eternally moved by the first mover. This other thing enters into
the explanation of the eternity of motion as an indispensable moved
mover. At that point, it was quite natural to wonder what kind of motion
this second mover undergoes and why. Hence, it was not really surprising to

and corruption is not this sort of moved mover. Without its presence, there would be no eternal cycle
of generation and corruption.
280 andrea falcon
see Aristotle turn his attention to this important question in Physics viii.7–9.
But once it is established that there is a genuine subject of motion, and that
this subject performs circular locomotion, which is the only motion that can
be single and continuous, we might be inclined to think that the explanatory
work has been completed. And yet Aristotle does not seem to think so,
because he returns to the cause of eternal motion. Why?
The interpreters who take Physics viii to be an argument in two stages
from the eternity of motion to the first unmoved mover have no trouble
explaining the final chapter of Physics viii. By their lights, Physics viii.10 is
the culmination of the investigation conducted in Physics viii. At first
sight, it would seem to be more difficult to find a satisfactory reading of this
chapter for someone who adopts the view I have been advocating, namely
that Physics viii as a whole is concerned with eternal motion. Recall,
however, that we do not have full knowledge of something unless we
have knowledge of its proper and adequate cause. The qualification
“proper and adequate” is the key to understanding the role that our chapter
plays in the argument of Physics viii. For Aristotle, it is never sufficient to
point to the relevant cause; it is also imperative to seek out the most
causally precise description of that cause. And it is only at the end of the
chapter that such a description is available to us.16
It cannot be just a coincidence that the second treatment of the cause of
eternal motion uses the proximate effect as its starting-point. This effect was
not available when Aristotle offered the first treatment of the cause. We
have seen that, in Physics viii.3–6, Aristotle reasoned from the remote effect,
which is to say from the fact that some things are at one time in motion and
at another time are at rest, to its ultimate cause, namely the first eternal
unmoved mover.17 It is only at the end of Physics viii.9 that the proximate
effect, namely the primary circular locomotion, is available. In Physics
viii.7–9, Aristotle is able to establish that the primary circular locomotion
is a single, continuous, eternal, and indeed infinite, motion. What seems to
be especially important is that this motion is infinite. At least twice
Aristotle returns to this feature.18 Also, in light of this fact, it cannot be
just a coincidence that at the beginning of Physics viii.10, in gathering
premises for his second treatment of the cause of eternal motion, Aristotle
establishes that nothing finite can impart an infinite motion, and that an
infinite power cannot reside in a finite magnitude.

16
This is how I understand the methodological recommendation to look for the most precise cause.
See, for instance, Phys. ii.3, 195b21–23.
17
See above, section v. 18 Phys. viii.8, 261b27 and 265a10.
 The argument of Physics viii 281
There is at least one other aspect of Physics viii.10 which is worth
mentioning: Aristotle’s focus on projectile motion, which presents a
prima facie difficulty for the thesis that everything that is moved is
moved by something: How can something be in motion when the mover
is no longer in contact with the thing that is moved? And yet there is a more
specific, and indeed pressing, reason to deal with this aporia in Physics
viii.10. But in order to appreciate why Aristotle engages in this excursus on
projectile motion, we need to recall, briefly, how he solves the aporia. His
solution is that the original mover gives the power of being a mover either
to air or to water: Both air and water are suited to being moved and to
moving themselves. In other words, projectile motion is explained as a case
in which the object is moved by a sequence of consecutive movers in
contact with one another. Now, one might try to apply this solution to
the case of eternal circular locomotion. In particular, one might try to
explain this motion as the result of an infinite succession of consecutive
finite powers transmitting motion from one to the other. Aristotle blocks
this explanatory route by making it clear that projectile motion merely
appears to be a case of continuous motion. A requirement for continuity is
that motion is caused by a single mover. Hence, positing an infinite
sequence of finite powers which jointly cause an infinite motion will not
do. The resulting motion will be infinite but will not be continuous. Yet
continuity is established as a key feature of eternal circular locomotion in
Physics viii.7–9.
We are finally in a position to turn to the second treatment of the cause
of eternal motion. This treatment is offered in the final stretch of text
(Physics viii.10, 267a21–b26). This treatment is not just a recapitulation of
what was said before. It has a new starting-point and does not end in the
same way. It goes as follows:
1. There must be a single continuous motion.
2. There must be a single subject of motion.
3. There must be a single mover.
4. This mover must be either moved or unmoved.
5. The chain of motion must end with a single unmoved mover.
6. This is the only mover which can always move because it is not
affected and its motion is effortless.
7. It moves with a uniform motion.
8. Because action at distance is not possible, this mover must be located
either at the periphery or in the middle of the universe.
9. It is located at the periphery.
282 andrea falcon
The treatment is not complete. A final step is missing. Aristotle invokes
two claims he has proven in Physics viii.10, namely that nothing finite can
impart an infinite motion, and that an infinite power cannot reside in a
finite magnitude. Since there cannot be an infinite magnitude, he is left
with the following conclusion:
10. The unmoved mover must be without magnitude, without parts,
and must be indivisible.
At this point we have reached the most precise possible description of the
(efficient) cause of eternal motion. This description is secured from within
natural philosophy and solely with the help of the conceptual resources
available to Aristotle in Physics viii. Note that the description of the
ultimate principle of motion is largely negative, which is exactly what is
expected, given that this principle is not a proper object of study for natural
philosophy. More directly, if one looks at what Aristotle says on the topic
of the first principle, one quickly realizes that there is not much of a
positive description beyond the fact that this principle is a mover. Sure,
the first principle is an unmoved mover. But this is a feature shared with a
number of other (lesser) principles. The analysis of self-motion offered in
Physics viii.5 suggests that souls (both human and animal souls) are
unmoved movers. One might reply that the first unmoved mover is eternal,
whereas these other unmoved movers are perishable. One might also add
that the first principle is absolutely unmoved, whereas the souls are
incidentally moved. Still, it is not clear that these qualifications get us
any closer to the essential nature of the first principle. For a positive
treatment of this principle we have to look elsewhere. In Lambda,
Aristotle has a great deal to say on the first principle. He describes it as a
substance whose essence is pure actuality. He also identifies it with nous
and the good. Clearly, these positive descriptions are not available in
Physics viii. Physics viii does not even refer to this principle as a divine
principle (a god). By contrast, in Lambda, Aristotle does not only refer to
the first principle as a god; he also identifies it with the good (Lambda 7,
1072b25, 28–29, 30).

VIII The scientific definition of eternal motion

I have argued that Physics viii is a single extended argument concerned
with eternal motion. The unity and integrity of this argument becomes
fully apparent as soon as the theory of science outlined in the Posterior
Analytics is applied to Physics viii. This theory requires Aristotle first to
 The argument of Physics viii 283
establish that there is eternal motion and then to look for a definition that
singles out the cause of the eternity of motion. To understand, however,
what gets defined, and thereby explained, we need to recall that the
argument advanced in Physics viii establishes not only the existence of
eternal motion, but also the existence of a single eternal motion. The
transition from the existence of eternal motion – in the sense of the
existence of one unbroken chain of motion with no beginning and no
end – to the existence of a single eternal motion is made at the end of Physics
viii.6. What gets defined, and thereby explained, is the latter, the single
eternal motion that is identical with the circular motion caused by the first
principle.
Admittedly, Aristotle does not give us such a definition. But at this point
it is not difficult to offer a definition that, in addition to giving us the most
precise causal description of the (efficient) cause of eternal motion, also
specifies the type of motion as well as the relevant subject of that motion:
Eternal motion is the primary circular locomotion of that which is moved
by the first eternal unmoved mover which is indivisible, without parts, and
without magnitude.
 Bibliography

A. Ancient, medieval, and Renaissance tradition

Aquinas, Thomas 1884. In viii Libros Physicorum Aristotelis (Opera Omnia, ed.
Leonina). Rome
Averroes 1562. Aristotelis De Physico Audito Libri Octo cum Averrois Commentariis
(vol. iv In Aristotelis opera quae extant . . . Averrois Cordubensis in ea opera
omnes qui a nos pervenere commmentairi). Venice (facsimile reprint Frankfurt
am Main: Minerva 1962)
Blackwell, R. J., R. J. Spath, and W. E. Thirlkel 1884. Commentary on Aristotle’s
Physics by St Thomas Aquinas. New Haven: Yale University Press
Bodnár, I., M. Chase, and M. Share 2012. Simplicius, On Aristotle’s Physics 8.1–5.
London, Ithaca, NY: Duckworth and Cornell University Press
Bruns, I. (ed.) 1887. Alexandri Aphrodisiensis praetor commentaria scripta minora:
De Anima liber cum mantissa. Commentaria in Aristotelem Graeca 2.1.
Berlin.
Coninbricenses (Coimbra Commentators) 1594. In Octos Libros Physicorum
Aristotelis. Lyons (facsimile reprint Hildesheim: G. Olms 1984)
Diels, H. (ed.) 1882. Simplicius, Aristotelis Physicorum Libros Quattuor Priores
Commentaria, in his Commentaria in Aristotelem Graeca, vol. ix. Berlin:
G. Reimer
(ed.) 1895. Simplicii in Aristotelis Physicorum Libri Quattuor Posteriores
Commentaria, in his Commentaria in Aristotelem Graeca, vol. x. Berlin:
G. Reimer
Hayduck, M. (ed.) 1891. Alexander of Aphrodisias, in Aristotelis Metaphysica
Commentaria. Berlin: G. Reimer
Hossfeld, P. (ed.) 1987–93. Albertus Magnus. Physica (Albertus Magnus, Opera
Omnia, ed. B. Gayer, vol. iv 1–2). Ashendorf
Hussey, Edward (ed. and trans.) 1983. Aristotle’s Physics Books iii and iv. Oxford:
Clarendon Press
Jowett, Benjamin (ed. and trans.) 1875. Timaeus: Plato. Translated with an
introduction by Benjamin Jowett. Oxford
McKirahan, R. 2001. Simplicius, On Aristotle’s Physics 8.6–10. London, Ithaca, NY:
Duckworth and Cornell University Press

284
 Bibliography 285
Pacius, Julius 1595. Aristotelis Naturalis Auscultationis libri viii. Frankfurt am Main
(facsimile repr. Frankfurt am Main: Minerva 1964)
1596. Aristotelis Naturalis Auscultationis. Frankfurt am Main (repr. Hildesheim:
G. Olms 1964)
Rashed, M. 2005. Aristote: De la génération et la corruption. Edition, traduction et
commentaire. Paris: Les Belles Lettres
(ed.) 2011. Alexandre d’Aphrodise, Commentaire perdu à la Physique d’Aristote
(Livres iv–viii). Berlin, New York: Walter de Gruyter
Ross, W. D. (ed.) 1928. The Works of Aristotle. Translated into English under the
editorship of W. D. Ross. Oxford: Clarendon Press
Schenkl, H. (ed.) 1900. Themistii in Aristotelis Physica Paraphrasis. Commentaria
in Aristotelem Graeca, vol. v. 2. Berlin: G. Reimer
Todd, R. B. 2008. Themistius, Aristotle, Physics 5–8. London, Ithaca, NY:
Duckworth and Cornell University Press
Vitelli, G. (ed.) 1887. Ioannis Philoponi in Aristotelis Physicorum Libri Quinque
Posteriores Commentaria. Commentaria in Aristotelem Graeca, vol. xii. Berlin:
G. Reimer
(ed.) 1887. Philoponus, Commentary on Aristotle’s ‘Physics’, Commentaria in
Aristotelem Graeca, vols. xvi–xviii. Berlin: G. Reimer
Zabarella, J. 1602. Opera Physica. Frankfurt am Main (facsimile repr. Verona:
Minerva 2009)
1607. De Rebus Naturalibus Libri xxx. Frankfurt am Main (facsimile repr.
Frankfurt am Main: Minerva 1996)
1608. Opera Logica. Frankfurt am Main (facsimile repr. Frankfurt am Main:
Minerva 1966)

B. Modern translations, monographs, and references

Anagnostopoulos, Andreas 2010. ‘Change in Aristotle’s Physics iii’, Oxford Studies
in Ancient Philosophy 39: 33–79
Anagnostopoulos, Georgios (ed.) 2009. A Companion to Aristotle. Chichester:
Wiley–Blackwell
Angioni, L. 2010. Aristóteles: Física i–ii. Campinas: Editora Unicamp
Balme, D. M. 1987. ‘Aristotle’s Biology was not Essentialist’, in Gotthelf and
Lennox (eds.), 291–312
Barnes, J. (ed.) 1984. Revised Oxford Translation: The Complete Works of Aristotle.
Princeton University Press
1993. Aristotle, Posterior Analytics. Oxford: Clarendon Press
Barnes, J., M. Schofield, and R. Sorabji (eds.) 1975. Articles on Aristotle, Volume i:
Science. London: Duckworth
Beere, J. 2009. Doing and Being – An Interpretation of Aristotle’s Metaphysics
Theta. Oxford University Press
Berti, E. (ed.) 1981. Aristotle on Science: The Posterior Analytics. Padua: Editrice
Antenore
286 Bibliography
Bodnár, I. 1997. ‘Movers and Elemental Motions in Aristotle’, Oxford Studies in
Ancient Philosophy 15: 81–117
2012. ‘Commentary on 246b3–246b20’, in Maso, Natali, and Seel (eds.), 73–84
Bolton, R. 1976. ‘Essentialism and Semantic Theory in Aristotle: Posterior
Analytics ii 7–10’, Philosophical Review 85: 514–544
1987. ‘Definition and Scientific Method in Aristotle’s Posterior Analytics and
Generation of Animals’, in Gotthelf and Lennox (eds.), 120–166
1991. ‘Aristotle’s Method in Natural Science: Physics i’, in Judson (ed.), 1–29
1997. ‘The Material Cause: Matter and Explanation in Aristotle’s Natural
Science’, in Kullmann and Föllinger (eds.), 97–126
2011. ‘Why Does Aristotle Need Four Causes?’, in Couloubaritsis and
Delcomminette (eds.), 27–46
Bonitz, H. 1961. Index aristotelicum. Hildesheim: G. Olms
1969. Aristotelische Studien: Fünf Teile in einem Band. First edition, Berlin 1862.
Berlin: Walter de Gruyter
Bostock, D. 2006. Space, Time, Matter, and Form: Essays on Aristotle’s Physics.
Oxford University Press
Bowen, A. C. and C. Wildberg (eds.) 2009. New Perspectives on Aristotle’s De
Caelo. Louvain: Brill Academic Publishers
Bowin, J. 2011. ‘Aristotle on Various Types of Alteration in De Anima ii.5’,
Phronesis 56: 138–161
Bradie, M. and F. D. Miller Jr. 1999. ‘Teleology and Natural Necessity’, in
L. D. P. Gerson (ed.), Aristotle, Critical Assessments. London: Routledge,
75–89; reprint from (1984) History of Philosophy Quarterly 1: 133–146
Broadie [Waterlow], S. 1982. Nature, Change, and Agency in Aristotle’s Physics.
Oxford: Clarendon Press
Broadie, S. 1990. ‘Nature and Craft in Aristotelian Teleology’, in Devereux and
Pellegrin (eds.), 389–403
2004. ‘On Generation and Corruption i.4: Distinguishing Alteration –
Substantial Change, Elemental Change, and First Matter in GC’, in De
Haas and Mansfeld (eds.), 123–150
2009. ‘Heavenly Bodies and First Causes’, in G. Anagnostopoulos (ed.),
230–241
Brunschwig, J. 1991. ‘Qu’est-ce que “La Physique” d’Aristote?’, in F. de Gandt and
S. Souffrin (eds.), La Physique d’Aristote et les Conditions d’une Science de la
Nature. Paris: Vrin, 11–40
Burnyeat, M. F. 1980. ‘Aristotle on Learning to be Good’, in Rorty (ed.), 69–92
Cambiano, G. 2012. ‘The Desire to Know: Metaphysics A1’, in Steel (ed.), 1–42
Carteron, H. (ed.) 1926–31. Aristote: Physique. 2 vols. Paris: Les Belles Lettres
Castelli, L. M. 2012. Aristotele: Fisica, Libro iv. Rome: Carocci Editore
Cavagnaro, E. 2002. Aristotele e il Tempo. Bologna: Il Mulino
Charles, D. 1984. Aristotle’s Philosophy of Action. London: Duckworth
1991. ‘Teleological Causation in the Physics’, in Judson (ed.), 101–128
2000. Aristotle on Meaning and Essence. Oxford University Press
(ed.) 2010a. Definition in Greek Philosophy. Oxford University Press
 Bibliography 287
2010b. ‘Definition and Explanation in the Posterior Analytics and Metaphysics’,
in Charles (ed.), 286–328
2010c. ‘Metaphysics Theta 7 and 8: Some Issues Concerning Actuality and
Potentiality’, in Lennox and Bolton (eds.), 168–197
2012. ‘Teleological Causation’, in Shields (ed.), 227–266
Charlton, W. 1970. repr. 1992. Aristotle’s Physics Books 1 & 2: Translated from the
Greek with Introduction and Notes. Oxford: Clarendon Press
Cherniss, H. F. 1944. Aristotle’s Criticism of Plato and the Academy. Baltimore:
Johns Hopkins University Press
Code, A. 1976. ‘The Persistence of Aristotelian Matter’, Philosophical Studies 29.6:
357–367
1987. ‘Soul as Efficient Cause in Aristotle’s Embryology’, Philosophical Topics
15.2: 51–59
1988. ‘Reply to Michael Frede’s “Being and Becoming in Plato”’, Oxford Studies
in Ancient Philosophy suppl. vol., 53–60
1995. ‘Potentiality in Aristotle’s Science and Metaphysics’, Pacific Philosophical
Quarterly 76.3–4: 405–418
1997. ‘The Priority of Final Causes over Efficient Causes in Aristotle’s PA’, in
Kullmann and Föllinger (eds.), 127–143
Coope, Ursula 2004. ‘Aristotle’s Account of Agency in Physics iii.3’, Proceedings of
the Boston Area Colloquium in Ancient Philosophy 20: 201–221
2005. Time for Aristotle: Physics iv.10–14. Oxford University Press
2007. ‘Aristotle on Action’, Aristotelian Society suppl. vol., 81.1: 109–138
2009. ‘Change and its Relation to Potentiality and Actuality’, in
G. Anagnostopoulos (ed.), 277–291
2012. ‘Commentary on 246a10–246b3’, in Maso, Natali, and Seel (eds.), 57–72
in press. ‘Aristotle, De Motu Animalium, Chapters 2 and 3’, in Primavesi and
Rapp (eds.)
Cooper, J. M. 1982. ‘Aristotle on Natural Teleology’, in Schofield and Nussbaum
(eds.), 197–222
1987. ‘Hypothetical Necessity and Natural Teleology’, in A. Gotthelf and
J. G. Lennox (eds.) Philosophical Issues in Aristotle’s Biology. Cambridge
University Press, 243–274
1996. ‘Reason, Moral Virtue, and Moral Value’, in Frede and Striker (eds.),
81–114
Corcilius, K. 2011. ‘Aristotle’s Definition of Non-Rational Pleasure and Pain and
Desire’, in I. Miller (ed.), Aristotle’s Nicomachean Ethics: A Critical Guide.
Cambridge University Press, 117–143
Corcilius, K. and P. Gregoric 2010. ‘Separability vs. Difference: Parts and
Capacities of the Soul in Aristotle’, Oxford Studies in Ancient Philosophy 39:
81–119
Couloubaritsis, L. and S. Delcomminette (eds.) 2011. La Causalité chez Aristote.
Paris, Brussels: Vrin/Ousia
Crisp, R. (trans.) 2000. Aristotle: Nicomachean Ethics. Cambridge University
Press
288 Bibliography
Crivelli, P. 2012. Plato’s Account of Falsehood: A Study of the Sophist. Cambridge,
New York: Cambridge University Press
Davidson, Donald 1970. ‘Events as Particulars’, Noûs 4: 25–32
De Haas, F. A. J. 2012. ‘Commentary on 247b1–247b13’, in Maso, Natali, and Seel
(eds.), 99–108
De Haas, Frans and Jaap Mansfeld (eds.) 2004. Aristotle: On Generation and
Corruption, Book i. Symposium Aristotelicum. Oxford: Clarendon Press
Denniston, D. 1934. The Greek Particles. Oxford University Press
Depew, D. 1997. ‘Etiological Approaches to Biological Aptness in Aristotle and
Darwin’, in Kullmann and Föllinger (eds.), 209–227
Devereux, D. 2011. ‘Aristotle on the Form and Definition of a Human Being:
Definitions and their Parts in Metaphysics Z.10 & 11’, Proceedings of the Boston
Area Colloquium in Ancient Philosophy 26: 167–196
Devereux, D. and P. Pellegrin (eds.) 1990. Biologie, Logique et Métaphysique chez
Aristote. Paris: Editions du Centre National de la Recherche Scientifique
Dirlmeier, F. (trans.) 1956. Aristoteles Nikomachische Ethik. Berlin: Akademie
Verlag
Dretske, F. I. 1980. ‘The Intentionality of Cognitive States’, Midwest Studies in
Philosophy 5: 281–294
Düring, I. (ed.) 1969. Naturphilosophie bei Aristoteles und Theophrast. Heidelberg:
Lothar Stiehm Verlag
Ebrey, D. unpublished ‘Aristotle’s Motivation for Matter’. UCLA dissertation
Falcon, Andrea 2005. Aristotle and the Science of Nature: Unity without Uniformity.
Cambridge University Press
Fazzo, S. 2013. ‘Heavenly Matter in Aristotle, Metaphysics Lambda 2’, Phronesis 58:
160–175
Fine, K. 1999. ‘Things and their Parts’, Midwest Studies in Philosophy 23.1: 61–74
Föllinger, S. (ed.) 2010. ‘Was ist “Leben”? Aristoteles’ Anschauungen zur
Entstehung und Funktionsweise von “Leben”’, Philosophie der Antike 27.
Frede, M. 1980. ‘The Original Notion of Cause’, in Schofield, Burnyeat, and
Barnes (eds.), 217–249
1990. ‘An Empiricist View of Knowledge: Memorism’, in Stephen Everson (ed.)
Companions to Ancient Thought, Volume i: Epistemology. Cambridge
University Press, 225–250
1996. ‘Aristotle’s Rationalism’, in Frede and Striker (eds.), 157–173
2000. ‘Introduction’, in Frede and Charles (eds.), 1–52
2004. ‘Aristotle’s Account of the Origin of Philosophy’, Rhizai 1: 9–44
Frede, M. and D. Charles (eds.) 2000. Aristotle’s Metaphysics Book Lambda.
Oxford: Clarendon Press
Frede, M. and G. Striker (eds.) 1996. Rationality in Greek Thought. Oxford
University Press
Freeland, C. 1987. ‘Aristotle on Bodies, Matter and Potentiality’, in James
G. Lennox and Allan Gotthelf (eds.), Philosophical Issues in Aristotle’s
Biology. Cambridge University Press, 392–407
 Bibliography 289
Freudenthal, G. 1995. Aristotle’s Theory of Material Substance, Heat and Pneuma,
Form and Soul. Oxford: Clarendon Press
Galton, A. and R. Mizoguchi 2009. ‘The Water Falls but the Waterfall does
not Fall: New Perspectives on Objects, Processes and Events’, Applied
Ontology 4.2: 71–107
Gauthier, R. E. and J. Y. Jolif (trans.) (1968) L’ Ethique à Nicomaque:
Introduction, traduction et commentaire. 2 vols. Louvain, Paris: Publications
Universitaires
Gill, M. L. 1980. ‘Aristotle’s Theory of Causal Action in Physics iii 3’, Phronesis
25.2: 129–147
1989. Aristotle on Substance: The Paradox of Unity. Princeton University Press
2009. ‘The Theory of the Elements in De Caelo 3 and 4’, in Bowen and
Wildberg (eds.), 139–162
Gill, M. L. and P. Pellegrin (eds.) 2006. Blackwell Companions to Philosophy, A
Companion to Ancient Philosophy. Malden, MA: Blackwell Publishing Ltd
Gill, M. L. and J. G. Lennox (eds.) 1994. Self-Motion: From Aristotle to Newton.
Princeton University Press
Godfrey-Smith, P. 1994. ‘A Modern History Theory of Functions’, Nous 28.3:
344–362
2014. The Philosophy of Biology. Princeton University Press
Gotthelf, A. (ed.) 1985. Aristotle on Nature and Living Things. Pittsburgh, PA,
Bristol (UK): Methesis
2012. Teleology, First Principles and Scientific Method in Aristotle’s Biology.
Oxford University Press
Gotthelf, A. and J. Lennox (eds.) 1987. Philosophical Issues in Aristotle’s Biology.
Cambridge University Press
Graham, Daniel 1999. Aristotle Physics Book viii. Translated with a Commentary.
Oxford: Clarendon Press
Grant, A. (ed.) 1866. The Ethics of Aristotle. 2 vols. London: Longmans, Green,
and Co.
Guthrie, W. K. C. 1946. ‘Notes on Some Passages in the Second Book of Aristotle’s
Physics’, Classical Quarterly 40: 70–76
Guyer, P. (ed. and trans.) and E. Matthews (trans.) 2000. Kant’s Critique of the
Power of Judgment (The Cambridge Edition of the Works of Immanuel Kant).
Cambridge University Press, 2000.
Halliwell, S. 2002. The Aesthetics of Mimesis. Princeton University Press
Hamelin, O. (trans.) 1931. Aristote: Physique ii, Traduction et Commentaire. Paris:
Vrin
Hardie, R. P. and R. K. Gaye 1930. ‘Physics’, in Ross (ed.), 315–446
Haslanger, S. 2003. ‘Persistence through Time’, in Loux and Zimmerman (eds.),
315–354
Heinaman, Robert 1994. ‘Is Aristotle’s Definition of Change Circular?’, Apeiron
27: 25–37
Hursthouse, Rosalind 2006. ‘The Central Doctrine of the Mean’, in Kraut (ed.),
96–115
290 Bibliography
Irwin, T. 1988. Aristotle’s First Principles. Oxford: Clarendon Press
1990. Aristotle’s First Principles. Oxford University Press
2006. ‘Aquinas, Natural Law, and Aristotelian Eudaimonism’, in Kraut (ed.),
323–341
Irwin, T. and G. Fine 1995. Aristotle: Selections. Indianapolis, IN: Hackett
Joachim, H. H. 1922. Aristotle. On Coming-to-be and Passing-Away: A Revised
Text with Introduction and Commentary. Oxford: Clarendon Press
Johansen, T. K. 2004. Plato’s Natural Philosophy. Cambridge University Press
2012a. ‘Capacity and Potentiality: Aristotle’s Metaphysics Theta 6–7 from the
Perspective of the De Anima’, Topoi 31.2: 209–220
2012b. The Powers of Aristotle’s Soul. Oxford University Press
Johnson, M. R. (2005). Aristotle on Teleology. Oxford University Press
Johnson, M. R. and D. S. Hutchinson 2005. ‘Authenticating Aristotle’s
Protrepticus’, Oxford Studies in Ancient Philosophy 29: 193–294
Johnson, W. E. 1921. Logic. Part 1. Cambridge University Press
Jones, B. 1974. ‘Aristotle’s Introduction of Matter’, Philosophical Review 83.4:
474–500
Judson, L. (ed.) 1991a. Aristotle’s Physics: A Collection of Essays. Oxford: Clarendon
Press
1991b. ‘Chance and “Always or For the Most Part” in Aristotle’, in Judson (ed.),
73–100
2005. ‘Aristotelian Teleology’, Oxford Studies in Ancient Philosophy 29: 341–366
Kahn, C. 1985. ‘The Place of the Prime Mover in Aristotle’s Teleology’, in Gotthelf
(ed.), 183–205
Katayama, Errol G. 1999. Aristotle on Artifacts. Albany: State University of New
York Press
2009. ‘Necessity and Artifacts in Aristotle’, unpublished manuscript
2011. ‘Soul and Elemental Motion in Aristotle’s Physics viii.4’, Apeiron 44.2:
163–190
Kelsey, S. 2003. ‘Aristotle’s Definition of Nature’, Oxford Studies in Ancient
Philosophy 25: 59–87
2008. ‘The Place of I 7 in the Argument of Physics i’, Phronesis 53.2: 180–208
2010. ‘Hylomorphism in Aristotle’s Physics’, Ancient Philosophy 30: 107–124
Kirwan, C. 1971. Aristotle’s Metaphysics, Books Γ, Δ, and E. Oxford: Clarendon
Press
Kosman, Aryeh 1969. ‘Aristotle’s Definition of Motion’, Phronesis 14.1: 41–62
1994. ‘Aristotle’s Prime Mover’, in Gill and Lennox (eds.), 135–154
2013. The Activity of Being: An Essay on Aristotle’s Ontology. Cambridge, MA:
Harvard University Press
Kostman, James 1987. ‘Aristotle’s Definition of Change’, History of Philosophy
Quarterly 4.1: 3–16
Kraut, R. (ed.) 2006. The Blackwell Guide to Aristotle’s Nicomachean Ethics.
Malden: Blackwell Publishing
Kühner, R. and B. Gerth 1892–98. Ausführliche Grammatik der griechischen
Sprache. 2 parts, 4 vols. Hanover: Hahnsche
 Bibliography 291
Kullmann, W. and S. Föllinger (eds.) 1997. Aristotelische Biologie: Intentionen,
Methoden, Ergebnisse. Stuttgart: Steiner Verlag
Lacey, A. R. (trans.) 1997. Simplicius: On Aristotle’s Physics 2. London, Ithaca, NY:
Duckworth and Cornell University Press
Lawrence, G. 2011. ‘Acquiring Character: Becoming Grown-Up’, in Pakaluk and
Pearson (eds.), 233–283
Lennox, J. G. 1982. ‘Teleology, Chance and Aristotle’s Theory of Spontaneous
Generation’, Journal of the History of Philosophy 20: 219–238
1984. ‘Aristotle on Chance’, Archiv für Geschichte der Philosophie 66: 52–60
2001a. Aristotle’s Philosophy of Biology. Cambridge University Press
2001b. Aristotle’s Parts of Animals Books i-iv. Oxford: Clarendon Series
2004. ‘The Place of Zoology in Aristotle’s Natural Philosophy’, in R. W. Sharples
(ed.), Philosophy and the Sciences in Antiquity. London: Ashgate, 58–70
2005. ‘Getting a Science Going: Aristotle on Entry Level Kinds’, in Wolters
(ed.), 87–100
2006. ‘Aristotle’s Biology and Aristotle’s Philosophy’, in Gill and Pellegrin
(eds.), 292–315
2008. ‘As if we were Investigating Snubness: Aristotle on the Prospects for a
Single Science of Nature’, Oxford Studies in Ancient Philosophy 35: 149–186
2009a. ‘De Caelo 2.2 and its Debt to De incessu animalium’, in Bowen and
Wildberg (eds.), 187–214
2009b. ‘Aristotle on Mind and the Science of Nature’, in Rossetto, Tsianikas,
Couvalis, and Palaktsoglou (eds.), 1–18
2010. ‘Aristotle’s Natural Science: The Many and the One’, in James Lesher
(ed.), From Inquiry to Demonstrative Knowledge: New Essays in Aristotle’s
Posterior Analytics, Apeiron 43.2–3 (June–September): 1–23
in press. Seeking and Knowing: Aristotle on Norms of Inquiry. Cambridge
University Press
Lennox, J. G. and R. Bolton (eds.) 2010. Being, Nature, and Life in Aristotle: Essays
in Honor of Allan Gotthelf. Cambridge University Press
Leunissen, M. 2009. ‘Why Stars Have No Feet: Teleological Explanations in
Aristotle’s Cosmology’, in Bowen and Wildberg (eds.), 245–271
2010. Explanation and Teleology in Aristotle’s Science of Nature. Cambridge
University Press
2012. ‘Aristotle on Natural Character and its Implications for Moral
Development’, Journal of the History of Philosophy 50.4: 507–530
2013. ‘Becoming Good Starts with Nature: Aristotle on the Moral Advantages
and the Heritability of Good Natural Character’, Oxford Studies in Ancient
Philosophy 44: 99–127
Lewis, F. 1988. ‘Teleology and Material/Efficient Causes in Aristotle’, Pacific
Philosophical Quarterly 69: 54–98
Loux, M. J. and D. W. Zimmerman (eds.) 2003. The Oxford Handbook of
Metaphysics. Oxford University Press
Mansion, A. 1946. Introduction à la Physique Aristotélicienne. 2nd edn. Louvain
and Paris: L’Institut Supérieur de Philosophie
292 Bibliography
Marmadoro, Anna 2007. ‘The Union of Cause and Effect in Aristotle: Physics iii
3’, Oxford Studies in Ancient Philosophy 32: 205–232
Maso, S., C. Natali, and G. Seel (eds.) 2012. ‘Reading Aristotle’s Physics vii.3:
“What is Alteration?”’ Proceedings of the International ESAP–HYELE
Conference. Las Vegas: Parmenides Publishing
Matthen, M. 2001. ‘Holistic Presuppositions of Aristotle’s Cosmology’, Oxford
Studies in Ancient Philosophy 20: 171–199
Meikle, Scott 1994. ‘Aristotle on Money’, Phronesis 39.1: 26–44
Menn, S. in press. ‘Physics ii and its Relation to Physics i–iv’
Millikan, R. 1989. ‘In Defense of Proper Functions’, Philosophy of Science 56:
288–302
Moravcsik, J. 1994. ‘Essences, Powers, and Generic Propositions’, in Scaltsas,
Charles, and Gill (eds.), 229–344
Morison, Benjamin in press. ‘Aristotle, De Motu Animalium, Chapters 4 and 5’, in
Primavesi and Rapp (eds.)
Morrison, D. 1985. ‘Separation in Aristotle’s Metaphysics’, Oxford Studies in
Ancient Philosophy 3: 125–157
Moss, J. 2011. ‘“Virtue Makes the Goal Right”: Virtue and Phronesis in Aristotle’s
Ethics’, Phronesis 56.3: 204–261
2012. Aristotle on the Apparent Good, Perception, Phantasia, Thought, and Desire.
Oxford University Press
Nagel, E. 1977. ‘Teleology Revisited: Goal-Directed Processes in Biology’, Journal
of Philosophy 74: 261–301
2012. Mind and Cosmos. Oxford University Press
Natali, C. 2012. ‘Commentary on 247b13–248a9’, in Maso, Natali, and Seel (eds.),
109–117
Nussbaum, M. 1978. Aristotle’s De Motu Animalium. Princeton University Press
Oehler, K. 1984. Der Unbewegte Beweger des Aristoteles. Frankfurt am Main:
V. Klostermann
Pakaluk, M. and G. Pearson (eds.) 2011. Moral Psychology and Human Action in
Aristotle. Oxford University Press
Pellegrin, P. 2000. Aristote, Physique. Paris: Flammarion
Penner, Terrence 1970. ‘Verbs and the Identity of Actions – A Philosophical
Exercise in the Interpretation of Aristotle’, in O. P. Wood and G. Pitcher
(eds.), Ryle: A Collection of Critical Essays. London: Macmillan, 393–453
Peramatzis, M. 2011. Priority in Aristotle’s Metaphysics (Oxford Aristotle Studies).
Oxford University Press
Polansky, R. 2007. Aristotle’s De Anima. Cambridge University Press
Prantl, K. (trans.) [1854] 1978. Repr. Aristoteles’ Acht Bücher Physik. Leipzig: Aalen
Primavesi, O. and C. Rapp (eds.) in press. ‘Aristotle’s De Motu Animalium’,
Proceedings of the XIX Symposium Aristotelicum. Oxford University Press
Quarantotto, D. 2005. Causa Finale, Sostanza, Essenza in Aristotele. Naples:
Bibliopolis
2010. ‘Aristotle on the Soul as a Principle of Biological Unity’, in Follinger (ed.),
35–53
 Bibliography 293
2014. ‘Aristotle’s Natural Teleology Seen from Above: A “Cosmogony” of the
Means–Goal Relation’, in M. Tuominen, S. Heinämaa, and V. Mäkinen
(eds.), Aristotelianism and the Critique of Modernity. Leiden, Boston: Brill
Reeve, C. D. C. 1992. Practices of Reason: Aristotle’s Nicomachean Ethics. New
York: Oxford University Press
2012. Action, Contemplation, and Happiness: An Essay on Aristotle. Cambridge,
MA: Harvard University Press
Roark, T. 2012. Aristotle on Time. Cambridge University Press
Rorty, A. O. (ed.) 1980. Essays on Aristotle’s Ethics. Berkeley: California University
Press
Ross, W. D. (ed.) 1924. Aristotle’s Metaphysics: A Revised Text with Introduction
and Commentary. Oxford: Clarendon Press
(ed.) 1936. Aristotle’s Physics: A Revised Text with Introduction and Commentary.
Oxford: Clarendon Press
(ed.) 1949. Aristotle’s Prior and Posterior Analytics: A Revised Text with
Introduction and Commentary. Oxford: Clarendon Press
1950. Aristotelis Physica, recognovit brevique adnotatione critica instruxit W. D. Ross.
Oxford: Clarendon Press
Rossetto, M., M. Tsianikas, G. Couvalis, and M. Palaktsoglou (eds.) 2009.
‘Greek Research in Australia’, Proceedings of the Biennial International
Conference of Greek Studies, June 2009. Flinders University Department
of Languages
Rowe, C. and S. Broadie (trans. and comm.) 2002. Aristotle: Nicomachean Ethics.
Oxford University Press
Sauvé Meyer, S. 1992. ‘Aristotle, Teleology, and Reduction’, The Philosophical
Review 101: 791–825
Scaltsas, T., D. Charles, and M. L. Gill (eds.) 1994. Unity, Identity, and
Explanation in Aristotle’s Metaphysics. Oxford University Press.
Scharle, Margaret 2008. ‘Elemental Teleology in Aristotle’s Physics ii.8’, Oxford
Studies in Ancient Philosophy 34: 147–183
Schofield, M. and M. C. Nussbaum (eds.) 1982. Language and Logos. Cambridge
University Press
Schofield, M., M. F. Burnyeat, and J. Barnes (eds.) 1980. Doubt and Dogmatism.
Oxford: Clarendon Press
Sedley, D. 1991. ‘Is Aristotle’s Teleology Anthropocentric?’ Phronesis 36: 179–197
2000. ‘Metaphysics Λ 10’, in Frede and Charles (eds.), 327–350
2007. Creationism and its Critics in Antiquity. Berkeley: University of California
Press
2010. ‘Teleology, Platonic and Aristotelian’, in Lennox and Bolton (eds.),
5–29
Shields, C. 2008. ‘Substance and Life in Aristotle’, in J. Mouracade (ed.), Aristotle
on Life, (Apeiron, special issue, 41.3), 129–152
2012a. The Oxford Handbook to Aristotle. Oxford University Press
2012b. ‘Being qua Being’, in Shields (ed.), The Oxford Handbook to Aristotle,
343–371
294 Bibliography
Solmsen, F. 1960. Aristotle’s System of the Physical World: A Comparison with his
Predecessors. Cornell University Press
Steel, C. (ed.) 2012. Aristotle’s Metaphysics Alpha: Symposium Aristotelicum. Oxford
University Press
Stewart, J. A. 1892. Notes on the Nicomachean Ethics of Aristotle. 2 vols. Oxford:
Clarendon Press
Stout, R. 1996. Things That Happen Because They Should. Oxford University
Press
Strohl, M. S. 2011. ‘Pleasure as Perfection: Nicomachean Ethics 10.4–5’, Oxford
Studies in Ancient Philosophy 41: 257–287
Taylor, C. C. W. (trans.) 2006. Aristotle’s Nicomachean Ethics Books ii–iv. Oxford
University Press
Torstrik, A. 1875. ‘Peri Tuchês kai tou Automatou’, Hermes 9: 425–470
Tracey, T. J. 1969. Physiological Theory and the Doctrine of the Mean in Plato and
Aristotle. The Hague: Mouton
Ugaglia, M. 2012. Aristotele: Fisica, Libro iii. Rome: Carocci Editore
Viano, C. 2012. ‘Commentary on 246b20–247a19’, in Maso, Natali, and Seel
(eds.), 85–98
Wagner, H. (trans.) 1989. Aristoteles: Physikvorlesung. Berlin: Akademie Verlag
Wardy, R. 1990. The Chain of Change: A Study of Aristotle’s Physics vii. Cambridge
University Press
Waterfield, R. 1996. Aristotle’s Physics (translation; with introduction and notes
by D. Bostock). Oxford University Press
Waterlow (Broadie), Sarah 1982. Nature, Change and Agency in Aristotle’s Physics.
Oxford: Clarendon Press
White, M. J. 2009. ‘Aristotle on the Infinite, Space, and Time’, in
G. Anagnostopoulos (ed.), 260–276
Whiting, J. 1986. ‘Human Nature and Intellectualism in Aristotle’, Archiv für
Geschichte der Philosophie 68.1: 70–95
2002. ‘Locomotive Soul: The Parts of Soul in Aristotle’s Scientific Works’,
Oxford Studies in Ancient Philosophy 22: 141–200
Wicksteed, Ph. H. and F. M. Cornford 1929. Aristotle, The Physics. 2 vols. The
Loeb Classical Library. Cambridge, MA: Harvard University Press
Wieland, W. 1970. Die aristotelische Physik. Göttingen: Vandenhoeck & Ruprecht
1975. ‘The Problem of Teleology’, in Barnes, Schofield, and Sorabji (eds.),
141–160
Williams, C. J. F. 1982. Aristotle’s De Generatione et Corruptione (translation
with notes). Oxford: Clarendon Press
Witt, Charlotte 1989. Substance and Essence in Aristotle. Ithaca, NY: Cornell
University Press
2012. ‘Aristotle on Deformed Animal Kinds’, Oxford Studies in Ancient
Philosophy 43 (winter): 83–106
2015. ‘As if by Convention Alone: The Unstable Ontology of Aristotle’s Ethics’,
in Devin Henry and Karen M. Nielsen, Bridging the Gap between Aristotle’s
Science and Ethics. Cambridge University Press.
 Bibliography 295
Wolters, G. (ed.) 2002. Homo Sapiens und Homo Faber. Festschrift Mittelstrass.
Berlin: Walter De Gruyter
Wright, L. 1973. ‘Functions’, The Philosophical Review 82.2: 139–68
1976. Teleological Explanations: An Etiological Analysis of Goals and Functions.
Berkeley, CA: University of California Press
Zeller, E. 1879. Die Philosophie der Griechen in ihrer geschichtlichen Entwicklung.
Part ii, section 2. Leipzig: R. Reisland
 Index

Action 202, 225, 231, 234 Embryology 59, 237

Activity 190–205, 237–9, 250, 254, 256, 258 Empedocles 42, 67, 89, 121, 140, 268–70
Actuality 37, 62, 187–204, 236–7, 254, 256, 282 Ends/Functions/Purposes (see causes, final)
Antiphon 33, 169, 171 Event 204
Aporia 21–2, 60, 89, 198, 201, 281 Explanation 164, 206, 265–83
archê (see first principle)
Art/Craft (analogy with nature) 2, 37, 49, 52, 54, First Principle 48, 163, 168, 192, 267–8, 273, 282
82, 105, 107, 128, 156, 162 Fitness 225, 230, 233
Atomists 121 Form (see also causes, formal)
Automata 52
Generation 26, 46, 53, 58, 91, 97, 99, 109,
Being 145, 162, 164, 166, 169 133, 144, 145, 156, 163, 183, 184, 229,
231, 276
Capacity 37, 63, 190–204, 218–23, 232, 236–44 Geometry 194, 219, 270
Causes/Causation 38
Efficient causes 55, 60, 123, 276 Habituation 225–44
Final causes 2, 23, 25, 28–9, 31, 33, 36, 42, 68, 82, Heavens/heavenly bodies 99, 206, 247, 278
88, 96, 107, 122, 138, 141, 174, 177, 203, 254, Homynomy 51
256 Hylomorphism 137, 151
Formal causes 121, 131
Material causes 29, 34, 51, 121 Imitation 225–35
Chance 66, 90 Inquiry 14, 19, 21, 24, 28, 166, 174, 270–1
Change Investigation (methods of) 3, 11–14, 20, 22, 28, 40,
Alteration/qualitative change 145, 153, 227, 162, 265–79
228, 231
Locomotion 145, 163, 173, 177, 210, 247, 252, Kinêsis (see change)
261, 272, 277–83 Knowledge
Perfection 228–30, 244 Scientific knowledge 15, 16, 19, 121, 122,
Vs. Process 186–205 145, 268
Substantial change 144, 147, 149 Practical knowledge 236, 239
Character 225, 227–44 Productive knowledge 53
Conceptual analysis 3
Cosmology 206, 223, 269 Life, living 168, 172, 185, 234
Luck 42, 66, 70, 80, 90
Definition 48, 190, 198, 266, 282–3
Democritus 100, 102, 268–70 Mathematics 20, 100
Demonstration 3, 12, 206, 228 Matter (see also causes, material) 148, 154, 157,
169, 191
Education (see Virtue, Habituation, Moral Metaphysics 202, 203, 207, 230
Development) Methodos (see investigation, methods of)
Elements 105, 113, 183 Moral Development 225–44

296
 Index 297
Motion 2, 30, 37 Priority
Self-motion 27, 44, 52, 107, 245–64, 272 In being 51, 136
Eternal, circular motion 177, 178, 180, 206, In causation 96, 254–5, 257
224, 247–50, 259, 263, 266–83 Process (see change)
Rectilinear motion 98, 178, 217, 277–83
Reproduction 111
Natural philosophy/physics
Boundaries of 3, 16, 19–22, 23, 38, 265 Science (see knowledge, scientific)
Methods of (see also investigation) 162 Soul (definition of, dualism with body) 36, 134
Nature/natural entities 2, 31, 32, 40, 47, 58, Space 186
88, 92, 113, 163, 169, 173, 266, 269 Spontaneity 66, 96
Necessity, conditional 23, 29 Sublunary Realm 26, 99, 164, 183, 261
Nous (Anaxagoras on) 101, 126 Substance (natural) 18, 34, 107, 164
Substantial being 16–17, 162
Ousia (see substantial being)
Teleology (see causes, final)
Parmenides 121, 160 Time 207, 208, 209, 213, 218, 268–9
Plato 36, 67, 105, 108, 121, 170, 171, 172, 246, 248, Tragedy 113
250–3, 257–8, 263
Political science 225, 232 Unmoved Movers 27, 206, 245, 247, 249–51,
Potentiality 60, 187–205, 218–23, 250, 253–7 260–2, 265, 268, 274–83
Pre-Socratics 170, 171
Prime Mover 27, 98, 163, 181, 245 Virtue 225–44