Embed Size (px)
Citation preview
TELEOLOGY IN ARISTOTLE AND CONTEMPORARY PHILOSOPHY OF BIOLOGY:
AN ACCOUNT OF THE NATURE OF LIFE
by
RICHARD J. CAMERON
B.A. St. Olaf College, 1991 M.A. University of Colorado Boulder, 1995
A thesis submitted to the faculty of the Graduate School of the
University of Colorado in partial fulfillment of the requirements for the degree of
Doctor of Philosophy Department of Philosophy
2000
This thesis entitled:
Teleology in Aristotle and Contemporary Philosophy of Biology: An Account of the Nature of Life
written by Richard James Cameron
has been approved for the Department of Philosophy
___________________________ Christopher Shields
___________________________ Graham Oddie
___________ Date
The final copy of this dissertation has been examined by the signatories, and we find that both the content and the form meet acceptable presentation standards for scholarly work in the
above-mentioned discipline.
iii
Cameron, Richard James (Ph.D., Philosophy) Teleology in Aristotle and Contemporary Philosophy of Biology:
An Account of the Nature of Life Thesis directed by Professor Christopher Shields
This dissertation undertakes the twofold task interpreting Aristotle's account of the
nature of life and evaluating that account in light of the findings of contemporary science and
philosophy of biology. I argue that Aristotle defines being alive teleologically: a is alive just in case
a possesses intrinsic teleological directedness. Further, this account of life reveals itself to be
philosophically astute and scientifically sound.
The dissertation's central chapters defend these general theses in both scholarly and
contemporary contexts. Given the central place of teleology in the analysans of life, chapters
three and four investigate Aristotle's commitments concerning teleology. I argue that Aristotle
has moderate, plausible and strong reasons for being a realist about teleological commitment,
that the scope of his commitment to teleology is compatible with his defining life teleologically,
and that his conception of the final cause is nonreductive; teleology is, for Aristotle, a sui generis
unreduced causal factor in the structure of the natural world.
This interpretation of Aristotle's ontic commitments vis-à-vis teleology raises serious
doubt concerning the contemporary adequacy of his view. I argue that these doubts are
unfounded. I argue in chapter five that the presumptive contemporary reductivist account of
teleology (the etiological account) fails to provide an adequate theoretical definition of teleology.
Chapter six argues that the failure of the best reductive account does not support eliminitivism
concerning teleology for two reasons. First, common objections to Aristotelian teleology are ill-
founded. Second, we have strong evidence from quantum mechanics that the ontology of the
physical world is � contrary to popular presupposition � congenial to the postulation of sui
generis Aristotelian teleology. If we accept the ontic authority of science rather the modern
presupposition that sui generis teleology is ontically or methodologically unacceptable in a
scientific conception of the world, then Aristotle's commitments reveal themselves to be, again,
plausible, moderate and sophisticated.
iv
With foundational work on the definition's central notion completed, chapter seven
returns to the exegetical task of uncovering the details of Aristotle's account of life. The result
indicates that Aristotle possessed a philosophically sophisticated and defensible conception of
life even in contemporary terms; Aristotle's account of life deserves serious contemporary
consideration.
v
Acknowledgements
It is a pleasure to thank a number of people and institutions for their contribution to
this project. I owe a special thanks to my parents, Bruce and Mary Ann Cameron, who both
encouraged me to excel and allowed me the freedom to choose my avenues for attaining that
goal; as is perhaps common with grown children, the benefits of the kindness and love they have
provided me throughout my development has become increasingly evident as I have grown and
matured. I am profoundly grateful.
My interest in Aristotle's metaphysical theses surrounding the problem of life's definition
took shape in a graduate seminar led by Christopher Shields. I am also deeply indebted to a
number of wise and caring teachers during the early stages of my graduate career (special thanks
to Claudia Mills, Jim Nickel, John Fisher and George Bealer). I must thank my students at the
University of Colorado as well for shaping and honing my skills in communicating clearly the
contours of philosophical debate.
I gratefully acknowledge the support of the George F. Reynolds Fellowship for a year's
support early in the researching and writing of this dissertation and the Department of
Philosophy for one semester's support through an Enrollment Enhancement Fellowship at the
end of the process.
Special thanks go as well to those who read and commented on early drafts of the
dissertation. Rachel Singpurwalla gave pointed and accurate criticism of an early draft of chapter
one. Members of my committee, Gabriela Carone and Robert Pasnau, offered criticism and
guidance that was absolutely essential to the final forms of chapters three and four respectively,
and each offered astute yet friendly criticism of the entire project as it has taken final shape. The
influence of Christopher Shields is felt on every page of the dissertation; I thank him deeply for
his patience and supportive guidance as I developed my views of life and Aristotle to some
degree independent of his own subtle views on Aristotle on life and homonymy. Professor
Shields has been a model advisor in all aspects of that many-faceted job � teacher, critic, mentor
vi
and friend. My views concerning excellence in life and learning derive in large part from
reflection on the model he has embodied in his interactions with graduate students and peers at
the University of Colorado and beyond.
Finally, special thanks for their friendship and support are owed to Richard Geenen and
Jennifer Everett. Both read the penultimate draft of the dissertation for both content and form
on a grueling schedule; their work in the dissertation's final stages improved the final product
immeasurably. Their influence extends further however � I thank Rich for his kind friendship,
good humor, and skill at counterexampling throughout our joint graduate careers. I have been
very fortunate indeed at the end of my graduate career to have found a partner with the
kindness, integrity and keen intellect Jen has shared with me. Her contributions to this
dissertation and the mental health of its author will of necessity go uncounted, but are warmly
remembered. Deepest thanks.
vii
Table of Contents
Abbreviations......................................................................................................................xi Chapter 1 The Problem of Life's Definition ........................................................................................ 1
1.1 Dissertation overview. .....................................................................................................................3
1.2 Life in contemporary biology and philosophy.............................................................................5 1.2.1 Biologists on life. .....................................................................................................................5
1.2.1.1 Do scientists need a definition of life? .......................................................................5 1.2.1.2 Must a definition wait on science? ..............................................................................6 1.2.1.3 What does biology seem to tell us about life?..............................................................9 1.2.1.4 A weaker notion of definition? ..................................................................................11 1.2.1.5 Why science has not shown life to have no definition. .........................................14 1.2.1.6 Summary........................................................................................................................16
1.2.2 Philosophers on life. .............................................................................................................17 1.2.2.1 Van Inwagen on life.....................................................................................................17 1.2.2.2 Churchland on life. .......................................................................................................19 1.2.2.3 Bedau on life. ................................................................................................................23 1.2.2.4 Opposition to conceptual analysis.............................................................................27 1.2.2.5 Epistemological vs. ontological projects. .................................................................30
1.2.3 Interim conclusions. .............................................................................................................32 1.3 Aristotle's promise..........................................................................................................................34
1.3.1 Was Aristotle's theory vitalistic? .........................................................................................34 1.3.1.1 What is vitalism?...........................................................................................................35 1.3.1.2 How vitalism was refuted. ..........................................................................................38 1.3.1.3 Are mechanism and vitalism the only alternatives? ................................................39 1.3.1.4 Was Aristotle a vitalist? ...............................................................................................41
1.3.2 Aristotle and material continuity.........................................................................................42 1.3.3 Theoretical or conceptual definition? ................................................................................44 1.3.4 Section summary. ..................................................................................................................46
1.4 Conclusion.......................................................................................................................................47
Chapter 2 Two Interpretations of Aristotle on Life............................................................................48
2.1 Locating Aristotle's theory in the de Anima. ...............................................................................48
2.2 Matthews on the Meaning of Life ................................................................................................54 2.2.1 Shields' Objections to Matthews' account.........................................................................57 2.2.2 Further objections to Matthews' account..........................................................................59
2.3 Shields on the Meaning of Life .....................................................................................................64 2.3.1 Shields on focal/associated homonymy. ...........................................................................64 2.3.2 Shields' analysis of life ..........................................................................................................70
2.3.2.1 The form of Aristotle's account of life. ....................................................................70 2.3.2.1.1 Establishing homonymy. ................................................................................70 2.3.2.1.2 Finding a core for life......................................................................................71 2.3.2.1.3 Defining the causal relation obtaining between core and derived cases.72
2.3.2.2 The substance of Aristotle's account. .......................................................................72
viii
2.3.3 Objections to the substance of Shields' account..............................................................73 2.3.4 Objections to the form of Shields' account. .....................................................................75
2.3.4.1 The core of life. ............................................................................................................76 2.3.4.2 The relation between core and noncore cases of living.........................................78 2.3.4.3 Section summary. .........................................................................................................82
2.4 Conclusion: prospectus for an account. .....................................................................................83
Chapter 3 The Extent of Teleology in Aristotle.................................................................................86
3.1 The regularities challenge. .............................................................................................................89 3.1.1 Against the coextensiveness of regularity and teleology. ................................................91 3.1.2 Restricting teleological regularities. ....................................................................................96 3.1.3 Summing up problems for the regularities challenge. .....................................................98
3.2 A more plausible reading of Physics ii.8. ......................................................................................98 3.2.1 The roles of rain and frequent winter rain. .................................................................... 100 3.2.2 The growth of teeth. .......................................................................................................... 104 3.2.3 The inference from breeding true to nonreductive nature.......................................... 106 3.2.4 The inference from natures to teleology. ....................................................................... 110 3.2.5 The dialectic of the argument: summary. ....................................................................... 111 3.2.6 What is Aristotle's argument for natural teleology?...................................................... 113 3.2.7 Summary of the argument thus far.................................................................................. 117
3.3 The natures challenge. ................................................................................................................ 118 3.3.1 God as an End in Aristotle............................................................................................... 124 3.3.2 Human well-being as an End in Aristotle ...................................................................... 128
3.4 Conclusion.................................................................................................................................... 135
Chapter 4 The Analysis of Aristotelian Teleology ........................................................................... 136
4.1 Ontological issues in contemporary scholarship.................................................................... 137
4.2 Eliminitivism, reductionism, or irreducibility?........................................................................ 139 4.2.1 Eliminitivist accounts. ....................................................................................................... 140 4.2.2 Realist accounts 1: reductive accounts............................................................................ 141
4.2.2.1 Reductions to the mental. ........................................................................................ 142 4.2.2.2 Reductions to the formal cause. ............................................................................. 143 4.2.2.3 Reductions to the efficient cause............................................................................ 147 4.2.2.4 Reductions to the material cause 1: the program view. ...................................... 152 4.2.2.5 Reductions to the material cause 2: irreducible potentials. ............................... 155
4.2.3 Realist accounts 2: nonreductive accounts.................................................................... 161 4.2.3.1 A caveat on nonreductive 'accounts'...................................................................... 161 4.2.3.2 Aristotle's commitment to sui generis final causality. ......................................... 163
4.3 Conclusion.................................................................................................................................... 166
Chapter 5 Teleological Reductionism: Is there a need for Aristotelian Teleology?...................................................................... 168
5.1 The modern problem of teleology............................................................................................ 169 5.1.1 Eliminitivism or realism?................................................................................................... 170
ix
5.1.2 Aristotelianism? .................................................................................................................. 172 5.1.3 Precursors to the standard line......................................................................................... 174
5.2 The etiological account of functions........................................................................................ 176
5.3 Objections to the etiological analysis. ...................................................................................... 178 5.3.1 The general form of the uniformity objection. ............................................................. 179
5.3.1.1 Three weak uniformity objections.......................................................................... 180 5.3.1.2 The etiological theorist's response.......................................................................... 182 5.3.1.3 Uniformity among the experts................................................................................ 183
5.3.2 The core problem with the etiological account............................................................. 188 5.3.3 Replies: direct and indirect................................................................................................ 197
5.3.3.1 The direct response: no history, no function. ...................................................... 197 5.3.3.2 The indirect response. .............................................................................................. 201
5.3.4 Necessary and sufficient conditions................................................................................ 203 5.3.5 Normativity. ........................................................................................................................ 206 5.3.6 Taking stock. ....................................................................................................................... 209
5.4 The value of the etiological account......................................................................................... 210
5.5 Conclusion.................................................................................................................................... 212
Chapter 6 An Ontology for Aristotelian Teleology. ......................................................................... 214
6.1 Objections to Aristotelian teleology......................................................................................... 215 6.1.1 Guilt by association............................................................................................................ 217 6.1.2 The methodological acceptability of Aristotelian teleology. ....................................... 219 6.1.3 Sui generis teleology and physical ontology...................................................................... 224
6.2 Emergentism and teleology's compatibility with contemporary science............................ 226
6.3 The ontology of the physical world.......................................................................................... 232
6.4 Naturalism. ................................................................................................................................... 238
6.5 Objections to strong ontological emergence. ......................................................................... 239 6.5.1 Off-hand dismissals. .......................................................................................................... 239 6.5.2 Is strong ontic emergence 'too mysterious'? .................................................................. 240 6.5.3 The causal closure of the physical. .................................................................................. 244
6.5.3.1 Evidence for causal closure. .................................................................................... 245 6.5.3.2 What system is causally closed when the physical is causally closed?............... 246 6.5.3.3 Section summary. ...................................................................................................... 249
6.5.4 Methodological worries about emergence. .................................................................... 249 6.5.5 Conclusion........................................................................................................................... 252
Appendix to Chapter Six: The ontology of emergent properties and laws............................................................... 254
6.6 The core sense of emergence. ................................................................................................... 254
6.7 Epistemological emergence. ...................................................................................................... 257
6.8 Ontological emergence............................................................................................................... 260 6.8.1 Novelty................................................................................................................................. 260 6.8.2 Emergent causal influence. ............................................................................................... 274
6.9 Conclusion, summary, and definitions..................................................................................... 278
x
Chapter 7 An Aristotelian Account of Life....................................................................................... 280
7.1 Teleology and life in Aristotle. .................................................................................................. 280 7.1.1 The (close) association of teleology and life in Aristotle. ............................................ 281 7.1.2 Evidence that teleology defines life................................................................................. 291 7.1.3 Puzzles. ................................................................................................................................ 295
7.1.3.1 Spontaneously generated living things................................................................... 295 7.1.3.2 God's life. ................................................................................................................... 298
7.1.4 Section summary. ............................................................................................................... 299 7.2 The unity and diversity of living things. .................................................................................. 300
7.2.1 Homonymies and synonymies. ........................................................................................ 301 7.2.2 A starting point: the parallel between figure and soul.................................................. 303 7.2.3 The account of unity.......................................................................................................... 310
7.2.3.1 Two modern models. ............................................................................................... 311 7.2.3.2 Aristotle's apparatus of account homonymy. ....................................................... 312
7.2.4 Applications. ....................................................................................................................... 322 7.2.4.1 Figure. ......................................................................................................................... 322 7.2.4.2 Soul.............................................................................................................................. 324 7.2.4.3 Life............................................................................................................................... 325
7.2.5 Section summary. ............................................................................................................... 328 7.2.6 The contemporary adequacy of the account's form. .................................................... 328
7.3 A first attempt at a definition. ................................................................................................... 329
7.4 An Aristotelian account of life. ................................................................................................. 335
7.5 Emergence in Aristotle............................................................................................................... 337 7.5.1 The emergence of teleology.............................................................................................. 338 7.5.2 The downward causal efficacy of teleology. .................................................................. 344 7.5.3 Section summary. ............................................................................................................... 345
7.6 Conclusion.................................................................................................................................... 346
Bibliography .................................................................................................................... 348 Index Locorum................................................................................................................ 361 General Index .................................................................................................................. 368
xi
Abbreviations
Cat. Categories Int. de Interpretatione APr. Prior Analytics APo. Posterior Analytics Top. Topics SE Sophistical Refutations Phys. Physics OH On the Heavens GC On Generation and Corruption Meteor. Meteorology DA de Anima PN Parva Naturalia Sense and Sensibilia On Memory On Sleep On Dreams On Divination in Sleep On Length and Shortness of Life On Youth, Old Age, Life and Death On Respiration HA History of Animals PA Parts of Animals MA Movements of Animals IA Progression of Animals GA Generation of Animals Met. Metaphysics NE Nicomachean Ethics MM Magna Moralia EE Eudemian Ethics Pol. Politics Rhet. Rhetoric
1
Chapter 1
The Problem of Life's Definition
The 'mystery' of consciousness today is in roughly the same shape that the mystery of life was before the development of molecular biology or the mystery of electromagnetism was before Clerk-Maxwell's equations. It seems mysterious because we do not know how the system of neuro-physiology/ consciousness works, and an adequate knowledge of how it works would remove the mystery. (Searle 1992, 101-2)
The philosophical literature is positively rife1 with claims similar to those made in the
epigraph, claims to the effect that the nature of life no longer presents interesting philosophical
questions. The field of Artificial Life � a field which bears essentially the relationship to biology
that Artificial Intelligence research bears to psychology2 � is currently bringing the nature of life
1 In addition to Searle's comment, see Searle (1997, 201), Cornman (1992, 5), Crane (1995, 4-5), McGinn (1991, 6, 8, & 45), & Chalmers (1996, 25). The view is commonplace in science as well. Boyce Resenberg says, "life is just so much chemistry and physics", and he quotes biologist Harold Erikson as saying, "The secret of life is not a secret anymore. We have known for twenty or thirty years now that life is not more mysterious than the chemical reactions on which it is based." Cell biologist Tom Pollard is quoted as saying (what may be something quite different) that "What molecular biologists have believed for two generations is now generally regarded as proved beyond any doubt. Life is entirely the result of physics and chemistry inside cells and among cells" (Rensberger 1996, 25. Emphasis added.). Against this view, see Kraemer (1984). 2 See Bedau (1992c); also Sober (1992) for necessary qualifications to the claim. It is perhaps best to let some of Artifical Life's (AL) leading practitioners describe the nature of AL research. Bonabeau and Theraulaz define AL thus: "We consider it as a general method consisting in generating at a macroscopic level, from microscopic, generally simple, interacting components, behaviors that are interpretable as lifelike" (1995, 303). Here is T. Ray on AL: "Artificial Life (AL) is the enterprise of understanding biology by constructing biological phenomena out of artificial components, rather than breaking natural life forms down into their component parts. It is the synthetic rather than the reductionist approach.... The umbrella of AL is broad and covers three principle approaches to synthesis: in hardware (e.g., robotics, nanotechnology), in software (e.g., replicating and evolving computer programs), and in wetware (e.g., replicating and evolving organic molecules, nucleic acids, or others).... I would like to suggest that software syntheses in AL could be divided into two kinds: simulations and instantiations of life processes. AL simulations represent an advance in biological modeling, based on a bottom-up approach, which has been made possible by the increase of available computational power.... In simulation, data structures are created that contain variables that represent the states of the entities being modeled. The important point is that in simulation, the data in the computer is treated as a representation of something else, such as a population of mosquitoes or trees. In instantiation, the data in the computer does not represent anything else. The data patterns in an instantiation are considered to be living forms in their own right and are not models of any natural life form. These can form the basis of a comparative biology. The object of an AL
2
and the problematic status of its analysis back into focus, however.3 This chapter's aim is to
focus attention on the philosophically rich character of questions concerning what it is to be
animate.
The following passage may serve to focus attention on the problematic status of our
conception of living:
There is no universally agreed definition of life. The concept covers a cluster of properties, most of which are themselves philosophically problematic.... Theorists differ about the relative importance of these properties, although it is generally agreed that the possession of most (not necessarily all) of them suffices for something to be regarded as alive. It is not even obvious that, as A-Life scientists assume, life is a natural kind. In other words, 'life' may not be a scientifically grounded category (such as water, or tiger), whose real properties unify and underlie the similarities observed in all those things we call 'alive'. (Boden 1996, 1)
The problem of the nature of life is at a troubling crossroad between philosophy and science;
scientists tend to find the question 'too philosophical,' whereas philosophers tend to find the
question 'too scientific'.4 This conjunction of philosophical and scientific aspects leads into a
nest of problems and arguments centering on core questions in the philosophy of biology, the
ontology of the physical world, and scientific methodology.
The thesis of this dissertation is that a proper understanding of Aristotle on the nature
of life yields an adequate contemporary account of life. Much of the plausibility of this thesis
hangs, of course, on the phrase 'proper understanding'. Aristotle is often invoked as an
archtypical vitalist; this dissertation attempts to debunk such a view. Aristotle's biology is justly
instantiation is to introduce the natural form and process of life into an artificial medium. This results in an AL form in some medium other than carbon chemistry and is not a model of organic life forms." (Ray 1995, 179-80) 3 As Bonabeau and Theraulaz (1995, 323) remark, "If you are a philosopher, AL will give you the opportunity to think about new issues in ethics, epistemology, and so on and will provide you with years of work to unravel the ontological status of life: you will be able to think over life as no other philosopher before." 4 As noted by Mark Bedau (1996, 496). This conflict is brought into the open by Lange: "Immunologist P.B. Medawar, Nobel laureate in 1960, says that discussions of what it is to be alive 'are felt to mark a low level in biological conversation' (1977, 7), whereas geneticist Joshua Lederberg, Nobel laureate in 1958, writes: An important aim of theoretical biology is an abstract definition of life. Our only consensus so far
3
eclipsed; this dissertation argues that this does not undermine the status of his account of life, for
our account of living needs to accommodate the possibility of nonbiological life just as Aristotle's
does. Aristotle's ontology is often thought to be incompatible with contemporary scientific
methodology; I argue that this view is incorrect.
The goals of this dissertation are therefore twofold. One goal is to articulate a
philosophically sophisticated and textually accurate account of Aristotle on the nature of life.
This project involves extended investigation of Aristotle's epistemological and metaphysical
grounding of the core notions involved in his account. Second, I seek to defend the view of life
that emerges from Aristotle's work in light of contemporary discussions of the reducibility of
teleology to evolutionary adaptation, the ontology of the physical world, and the presuppositions
of a successful scientific methodology.
1.1 Dissertation overview.
My strategy is to ease into the deep philosophical issues surrounding Aristotle's account
of life through the relatively narrow scholarly elucidation of his views. The goal of this
introductory chapter is the modest one of bringing to light a set of issues which mark the status
of life as an area ripe for contemporary philosophical investigation. Through critically discussing
a number of contemporary accounts of life, I argue that Aristotle holds a surprisingly
sophisticated and interesting set of theses in his philosophy of nature; these commitments are
currently unrepresented in discussions of the nature of life and amply justify further interest in
and investigation of Aristotle's account.
In chapters two through four I begin in earnest the scholarly task of uncovering
Aristotle's core theses concerning life and investigating the philosophical presuppositions of the
account. Chapter two critically surveys the contemporary secondary literature on Aristotle's
account of life. I endorse the consensus view that Aristotle's conception of life is teleological,
is that such a definition must be arbitrary [since] life has gradually evolved from inanimate matter... (1960, 394)" (quoted in Lange 1996).
4
but separate this core commitment from the details of the particular accounts. Chapter three
defends the provisional teleological account of life by arguing that the scope of Aristotelian
teleology is compatible with an account of life in teleological terms. In the course of this
defense, I offer a reading of Aristotle's central argument for teleology in Physics ii.8 on which
Aristotle's grounds for postulating natural teleology are moderate, plausible, and widely shared
today. Chapter four raises the specter of the deep philosophical and methodological questions to
come, however, for I argue in chapter four that Aristotle's view of the ontic status of final
causality is in deep conflict with the presumptive modern realist approach to biological teleology.
I argue (against the present trend in the secondary literature on Aristotle) that Aristotle takes the
final cause to be a sui generis real causal factor in the structure of the world.
Chapters five and six address the contemporary status of commitment to sui generis
teleology. In these chapters I argue against the consensus theses that (1) the ontology of the
physical world has no room for sui generis teleology, (2) the methodology of science cannot
survive its postulation, and (3) an adequate realist yet reductivist account of teleology � or at
least the broad outlines of such an account � currently exists. I argue in chapter five that
contemporary philosophy of biology provides no adequate reductivist account of teleology, but
that the grounds that encourage theorists to be realists rather than eliminitivists concerning
teleological commitment survive the failure of these reductive accounts. I argue further that
methodological worries about the postulation of sui generis teleology are misguided and may
reasonably be allayed � it is no part of this dissertation's thesis that a priorist science should
make a comeback.
Chapter six argues that the ontological presuppositions which ground methodological
and scientific worries concerning sui generis teleology and the drive for reductivist accounts of
teleology (and mind) are highly suspect; we have good theoretical grounds for supposing that
there exists a coherent ontology for the physical world in which sui generis teleology finds a
natural home. Further, we have strong evidence from our best science (physics) that this
5
ontology is the ontology of the actual world. Chapter seven pulls together the scientific,
philosophical and scholarly threads of the dissertation and presents a formal Aristotelian
definition of what it is to be alive. Chapters five and six therefore constitute the contemporary
philosophical defense of the core philosophical commitments of Aristotle's account of life.
Chapter seven is the culmination of the scholarly task begun in chapters two through four of
explicitly formulating and defending Aristotle's account of life.
I turn now to the task of the present chapter: displaying the problematic status of the
nature of life in contemporary thought and the promise of an Aristotelian account of life given
the contemporary intellectual milieu.
1.2 Life in contemporary biology and philosophy.
This section constitutes an extended introduction to the problem of life's definition. I
critically survey a number of modern approaches to the nature of life, raising and defining deep
philosophical questions surrounding the issue and the philosophical and scientific
presuppositions of such accounts of life as exist. My thesis is the moderate one that the nature
of life is a more difficult and a more interesting philosophical problem than most modern
treatments suggest.
1.2.1 Biologists on life.
Given the wild success of biology since Darwin, it is attractive to think that we may
simply look to the contemporary (flourishing) science of life for a solution to our problem. This
subsection's goal is to display the problematic nature of this initially promising idea.
1.2.1.1 Do scientists need a definition of life?
The epigraph that opens this chapter suggests that contemporary biology already reveals
the nature of life, but this assumption is highly problematic. Surprisingly, "The problem is that
biology seems to have little to tell us about what it is to be alive" (Sober 1992, 375). We expect
6
biologists to answer an intuitively obvious question � what is it that distinguishes the animate
from the inanimate? � but we find that working biologists provide no answer.
The explanation of this puzzle appears to be this: biologists have sufficient work
investigating the detailed functioning, structures, and evolutionary histories of things that we all
intuitively consider to be alive (along with troubling and interesting 'borderline cases'), and they
need no explicit account of life to continue this research. Our intuitive grasp of what things are
alive is sufficient to ground the science of life without any satisfactory analysis of what it is to be
alive.5 The first problem with turning to biology for an account of the nature of life, then, is that
in the theoretical areas in which biology has been so wildly successful in the twentieth century
scientists operate under the guidance of an intuitive understanding of what things are alive rather
than seeking a general definition of life. This constitutes no problem whatsoever for the
confirmation of the views about the functioning, structure, and evolutionary histories of
organisms that have developed in this thriving field, but it does constitute a problem for the
naïve view that the success of the science of life must have included or been based upon an
explicit account of life.
1.2.1.2 Must a definition wait on science?
Nevertheless we may have reason to suppose that we must look to empirical research to
provide the central theoretical grounds for a proper definition of life. This view is widely held by
computer scientists and theoretical biologists working in the field of Artificial Life.6 They
motivate the centrality of their work to the definition of life through raising a purported problem
for more traditional approaches. The problem is supposed to be that biologists have only one
5 Mark Bedau and Packard note, "What is life? How can it be recognized? In an everyday context, these questions seem tantalizingly clear�a cat is alive and a rock is not" (1991, 457). John Searle says, "Of course, biologists do not need to be constantly thinking about life, and indeed, most writings on biology need not even make use of the concept of life. However, no one in his right mind denies that the phenomena studied in biology are forms of life" (Searle 1992, 227-8). Kim Sterelny and Paul Griffiths say, "biologists do not need a definition of life to help them recognize what they are talking about" (1999). 6 On Artificial Life, see above, n.2.
7
example of life to work with: life as it has evolved here on earth. Christopher Langton has stated
the problem vividly: "theoretical biology has long faced the fundamental obstacle that it is
impossible to derive general principles from single examples" (1989/ 1992/ 1996, 39).7 8 We
may thus formulate a first problem concerning the nature of life, the small sample problem.
The Small Sample Problem (SS): is the problem of determining from a small sample (in this case a sample of one) which features of some thing (in this case, life) are essential to that thing.
On this view there is a reason that traditional biology has not explicitly answered our question:
science is inductive, its method is effective only where we have large numbers of cases to survey.
But there is only one kind of life9 � biological life � and so scientific method has up until now
been powerless to provide us with a definition.
Artificial Life research is explicitly offered as a solution to this problem by many of its
proponents. Thus, Langton says that "our only alternative is to try to synthesize alternative life-
forms ourselves" (1989/ 1992/ 1996, 39). The study of AL is, on this construal, partially
motivated by the desire to generate a large enough sample of types of life to overcome the
(contingent) poverty of examples currently within view and to open the problem of life's
definition to proper (i.e., inductive) scientific treatment.10
The small sample problem is a pseudo problem, however, for two reasons. First, the
problem appears to presuppose an incorrect view of the scientific method; definitional strategies
7 Eliott Sober says, "The point is that there is little in the way of a principled answer to the question of which features of terrestrial life are required for being alive and which are accidental" (1992, 375). 8 Thomas Ray illustrates the problem by suggesting that our limited grasp of the nature of life prevents us from recognizing the "sentient properties of plants" (1979; cf. 1992b; 1995, 183; Strong and Ray1975). 9 Note that the problem is not that we lack a sufficient number of living things to study; there are plenty of those. What we lack are types of life itself on Langton's view. 10 Langton explains, "Among all of the things that Artificial Life is or will come to be, however, it is probably safe to say that the field as a whole represents an attempt to increase vastly the role of synthesis in the study of biological phenomena. Synthesis has played a vital role in the grounding of many scientific disciplines, because it extends the empirical database upon which the theory of the discipline is built beyond the often highly accidental set of entities that nature happened to leave around for us to study." (Langton 1995, ix)
8
do not require types of life before they may succeed. On the view which generates the small
sample problem the fact that science proceeds inductively entails that we must have types of life
to compare before the definition of life can be approached scientifically. But in the sense in
which this is true we already possess types of life, for we know of an enormous range of types of
living things and we can survey these cases for commonalties and differences. Thus, science is
able to discover the chemical essence of water not because we have different types of water to
survey but because we have different types of molecules (and atoms). In the only sense in which
types of life are required for traditional definitional strategies to succeed we already have
numerous types of life and need not wait on the results of AL.
Second, Langton's proposed solution falls prey to a dilemma.11 The dilemma is this: if
our grasp of life is sufficient to allow us to distinguish which properties in AL models are
relevant to a study of the nature of life, then AL is not the only method available for studying the
nature of life for the methods that underpin our selecting between definitional and accidental
properties of life in AL models would be sufficient to approach questions of life's account even
in the absence of the models. On the other hand, if our intuitive grasp of life is insufficiently
robust to distinguish relevant properties in biological phenomena as they currently exist and as
they may begin to make appearances in AL models, then AL cannot overcome SS any more than
its more traditional competitors. The upshot is that if the small sample problem is a genuine
problem, AL cannot help us solve it, whereas if it is not a problem, then traditional philosophical
and scientific methods working on the basis of cases of living things with which we are already
familiar can proceed in the absence of AL models.12 Luckily we have reason (given in the
11 Arguments of this general type are offered by Bonabeau and Theraulaz (1995), Emmeche (1992, 468), and Nagel (1986, 24). 12 This does not mean that AL models may have no use for theorists concerned to define life; my point in the argument above is not to dismiss the value of the study of AL but to undermine the very strong claim made by some of AL's proponents that only AL can provide the grounds for an adequate definition of life.
9
preceding paragraph) to suppose that our grasp of life is sufficient for productive work on the
nature of life both within and outside of AL research.
The conclusion of this section is a rather moderate one: AL may be a useful tool in our
search for a definition of life, but it need not constitute the essential cornerstone of an adequate
definition of life.
1.2.1.3 What does biology seem to tell us about life?
Perhaps the core difficulty with turning to biologists for insight into the general nature
of life can be illustrated by turning to one biologist's account of the lessons of modern biological
inquiry. In The Selfish Gene, Richard Dawkins says of our subject, "Words are only tools for our
use, and the mere presence in the dictionary of a word like 'living' does not mean it necessarily
has to refer to something definite in the real world" (1989, 18). Dawkins offers an explanation
of the origin of life in terms of the tendencies of collections of atoms to settle into relatively
stable states. Soap bubbles are spherical because this is a stable configuration for thin films filled
with gas; the same shape is stable for water in weightless environments but not on earth. On
Dawkins' explanation of life's origin, the tendencies that atoms and molecules have to stabilize in
different environments explains the genesis of the first simple replicators as well as the
subsequent behavior (including evolutionary change) of those replicators in their environment.13
Dawkins concludes from these reflections concerning life's continuity with nonlife that
in some real sense biology has revealed to us that there is no such thing as what it is to be alive,
13 The same position is stated with brevity and clarity by Thomas Ray: "Evolution by natural selection.... exploits any inherent self-organizing properties of the medium [into which it enters] and flows into natural attractors realizing and fleshing out their structure.... Evolution is both a defining characteristic and the creative process of life itself. The living condition is a state that complex physical systems naturally flow into under certain conditions. It is a self-organizing, self-perpetuating state of autocatalytically increasing complexity" (Ray 1995, 181). Boyce Rensberger explains, "Crystallization... happens because certain molecules have the innate property of organizing themselves into predictable arrangements entirely without guidance from an outside intelligence.... The key point is that no outside force need guide the events.... This concept of self-assembly is fundamental in today's cell and molecular biology. Under the right conditions, atoms or small molecules automatically link... to form larger structures of absolutely predictable shape. Life's chemistry generally involves very large, highly complex molecules such as the proteins, but these too assemble themselves into still larger, predictable structures" (Rensberger 1996, 22).
10
for there are no parts or processes of living things which are distinctive of all and only living
things. Dawkins is happy to employ the term 'life' throughout his book, and to use it to apply to
all sorts of things�genes, organisms, populations. Presumably, however, his intent is that a
strict reading of any of these uses implies nothing concerning the existence of a property that all
and only things classified as 'living' share.
The general position staked out by Dawkins in this illustration � a position that is widely
accepted as correct � contains two14 strands which require separating. The first strand is the
empirically supported view that the emergence and nature of life is continuous with and the
result of the same processes (iterated many, many, many times) that account for the formation
and persistence of soap bubbles.15 The second strand of the position, a strand which is strongly
correlated in contemporary work with the first, is the view that because of the continuity we are
entitled to conclude that 'life' indicates no natural kind, there is no 'essence' of being alive for
science to find, no real (rather than merely conceptual) definition16 to elucidate. Let us formalize
these theses:
The continuity thesis (CT): Living things (i) are materially continuous with nonliving things, in the sense that at the borderline of the two classes they fade indistinguishably into one another, and (ii) living things contain no features or properties (except structural properties related to the complexity of their organization) not had by the same material properties and entities outside the context of living things.17
CT entails eliminitivism: There being a real or theoretical definition of life is incompatible with the truth of the continuity thesis.
14 I argue below that there is actually a third distinct strand in the view as expressed by Dawkins and accepted by many contemporaries � a presumption of mechanism. See below pp. 42 ff., for the relevance of this further presupposition of the view. 15 As we will see, Aristotle himself holds this view. 16 For the distinction between real or theoretical definitions and conceptual analyses as it is traditionally employed in this literature, see below, p. 19. 17 cf. Campbell, "The unraveling of the DNA maze has enabled us to discover that even growth, replication, and the transmission of hereditary characteristics are processes of an ordinary chemical kind. The molecules involved are more complex, but the atoms gain no new nature in the living organism." (Campbell 1970, 26)
11
The most charitable reading of the epigraph that opens this chapter is perhaps that it was this
idea, this scientifically grounded scepticism concerning the existence of a unique nonstructural
property or activity had only by living things, to which philosophers refer when they express the
view that modern of biology has revealed to us the nature of life.18
Before criticizing the second thesis, I turn in what follows to address a popular form of
'definition' of life which appears to be motivated by the acceptance of these theses.
1.2.1.4 A weaker notion of definition?
In the face of this eliminitivism about life, many theorists � both philosophers and
scientists � have retreated to the position that while science may have shown us that life has no
'strict' definition given in terms of necessary and sufficient conditions, it may nevertheless have a
weaker form of 'definition', a cluster definition. The cluster theorist hopes to specify some set of
properties that at best 'approximates' necessary and sufficient conditions for life.19 No one
theorist takes her cluster to state necessary and sufficient conditions, but cluster theorists fall
along a scale from the very anti-essentialist to the 'more' (??) essentialist.
Here is a sampling of the lists of properties that Mark Bedau (1996, pp. 335-6) cites as
typical of the offerings of cluster theorists:
Self-reproduction; information storage of self-representation; metabolization; functional interactions with the environment; interdependence of parts; stability under perturbations; the ability to evolve. (Farmer and Belin 1992, 818)
18 This view is taken by some to have grand consequences for our theories of life and even for philosophy more generally. Dan Dennett claims that Darwin's key insight, the insight that founded modern biology, implied that our intuitions concerning essentialism are fundamentally flawed. Thus, from the Darwinian discovery of the truth of the continuity thesis, we ought to draw morals about the ontology of the world and our epistemic capacities. In a different context Dennett makes this remark, which carries over without distortion to his views on our topic: "The Darwinian perspective lets us see with unmistakable clarity why there is no hope at all of discovering a telltale mark, a saltation in life's processes, that 'counts'.... [T]here is no 'natural' way to mark the birth of a human 'soul,' any more than there is a 'natural' way to mark the birth of a species. (1995b, 513) 19 The only sense I can give to the notion of approximating necessary and sufficient conditions is epistemological: the clusters are imperfect epistemic mimics of genuine necessary and sufficient conditions.
12
Telenomic or purposeful behavior; autonomous morphogenesis; reproductive invariance. (Monod 1972)
Self-reproduction, genetics and evolution; metabolization. (Crick 1981)
Metabolism, self-reproduction; mutability. (Küppers 1985)
Metabolism; having parts with functions. (Maynard Smith 1986)
Self-reproduction; the capacity for open-ended evolution.20 (Ray 1992a)
1. All levels of living systems have an enormously complex and adaptive organization. 2. Living organisms are composed of a chemically unique set of macro-molecules. 3. The important phenomena in living systems are predominantly qualitative, not quantitative. 4. All levels of living systems consist of highly variable groups of unique individuals. 5. All organisms possess historically evolved genetic programs which enable them to engage in 'teleonomic' processes and activities. 6. Classes of living organisms are defined by historical connections of common descent. 7. Organisms are the product of natural selection. 8. Biological processes are especially unpredictable. (Mayr 1982)
The main appeal of cluster definitions stems from reflections on the continuity thesis and its
supposed dire consequences for more robust definitions plus reflections concerning our
intuitions about life. The vagueness of the cluster theories allows such accounts to
accommodate borderline cases of life � things that we just aren't sure, intuitively, how to
categorize as animate or inanimate, e.g. viruses.21 When our intuitions clash with what the
20 Ray (1995, 180) comments on the cluster definitions: "Most approaches to defining life involve assembling a short list of properties of life and then testing candidates on the basis of whether or not they exhibit the properties on the list. The main problem with this approach is that there is disagreement as to what should be on the list.... I prefer to avoid the semantic argument and take a different approach to the problem of recognizing life.... Rather than create a short list of minimal requirements and test whether a system exhibits all items on the list, one could create a long list of properties unique to life and test whether a system exhibits any item on the list." Two brief comments on Ray's attitude. First, he takes the issue of life's definition to be mainly the epistemological issue of getting a scientifically clear criterion up and running for recognizing instances of living things. Indicative of this, Ray considers the main problem with such definitions to be the disagreement people have about what should be on the lists rather than the fact that the lists appear to give neither necessary nor sufficient conditions for being alive. Ray's question should be distinguished from the metaphysical question of the actual analysis of the property of life more clearly than it is. Second, Ray's second "softer, more pluralistic approach" (181) to defining life should remind us of Aristotle's statement that life "has more than one sense, and provided any one alone of these is found in a thing we say that thing is living�viz. thinking or perception or local movement and rest, or movement in the sense of nutrition, decay and growth" (DA 413a23-5). The problems for this sort of 'definition' are treated helpfully by Matthews (1992) and Shields (1999, chapter 7, 'The Meaning of Life'). 21 Spafford (1995, §6 pp. 259-62) applies the list provided by Farmer and Belin (above) point by point to the case of computer viruses. Spafford concludes the computer viruses are not alive (262) after carefully considering how they do on each member of the list along with 'other behavior' (§6.10) but makes no
13
cluster theorist's list tells us, our intuitions come out on top.22 The clusters are left vague and
imprecise for the purpose of accommodating our unruly intuitions about which things are
alive.23
The vagueness of such 'definitions' contains, however, both their main appeal and the
seeds of their ultimate failure. Cluster theorists offer neither necessary nor sufficient conditions
for being alive, but offering such conditions constitutes a minimal necessary condition for
providing an adequate definition;24 there is a clear sense in which we do not know what
something is unless our definition gives us necessary and sufficient conditions. As a further
failing, such 'definitions' fail to give any informative explanation for the unity of the cluster of
properties cited in the 'definition' � why these properties and no others? Why all of them and
not simply some of them or only one?
Gareth Matthews expresses the frustration engendered by so-called 'cluster definitions'
in commenting on one such definition offered by The World Book Encyclopedia.25 Matthews
comments that,
explicit comment on the list provided by Farmer and Belin except that it is a "very reasonable list of properties associated with life" (260). 22 At least in some cases; as with Churchland's handling of counterexamples to his two definitions, there is a bit of the ad hoc in the way that cluster theorists treat the evidence of their preanalytic intuitions about what things are alive. One thing that is clear is that the cluster theories themselves provide no principled way of making decisions in particular hard cases. 23 Mary Anne Warren endorses the cluster 'definitions.' She says, "It should not surprise us that there is no single (or multiple) necessary and sufficient condition for the proper application of the ordinary concept of life.... Basic practical concepts, such as that of life, develop through many generations of experience. Consequently, such concepts often lack the clarity and simplicity that are desirable in, for instance, mathematical or scientific theories.... [T]he complexity and the unclear boundaries of many of our ordinary concepts cannot readily be defined away, except at the cost of substituting a different concept for the original one." (26) After surveying and criticizing Paul Taylor's attempt to do better (in Taylor (1986)), she comments that she thinks "it best to stay with a definition similar to Webster's, i.e. one that lists characteristic features that can serve as criteria of life, but that does not attempt to resolve in advance all possible uncertainties about what ought to count as a living thing" (30). All page references are to Warren (1997). 24 Thus, F is defined by X, Y, and Z only if X, Y, and Z provide necessary and sufficient conditions for being F. 25 Matthews cites the 1986 edition of this encyclopedia. Upon checking the reference I am unsure that his comment reveals a charitable reading of the text.
14
One doesn't have to be much of a Platonist to become concerned and puzzled about discussions like this. How can it be that only 'nearly all', and not simply 'all', living things have the characteristics that 'outline the basic nature of living things'? And how can a list of characteristics such that only 'nearly all' living things have those characteristics and some non-living things also have them be a list that outlines 'the basic nature of living things'?" (1992, 185)26
Cluster definitions appear not to provide us what we look for in definitions � genuine accounts
of the unity of some concept or property.
I have argued that cluster definitions fail as definitions, but this is of course not to argue
for the stronger claim that cluster definitions fail to do any important work. I see no reason to
doubt that cluster definitions contain theoretically important lists of properties and processes
which are frequently used (tacitly or explicitly) as criteria for being alive. Making such lists of
properties and processes explicit may ground both scientific research into the physical bases for
such properties and processes, and their gathering may constitute an essential first step along the
way to an adequate definition of life. What a definition must do with such clusters, however, is
find a property or properties which explains the unity of these sets of properties and provides
genuine necessary and sufficient conditions.
That making such clusters explicit has scientific value and value on the road to giving an
account or definition of biological life is not sufficient for rendering the clusters themselves
definitions. As accounts they are inherently unstable, unclear, and do not explain the
phenomena they seek to unify.
1.2.1.5 Why science has not shown life to have no definition.
The retreat to cluster definitions is invoked in response to the views set out above by
Dawkins. Science, it is claimed, has shown us two things: that life is materially continuous with
nonlife, and that this entails that life cannot have a 'strict' definition � there is no property
shared by all and only living things. I argue in this section, however, that this retreat to the
unstable weaker notion of a cluster 'definition' is unmotivated; science has not established the
26 Shields (1999, chapter 7) also addresses the problems with cluster definitions.
15
dual theses Dawkins invokes. Dawkins' view is composed of an empirical and a philosophical
thesis; the philosophical thesis disintegrates upon examination leaving the empirical core of the
thesis untouched.
It must be emphasized that the empirical thesis that life is continuous with nonlife is not
challenged by my rejection of Dawkins' view.27 Rather, the important but infrequently noted
problem with the view that science has proven eliminitivism about life is that it contains a false
view of what material continuity entails for definability. According to Dawkins' eliminitivism,
continuity of phenomena entails the absence of a real property to be defined, but this does not
follow. Intuitively, we know that in the color spectrum the colors shade indistinguishably into
one another, but we also recognize that this does not mean that red is not distinct from yellow, no
matter how difficult it may be to define that difference. The case that continuity fails to entail
eliminitivism can be made again with a case from the philosophy of science.
Elliott Sober (1980, 166-7), arguing in response to David Hull's (1965) claim that the
Aristotelian requirement that species have essences specifiable by nontrivial necessary and
sufficient conditions conflicts with material continuity, contends that essentialism can and does
coexist quite peacefully with the truth of continuity theses. Sober provides an example from
chemistry; there is no 'saltation'28 � no plausibly specifiable particular moment � in the
transmutation of nitrogen to oxygen when the nucleus of the atom ceases to be a nitrogen
nucleus and becomes the nucleus of an atom of oxygen. Nevertheless, chemical kind
essentialism is widespread and defensible: nitrogen just is the chemical kind with the atomic
number of 14. The continuity thesis does not undermine this position; as Sober argues: "This
really does not matter, as long as the vagueness of 'nitrogen' and that of 'atomic number 14'
27 Below I suggest that the view I call here the continuity thesis does not itself constitute the empirically supported core of what science has revealed about the material continuity of living things with nonlife. This version of the continuity thesis in fact contains philosophical presuppositions concerning the ontology of the physical world which rule out possible scientifically respectable views concerning the nature of life. 28 To employ one of the inessentialist's favorite terms. See Dennett (1995b).
16
coincide" (Sober 1980, 167).29 This example vividly illustrates the compatibility of essentialist
real definitions with physically continuous processes in the area (chemistry) where essentialism
has, perhaps, its widest contemporary acceptance,30 and the same point carries over without loss
to the biological case.31
The view that science has shown us the truth of eliminitivism about life rests then on a
confusion between the legitimate and respectable empirically justified results of scientific
research and the warrant for a peculiar (and false) view of the relation between continuity and
definability. Closer attention to the philosophical aspects of life's definition reveals that science
has not proven eliminitivism about life.32
1.2.1.6 Summary.
I have argued that one reason to suppose that those seeking an account of life must wait
on empirical science, i.e., Artificial Life, fails. I have argued, further, that the unprecedented
progress of biological science in the last century has essentially left our problem untouched for
two reasons. First, progress in biology has been in the area of detailed theoretically grounded
understanding of biological structure, functioning, and evolutionary history, but the investigation
of these aspects of living things proceeds on the basis of an intuitive understanding of life rather
than on an explicit definition. Second, the widespread view that an eliminitivist solution to our
problem follows from one of the empirically established results of modern biology � material
29 See also Balme's translation and notes for Aristotle's History of Animals (1991, pp. 60-1 note a and pp. 72-3 note a). See also Bedau (1992d, pp. 272-3). Polanyi (1968, 1310) gives a nice illustration of this point. Sounds I emit in speaking convey meaning. But, those sounds can be distorted by other sounds in varying degrees such that at one end of the scale, my meaning continues to be conveyed quite clearly while at the other end of the scale my utterance fails to convey its meaning since its physical carrier has been so utterly degraded. 30 Scientific essentialism's favorite examples of scientifically discoverable essences are mainly chemical in nature. 31 Chalmers (1996, pp. 53-4 & 77) contains another fruitful discussion of this point. 32 Indeed, those who are historically informed may have wondered about this association from the start since Aristotle held a version of the continuity thesis and held the view that life had a definable essence. As the argument of the text shows, Aristotle's position was not based on a crude scientific error but a shrewd philosophical insight into the compatibility of these two facts.
17
continuity � is incorrect; material continuity is compatible with the existence of robust
definitions. In this light we can see common cluster 'definitions' of life for what they are: useful
lists of criteria for being alive that nevertheless fall short of offering an account of life.
A small number of philosophers, apparently recognizing these facts, have attempted to
provide illuminating and scientifically respectable accounts of life. We turn now to a critical
survey of such positions.
1.2.2 Philosophers on life.
While in general it seems to be true that philosophers have avoided questions
concerning the nature of life,33 some have written about the problem, and indications are that
more voices will be heard in the years to come. In this section we survey the accounts of life
offered by three contemporary philosophers � Peter van Inwagen, Paul Churchland and Mark
Bedau.
1.2.2.1 Van Inwagen on life.
Van Inwagen's analysis of life, offered in his book Material Beings,34 is perhaps least
satisfactory of the of the three views I survey in this section. Despite this fact, van Inwagen's
account is worth exploring because his attempted account of life is guided by his intuitions about
the identity conditions of living things. As we will see, van Inwagen's reliance on traditional
philosophical methods sets his account apart from the lion's share of contemporary work on life
and places him methodologically closer to Aristotle than to other contemporary theorists.
Van Inwagen believes that it is in the end "the business of biology" to provide a full
account of life (84); nevertheless, he thinks that he has 'something useful' (84) to say in response
33 Lange writes that our question "has recently been treated with benign neglect by philosophers. (It goes entirely unmentioned in current philosophy of biology texts�e.g., Rosenberg 1985, Sober 1993.)" (Lange 1996, 225). Hull and Ruse (1998) also neglect the topic; Hull (1974, ch. 5), however, does discuss life, vitalism, and mechanism. The topic receives narrow attention in Hempel (1966) and Sterelney and Griffiths (1999). 34 Unless otherwise noted, all page numbers and citations in this sub-section refer to the text of van Inwagen (1990).
18
to our question, and develops his contribution in the form of a lengthy analogy. The upshot of
van Inwagen's analogical approach is that lives are reasonably well individuated, jealous, self-maintaining
events.
Let us begin with the notion of a self-maintaining event;35 a storm is an example of a self-
maintaining event in van Inwagen's sense. We can individuate a storm as a thing that persists
over time (think of the red spot on Jupiter), but, at the same time, its composition is constantly
changing: certain particles are entering and others leaving the storm constantly. Lives are like
that; living things frequently ingest, digest, and slough off waste products from the materials
necessary for life.
According to van Inwagen, lives are also reasonably well individuated and jealous. Lives, van
Inwagen claims, are relatively well individuated compared to flames; whereas our intuitions are
no guide in telling us whether the flames of the forest fire are the same flames as the flames of the
match that started it, van Inwagen is confident that lives are well individuated enough for us to
know that artificially joined Siamese twins would have two lives rather than one.
The example of the artificial Siamese twins also illustrates the jealous nature of lives when
taken with a new contrasting case � lives are jealous whereas waves are not. What this means is
that waves can pass through each other, and when they do they share the activities of the same component
molecules. Lives on the other hand do not do this according to van Inwagen. The artificial
Siamese twins do not share the same life; the activities of the simple molecules which compose the
two persons are distinct activities whereas the activities of the simples which compose the two
waves as they pass through each other are identical.
My objection to van Inwagen's account is not with his description of the output of his
intuitions; I have a great deal of sympathy for his judgements in these cases. My problem is
rather that this isn't an account of life at all, it is simply a reporting of intuitions. We know (let us
35 There is a certain vagueness in what van Inwagen means by calling lives events which he does not clear up (see Jay Rosenberg's , criticisms); however, I will concentrate on a different problem.
19
agree) that lives are reasonably well individuated compared to flames, but we don't know why.
We know (let us agree) that lives are jealous whereas waves are not, but not why. Van Inwagen
may of course respond to this criticism by noting that it is the scientists, not the philosophers,
who will eventually reveal to us why these things are true of lives, but if he applies this strategy
then it is unclear what he supposes that his own account accomplishes. If traditional
philosophical methods have something to contribute to our search for life's definition, then it
must be more robust than this survey of features intuitively associated with being alive. Genuine
analyses must do more than report on the contours of our intuitions, they must seek to root out
the causes that underlie those intuitive judgements and formulate them into coherent accounts.
Van Inwagen's analysis fails to do this, and for that reason fails to provide us with the account
we desire.
1.2.2.2 Churchland on life.
Paul Churchland takes another tack. Where van Inwagen's account failed to systematize
our intuitions, Churchland opts for systematization of phenomena at the expense of the
preservation of our intuitions,36 at least in his 'more penetrating' definition of life. In the
philosophical background here is a common distinction between the tasks of seeking a conceptual
analysis and a theoretical definition. On a common view, conceptual analysis is conceived as an
investigation into the structure and relations among the concepts we employ, while real or
theoretical definition is conceived of as an investigation of actual (often scientifically revealed)
nature of the entity or property that is the referent of our words and concepts in the world.37
36 The search for definitions that can serve as scientifically acceptable real definitions and intuitively acceptable 'conceptual analyses' is not incoherent, and Churchland's project of seeking one type of definition to the exclusion of the other may present a false dichotomy. For example, our intuition is that water and other stuffs ought to belong to chemical kinds susceptible to theoretical definition. This intuition is not undermined, then, by the scientific identification of H2O as the compositional stuff underlying water � our intuitive conception of water is in line with this scientific finding. As we will see, Aristotle believes that life has an account which is acceptable on both grounds as well. 37 Neander (1991a) and Millikan (1989) discuss and employ this set of distinctions (see also Bigelow and Pargetter1987, 257). Shields (1999, chapter 3) elucidates a conception of analytic work on which these two projects come together.
20
The distinction arises from the intuitive idea that our concepts may fail to capture the way the
world is; we may give a conceptual analysis of phlogiston describing the concept's meaning to
theorists who employed it, but we cannot give its theoretical definition because there is no thing
in the world corresponding to the concept for that real definition to capture.
Given this distinction, Churchland seeks to give a theoretical definition of life, an
account that need not be acceptable to our (untrained) intuitions. Churchland offers two
analyses of life in his book Matter and Consciousness, one simple and one "more penetrating" (172-
3).38 The simple characterization takes self-replication to be the defining feature of life:
x is alive =df x has the capacity for self-replication. (172)
Churchland rejects this definition on the basis of the intuitive idea that while even extremely
simple molecules may satisfy the analysans in certain circumstances, such molecules are not alive
in those circumstances.39
Churchland does think that a more penetrating definition of life can be given, however.
x is alive =df x is a semiclosed physical system and x exploits the order it already possesses and the energy flux through it in such a way as to maintain and/ or increase its internal order. (173)
Churchland notes that on this definition, beehives, termite colonies, cities, the biosphere and
candle flames are all living things. If our project were the project of conceptual analysis rather
than theoretical definition, these examples would constitute counterexamples to the proposed
definition. Churchland does not see them that way, however; he prefers instead to learn a lesson
from these purported counterexamples.
The wiser lesson is that living systems are distinguished from nonliving systems only by degrees. There is no metaphysical gap to be bridged: only a smooth slope to be scaled, a slope measured in degrees of order and in degrees of self-regulation. (173)
38 All references in this section, unless otherwise noted, are to Churchland (1988). 39 This basis for rejecting the definition appears inconsistent with Churchland's disavowal of conceptual analysis. I comment on the point in the main text, below.
21
Churchland thus appeals to the thesis of the material continuity of living things both as support
for the idea that conceptual analyses of life will fail and as the basis for a defense of a theoretical
definition that yields some counterintuitive results.
One obvious question that should be raised by Churchland's two definitions and his
distinct attitudes towards them has to do with the asymmetry of his reaction to the
counterexamples offered to the definitions. Counterexamples were allowed to overturn the first
definition but not the second although there are no principled differences between the
definitions themselves as concerns their relation to the purported counterexamples. Handling
the purported counterexamples differently in the two cases thus appears ad hoc and unmotivated.
Charitably interpreted, perhaps the best thing to say in Churchland's defense is that it was not the
counterexamples to the first definition that drove his rejection of it at all; rather, it was the fact
that a theoretically 'more penetrating' account of the phenomena of life was available, and that
richer conception of life deserves top billing. Counterexamples should (if Churchland is not
simply picking and choosing among definitions in an unprincipled manner) count for nothing
against either definition; theoretical biological concerns drive Churchland's search for definitions.
We may note further that Churchland's handling of the second definition of life appears
to commit him to the view that while material continuity is compatible with theoretical
definition, it is incompatible with conceptual analysis � material continuity is after all the reason
we are given for not seeking conceptual analysis. We may thus propose the following revised
thesis on the basis of Churchland's response to the second definition.
CT entails the failure of conceptual analysis: There being an account of life which is intuitively acceptable is incompatible with the truth of the continuity thesis.
Unfortunately, this thesis fairs no better than the earlier thesis that the continuity thesis entails
eliminitivism about life.
Just as there may be a real definition of life compatible with the truth of the continuity
thesis (see above, p. 15 ff.), so there is no incompatibility between continuity and conceptual
22
analysis. Thus, being bald and being hirsute are materially continuous, but we may give an intuitively
acceptable analysis of the concepts: the former indicates having little or no hair whereas the latter
indicates having a lot of hair. These conceptual analyses are adequate insofar as they mirror in their
accounts the continuity between the things in the world. Speaking linguistically, the terms and
their definitions are vague in their reference to the same cases and clear in their reference or
failure of reference in the same cases and the conceptual analysis is adequate (at least in this
respect). If such a conceptual analysis can be given of being bald despite material continuity, we
have no reason yet to believe that life cannot be given an adequate conceptual analysis given
material continuity.
Of course, Churchland may not be ruling out giving a conceptual analysis of life on the
grounds of this thesis, but we should note that no other reason or indication of a reason is given
for his rejection of the project. Churchland and others may of course feel themselves committed
to general philosophical theses about the poverty of conceptual analysis, but the point to be
made here is the simple one that those general views concerning conceptual analysis constitute a
separate issue than the question of whether or not conceptual analysis is compatible with
empirically confirmed material continuity. The project of giving an account of life that does
justice to our intuitions about life would be undermined decisively if there were an
incompatibility here, but that there is not shows that narrow concerns to offer a scientifically
acceptable account of life have not yet ruled out also attempting to offer an account which is at
the same time acceptable to intuition.
We may thus offer the following theses.
CT plus theoretical definition: The truth of the continuity thesis is compatible with the search for a theoretical or real definition of life.
CT plus conceptual analysis: The truth of the continuity thesis is compatible with the search for a conceptual analysis of life; CT is compatible with there being an intuitively acceptable account of life.
CT plus a conceptual analysis which is a real definition: The truth of the continuity thesis is compatible with the search for a conceptual analysis of life which at the same
23
time constitutes a scientifically acceptable real definition of the property of being alive.
Our conclusion thus far is that none of these optimistic theses concerning life's definition is
undermined by narrow concerns to preserve what we know to be biological fact; the search for
each type of account � or one that satisfies both simultaneously � is not undermined by
biological facts.
1.2.2.3 Bedau on life.
Mark Bedau (1996) takes Churchland's idea of offering only a theoretical definition of
life to a new extreme. On Bedau's approach, there are presumed to be strong enough grounds to
show that all accounts sticking too close to our intuitive conception of life � whether they
purport to offer theoretical or conceptual analyses of life � are bound to fail. This in Bedau's
view justifies our taking a radical new approach to giving a theoretical definition of life, one only
very loosely moored in what we ordinarily think of as living.
Bedau (along with his co-author for one paper, Packard), introduces the approach as
follows.
What is life? How can it be recognized? In an everyday context, these questions seem tantalizingly clear�a cat is alive and a rock is not. But formalizing this distinction is difficult, especially if the formalization is to be used in empirical measurements. (Bedau and Packard1991, 457)
Because of the difficulties confronting a project to correspond to our 'tantalizingly clear' intuitive
conception of life � difficulties that must be taken to be quite fatal to all such projects � Bedau
and Packard suggest a "gestalt switch" to a view of life that is "more global, statistical" in
perspective.
So, rather than try to define what it is for an individual 'microscopic' organism to be alive, our concern is with what it is for a 'macroscopic' system (population of organisms) to exhibit the property of indefinitely ongoing life. (1991, 457)
Conceptual analysis and even theoretical definition remain too closely tied to 'life' as applied to
individual organisms to succeed; the solution is to cease thinking of 'life' as applying primarily to
individuals and start thinking of systems as the primary referents of 'life'.
24
Once we have made this 'gestalt switch' it may be possible to discover a property that
will be available to empirical measurement and verification (345) and that will explain the unity
of the diversity of life which we actually see (and might possibly see in AL).
The essential principle that explains the unified diversity of life seems to be this suppleness of the adaptive processes � its unending capacity to produce novel solutions to unanticipated changes in the problems of surviving, reproducing, or, more generally, flourishing. (338, cf. 333)
The essential nature of life is revealed in the property (crudely speaking) of evolving. This
property is something that applies in the first instance to populations of living things rather than to
individual organisms. Populations are alive in a primary sense according to Bedau, individual
organisms are only alive in a derivative sense.
This last-mentioned consequence is perhaps the most striking of the counterintuitive
results of Bedau's definition. Bedau is emphatic that his account of life applies in the primary
instance to populations rather than individual organisms, which are only alive � in his sense �
derivatively. One might plausibly object that Bedau has simply made a category mistake here: in
the sense in which we were asking questions about the nature of life it is organisms that are alive in
the primary sense, not populations. Any definition that implies that populations are alive in a
primary sense fails as a definition of life, whatever else might be.
Of course, if this is indeed a category mistake, we must remember that it is one that is
consciously chosen from a considered philosophical position.40 Bedau's reply to the above
objection is simple and direct: "this objection has no force for those who are seeking the
fundamental explanation of the diversity of living phenomena" (340). Bedau insists that his
project is not to capture our intuitions. He is not doing 'conceptual analysis' on the meanings we
all have in mind when we use the term 'life', and in his view it is precisely our weddedness to this
tantalizing, intuitive, yet vague and undefined conception of what it is to be alive that stands in
40 Although, of course, we could also give Bedau too much credit here; it must be remembered that he has offered us no reasons � theoretical, conceptual, or otherwise � to suppose that we must take the drastic 'gestalt switch' he recommends before we can have an adequate theoretical definition of life.
25
the way of our accepting the benefits of his theoretical approach to life's nature. Thus, Bedau
asserts that the counterintuitive consequences of his definition fail to count as evidence against
it.
In evaluating Bedau's solution, we need to be absolutely clear that we are not calling into
question the potential (although as yet largely unproven) theoretical merits of his account.41 I
will offer no objection to Bedau's account that challenges the scientific value of his explication of
supple adaptation.
This having been said, however, it is tempting to reply to Bedau that he does not address
the original objection. To object that Bedau has not defined life, whatever else he has done, is
not in conflict with supposing that what he has done is valuable in the service of other
theoretically valuable projects in biology. The theoretical project that is presently at stake,
however, is the definition of life, and with regards to that project it seems that Bedau's analysans
is irrelevant. Let us grant that the notion of supple adaptation will prove invaluable to biology's
search for unity in the diversity of life (as ordinarily understood); even so, this fact alone is
insufficient to motivate the extreme linguistic revisionism Bedau suggests when he suggests that
we take 'life' now to mean his notion of supple adaptation. If Bedau is right both about the
unavailability of less radical definitions of life, and also about this concept's theoretical value,
then why not simply suppose that life as we understand it has been eliminitively reduced from
our scientific ontology and that our science has gained a new (and previously unknown)
explanatory concept in supple adaptation? Nothing in Bedau's work suggests that we either need
to or should accept the extreme linguistic revision he proposes.
41 Mitchel and Forrest (1995) have this to say about the work of Bedau and Packard: "The important contribution of Bedau and Packard's paper is the attempt to define a macroscopic quantity such as evolutionary activity. It is a first step at such a definition and the particular definition of gene usage is no doubt too specific to the Strategic Bugs model, in which the relationship between genes and behavior is completely straightforward. In more realistic models it will be considerably harder to define such quantities. However, the formulation of macroscopic measures of evolution and adaptation, as well as descriptions of the microscopic mechanisms by which the macroscopic quantities emerge, is essential if artificial life is to be made into an explanatory science and if it is to contribute significantly to real evolutionary biology." (275-6)
26
My reply to Bedau is thus that he has elucidated a property which may be true
(indirectly) of living things (as well as being true of and applying directly to non-living things)
and which may be scientifically useful for various sorts of research42 without his having given us
an analysis of what it is to be alive. The conceptual shift Bedau is asking us to take is simply too
drastic; the fact that his account applies primarily to the wrong sorts of things is sufficient to motivate
the rejection of his account as an account of life.43
The intuitive idea behind this rejection of Bedau's account is that while we do not
require that our theoretical definitions capture every aspect of our intuitive understanding of
everyday notions, there are some constraints on when we can legitimately say that X, Y, and Z
account for A. This notion of 'this far and no further' may be explicated by distinguishing
between categorial and noncategorial features of our notions.44 Thus, the number two is
(intuitively) categorially an abstract object; it is not a particular thing; this is why nominalists deny
that there is any such thing as a number two when they insist that all our talk about the number
two can be accounted for in terms of talk about particulars � they eliminitively reduce numbers
out of our ontology.45 Let us formulate this idea into a constraint on adequate definitions.
The categorial constraint on proper definitions: Roughly, all definitions of F must minimally capture features of F which place it in fundamental ontological categories. While we may allow that theoretical definitions of everyday notions may fail to
42 Cf. Bedau 1996, pp. 345-54 in this regard. 43 Faced with a similar opponent, one who stubbornly retains a definition which has been shown to fly in the face of common usage on the grounds that if common usage doesn't meet the requirements of the definition then so much the worse for common usage, Aristotle says, "one may retort to such a man that though in some things one must not speak with the vulgar, yet in a question of terminology one is bound to employ the received and traditional usage and not to upset matters of that sort" (Top. vi.10 148b16-22). Coincidentally, one of the main examples Aristotle has had in mind in his general discussion of method in Topics vi.10 is the definition of life. 44 I offer no analysis of this distinction but simply take it to be basic. The distinction nevertheless enjoys broad acceptance and use in traditional metaphysics. See chapter five for a more thorough discussion. 45 Examples of what I mean to indicate by categorial features of our notions may be expanded easily. Ideas are categorially such that they cannot sleep or be green, but they are mental things. My chair is a physical object and not an abstract object or a mental item such as an idea. Etc. See chapter five for more on this distinction.
27
include every aspect of our preanalytic notion of an F, there are limits to what even a theoretical definition may revise out of our ordinary conception.
This indicates in a rough way the principle on which I reject Bedau's analysans as a definition of life.
If there is no theoretical or real definition of life which captures our basic intuition that life
belongs primarily to individual organisms and only derivatively (if at all) to populations, then we
ought to accept eliminitivism about life's real definition.
There is, however, another response to Bedau's suggestion, and while carrying it off is
much more ambitious, it employs fewer presuppositions. The response I have in mind is simply
to construct a scientifically and intuitively adequate definition of life. We may show such a task
to be possible by making it actual, and by doing so we will take away all plausibility to the claim
that Bedau's 'gestalt switch' about life is advisable. It is crucial to remember, after all, that while
Bedau presupposes that the obstacles in the way of a definition more faithful to our everyday
notion of life are extreme, he neither makes clear what those obstacles are taken to be nor offers
any principled reason why the task might not be done despite the obstacles. There is a strong
sense, then, in which Bedau's gestalt switch is unmotivated, and this lends us another reason not
to follow Bedau's radical linguistic revisionism.
1.2.2.4 Opposition to conceptual analysis.
None of the philosophers or biologists surveyed thus far has had ambitions to give an
account of life which satisfies at the level of our intuitions.46 Much of this stems from the
acceptance of a school of philosophy in which conceptual analysis is, for reasons beyond the
scope of this dissertation, out of vogue.47 Adherents of this view accept the naked mind thesis:
46 Recall that even van Inwagen, who utilized intuitions in what he had to say about life really failed to offer any theoretically unified account of life � his discussion amounted to little more than a survey of some of our intuitions. Even van Inwagen appeared ready to appeal, in the end, to what science will eventually tell us about living things for the theoretical account which unified his intuitions about life. 47 Millikan (1989, 297) expresses a particularly vehement rejection of conceptual analysis; she claims that it is "a confused program, a philosophical chimera, a squaring of the circle, the misconceived child of a mistaken view of the nature of language and thought." The defense of this strident assessment takes the
28
The naked mind thesis (NM): our intuitions about x must be constantly constrained by hard empirical data concerning x if we are to make any progress in scientific (or philosophical) pursuits.48
and have taken what has been called the 'empirical turn'.49
One disturbing result of this brand of philosophical analysis that is important to note is
its supposed easy victories in substantive debates. Thus, Margaret Boden (1996), objects that
according to Bedau's position, creationism is analytically false. Life as Bedau defines it is only
instanced by things that supply adapt (i.e. evolve), but if creationism is true, or if it could have been
true, then this definition cannot be correct. Granting that scientific creationism is a deeply flawed
scientific program, should our definition of life make scientific creationism analytically false?50 Is
the issue not a (lopsided) scientific one? In a similar vein, someone might ask whether or not it
would be true if there was a God, or there were angels, that such beings would be alive. Such
beings seem intuitively to be both possible and alive,51 but again, Bedau's definition makes such
intuitions analytically false because the beings are not members of evolving populations.
Bedau does not directly respond to such criticisms, but it is tolerably clear what he
would say. To similar counterexamples to his definition (ex: a world in which there was and only
form of her positive work on the nature of language and thought, see especially (1984). Kornblith (1998) offers a less strident account of the role of intuition in naturalized philosophy. 48 The phrase derives from Dennett (1995a). For adherents and expressions of the thesis, its historical antecedents in Engels (1880/1978) and Mach (see Misak 1995, 34) and its contemporary supporters see Bedau (1992c), Dennett (1989), Kornblith (1998), Matthen (1991), Millikan (1989) and many others. 49 I borrow the phrase from Bedau (1998, 135). The so called 'empirical turn' is analogous to the linguistic turn taken by philosophers earlier in this century. As Richard Rorty notes in the introduction to his wonderful collection of essays on the linguistic turn, philosophy is marked by attempts to transform itself into science through the development of new methods. In the face of endless philosophical disagreement, philosophers of the linguistic turn tried to find objective methods by falling back to the analysis of language (Rorty 1992a; Rorty 1992b). In more recent history some philosophers, disheartened by the failure of the linguistic turn and continuing disagreement in philosophy, turn to the empirical methods of science itself to settle (what were traditionally but incorrectly thought to be) distinctively philosophical questions in metaphysics, epistemology and ethics. The 'empirical turn' is roughly equivalent to the Naturalization of philosophical thinking which heartened Daniel Dennett so in his introduction to Millikan's book (cited above). 50 On the status of 'scientific' creationism in its various current manifestations, see especially Futuyma (1983), Kitcher (1982), and, more recently, Miller (1999).
29
would be one organism) he replies, "I take it that these fantasies are just that�fantasies, with no
bearing on the true nature of any form of life that we could discover or synthesize" (341).52
Bedau gives such counterexamples no evidential weight; creationism is false, there are no worlds
where there was and only would be one organism, and 'fantasies' about God and angels are and
should remain irrelevant to science. What we want is the correct account of life as it really is in
the actually obtaining world.
This response is a red herring, however. The objector notes that on this philosophical
position a (lopsided) scientific debate is turned into an analytic falsehood. This point is not
undermined by noting that the view in question is false and is a mere fantasy; that much may be
granted by the objector. The point is that this philosophical approach achieves through theft
what must be achieved through honest toil.
Further, Bedau's response rests on the undefended presupposition that intuition
accommodating accounts inevitably straight-jacket science. But of course when it comes to life
none of our authors bothers to actually argue for this thesis. What we want is some principled
reason to ground the supposed impossibility of providing a theoretically acceptable definition of
51 See Matthews (1977) for an illuminating discussion of the connections between the concepts of living and consciousness. 52 It is interesting (or, rather, puzzling) to compare Bedau's statements here with the defenses he gave of his earlier work on the nature of teleology. Considering formally identical complaints about counter-intuitive possibilities from his opponents and the suggestion that maybe getting it empirically correct was good enough, Bedau remarks that "It is true that, by accident, some systems-theoretic condition might happen to be true of all and only the actual goal-directed systems because no appropriate pseudo-goal system exists. But an appropriate pseudo-goal system might have existed and might come to exist. Therefore, any systems-theoretic condition that happened to be true for the moment would be just an accident." ( cf. also Bedau 1986, 489; Bedau 1992b, 32, emphasis added; Bedau 1992d, pp. 285-6) My sympathies lie with the answers to these objections given by the 'early Bedau' rather than the 'late Bedau'�modal intuitions are relevant in the search for adequate definitions; reading Bedau's work (late '80s & early '90s vs. mid '90s) leaves me at a loss to explain the shift in his argumentation. However, Bedau may be more sympathetic to the statement of a couple of researchers whose outlooks are in greater sympathy to his own. Langton writes, "It was only within the context of this much larger set of 'possible' chemical compounds [developed through synthesis] that researchers were able to see beyond the accidental nature of the 'natural' chemical compounds, and glimpse the regularities in the constitution of matter. To have a theory of the actual, it is necessary to understand the possible." (Langton 1995, ix-x). Karen Neander suggests that "A theoretical definition should... elucidate the theoretical role that the notion has enjoyed, and hypothetical cases can be instrumental in this respect" (Neander 1991a, 324-5).
30
life that is at the same time intuitively acceptable � a definition that would capture the
conceptions of life held in common among creationists and evolutionists.
My point is simple and straightforward. An account that purports to offer a conceptual
analysis of life but that also attempts to reveal the real nature of the property of being alive must
pass the muster of the legitimate, empirically established conclusions of the science of life,
biology. It is important to note, however, that no specific rather than perfectly general
philosophical reasons to suppose that this task cannot be accomplished have been offered.53
1.2.2.5 Epistemological vs. ontological projects.
One final dimension of contemporary approaches to life deserves treatment. Recall that
Bedau and Packard motivated their 'gestalt switch' in defining life as follows:
What is life? How can it be recognized? In an everyday context, these questions seem tantalizingly clear�a cat is alive and a rock is not. But formalizing this distinction is difficult, especially if the formalization is to be used in empirical measurements. (Bedau and Packard 1991, 457)
Thomas Ray approaches the problem in the same way. He says,
Most approaches to defining life involve assembling a short list of properties of life and then testing candidates on the basis of whether or not they exhibit the properties on the list. The main problem with this approach is that there is disagreement as to what should be on the list.... I prefer to avoid the semantic argument and take a different approach to the problem of recognizing life. (Ray 1995, 180)
Mary Anne Warren, supporting cluster definitions of life, writes,
I think it best to stay with a definition similar to Webster's, i.e. one that lists characteristic features that can serve as criteria of life, but that does not attempt to resolve in advance all possible uncertainties about what ought to count as a living thing. (Warren 1997, 30)
All of these approaches to life's definition share one common assumption, the assumption that
an adequate definition of life will help us resolve questions concerning so called 'borderline cases'
of living things (ant colonies, self-replicating molecules, viruses�computer and otherwise, etc.).
53 See Bedau (1996, 354), Sober (1992, 375-6), Harnad (1995, 298) and Warren (1997, 26 & 30).
31
This assumption places an epistemological demand on adequate philosophical or
scientific definitions. Call this requirement the decision procedure requirement.
The decision procedure requirement (DPR): A definition of X, given in terms of necessary and sufficient conditions, is adequate only if it provides us with some objective method for determining, in all cases, whether specific things are X.
The decision procedure requirement places a non-trivial requirement on definitions. DPR as
applied to the case of life requires that adequate definitions given in terms of necessary and
sufficient conditions yield objective decision procedures for determining, in borderline cases,
whether something is alive or not.
This presupposition is highly problematic on philosophical grounds given its conflation
of epistemological and ontological projects. Definitional projects as such need not be held
hostage to our desire for epistemically distinct lines between the kinds of things there are in the
world. No definition of being bald must provide decision procedures for determining in every
case when something is and is not bald; the more natural thing to say is that if a definition of
baldness did give such precise results it would for that reason fail as a definition of baldness. The
analysans of baldness should follow our uncertainty about cases, it should not rule decisively on
cases where the application of the term is uncertain. We may have strong pragmatic or
methodological grounds for desiring clear and measurable lines between things that count as
alive and those that do not, but there is no reason that the ontology of the world or our
definitional practices must cooperate with our methodological and pragmatic hopes.
Acceptance of the continuity thesis, then, places no constraint on us to accept the
inappropriate decision procedure requirement on adequate definitions. We can allow a correct
and informative specification of a thing's essence to be epistemically indeterminate in the results
it yields in borderline cases. Recognizing this we may offer a competing thesis concerning
adequate definitions of life.
The No-Epistemological-Baggage Thesis (NEB): A definition of life need carry no epistemological baggage. Our definition must tell us what conditions a thing must satisfy in order to be alive; however, adequate definitions of life need not
32
yield objective decision procedures for determining in hard cases whether or not something is alive.54
The decision procedure requirement mixes legitimate metaphysical constraints on the proper
analysis of a property or concept with improper epistemological constraints. The no-
epistemological-baggage thesis distinguishes properly between the metaphysical tasks we
legitimately expect our definitions to satisfy and improper epistemological or pragmatic concerns
about the exact class of things in the world to which our definition applies. As a result, the
prospects for defining life brighten considerably.55
1.2.3 Interim conclusions.
The time has come to regroup and take stock of our rather brief survey of the problem
of the nature of life in contemporary philosophy and science. In addition to looking critically at
a number of interesting accounts and approaches to the nature of life we have extracted and
critically examined a number of theses and problems underlying contemporary work. One major
point of this exercise has been simply to establish that the definition of life in contemporary
philosophy and biology continues to reveal important, interesting, and unresolved philosophical
problems. A second point of the exercise has been to ground the claim that there is important
and interesting space for a contemporary Aristotelian account of life.
We may draw several conclusions from this brief survey. First, although the largest
portion by far of contemporary work on life proceeds on the assumption that life may be given
an informative account only along the lines of a theoretical definition, we have been offered no
proximate reasons to suppose that respecting the legitimate results of biological science requires
54 In Paul Humphreys' words, "The slogan is: ontology first, epistemology second" (Humphreys 1995, 112). 55 It is worth recalling at this point the truth of a claim made earlier (see p. 6): the intuitive grasp biologists have on which things are living and which are not is sufficient to provide them with bountiful opportunities for fruitful study. The demise of DPR might deprive biologists and philosophers of a hope they had of gaining easy answers to previously hard questions, but biology will go on without noticing. Sterelny and Griffiths comment, "we do not see how a definition of life is likely to help us with odd and
33
abandoning the project of conceptual analysis. The strongest possible argument in favor of the
compatibility of these two projects is to successfully complete it; one goal of this dissertation is
to establish that Aristotle's account of life provides just such a scientifically and intuitively
acceptable account.
Second, we have noted that the core scientific motivation for seeking definitions that are
at best theoretical definitions stems from the acceptance of the continuity thesis (roughly, the
thesis that living and nonliving things are materially continuous). But as we have seen, the truth
of the continuity thesis is compatible with there being definitions that satisfy the constraints of
projects of conceptual analysis, theoretical definition, or both. Further, as I will argue below, the
continuity thesis as it guides contemporary work itself fails to capture the empirically established
core of what we have learned from biology, for the continuity thesis itself is the conjunction of a
scientifically established thesis (the thesis of the material continuity) and a mechanistic
assumption about the ontology of the physical world (see below). Scientifically respectable
accounts of life need not even accept the continuity thesis because of this philosophical
commitment, but all accounts must accept the weaker empirical core of the thesis.
My argument in what follows is that Aristotle holds both interest and promise for his
unique and subtle position on the range of issues outlined above. Aristotle is committed to the
empirical core of the continuity thesis and he undertakes the ambitious project of providing an
account of life that satisfies constraints from both intuition and scientific theory. Aristotle's
approach to life therefore constitutes a fresh and potentially powerful contender in the
burgeoning field of accounts of life and serves as a philosophically sophisticated counterbalance
to contemporary accounts that operate under controversial and undefended philosophical
presuppositions.
hard-to-classify cases. . . . The adequacy of the definition is settled by our view of the case, not vice versa." (1999)
34
1.3 Aristotle's promise.
Given the tightly connected system of problems concerning an adequate definition of
life brought to light in the previous section we may appreciate the philosophical acumen with
which Aristotle approached our question. It is both a blessing and a curse that Aristotle's
account of life is intertwined with the core of his metaphysics. It is a blessing in that it is this
very feature of Aristotle's thought that makes his contribution so rich in the philosophical
resources it can bring to bear on the sorts of issues raised above. It is a curse in that obscurities
in the interpretation both of Aristotle's metaphysics and his account of life threaten to diminish
the contributions his thought may make to the ongoing debate concerning the nature of life.56
My goal in this dissertation is to develop a textually supported interpretation of Aristotle's
account of life.
Before moving to Aristotle's promise in this regard, however, we must allay concerns
arising from Aristotle's historical associations with vitalism.
1.3.1 Was Aristotle's theory vitalistic?
Our Darwinism separates us from Aristotle so thoroughly that we must be non-Aristotelians. (Falk 1995, 311)
Aristotle is considered by at least one commentator to be a patron saint of vitalism,57
but vitalism is dead.58 My project may thus be perceived as doomed from the start. I would like
56 It is hard not to feel some sympathy for Hobbes' famous complaint: "And indeed, that which is written there [in Aristotle's Metaphysics] is for the most part so far from the possibility of being understood, and so repugnant to natural reason, that whosoever thinketh there is anything to be understood by it, must needs think it supernatural" (1994, xlvi.14, 458). It is not unreasonable, however, to feel more optimism about Aristotle's contribution than Hobbes would allow. 57 According to Beckner (1967), "Aristotle established four traditions that, it can be said, virtually determined the course of subsequent critical vitalism: he identifies what has been called here the Life [i.e., the substantial entity which animates an organism] of an organism with its psyche; he locates the purposive activity, organic unity, and embryological development as the phenomena that vitalism must take most seriously; he argues that the activities of the part must be understood by reference to the form of the whole and that morphogenesis must be understood by reference to the form of the adult; and finally, he describes the manner of the psyche's influence on its organism as formal, not efficient, causation. In
35
to address this worry by seeking clarity about the content of a charge of vitalism.59 Given that we
do not want to be vitalists, why do we want not to be vitalists? Just what is contained in the
charge that a theory a theory of life's nature is or is not vitalistic, and what makes such charges
stick? Why is the charge of vitalism fatal to the theory against which it is justly raised?
1.3.1.1 What is vitalism?
Vitalism is best introduced with reference to the rival thesis of mechanism. Ernst Mayr
distinguishes usefully between two distinct versions of a mechanist view of life. Strong
mechanists claim that organisms are 'nothing but' machines, "the workings of which can be
explained by the laws of mechanics, physics, and chemistry" (1982, 51). However, Mayr notes
that contemporary mechanists modify this view "rather drastically." The contemporary
mechanist accepts that:
there is nothing in the processes, functions, and activities of living organisms that is in conflict with or outside of any of the laws of physics and chemistry.... [But] they do not accept... that animals are 'nothing but' machines.... The phenomena of life have a much broader scope than the relatively simple phenomena dealt with by physics and chemistry. (Mayr 1982, 52)60
short, critical vitalism after Aristotle takes the soul as the model of the Life and attributes to Life the power of achieving and maintaining organic form." (254) 58 Beckner (1967) thinks that vitalism is irrefutable as a general thesis, but this would be bad enough for the any vitalistic theory if there were no explanatory gains to be gotten from the hypothesis. 59 I am not alone in seeking such further clarity. See also C.D. Broad, "I think one feels that the disputes between Mechanists and Vitalists are unsatisfactory for two reasons. (i) One is never quite sure what is meant by 'Mechanism' and by 'Vitalism'; and one suspects that both names cover a multitude of theories which the protagonists have never distinguished and put clearly before themselves. And (ii) one wonders whether the question ought to have been raised long before the level of life.... The question: Is chemical behaviour ultimately different from dynamical behaviour? seems just as reasonable as the question: Is vital behaviour ultimately different from non-vital behaviour" (1925, pp. 43-4). Carl Hempel says, "the issue can be fruitfully discussed only if the meaning of the opposing claims can be made sufficiently clear to show what sorts of argument and evidence can have a bearing on the problem and how the controversy might be settled" (1966, 101). 60 Building on Broad's account (1925, 45-6), McLaughlin gives this definition of Comprehensive Mechanism: "What we may call 'Comprehensive Mechanism' holds if and only if the following four conditions are met. First, every object is or is entirely made up of elementary material particles. Second, the force-generating properties are possessed by (at least) some kinds of elementary material particles. Third, the value of any force-generating property of a whole is determined, in accordance with a compositional principle, by the values of that sort of property for at least some of its parts. Fourth, forces combine by a principle of vector addition; such principles are themselves compositional principles: the value of the force exerted by a whole is determined in accordance with such a principle by the value of that force as it is exerted by the
36
Building on this account, we may distinguish between strong and weak mechanisms as follows.
Strong mechanism: Organisms are nothing but chemical machines � all their properties and parts are wholly reductively identifiable with and explicable in terms of the laws of physics and chemistry as they apply outside of living systems.
Weak mechanism: Organisms are 'something more' than chemical machines; although their activities do not conflict with the laws of physics and chemistry, at least some of their properties or parts are not wholly reductively identifiable with or explicable in terms of the laws of physics and chemistry as they apply outside of living systems.
It is worth pausing over these definitions for a moment before moving on to elucidate
corresponding vitalistic doctrines.
The point that concerns me here is to note that Mayr's definition of contemporary or
weak mechanism is fully compatible with that paradigm of nonphysicalist doctrines, Cartesian
dualism. No Cartesian dualist about the mind must claim that mental interactions are in conflict
with laws of physics (rather than compatible with them when fully understood), nor must she
affirm that minds are outside the influence of physical and chemical laws. As we will see below
(in the appendix to chapter six), distinguishing between the physical and the nonphysical is an
incredibly thorny job, surprisingly difficult to complete successfully without simply resorting to
stipulative definitions.61 Mayr's failure to exclude such paradigm nonphysicalist views as
Cartesian dualism in his account of mechanism (!) is just one indication of the thorniness of the
philosophical problems we are approaching, and this should put us on guard against any too
hasty judgements against a theory for supposed 'vitalistic' contamination. Capturing the set of
distinctions that interests us in this area of physical ontology is no easy matter.
Moving ahead, just as Mayr distinguishes between crude and sophisticated mechanism,
so we may distinguish between two forms of vitalism. Mayr labels as "extreme vitalism," the
components of the whole." (McLaughlin 1992, 77) As McLaughlin notes, this version of mechanism is false if compositional principles employ nonlinear functions (77n36). However, certain contemporary mechanists appear to want to use the modeling of such nonlinear functions in explanations their explanations, see for example Bedau (1997). 61 Or sophisticated moves which amount to mere stipulation.
37
view that organisms are "completely controlled by a sensitive if not thinking soul" (1982, 114).
This seems to suggest a view in which living things are governed entirely by souls or entelechys to
the exclusion of material or physical causes and explanations. On extreme vitalism, we can expect
laws of nature to be broken and overridden.
Mayr also speaks of much less radical views as belonging to the "vitalistic school."
Distinguishing the defining characteristics of this 'sophisticated' vitalism requires a bit of work.
According to Mayr, these vitalists believe "that there are processes in living organisms which do
not obey the laws of physics and chemistry" (Mayr 1982, 52). According to a recent
encyclopedia of philosophy, "Vitalism holds that living organisms are fundamentally different
from non-living entities because they contain some non-physical element or are governed by
different principles than are inanimate things" (Bechtel and Richardson 1998, 639). Carl Hempel
writes,
According to neovitalism, [the teleological characteristics of living things] do not occur in nonliving systems and cannot be explained by means of the concepts and laws of physics and chemistry alone; rather, they are manifestations of underlying teleological agencies of a nonphysical kind, referred to as entelechies or vital forces. Their specific mode of action is usually assumed not to violate the principles of physics and chemistry, but to direct the organic processes, within the range of possibilities left open by the physico-chemical laws, in such a way that, even in the presence of disturbing factors, embryos develop into normal individuals, and adult organisms are maintained in, or returned to, a properly functioning state. (1966, 71)62
Elliot Sober adds a condition on vitalism such that vitalists maintain, while mechanists (or
physicalists) deny that "two objects could be physically identical even though one of them is alive
while the other is not" (2000, 23-4).
While there is some discrepancy between these statements, the consistent theme of this
weaker sense of vitalism is that vital forces are (in some sense) independent of physical laws.
62 According to Hempel, the problem with positing such entities from a scientific standpoint is that such a posit is not "definite enough to permit the derivation of specific implications concerning the phenomena that the theory is to explain.... [A]ll that the neovitalist doctrine enables us to do is to make the post factum pronouncement: 'There is another manifestation of vital forces!'; it offers us no grounds for saying 'On the basis of the theoretical assumptions, this is just what was to be expected�the theory explains it!'" (1966, 72)
38
This independence is either explained by vital forces' ability to break physical laws (as in strong
vitalism), or a claim to the effect that vital forces fail to depend for their existence upon the
purely physical interactions of the physical parts of organisms (weak vitalism). If this is correct,
we may distinguish between two vitalistic theses.
Strong vitalism: That property or entity in virtue of which living things are alive is (i) irreducibly distinct from its physical properties and parts, (ii) to some extent determines the course of physical events in living things, and (iii) breaks or overrides physical laws in the course of (ii).
Weak vitalism: That property or entity in virtue of which living things are alive is (i) irreducibly distinct from its physical properties and parts, (ii) to some extent determines the course of physical events in living things, and (iii) fails to depend for its existence on the interactions of such physical parts of living things as there may be.
With these definitions in hand we are in a position to understand why, given contemporary
scientific knowledge, it is unacceptable to be a vitalist.
1.3.1.2 How vitalism was refuted.
The progress of biology in the last century has seen the advent of methods and theories
sufficiently robust to make any claims that physical laws are broken in living things implausible as
well as undermining any plausibility even for the weaker claim that living things somehow fail to
depend on their physical parts for their being alive. This dissertation accepts in no uncertain
terms that it is a condition of adequacy on any theory of the nature of life that one respect these
results and the unifying explanatory role63 that Darwin's theory of evolution has provided to our
63 In Dobzhansky's famous words, "Nothing in biology makes sense except in the light of evolution" (Dobzhansky 1973). Mayr notes, "Perhaps the most impressive aspect of current biology is its unification. Virtually all the great controversies of former centuries have been resolved" (Mayr 1982, 131). Ruse (following Whewell) argues that the theory of evolution is 'the best kind of science' in that it brings many disparate areas of inquiry under on principle: the principle both illuminates the sub-disciplines and the sub-disciplines combine to give credence to the principle. "Convergence on a common principle convinces us that we have moved beyond coincidence" (Ruse 1998, cf. esp. pp. 2-4). Kitcher says, "When we look at the last 120 years of the history of biology, it is impossible to ignore the fecundity of Darwin's ideas. Not only have inquiries into the presuppositions of Darwinian histories yielded new theoretical disciplines (like population genetics), but the problem-solving strategies have been extended to cover phenomena that initially appeared troublesome.... The comparison [with Newton's physics] is apt.... In both cases, we find a unified theory whose problem-solving strategies are applied to illuminate a host of diverse phenomena. Both theories offer problem solutions that can be subjected to rigorous independent checks. Both open
39
understanding of living things.64 Vitalism in both its strong and weak forms, then, is well and
truly dead, and we can appreciate the grounds for this judgement.
1.3.1.3 Are mechanism and vitalism the only alternatives?
Nevertheless, we can see why it is that the terms 'mechanism' and 'vitalism' are slippery;
each thesis shades off into theses which have not been empirically refuted and remain live
options. Thus, plausible property dualisms maintain that the mental depends for its existence (at
least in cases with which we are most familiar) on the physical and interacts with the physical
without breaking any physical laws.65 Given this, such property dualisms avoid vitalistic charges
on the account given above.66 By the same token, however, property dualism is compatible with
a commitment to the formulation of weak mechanism given above, and indeed most property
dualists would endorse weak mechanism as stated.
Of course, the fact that property dualism is compatible with this form of weak
mechanism establishes nothing substantive about its acceptability as a thesis in the philosophy of
mind; my point here is rather that it shows something important about defining vitalism and
mechanism. If we are to capture the distinction between property dualisms, vitalisms, and
mechanisms, then we need to draw finer grained distinctions between our concepts.
up new lines of inquiry and have a history of surmounting apparent obstacles. The virtues of successful science are clearly in display in both.... Evolutionary theory is not simply an area of science that has had some success at solving problems. It has unified biology and has inspired important biological disciplines." (Kitcher 1982, 53-4) 64 According to Mayr, it was the recognition "that all manifestations of development and life are controlled by genetic programs" (Mayr 1982, 106) that finally put an end to any controversy between vitalism and (crude) mechanism. As Boyce Rensenberg says, "modern biology confirms the view that all the phenomena that together constitute life can be understood in terms of chemistry and physics" (Rensberger 1996, 21). 65 Perhaps along the lines suggested by Hempel's remarks, quoted above, perhaps in other ways. 66 They avoid this trivially, of course, because they are not meant as accounts of the property in virtue of which things live. The point is simply that the ontological conditions stated in vitalistic theories that make it plausible to suppose that vitalism has been scientifically refuted do not likewise make it plausible to suppose that property dualism along the lines indicated above has been scientifically refuted.
40
I approach this issue by strengthening our definition of weak mechanism so that it rules
out at least some nonepiphenomenalist versions of property dualism.67 How we choose to
precisify the definition of weak mechanism, above, such that it excludes at least strong
nonepiphenomenalist versions of property dualism without collapsing back into strong
mechanism is a technical decision.68 I choose the following precisification.69
Sophisticated mechanism: (i) living things are wholly composed of physical parts; (ii) some parts or properties of living things are not reductively identifiable with underlying physical parts and properties but are reductively explicable in those terms; and (iii) no property not reductively identifiable with underlying physical parts or properties has 'its own' causal powers; the causal powers of organisms are exhausted by the causal powers of their physical parts.
Such a definition allows us to distinguish a 'middle way' between mechanists and vitalists, and
allows us to deny that property dualists are mechanists despite the fact that they accept weak
mechanism. This 'middle way' is occupied by emergentist accounts of life and mind.
Emergentist accounts: (i) living things are wholly composed of physical parts; (ii) all parts and properties of living things are causally or logically dependent for their existence on the interactions of their physical microparts; (iii) not every part or property of living things is a part or property fully reductively identifiable with the parts and relations of those same materials outside of living things; and (iv) the 'emergent' parts or properties noted in (iii) exert 'their own' causal influence over the course of their microparts' careers without breaking laws governing the behavior of microparts and properties.
As I have drawn these definitions, the crux of this issue between sophisticated mechanists and
emergentists has to do with the problem of 'downward' causation.70
67 I choose this option instead of the option of weakening a definition of vitalism so that strong nonepiphenomenalist versions of property dualism satisfy its definition because I believe there are strong constraints on the contemporary usage of 'vitalism' to the effect that vitalism, whatever it is, must have been shown empirically false by the progress of biology in the twentieth century. But, property dualism has not been shown false by the progress of biology in the twentieth century, and so property dualism is not a form of vitalism, and it would be misleading to characterize it as such. Best to follow the intentions of property dualists and seek to find a 'middle way' between Cartesian dualisms and strict physicalism. 68 Cf. Horgan (1994, 471 & 474). 69 I defend the terms in which this definition is framed in the appendix to chapter six of this dissertation. 70 'Downward causation' is typically contrasted with 'upward determination'. Water's properties (liquidity, etc.) are thought to be determined in an 'upward' fashion by the properties of water's micro-parts. Whether there are any genuine instances of 'downward' causation where 'high level' properties of wholes
41
Both sides of the debate between sophisticated mechanists and emergentists accept
upward determination. Both sides accept, that is, that higher level properties and parts of
organisms, whether reducible to their physical bases or not, depend for their existence on micro
properties and parts and their activities. Both sophisticated mechanists and emergentists accept,
in addition, that there are no activities in embodied living beings which break physical law.
Emergentism and vitalisms accept, but sophisticated and strong mechanisms reject, downward
causation.
1.3.1.4 Was Aristotle a vitalist?
We must now return from these conceptual issues and bring our results to bear on the
question of Aristotle's supposed commitment to vitalism. I accept that vitalism is a theory that
has suffered empirical refutation in the twentieth century; thus, if Aristotle endorses a vitalistic
account of life then his account has been falsified and cannot offer an adequate contemporary
account of life. While I cannot argue the claims in sufficient depth here (see further chapter
seven), I will outline the case for the plausible view that Aristotle was committed not to vitalism
but to some form of emergentism.
The case proceeds as follows. Aristotle accepts both the irreducibility and the
downward causal efficacy of certain forms (see Phys. i.6 189a28-33, GA ii.4 738b19-26; see also
GA i.20 729a9-14, i.21 729b6-21, iv.1 765b10-15); he is therefore no strong or sophisticated
mechanist. Further, he accepts upward determination (see GC i.10 327b23-7, DA i.5 409b28-8);
he is therefore no strong vitalist. Aristotle accepts that downward causation breaks no lower
level material laws (see GA v.8 789a8-b8, 789b1-15, APo. ii.11 94b28); he is therefore no weak
vitalist. Aristotle appears, then, to be an emergentist weak mechanist.
If this hypothesis is correct, then we have an understanding both of why Aristotle's view
has been taken to be vitalist and of why it is that his account may nevertheless constitute a
influence the course of micro-events is a vexed issue which I discuss in depth in chapter six of this dissertation.
42
scientifically acceptable approach on the contemporary scene. Confusion about Aristotle's status
as a vitalist or a mechanist comes from insufficient attention to the complex metaphysical
differences between the views; sloppy accounts of mechanism or vitalism fail to capture the
distinctive senses of either thesis. Aristotle's approach holds contemporary promise, moreover,
because contemporary accounts of life proceed on the philosophical � not scientific �
assumption of sophisticated mechanism. Nothing about the progress of science has shown
emergentism false; indeed, we have positive evidence from contemporary science that some
form of emergentist ontology obtains in the actual world.71
1.3.2 Aristotle and material continuity.
Aristotle's approach rejects the philosophical presuppositions of contemporary accounts
of life in another scientifically acceptable and philosophically astute way. This comes to light
through noting that Aristotle seems to have held a surprisingly modern version of a continuity
thesis while rejecting what is today almost universally assumed to be a consequence of that view.
Aristotle tells us that,
Nature passes from lifeless objects to animals in such unbroken sequence, interposing between them beings which live and yet are not animals, that scarcely any difference seems to exist between two neighboring groups owing to their close proximity. (PA iv.5 681a11-15)72
Nature proceeds little by little from things lifeless to animal life in such a way that it is impossible to determine the exact line of demarcation, nor on which side thereof an intermediate form should lie. Thus, next after lifeless things comes the plant, and of plants one will differ from another as to its amount of apparent vitality; and, in a word, the whole genus of plants, whilst it is devoid of life as compared with an animal, is endowed with life as compared with other corporeal entities. Indeed, as we just remarked, there is observed in plants a continuous scale of ascent towards the animal. So, in the sea, there are certain objects concerning which one would be at a loss to determine whether they be animal or vegetable. For instance, certain of these objects are fairly rooted, and in several cases perish if detached; thus the pinna is rooted to a particular spot, and the razor-shell cannot survive withdrawal from its burrow. Indeed, broadly speaking, the entire genus of testaceans have a resemblance to vegetables, if they
71 These claims are argued below in chapter six of this dissertation. 72 Unless otherwise noted, all translations are from those in the Revised Oxford Translation (ROT) edited by John Barnes (1984).
43
be contrasted with such animals as are capable of progression. In regard to sensibility, some animals give no indication whatsoever of it, whilst others indicate it but indistinctly. Further, the substance of some of these intermediate creatures is fleshlike, as is the case with the so-called ascidians and the sea-anemones; but the sponge is in every respect like a vegetable. And so throughout the entire animal scale there is a graduated differentiation in amount of vitality and in capacity for motion. (HA viii.1 588b4-23; see also OH i, 279a29)
It may appear from these selections that Aristotle held a straightforward version of the
continuity thesis as expressed above (see p. 10).
This supposition runs into trouble, however, given that Aristotle fails to endorse the
type of account of life offered by contemporary nonsceptics concerning life's definition. Note
that the continuity thesis (as stated) is compatible only with mechanist accounts of life.73 I have
formulated the theses along these lines to reflect the fact that, for a number of philosophical
reasons to be explored in this dissertation,74 contemporary theorists have exclusively defended
strong or sophisticated mechanist accounts of life. Expressions of the continuity thesis have
been framed exclusively in these terms in contemporary debate as a result.
If we supposed that the only options in giving an account of life were vitalist accounts
(strong or weak) and mechanist accounts (strong or sophisticated), then we would be justified,
perhaps, in making this association between the continuity thesis and mechanist accounts of life
given the refutation of vitalism by contemporary biological investigation. Given the availability
in logical space of emergentist accounts of life, however, there is interesting and scientifically
respectable territory left to be occupied given the acceptance of the empirical core of continuity
thesis. Two main burdens of this dissertation consist in establishing that Aristotle filled at least a
73 While the current formulation of the continuity thesis rules out sophisticated mechanism, a weaker version of the continuity thesis could be formulated compatible with contemporary work on the nature of life. 74 See chapters five and six of this dissertation for in depth treatment of modern philosophical and scientific approaches to the methodological and ontological concerns that drive a wedge between Aristotelian and modern approaches to our problem.
44
part of this unoccupied logical space75 and that such a view is not only a respectable contender
on the contemporary scene but has much to recommend it.76
The same strains in Aristotle's thought which lead interpreters to attribute vitalistic or
emergentist positions to Aristotle force us to deny that he held the modern version of the
continuity thesis defined above (see p. 10). But, History of Animals viii.1 588b4-23 and Parts of
Animals PA iv.5 681a11-15 establish that Aristotle held the empirical core of the continuity
thesis. We may formulate this core as follows:
The thesis of material continuity (TMC): Living things are materially continuous with nonliving things, in the sense that at the borderline of the two classes they fade indistinguishably into one another.
This thesis constitutes the empirically supported core of the continuity thesis, and, unlike the
continuity thesis (as formulated above) makes no commitment to strong or sophisticated
mechanism. This general formulation of a continuity thesis makes no commitments to
philosophical claims concerning the compatibility or incompatibility of material continuity with
the existence of irreducible properties possessing novel causal influence, and it is this weaker
continuity thesis that science has established and which we should attribute to Aristotle.
1.3.3 Theoretical or conceptual definition?
Given that Aristotle will be no sceptic concerning the account of life, we may wonder
whether the account he offers will attempt an adequate conceptual analysis or be in the form of a
theoretical definition � or whether his account will attempt to supply both.
Aristotle's commitments in terms of both philosophical and scientific methodology
strongly suggest that he does not support any strong version of the naked mind thesis (see
above, p. 28). Aristotle's statements concerning scientific methodology all declare that he is
75 The bulk of this dissertation is devoted to establishing this thesis; see especially chapters two through four and chapter seven. 76 See especially the arguments of chapters five and six.
45
liberal, if judicious, in the types of evidence he is willing to countenance in theory construction.77
At times he allows empirical observation to over-ride common sense beliefs78 while at other
times he allows common-sense and intuition to assert themselves over theories that are too far
removed from the facts.79 If these positions are correct, then there is support for the claim that
at least part of Aristotle's goal in offering a definition of life will be to capture the core
commitments of the everyday notion of being alive (see Met. ii.1 993a27-b8, etc.) even while he
seeks a theoretically sound scientific and philosophical position on the subject.
Whether or not life has an account satisfactory to philosophically and scientifically
informed common sense is discovered only through formulating and testing such accounts.
Aristotle criticizes the theories of previous thinkers (see especially his criticism of Dionysius'
definition of life in Topics vi 10, 148a23 ff.) but does not despair of giving a philosophically and
scientifically adequate definition of life�instead he offers one (which requires a great deal of
unpacking) at de Anima ii.4 415b13.
77 In my view Aristotle accepts intuition (nous) as a necessary component of any satisfactory epistemology. This view is heavily embattled in contemporary literature on Aristotle's epistemology: see Lesher (1973), Hamlyn (1976), Burnyeat (1981), and Engerg-Pedersen (1979). Irwin (1988) supports the position for Aristotle's early works, but denies it for Aristotle's later philosophical works. I argue that Irwin does not establish this switch on either textual or philosophical grounds in Cameron (1998). This is not the place, however, to defend Aristotle's epistemological commitments in greater detail. For purposes of this dissertation Aristotle's commitment to dialectic as a way to first principles in science is sufficient to ground my claims (see Top. i.2 101a25-b4). 78 See in this regard especially Aristotle's comments on which opinions it is necessary to examine in ethical and dialectical contexts. Here are a couple of examples: "Dialectic does not construct its deductions out of any haphazard materials, such as the fancies of crazy people, but out of materials that call for discussion" (Rhet. i 1356b35) and "Educated men lay down broad general principles; uneducated men argue from common knowledge and draw obvious conclusions. We must not, therefore, start from any and every opinion, but only from those of definite groups of people � our judges or those whose authority they recognize" (Rhet. ii 1395b30; cf. also Rhet. ii. 12-14 for a description of the typical character traits of a young man, an old man, and a man in his prime, as well as 105a3, 1094b27 ff., 1095a28, 1096a9, 1179b25, 1214b29 ff., 1217a9, 1364b12 ff., 1394b7, etc.) 79 Theories are 'too abstract' when they are too far removed from the principles or phenomena at hand (GA ii. 747b29; cf. 698a13, 748a8, & 1235a2). See Meteorology i.13 for an 'ingenious' scientific account that is 'plainly false' and compared to which the 'unscientific account' is better. Note also, "If the axiom is relevant but too implausible, the answerer, while admitting that if it is granted the conclusion sought follows, should yet protest that the proposition is too silly" (Top. viii 160a6) and "the argument of Melissus is gross and offers no difficulty at all: accept one ridiculous proposition and the rest follows � a simple enough proceeding" (Phys. i 185a10; also 193a4 ff., 1218b22, 1006a5, 1217a9, & 163b13).
46
Further, regardless of one's position on intuition and the role of 'conceptual analysis'
either in Aristotle or in contemporary philosophy, Aristotle's procedure of attempting to offer a
satisfactory analysis of life is certainly not misguided at this point. No theorist we have
investigated80 has offered any reason other than the thesis of material continuity (and, perhaps
implicitly, the perceived failures of past attempts) to believe that a definition of life that satisfies
both our everyday intuitions and legitimate scientific constraints cannot be given. We have
already seen, indeed, that the thesis of the material continuity of living things � the empirical
core of the continuity thesis � while true gives no support for the claim that life has no
intuitively plausible definition.81 In the absence of such a principled reason, surely the attempt to
satisfy the strong demands of such an account has both philosophical and scientific interest.
It is plausible, then, to suppose that Aristotle sees no conflict between the goals of
arriving at a considered theoretical position which also meets constraints set by our everyday
notions concerning life. Aristotle would appear to seek both a conceptual analysis and a
theoretical definition of life.82
1.3.4 Section summary.
In this section I have argued that Aristotle's view on life holds contemporary interest
and promise for a number of complementary reasons. Aristotle's view is plausibly construed as
unrefuted by the scientific evidence that succeeds in refuting vitalism, yet Aristotle's view is
unique on the contemporary scene for offering a nonmechanistic account of life. If this is
Aristotle's position, then he occupies an intrinsically interesting and currently unoccupied
location in the space of possible accounts of life. Further, Aristotle's acceptance of the empirical
80 See Bedau (1991, 457), Sober (1992, 375-6), Warren (1997, 26), and Harnad (1995, 298) quoted above. 81 We do, after all, find the essentialist account of chemical kinds in terms of their atomic numbers highly intuitively satisfying given that our intuitions indicate that these are compositional kinds whose natures will be revealed to us, if at all, not through a priori investigation but through empirical research. 82 This view is argued to be a pervasive feature of Aristotle's philosophical and scientific approach by Shields (1999).
47
core of the continuity thesis without its philosophical presupposition of mechanism displays
another respect in which Aristotle is unique, sound, and sharp in his philosophical commitments.
Finally, Aristotle's project is ambitious and interesting in that he promises to provide an account
of life that is both scientifically and intuitively acceptable, bucking the contemporary trend to
abandon without argument the attempt to satisfy intuitional constraints.
1.4 Conclusion.
Our brief survey of the contemporary literature on the nature of life has revealed an area
of investigation rich in substantive philosophical and scientific problems remaining to be
resolved. Furthermore, while a number of distinctive positions in logical space have been filled
by contemporary theorists, it remains the case that positions that hold both philosophical and
scientific promise have not yet received a developed treatment.
48
Chapter 2
Two Interpretations of Aristotle on Life
Recent scholarship has produced two important contributions to the interpretation of
Aristotle's account of life: Gareth Matthews' 'De Anima 2.2-4 and the Meaning of Life' and
chapter seven, 'The Meaning of Life' of Christopher Shields' recent book Order in Multiplicity.1
Both of these accounts take their start from perceived inadequacies in such modern
interpretations of life as there are,2 and each attempts to ground a sound contemporary account
of life in an interpretation of Aristotle.
My dissertation takes up both of these goals. In this chapter I critically discuss each of
these two important accounts as an introduction both to Aristotle's doctrines and to motivating
my own approach. I argue that while neither account will do either as a definition of life or as an
interpretation of Aristotle, the central substantive insight into Aristotle's account of life
contained in each essay is correct. Further, the shortcomings of each account set the stage for
the work which remains to be done in the service of an adequate Aristotelian account of life.
2.1 Locating Aristotle's theory in the de Anima.
At the start of the de Anima, Aristotle's treatise on the soul, he tells us that one of the
reasons to be interested in a study of the soul is that "the soul is in some sense the principle
[archê]" of living things (402a6). At 403b23-8 Aristotle notes that what has soul has been
traditionally distinguished from what does not have soul by the possession of movement and
sensation.3 The book closes with a number of questions to be resolved in book ii: Are the
1 Matthews (1992), Shields (1999). 2 On the contemporary status of accounts of life, see my chapter one. 3 Incorporeality is added to the list at DA 405b11; cf. MM i.4 1185a17.
49
attributes of the soul4 attributes of the whole soul, or of parts of it? Does life depend on the
whole soul, or only a part of it? "Or has it some quite other cause?" (DA i.5 411a24-b4)
Aristotle begins a fresh start in book ii of the de Anima, quickly moving through a
number of general accounts of the soul.
The soul must be a substance (ousian) in the sense of the form ((h)ôs eidos) of a natural body having life potentially within it. But substance is actuality, and thus soul is the actuality of a body as above characterized. (DA ii.1 412a19-22)
The soul is an actuality of the first kind of a natural body having life potentially in it. The body so described is a body which is organized (ê(i) organikon). (DA ii.1 412a27-9)
If, then, we have to give a general formula applicable to all kinds of soul, we must describe it as an actuality of the first kind of a natural organized body (entelexeia (h)ê prôtê sômatos phusikou organikou). (DA ii.1 412b4-6)
What is soul? It is substance (ousia) in the sense which corresponds to the account of a thing. That means that it is what it is to be (to ti ên einai) for a body of the character just assigned. (DA ii.1 412b10-1; see also DA ii.2 414a28; PA i.1 641a25-9; GA ii.4 738b27; Met. vii.10 1035b14-31, vii.11 1037a5, viii.3 1043a35)
Aristotle believes that he has given a satisfactory general definition, but moves on in de Anima ii.2
to investigate the matter from the point of view of what is more observable to us in order to
elucidate the cause of the definition (DA ii.2 413a11-20).
From this point of view, it turns out that the soul is the principle (archê) of living things
and is defined by (horistai) powers it possesses (DA ii.2 413b11-13). These powers are specified
with reference to the fact that the soul is the cause of life.
Let us say, then, in taking up a new starting point for our inquiry, that what is ensouled is distinguished from what is unensouled by living. But living is spoken of in many ways, and if even one of these belongs to something, we say that it is alive, that is: thought; perception; motion and rest with respect to place; and further motion with respect to nourishment, decay and growth. For this reason all plants too seem to be alive. (DA ii.2 413a20-6. Translated by Shields.)
4 The attributes are listed here as knowing, perceiving, opinion, desiring, wishing, appetite in general, local movement, growth, maturity and decay.
50
Given Aristotle's apparent commitment in this passage to the claim that something is alive just in
case it possesses one of a number of capabilities, one might suppose that Aristotle here offers a
cluster definition of life.
Cluster definitions seek to indicate a number of different properties associated with the
definiendum, and 'define' it as possessing one or another of the indicated properties.5 Thus, we
might suspect that Aristotle's account of life is:
x is alive =df x has one of the following powers: (a) thinking; (b) perception; (c) local movement and rest; (d) the movement of nutrition, growth, and decay.
One problem6 with attributing such a definition to Aristotle is that Aristotle does not typically
offer disjunctive lists as definitions; his standard form of definition instead involves the
specification of genus and differentia.
Further, such a definition will inherit the main problem associated with any cluster
definition, the problem of unity: Why are these powers on the list and others not?
The problem of unity (PU): A condition of adequacy on an adequate definition of a property, P, is that it explain the unity of the items which fall into the extension of P: what is it that makes them all P?
If we cannot find the answer to this question, it seems we will not have given an explanatory
account of life (i.e., the account is too weak). In fact, if we have reason to believe that this cannot
be done, we may well abandon the project of definition altogether and become sceptics
concerning the possibility of defining life. If we can satisfy the requirement stated in (PU), then
it seems as though that answer � and not the cluster of properties referred to � will be the
definition of life (i.e., the account will thus be too strong to remain a cluster account).
As a next attempt at Aristotelian exegesis, we might attempt to find unity in the cluster
of properties Aristotle indicates by noting that he believes these powers of the soul to be
5 On cluster definitions, see chapter one and Mark Bedau (1996). 6 In what follows I am covering ground which has already been well covered by Matthews (1992) and Shields (1999); I make no claims for originality in this portion of my brief introduction to Aristotle's account of life.
51
arranged hierarchically. The power of nutrition can be had without sensation or thought (as in
plants), but the powers of thought and sensation (at least in biological organisms) cannot be had
without the power of nutrition (see DA ii.2 413a32-4, 413b4-13, ii.3 414a29-415a12; PN 454a11-
454a18, 474b10-12). Perhaps, then, the original list jumbles together some marks of life
(sensation, thought, local movement) while the definition of life is in terms of one member of the
original list: the power of nutrition. This hypothesis appears supported in the texts. Aristotle
says that it is the power of nutrition that "is the originative power the possession of which leads
us to speak of things as living at all" (DA ii.2 413b1-2). Further,
By life we mean self-nutrition and growth and decay. (DA ii.1 412a14)
The nutritive soul is found along with all the others and is the most primitive and widely distributed power of soul, being indeed that one in virtue of which all are said to have life. (DA ii.4 415a24-5)
Generation is the initial participation, mediated by warm substance, in the nutritive soul, and life is the maintenance of this participation. (PN 479a28-9)
The nutritive soul ... is the nature [hê phúsis] of every organism (GA ii.4 741a1).
Such a definition would appear to solve our problems since it can be formulated in canonical
form, and since it gives unity to our original cluster: having the nutritive soul explains a thing's
being alive, but sense perception, thought and local movement are marks or signs of life that are
in some way connected to this power.
This hypothesis fails as well, however, for two reasons. The first reason is that the
definition falls prey to another problem, the problem of circularity.
The circularity problem (CP): Definitions of F must be stated in terms of concepts are explanatory of being F.7
The definition of life in terms of the possession of the nutritive soul appears not to avoid this
problem, for, as Aristotle says, "it is by this function of absorbing food that this psychic power is
7 Defining circularity is a difficult matter given the possibility of small families of terms � such as modal, ethical or physical terms � that may only be defined by displaying the interrelations between the concepts within the families. Still, the danger is tolerably clear.
52
distinguished from all the others" (DA ii.4 416a20-1). However, Aristotle's analysis of food
reveals that we cannot define food or the power to take it in independently of living.
Since nothing not alive can be fed, the ensouled body (qua ensouled) would be the thing fed, and thus food does not merely happen to be related to what is ensouled. Food has a power which is other than the power to increase the bulk of what is fed by it; so far forth as what has soul in it is a quantum, food may increase its quantity, but it is only so far as what has soul in it is a 'this-somewhat' or substance that food acts as food. (DA ii.4 416b10-13, first sentence my translation. cf. GA 335a15-6)
Somehow we need to distinguish the taking on of nutrition that sponges undergo in wiping up
kitchen spills and the feeding that fires undergo from the feeding/taking on of nutrition that
living things do. However, it isn't clear at this stage how that distinction can be drawn without
baldly distinguishing between the sort of taking on of nutrition which living things engage in vs.
that of nonliving things. Such an expedient obviously falls prey to CP.
Because the proposed Aristotelian definition of life in terms of the nutritive soul falls
victim to CP we hope that Aristotle does not endorse it. We have strong reason to believe that
he doesn't, thankfully, since Aristotle is unabashed in supposing god to be alive in virtue of god's
possession of nous.
Life also belongs to God; for the actuality of thought is life, and God is that actuality; and God's essential actuality is life most good and eternal. We say therefore that God is a living being, eternal, most good, so that life and duration continuous and eternal belong to God; for this is God. (Met. xii.7 1071b26-30)
Everything which has a function exists for its function. The activity of God is immortality, i.e. eternal life. (OH 2.3 286a8)
If you take away from a living being action, and still more production, what is left but contemplation? Therefore the activity of God, which surpasses all others in blessedness, must be contemplative. (NE 10.8 1178b20)
Our intuitions about living, like Aristotle's, allow that it is possible for there to be non-biological
life, and Aristotle's account takes this possibility seriously.
This feature of Aristotle's account highlights the fact that his approach to the definition
of life is set apart from contemporary accounts not only substantively, but also methodologically.
Contemporary accounts abandon the evidence of intuition as reflecting outmoded ways of
53
thinking whose death came with Darwin's revolution in biology.8 Part of the project of this
dissertation is to show that this verdict is out of place: we can have both a scientifically
acceptable biology and an adequate account of life which respects longstanding, widely shared,
and dialectically firm intuitions concerning what is alive. In such a situation, it is a benefit that
Aristotle's account will meet the constraints not only of science, but also of scientifically and
philosophically informed intuition.
On this way of proceeding, however, we escape the frying pan of attributing a circular
definition of life to Aristotle only to fall into the fire of a renewed and strengthened problem of
unity.
The radical problem of the unity of definition (RPU): A condition of adequacy on an adequate definition of a property, P, is that it explain the unity of the actual and possible items which fall into the extension of P. Adequate definitions must respect our modal intuitions about the possible extensions of properties.
Applied to the problem of the definition of life, RPU plus the intuition that non-biological forms
of life (e.g., god) are possible yields the problem of finding a definition of life which will cover
biological and non-biological cases. Aristotle's account of life, if it is to be satisfactory in its own
terms must apparently be weak enough to avoid CP but strong or rich enough to meet
conditions of adequacy concerning the unity of definitions that account not only for the actual,
but also the possible diversity of life.
Finally, we may note that Aristotle is committed to theses concerning the definition of
life that appear to be in tension with the radical problem of unity just described. In addition to
needing to discover the unity in the diverse phenomena life presents to us both empirically and
in our modal intuitions concerning the possible extension of living beings, Aristotle maintains
that there is an important sense in which life is multivocal; what it is to be alive is neither exactly
8 See chapter one's discussion of the naked mind thesis for a defense of this claim. Peter van Inwagen's (1990) account is a possible exception.
54
the same across its instances, nor do the phenomena present us with radical disunity.9 As
Aristotle says,
[W]hat is ensouled is distinguished from what is unensouled by living. But living is spoken of in many ways. (DA ii.2 413a21-2. trans. Shields.)
[L]ife seems to be not one kind of thing only ((h)ê de zôê ou kath (h)en eidos dokei legesthai), but one thing in animals and another in plants. (Top. vi.10 148a29)
Thus, we have a constraint on a proper understanding of Aristotle that pulls in the opposite
direction as the radical problem of unity; we must reconcile and explain robust forms of both
unity and disunity between the life of living things. We may thus add to our list of problems the
problem of multivocity.
The problem of multivocity (PM): Life, as Aristotle conceives of it, is multivocal. There exist distinct ways of being alive which any adequate account of Aristotle on the nature of life must distinguish.
This brief survey of prima facie promising accounts of life basing themselves in Aristotle's
text reinforces the conclusion of chapter one that there exist rich philosophical problems
surrounding an adequate account of life. Part of our motivation for turning to Aristotle in
approaching the problem is that Aristotle embeds his account of life in a rich philosophical
system which may be able to resolve the thorny problems surrounding life's definition. It is
unsurprising then, that simple answers which suggest themselves in Aristotle's texts fail. We
must look more deeply if we are to find an adequate Aristotelian account of life.
2.2 Matthews on the Meaning of Life
Matthews' account of life takes its central motivation from reflection on the radical
problem of unity for life's definition posed by god's (possible) life. If the only (possible) living
things were natural organisms, it might be possible for us to give an account of life in terms of
9 As we will see, the fact that Aristotle recognizes diversity or multiplicity in ways beings may live is easy to establish textually. That he recognizes fundamental unity is not simply intuitively plausible, but plausible as a general impression based on wide familiarity with his theoretical writings on life and living things. The claim of unity, is, however, more difficult to substantiate with specific texts, and more difficult to elucidate.
55
self-nutrition, the most basic and widely shared psychic function.10 The problem created by the
fact of (the possibility of) god's life is so drastic, however, that following Aristotle's
methodological practice Matthews turns to "try a very different approach" (190).11
Matthews' new beginning starts from the insight that Aristotelian individuals "naturally
act so as to preserve their species" (190, citing DA ii.4 415a27-b2). From this insight Matthews
develops his account of Aristotle around the notion of a 'psychic' or 'species preserving' power,
which is then employed in a definition of what it is to be alive.
x is a psychic power =df there is a species s, such that, for x to be preserved, individual organisms that belong to s must, in general, exercise x (191).
x is a psychic power for species s =df for s to be preserved individual organisms that belong to s must, in general, exercise x (191).
x is alive =df there is a species s, and a psychic power p, such that x belongs to s, p is a psychic power for species s, and x can exercise p (191).
Matthews explains, "what it means to say that an organism is alive is that it can exercise at least
one psychic power; that is, at least one of the powers that organisms of its species must, in
general, be able to exercise for the species to survive" (191).
Matthews believes that this account of life is Aristotle's, and that it solves three
problems any account must overcome: the circularity problem, the problem of the special role of
self-nutrition in the account of (biological) life � a problem that includes the special problem of
My own account of the unity of life is offered in chapter seven, where I take as my key text de Anima ii.3 414b20-415a1. I canvass the solutions of Matthews and Shields in the main text, below. 10 Of course we would have to get over the circularity problem to make this solution work. Can we give a non-circular, informative, scientifically adequate definition of the power to take on nutrition that includes all and only living things (i.e. includes plants and animals but excludes (household) sponges)? Aristotle scholars would have the additional task of explaining away Aristotle's apparent commitment to the homonymy of life as well. 11 Unless otherwise specified, all page numbers in this section are from Matthews (1992).
56
Aristotle's insistence on the multivocity or homonymy12 of life, and the radical problem of unity
created by (the possibility of) god's life.
While Matthews admits that there is a 'certain circularity' in his account, he does not
believe that the circularity is vicious. Matthews explains the circularity by noting that a specie's
preservation will involve keeping individual organisms in existence, but keeping individual
organisms in existence just is keeping them alive. Thus, a reference to keeping things alive sneaks
into Matthews' account of what it is to be a species preserving power, the central notion in his
account. Matthews does not believe, however, that this circularity is vicious. For something to
be alive is just for it to have a 'self-perpetuating power', or, more carefully, for it
to be able to exercise one power in a list such that it is necessary for individuals of [its] species, in general, to be able to exercise those powers for there to go on being individuals in that species that can exercise one or more of those powers. If we are justified in supposing that dead cats, dead trees, and dead human beings can't exercise any powers at all, then the circularity in the proposal is not, I think, objectionable. (192)
Thus, Matthews believes that his account escapes any vicious circularity.
Further, Matthews believes that his account explains both why self-nutrition plays such a
central role in Aristotle's account of life and why, on Aristotle's account, life is 'said in many
ways.' Among mortal living things the most basic species-preserving power is self-nutrition, and
this accounts for its special place in Aristotle's account of life. "Still, what exactly having species-
preserving, or psychic, powers amounts to varies from species to species. And in this way
something like [the homonymy of life] is also true" (193).
Finally, Matthews believes that his account can handle (the possibility of) god's life
despite the fact that the account is framed in terms of the possession of species preserving
powers. "To preserve their species," according to Matthews, non-mortal beings such as god
"need only preserve their existence by continuing to engage in whatever activity is essentially
12 I will frequently use the terms 'multivocal' and 'homonymous' and their variants interchangeably in what follows; in most contexts this should not cause serious confusion. When I employ the technical vocabulary surrounding Aristotle's theories concerning homonymy, however, I generally employ the later term.
57
theirs" (193). Thus, Matthews provides us with a gloss on his account which shows that and
how it applies even in the case of a living but non-biological god who is the member of no
species.
Matthews' account is important for any investigator into Aristotle's account of life
because of the seriousness with which he raises and attempts to deal with the major problems
confronting an account of life (Aristotelian or modern). Matthews attempts to deal with the
circularity problem, the problem of finding unity in a diverse set of 'marks' of life, and the
problem posed by the possibility of non-biological life (e.g., gods, angels, or devils). Further, the
guiding insight of Matthews' account, Aristotle's commitment to the teleological character of life
(415a27-b2), will form the cornerstone of my eventual account of Aristotle on the nature of life.
Despite the depth and insight of Matthews' account, however, the details of his analysis are
fundamentally flawed.
2.2.1 Shields' Objections to Matthews' account
Christopher Shields offers three criticisms of Matthews' account. Shields claims to
uncover a circularity problem different than that addressed by Matthews; he argues that
Matthews' account is too 'biocentric', and he argues that Matthews' does not adequately capture
Aristotle's commitment to the homonymy of life.
Shields believes that Matthews' account involves a circularity in that "the having of a
psychic power is simply the manifestation of a capacity of living things as such" (1999, 183).
Shields thus believes that being alive is a more basic property than the property of being a
psychic power; being a psychic power must be explained (Shields contends) in terms of being
alive rather than the other way around. Shields and Matthews therefore disagree over which
concept is explanatorily basic: being a psychic power or being alive.
Second, Shields argues that Matthews' account is too biocentric; Matthews' account
hasn't adequately addressed the problem which was supposed to have set Matthews off on a 'new
beginning' � the problem of (the possibility of) god's life. The problem is that god is not a
58
member of any species and does not act or think in order to reproduce or perpetuate a species.
According to Shields, Matthews' attempt to extend the account to include god fails to capture
Aristotle's striking claim that 'life is the actuality of mind' (Met. xii.7 1072b24-30, trans. Shields).
Finally, and most importantly for the account that he will be offering, Shields objects
that Matthews' account does not adequately reflect Aristotle's commitment to the homonymy of
life. Shields embeds his interpretation of Aristotle's account of life in a framework for
understanding homonymy claims generally which is rigorously developed and defended
throughout part one (chapters one through four) of Order in Multiplicity.13 Shields argues that
Matthews hasn't offered a homonymous account at all, but has instead put forward a second
order univocal account. On such second order accounts, even though realizations of the
property specified in the account (e.g., being in pain, being a poison, being a mind, being alive on
Matthews' account) may be distinct first order physical properties or structures, it is nevertheless
the case that the property instantiated by these distinct physical bases is univocal across all of its
applications. Given the rigorous theoretical background of Shields' approach to homonymy he
is understandably taken aback by the looseness of Matthews' assertion that "something like" the
homonymy of life is true on his account.
Shields' last objection points to the limitations of an article-length treatment of
Aristotle's conception of life. One of the great advantages of Shields' treatment of Aristotle on
life is that he self-consciously embeds the account within a thorough study of one of Aristotle's
central philosophical concepts. Such an approach properly reflects the central place life and
living things play in Aristotle's mature metaphysics and pays due respect to the difficulty of the
problems surrounding life's definition.
13 Shields' account of homonymy is given a more thorough exegesis below, pp. 64-70.
59
2.2.2 Further objections to Matthews' account.
While Shields gleans the correct moral for an Aristotelian account of life from the
looseness of various points in Matthews' account, I do not believe that he has diagnosed the core
of the problem with Matthews' account. Shields believes that Matthews strayed from his
intended target in losing sight of the need to account for (the possibility of) god's life.14 I
believe, however, that Matthews' account strays from its motivating insight (DA 415a27-b2)
prior to this point (i.e., even in the account of biological life). Further, the failure I elucidate below
explains the truth contained in both Shields' circularity objection and his observation of the
unsatisfactory character of Matthews' extension of his account to cover god.15
Allow me to illustrate. The central concept in Matthews' account of life is the concept
of a psychic power. Here is his definition again:
x is a psychic power for species s =df for s to be preserved individual organisms that belong to s must, in general, exercise x (191).
The first problem for Matthews' account is that on his definition it turns out that the heart's
ability to make thumping noises is a psychic power of creatures with hearts.16 This is made
apparent by plugging 'ability to make thumping noises' into Matthews' account of a psychic
power: for creatures with hearts to be preserved it must in general be the case that their hearts
exercise the power to make thumping noises (this is true because when hearts perform their
function, circulating blood, they inevitably make thumping noises). If too many hearts cease to
exercise the power to make thumping noises, the species will not be preserved. But surely the
14 To rectify this oversight, Shields centers his account around Metaphysics xii.7 1072b24-30 and "the striking locution that 'life is the actuality of mind'" (Shields 1999, 183). Taking this phrase seriously, the core of Shields' account will be the notion of a native intentional system. 15 Issues surrounding the homonymy of life are resolved in this dissertation first through an critical investigation of Shields' account in this chapter, below, and then in chapter seven through a positive account of both the unity and diversity Aristotle claims to find in the phenomena of living beings. 16 See DA iii.9 432a-b4.
60
heart's ability to make thumping noises is not a psychic power � it is instead a mere
concomitant of a psychic power.
The example I've just given is framed in modern terms that foreshadow my diagnosis of
this problem with Matthews' account. It is important to note, however, that despite the blatantly
anachronistic nature of the example Aristotle draws the same distinction in his own terms. Thus,
with Aristotle we must distinguish between the definitory powers of the nutritive soul and the
inevitable concomitants of those powers (such as the production of useless residues and the
eventual death and disintegration of the organism � see PA iv.2 677a12-31, Pol. vii.8 1328a24-6,
GA v.1 778a32-b6, b11-19). Generalizing from this case, my first conclusion is that Matthews'
account fails because it cannot distinguish psychic powers (i.e. pumping blood, the activity of the
nutritive soul) from accidental concomitants of psychic powers (e.g., making thumping noises,
producing useless residues).
Another, related failure may be illustrated with a different example. I begin by noting
that in order to save his definition from circularity Matthews must intend us to understand
'organism' in his series of definitions not to mean 'biological entity' but to mean, roughly and
more generally, 'thing.'17
x is a psychic power for species s =df for s to be preserved individual organisms [i.e. things] that belong to s must, in general, exercise x (191).
Framing the account in terms of organisms already picks out the phenomena we desire to explain
� we would appear to be using a prior understanding of what it is to be alive when we select
just those things we want our definition to apply to.
Once this basic point is exposed, however, we see immediately that the account is far
too broad. In order for sounds to preserve themselves in echo chambers they must, in general,
exercise their power to bounce off of walls. In order for radioactive isotopes to remain the
61
isotopes they are, they must decay at a given rate (when they cease to do this, they have become
a different isotope).18 In order for chairs to preserve themselves in existence, their parts must
exercise their powers for retaining structural integrity. Generalizing from these cases, we may
conclude that for any continuing subject there will be some list of powers such that subjects of
that kind must exercise those powers to continue in existence; further, any such power may play
the role of a 'psychic power' in Matthews' account of life. Matthews' account of life is therefore
far too liberal; its central concept (if understood in a non-circular fashion) applies to all
continuing subjects and not simply to living things. If Matthews' account is not circular then it is
far too broad.19
As evidence that this problem genuinely arises for Matthews' account, note that the fault
I've just illustrated explains Matthews' own extension of his account to cover god's life.20 The
prima facie problem Matthews' account had with god is that while god is alive, it is the member of
no species while Matthews' account is framed in terms of species-preserving powers. Matthews'
solution to this problem is to extend his account along the lines marked out in the preceding
criticism. He argues that god is alive just in case it exercises powers necessary for it to remain in
existence (193; quoted above); but this condition will by satisfied by any continuing subject (alive
or not),21 just as I argued in the last paragraph.
I conclude that something has gone terribly wrong with Matthews' account. To
summarize, the account is too weak in two ways. First, it fails to distinguish between psychic
17 We must make a similar move for 'species' in Matthews' definitions; 'species' must be read throughout such that it means only 'kind' if we do not wish to unconsciously restrict the account's applications to biological kinds. 18 I owe this example to Robert Pasnau. 19 It may be objected that none of the counterexamples I raise truly 'exercises' the powers I've listed. The problem with this attempted reply, however, is that it reintroduces the problem of circularity into the account. Which things 'exercise' powers in the appropriate way, after all? Living things. I owe this objection independently to comments by Christopher Shields and Robert Pasnau. 20 And, by the same token, Shields' objection that Matthews' account is 'too biocentric'.
62
powers and the accidental concomitants of psychic powers. Second, on one interpretation it falls
into the circularity problem (if 'organism' is understood to restrict the account to biological
organisms). On the interpretation which avoids the circularity problem, however, the account is
far too liberal since on this reading all continuing subjects turn out (incorrectly) to have the
'psychic powers' definitive of life.
My diagnosis of the cause of these failures is that Matthews lost the thread of his
motivating insight into Aristotle's account of life early on. Matthews took his start from
Aristotle's comment that:
. . . for any living thing that has reached its normal development and which is unmutilated, and whose mode of generation is not spontaneous, the most natural act is the production of another like itself, an animal producing an animal, a plant a plant, in order that [hina], as far as its nature allows, it may partake in the eternal and divine. That is the goal towards which all things strive, that for the sake of which [ekeinou heneka] they do whatsoever their nature renders possible [hosa prattei kata phusin]. (DA ii.4 415a27-b2)
In this passage Aristotle asserts the teleological character of life. The most natural act for any
living thing is its action for the sake of participation in the eternal and divine (cf. Plato's Symposium
206e-7a).
Aristotle claims that living things act for the sake of preservation in the only way they can.
Matthews' first gloss on Aristotle's commitment to teleology is to gloss it as a thing's acting to
preserve its species. While this phrase is open to a teleological reading, it is also open to a
nonteleological reading in a way that Aristotle's unequivocal statement is not. Matthews' next
step in the cleansing process removes all teleological taint from the idea that living things 'act to
preserve' their species. The result of this cleansing is Matthew's final account, where no
teleological concepts have any place. In Matthews' final account, 'acting for the sake of' kind
preservation is reduced to the mere possession of powers that do in fact result in kind
preservation when exercised; exercising those powers need not be teleologically directed towards
21 Again, this is true so long as we do not understand 'exercises' in such a way that it introduces a circular reference to the 'exercising' of powers that only living things can do.
63
the end of kind preservation (191-2).22 Matthews' account, in effect, offers us a reductive
analysis of teleological commitment in terms of the possession of (one on a list of) necessary
conditions for existence,23 and his account inherits all the problems of this inadequate analysis of
teleological commitment.24
This inadequate account of teleological commitment explains the account's failure to
distinguish psychic powers (e.g., the exercise of the nutritive soul) from accidental concomitants
of psychic powers (e.g., the production of residues or death). The problem besetting Matthews'
account in this regard is well known in the literature on biological teleology,25 and its solution is
considered by contemporary philosophers of biology a necessary condition on the adequacy of
an account of teleology. In addition, the elimination of teleological elements from Matthews'
final account also explains why it is possible to extend the account to non-teleological entities
that are plainly not alive such as sounds, radioactive isotopes, and chairs. Such entities (indeed
any continuing subject) will exercise powers necessary for their preservation; none will do so for
the sake of their preservation.
How did Matthews' account stray so far from its guiding insight in Aristotle's
commitment to the teleological character of life? It would appear that Matthews � perhaps
driven by the thought that an irreducibly teleological account of life would not have been
scientifically adequate � attempted to 'sanitize' teleological commitment out of Aristotle's
22 Problems of this sort concerning teleology are dealt with at greater length and more explicitly in chapters three and four of this dissertation. I refer readers with questions concerning my objections to Matthews' account to those discussions for further analysis and defense of the claims made briefly here. 23 The reductive account follows that of Hempel (1965). For a brief evaluation of the problems with this account of teleology, see Buller (1999b). 24 We ought to note, further, that Matthews supplies no textual basis for supposing that Aristotle thought of teleology in this reductive way. This shortcoming will appear again and again in reductivist accounts of the ontic basis for Aristotle's commitment to teleology. See chapter four of this dissertation for details. 25 How are we to distinguish between functions and the accidental concomitants of functions? I believe Hempel first raised the problem.
64
account.26 Whatever his motivations, we can see that Matthews' very first exegetical comments
on de Anima ii.4 415a27-b2 lead us away from Aristotle's text.
This survey of Matthews' account vindicates my claim that any satisfactory account of
Aristotle on the nature of life must embed itself in deep metaphysical discussions at the heart of
Aristotle's metaphysics � in this case, discussions of the proper analysis of his teleological
commitment. The failure of Matthews' account to adequately capture Aristotle's commitment to
the teleological character of life leads to the account's eventual failure; any adequate teleological
account of life must do better.
2.3 Shields on the Meaning of Life
As I have already noted, one of the attractive features of Shields' account is the fact that
it embeds an account of life in a rich metaphysical and philosophical context: a book-length
treatment of Aristotle's notion of focal or associated homonymy.27 In what follows, therefore, I
will first explain Shields' conception of focal homonymy in Aristotle. Second, I will illustrate
Shields' application of the concepts to the case of life. Finally, I will end with criticisms of
Shields' view; I do not believe that his analysis solves the problem of Aristotle's commitment to
the multivocity of life.
2.3.1 Shields on focal/associated homonymy.
As is well known, Aristotle begins his Categories by distinguishing between being F
homonymously and being F synonymously.
When things have only a name in common and the definition of being which corresponds to the name is different, they are called homonymous. Thus, for example, both a man and a picture are animals. These have only a name in common and the
26 Note that in discussing Dawkins (1989) Matthews claims contemporary adequacy for his account of life. He says that "if, for whatever scientific or non-scientific reasons, we want to deal with the threats to incoherence posed by encyclopaedia entries under 'life' of the sort I began this discussion with [i.e. cluster definitions of life], the best way to do so, I think, is to appeal to the picture I have constructed from DA 2.2-4" (Matthews 1992, 193). 27 As I use these terms, associated homonymy picks out a broader class of which focal homonymy is a type. This use diverges from that of Shields (1999).
65
definition of being which corresponds to the name is different; for if one is to say what being an animal is for each of them, one will give two distinct definitions. When things have the name in common and the definition of being which corresponds to the name is the same, they are called synonymous. (Cat. 1a1-7)
The concept of homonymy comes to play a central, and enriched, role when Aristotle introduces
the science of being qua being in Metaphysics iv.1.
In explaining the possibility of a science of being Aristotle notes that homonymy is
compatible with the unity required for a science so long as there is a core sense of being F
around which other ways of being F are organized.
There are many senses in which a thing may be said to 'be', but they are related to one central point, one definite kind of thing, and are not homonymous.28 Everything which is healthy is related to health, one thing in the sense that it preserves health, another in the sense that it produces it, another in the sense that it is a symptom of health, another because it is capable of it. And that which is medical is relative to the medical art, one thing in the sense that it possesses it, another in the sense that it is naturally adapted to it, another in the sense that it is a function of the medical art � and we shall also discover other things said in ways similar to these [homoiotropos de kai alla lepsometha legomena toutois] � so too is being said in many ways, but always relative to some one source [pros mian archên-following Shields for the last two phrases]. (Met. iv.2 1003a33-b6)
It would seem, then, that to give an adequate account of Aristotle's analytical tools in this area,
we must distinguish three ways Aristotle recognizes in which x and y can both be F.
x and y are synonymously F iff (i) x is F and y is F and (ii) the accounts of being F in 'x is F' and in 'y is F' are the same.
x and y are homonymously F iff (i) x is F and y is F and (ii) the accounts of being F in 'x is F' and in 'y is F' are different.
Homonymy may take two forms.
x and y are discretely homonymously F iff (i) x is F and y is F; (ii) the accounts of being F in 'x is F' and in 'y is F' are different; and (iii) the accounts of being F in 'x is F' and in 'y is F' have nothing important in common.
28 Shields (1999, §1.3) argues convincingly that, despite appearances, Aristotle is not denying the homonymy of being at this point, but is instead distinguishing between two forms of homonymy: discrete and associated homonymy. Aristotle here denies that being is a discrete homonym (like 'animal' from Categories 1a1 ff.) but affirms that being is an associated homonym like health. These distinctions are drawn and further clarified in the main text, below.
66
x and y are associated homonyms F iff (i) x is F and y is F; (ii) the accounts of being F in 'x is F' and in 'y is F' are different; and (iii) the accounts of F in 'x is F' and in 'y is F' have something important in common.
Intuitively, the distinction between the two sorts of homonymy, discrete and associated, may be
illustrated as follows.29
'Bank' displays discrete homonymy in 'Sara ate on the bank [of the river]' and 'Sara
deposited a sum in the [savings] bank'. It is a mere accident of language that the one word,
'bank,' applies to these two very different things. However, it is no mere accident of language,
according to Aristotle, that both my enchilada and I are (and are called) healthy. Although what
it is for each to be healthy is different, our being healthy is no mere accident of language but
instead reflects a deep relationship between the two: my being healthy is my being in a sound
bodily condition whereas my enchilada's being healthy is its being productive of or good for
maintaining a state of sound bodily condition in things like me. There is, thus, an important
definitional connection between my being healthy and my enchilada's being healthy (one sort of
health is productive of the other sort of health) whereas there is no important connection between a
river bank's being a bank and a savings bank's being a bank.
This much is standard in the literature on focal homonymy. Shields' unique and
important contribution to the literature lies in his analysis of what it is for the connections in
question to be (or not be) 'important.' If someone were to claim that the definition of 'savings
bank' was 'financial institution within five hundred miles of a river bank' we would not want to
say that the person had established a genuine association between the two senses of 'bank' even
if it were true that all savings banks were within five hundred miles of river banks. But how are
we to draw the distinction between genuine and spurious cases of connection in a principled
way?
29 Shields actually distinguishes three basic cases of homonymy, distinguishing as he does between two forms of discrete homonymy (1999, pp. 29-31). I will ignore this distinction until it becomes important again in chapter seven.
67
Shields' solution (following Cardinal Cajetan, De Nominum Analogia 2.9) is to restrict
legitimate forms of connection to Aristotle's four causes. Shields distinguishes between core and
non-core senses of being F30 and restricts genuine cases of associated homonymy to cases where
the non-core instances of F stand in a relation of asymmetrical definitional dependence upon
core ways of being F. Further, to be legitimate, this asymmetrical dependence must take the
form of one of Aristotle's four causal relationships. Thus, in the example above, my being
healthy is a core instance of health and my enchilada's being healthy � while not a core instance
of health � is an efficient cause of health in the core sense: eating the enchilada helps to
produce/maintain my (core) health. Since there exists a four-causal definitional dependence of
the non-core case of being F upon the core case of being F we have an instance of associated
homonymy, 'focal homonymy'.31
One problem for this account, worth exploring because of the central role it will play in
Shields' account of life, is that "None of Aristotle's illustrations [of focal homonymy] is an
obvious instance of a formal causal relation" (114). We may, therefore, agree with the
Shields/Cajetan proposal for delimiting genuine cases of focal homonymy, but restrict it to three
out of the four causes Aristotle mentions. For, if being F is the formal cause of a's Fness and of
b's Fness, is it not the case that a and b are F in exactly the same sense and hence that they are F
synonymously and not homonymously? Shields remarks, "This conclusion would restrict FCCP [his
account of focal homonymy] by precluding the possibility that a non-core homonym could ever
stand in a formal causal relation to its core instance" (115).
30 'Senses' as Shields uses the term are not exclusively linguistic or propositional items, but instead come in 'shallow' and 'deep' varieties. Shallow senses are available to competent speakers of the language simply through reflection on their knowledge of the language. 'Deep' senses are identical with properties and may be revealed only after rigorous philosophical or scientific analysis. For an Aristotelian parallel, see Aristotle's criticism of Democritus's contention that "it is evident to every one what form it is that makes the man, seeing that he is recognizable by his shape and color" at PA i.1 640b30-41a17. 31 In chapter seven I distinguish a further type of associated homonymy, one which I claim to be recognized by Aristotle and to be crucial to understanding the form his definition of life takes. Shields does not recognize this form of associated homonymy in his work, and hence equates what I am calling 'focal' and 'associated' homonymy.
68
Shields believes that such a conclusion "rests on an unduly narrow conception of formal
causation," however. The assumption that "formal causation is sufficient for univocity. . . . is
false" (116). Shields illustrates its falsity through Aristotle's account of perception. On this
theory, in cases of perception the sense faculty receives the form of an object of perception
without its matter (DA ii.12 424a18-24, 424a32-b3, iii.2 425b23, iii.8 431b28-32a2). In such a
case the object of perception comes to be F (i.e. white) and the sense faculty comes to be F (i.e.
white) � but they are F differently. The sense faculty does not exemplify Fness, as the object of
perception does. Rather, the sense faculty encodes Fness; it is "in a state corresponding to - fully
representative of" Fness (116).
Encoding F, Shields maintains, is a form of being F, both in Aristotle and in fact.32
When a table becomes red (by being painted red) and a perceiver's sense faculty (sense datum,
phenomenal experience?) becomes red (by perceiving the table being painted red) both the
perceiver and the table take on the form of redness,
but it should not be inferred that they are red in the same way. Because of the possibility of distinct types of formal realization, it is possible for Aristotle to reject any direct inference from formal sameness to synonymy. It is therefore possible for him to hold that a and b are F, that a's being F is a formal cause of b's being F, and for a and b nevertheless to be non-synonymously F. It is, consequently, possible for a's being F to be a formal cause of b's being F, even while a and b are homonymously F. (117)
Even if we accept everything up to this point, however, this result would not establish that there
are cases of focal homonymy such that core and derived instances of F are related by formal
causation.33 All Shields has established thus far is that formal causation is compatible with
homonymy. Shields must now go on to show that there are cases of focal homonymy �
recognized correctly as such by Aristotle � which involve formal causation between core and
non-core instances.
32 I disagree. While there may be cases in which it might be appropriate, albeit strained, to claim that an encoding of F is itself F, there are clear cases where this does not hold. The sentence 'The table is red' encodes red, but the sentence itself is not red. 33 Shields recognizes this point explicitly; see p. 117.
69
Shields thinks that 'perhaps' Aristotle illustrates a case of this sort in a difficult passage at
Metaphysics iv.2 1003b2-3. The case concerns things which are medical by being 'well-constituted
relative to' medicine. Shields suggests that the different cases of being well constituted relative to
medicine may reflect the difference between medical doctors being medical through their craft
training and folk-healers being well-constituted relative to medicine by realizing the form which the
medical doctor has in virtue of her training, but "incompletely or inchoately, or at any rate
differently from the way a trained physician manifests it" (118).34 From this example, Shields
concludes:
[I]t seems clear minimally that Aristotle expects some non-core homonyms to stand in formal causal relations to core homonyms. Further, he is justified [by the argument involving encoding] in assuming that formal causation is not by itself sufficient for univocity. We can therefore add formal causation to material, efficient, and final causation as an instance of the kind of relation Aristotle expects core-dependent homonyms to realize towards one another. (118)
Shields proceeds to use this analysis of focally connected homonyms to give an account of life
on which life is a focally connected homonym; god's life is the core case of life, and all other
forms of life (plant, animal, and human) are differently formally caused to be alive by god's
living.
34 At footnote 10 on page 112 Shields explains the most likely meaning of the passage in question (1003b2-3) as being that "someone who has the right form of intelligence together with other appropriate dispositions will count as medical even prior to training (sometimes adults refer to children as 'scientists' while faculty members call certain undergraduates 'professors' or 'senators' in much this sense)" (112). If this is the sense of being F at stake, it appears to be roughly the sense in which someone with a strong natural aptitude for music might be said already to be a pianist before taking lessons. In this sense we might say that x is F iff (i) x has a first level potentiality for being F and (ii) x has already displayed an aptitude for acquiring second level potentialities of the F sort (see NE i.9 1100a3-4). This seems, however, to be a different sense than the one used in the text at p.118 where folk healers are said to have medicine 'inchoately.' Shields does not give us any guidance on how we are to reconcile these two seemingly distinct accounts.
On first and second level potentialities, see especially DA ii 417a22-31, but also 417b31 ff., 412a9, 412a22, 412b27, & 429b6. A human with normal language capacities who does not know German has a first level potentiality for speaking German (whereas a rock does not have such a potentiality) and a speaker of German not currently speaking German has a second level (developed, ready to be actualized at will) potentiality for speaking German.
70
2.3.2 Shields' analysis of life
Shields takes as his two fundamental starting points in interpreting Aristotle's account of
life Aristotle's claim that life is homonymous (Top. vi.10 148a26-31, DA ii.2 413a20-6) and
Aristotle's "striking locution that 'life is the actuality of mind'" (183, see Met. xii.7 1072b24-30).
In order to establish his reading, Shields must establish three theses concerning the type of
account of life (synonymous, homonymous, etc.) Aristotle offers and must then fill in the
substance or details of this form with a specific Aristotelian account of life. Formally Shields
must show both that and why Aristotle considers life to be homonymous; next, he must establish
that life is a focal homonym by finding a plausible core way of being alive and then by
establishing the nature of the particular four causal relation that obtains between the core and
derived ways of being alive. Substantively, Shields must formulate specific definitions of both
core and derived senses of being alive.
2.3.2.1 The form of Aristotle's account of life.
2.3.2.1.1 Establishing homonymy.
Aristotle does say that life is 'spoken of in many ways' (DA ii.2 413a21-2) and that it
'seems not to be spoken of according to one form' (Top. 148a29), and these are standard ways of
indicating that he takes a thing to be homonymous. For this reason Shields is on firm exegetical
ground in attributing to Aristotle a thesis concerning the homonymy of life.35
Further, Shields develops an argument for the homonymy of life from a key passage in
the de Anima. Aristotle says,
The soul is the cause and source of the living body. But these <cause and source> are spoken of in many ways. Similarly, soul is a cause according to the ways delineated, which are three: it is a cause <as> the source of motion, <as> that for the sake of which, and as the substance of ensouled bodies. That it is a cause as substance is clear,
35 What type of homonymy, discrete or associated, and how these types are further analyzed is, of course, subject to critical debate. I engage in this level of debate with Shields over the type of associated homonymy Aristotle invokes both here and in chapter seven.
71
for substance is the cause of being for all things, and for living things being is life, and the soul is also the cause and source of life. (DA 415b8-14; trans. Shields)36
Taking his cue from the fact that when Aristotle says that something is the being of something
else "Aristotle means to highlight what is essential to Fs qua being F" (187; citing APr. 67b12,
Top. 133b33, DA 431a9),37 Shields mounts the following argument on Aristotle's behalf.
(1) For living things, to be is to be alive. [DA ii.4 415b13; PN 467b12-25]
(2) Being for Fs is what is essential to Fs. [An.Pr. ii.21 67b12, Top. v.4 133b33, DA iii.7 431a9]
(3) So, in the case of living things, essence is identical with life.
(4) The essence of Socrates is not identical with the essence of Pavlov, the dog; nor is it identical with the essence of my rose bush; nor again is it identical with the essence of god, which is complete actuality.
(5) Hence, life is not the same for Socrates, Pavlov, my rose bush, and god.
(6) Hence, 'life' is non-univocal. (187)
We thus have an argument for the homonymy of life in Aristotle growing from central elements
of Aristotle's metaphysics. Aristotle's essentialist theses about living things38 leads naturally, via
the thesis that for a living thing its being/essence is to live, to the homonymy of life.
2.3.2.1.2 Finding a core for life.
Shields notes that if Aristotle considered life homonymous, then it is plausible that he
considered it to be focally homonymous, since "life is a clear candidate for association; one is
36 Shields also cites the following two passages in support of his view: PN 467b12-25, which claims that among living animals "there must be a single identical part in virtue of which they live and are called animals" and Physics viii.4 255a6-10 which contains the statement that having motion derived from itself "is a characteristic of life and peculiar to living things." The relevance of Physics viii.4 255a6-10 is not clear to me in this context. See also Met. viii.3 1043b2, "soul and to be soul are the same"; GA ii.4 738b25, "the soul is the substance of a particular body"; and fragment B86. 37 See also Top. i.5 133b31-6; Phys. i.3 186a30-1, iii.3 202b9-10, iii.3 202b20-1, iv.2 219a20-1, 219b18-9, v.5 229a19, 229a28, vii.8 262a20, vii.8 263b7-8, 263b12-4; OH ii.14 296b8 ff., iv.4 312a19-21, iv.5 312a30-2; GC i.5 320b13-4, 320b23-5; DA iii.4 429a10-1, iii.7 431a13-4, 431a28, iii.9 432a20, iii.10 433b24; PN 449a14-9, 467b26-7; Met. vii.8 1033b31, viii.6 1045b18-9, xii.10 1075b4-5 and xiii.2 1077b1-4. 38 Thus, roughly, the essence of being a plant is the possession of a nutritive soul; the essence of being an animal is the possession of a sensitive soul; the essence of being a human is the possession of a rational soul.
72
bound to believe, even if one accepts its non-univocity, that 'life' is not a mere or accidental
homonym" (188). This expectation is borne out, Shields argues, given the fact that Aristotle
identifies life as the actuality of mind (Met. xii.7 1072b26-7). Given that Aristotle accepts that
non-minded beings (i.e., plants) live, this identification cannot contain the literal truth of
Aristotle's considered account of what it is to be alive, Shields argues. Rather, "because he thinks
life is non-univocal but that the various applications of life are focally related, Aristotle asserts
this identification as the core of life" (189). Thus, Aristotle identifies one way of being alive �
god's way � as the core way of being alive in this passage.
2.3.2.1.3 Defining the causal relation obtaining between core and derived cases.
If life is a focal homonym and god's life is the core way of being alive, then given the
Shields/Cajetan account of focal homonymy there must exist a four-causal relation tying core
and derivative cases of being alive together. Shields argues on the basis of the fact that the
relation cannot be plausibly considered to be either efficient or material (with a footnote about
the final cause), that the tie between core and derived instances of life in Aristotle involves the
formal cause.
2.3.2.2 The substance of Aristotle's account.
Given Shields' arguments that the definition Aristotle offers will involve a formal four-
causal relation between core and derived instances in a focally homonymous definition, he must
round out his account by filling this account structure with particular definitions of core and
derived ways of being alive. As I mentioned above, Shields takes his cue in elucidating the core
of Aristotle's account of life from Aristotle's striking claim that 'life is the actuality of mind.' In
light of this statement, Shields believes that Aristotle treats "intentionality as the core of living"
(192). The core account of being alive, then, is as follows:
73
x is alive =[core homonymous definition] x is an intentional system.39
Intentional systems are in Shields' terminology, we should note, systems "whose nature is
accurately explained by [their] functions and whose behaviour is reliably predicted by
understanding [their] behaviour in terms of [their] directing [their] energies towards those ends"
(192, cf. 189). Intentional systems, then, are teleological systems.40
On this account, biological entities are alive in virtue of standing in distinct (but focally
related) formal causal relations to god's manner of living. In Shields' terms, god's life is that of a
pure, enriched, everlasting, seamlessly actualized, complete, supreme and serene intentional
system41 � we may indicate this set of features by calling god a PEAC intentional system.
Plants, animals, and humans are formally caused to be intentional systems by god's life, but
despite this formal causal tie between god's life and mortal living things, such biological living
things are not alive either in the PEAC way or all in the same way � although we are never told
in exactly what sense they are intentional systems. Instead, each type of mortal living thing has
its own distinct way of living, related but not identical to god's way of living.42
2.3.3 Objections to the substance of Shields' account.
Shields argues that in Aristotle's view god's way of living is the core way of living and
that other ways of living stand in focally homonymous relations of formal causal dependence on
god's core way of living. Shields has two distinct ways of spelling this out. On the first, god is
alive in the sense that it is a PEAC intentional system, where being a PEAC intentional system
39 The account will require some revision in response to purported counterexamples. Guided missiles are intentional systems but not alive. Shields' response to this is to distinguish between things with native (underived, basic) intentional states and things that have derived intentional states � only native intentional systems are alive. See also the parallel treatment of the same counterexamples offered by Taylor (1986, chapter 4 §3 esp. pp. 123-4) for teleological systems. 40 Note, however, that this understanding of intentional systems undermines Shield's claim that on his account (and Aristotle's) "intentionality is the core of life" (192). 41 Shields uses each of these terms to describe god's life on pp. 189-93.
74
involves that system's being a pure, enriched, everlasting, seamlessly actualized, complete,
supreme and serene intentional system. Borrowing terminology from Shields' discussion of
formal focal connections, we may suppose that other types of living things are only intentional
systems "incompletely or inchoately" (118).
This account cannot be allowed to stand as it is, however, since the relations it specifies
are at base metaphorical. We do not understand in nonmetaphorical terms either what it is to be
a PEAC intentional system, or how derived systems derive their incomplete or inchoate
intentionality from the core way through relations of formal causation. Indeed, it is unclear that
incomplete or inchoate possession of a form would be homonymous possession of a form at all.
This problem may be overcome, however, by Shields' more specific statements
concerning intentional systems. On Shields' terminology, intentional systems are systems that
possess teleological directedness (pp. 189 and 192). Using this terminology, Shields suggests that
living things live differently in the sense that they possess different ends or goals, "a living thing's
ends partially determine what kind of thing the entity in question is, in line with FD, the
functional determination thesis" (189-90). If it is the diverse ends of living things that
distinguishes their ways of living, then we may have a way of understanding how it is that god's
life is meant to be different from the lives of mortal creatures on Shields' interpretation: god's
ends are different from the ends of mortal living creatures, and different types of creatures also
have different types of ends.
One problem with this account is that it does not help us to understand the focal
homonymy of life any better than the metaphorical expression attempted above. That I have
different ends than my dog does not suggest that my dog and I have ends only homonymously.
Instead, we have ends univocally, although those ends differ. If this is how we are to understand
the differences between different ways of living on Shields' account, then the gain in our
42 The fact that plants are not alive in exactly the way that god is alive explains, by the way, how it is both that intentionality can form the core of life and that plants are alive even though plants do not, either in
75
understanding is accompanied by the abandonment of plausible explanations of focal
homonymy.
Because of the unclarity in Shields' account concerning what relation is meant to obtain
between core and derived ways of being alive I do not believe that he has solved the interpretive
puzzle raised by Aristotle's theses concerning the unity and diversity of living things. As we will
see in the next section, I believe this shortcoming in Shields' account reflects deeper problems
with the form of analysis he believes Aristotle endorses.
Despite this point, however, I believe that Shields' teleological specification of ways of
living in Aristotle contains a second43 invaluable insight into Aristotle's account of living.
Matthews provided the first key insight: the centrality of teleological directedness to life. Shields'
explanation of the diversity of living beings in terms of their possession of distinct ends is the
second key insight. On my account, then, Shields develops Matthews' substantive insight that
life is at base teleological in the correct direction. Shields' account of life contains a key insight
into the substance of Aristotle's account of life but fails to explain Aristotle's claims concerning
the unity and diversity of life because it fails to specify the correct form for such an account.
2.3.4 Objections to the form of Shields' account.
In this section I argue, pace Shields, that in Aristotle's view life is not a focal homonym.
Shields is correct that in Aristotle's view life is 'spoken of in many ways' (DA 413a21-2 & Top.
148a29); he is incorrect, however, in his analysis of what this comes to with life and other similar
cases. I make my positive case for understanding this locution as it applies to Aristotle in
chapter seven; here I restrict my comments to showing that Aristotle did not conceive of life as a
case of formally related focal homonymy.
Aristotle's view or in fact, think. 43 The third portion of the solution comes in my positive account of the form Aristotle's definition of life takes, developed in chapter seven. In the text, above, I go on to argue that Shields' conception of the form of Aristotle's definition of life is unsupported and incorrect. In chapter seven I make the case for a new understanding of Aristotle's definition of life.
76
In Shields' view there are three exhaustive and mutually exclusive options for the form
of definition Aristotle offers: synonymous, discretely homonymous, or focally homonymous.
Shields then argues that since life is neither synonymous nor discretely homonymous it
constitutes a case of focal homonymy. This case is bolstered by his identification of what
Aristotle took to be the core way of living and of the relations of causal dependence Aristotle
took to hold between core and noncore ways of living. In what follows I argue against these two
supporting arguments. I argue first that Shields fails to identify a core way of being alive, and
second that he fails to specify a plausible form of causal dependence between core and noncore
cases of being alive. In chapter seven I argue that there is textual basis for a fourth form of
definitional relation which combines features of synonymy and discrete homonymy, and that we
should be thankful that such a fourth category exists for independent philosophical reasons.
2.3.4.1 The core of life.
If Shields can discover a plausible textual basis for Aristotle's commitment to life's
having a core sense upon which derivative senses depend, then he will have a correspondingly
strong case for the claim that Aristotle considered life to be a focal homonym. Without such a
textual basis, however, the claim rests solely on interpretive commitments concerning the modes
of definitional analysis open to Aristotle.44 Shields concludes that god's way of living is the core
form of life through the following argument (pp. 188-9).
1. Aristotle identifies life as the actuality of mind (Met. xii.7 1072b26-7).
2. Aristotle allows that some living beings, i.e. plants, are alive but have no mind.
3. If plants live but have no minds, then life cannot be literally identical to the actuality of mind.
4. Therefore, god's way of living (i.e., the actuality of mind) must be the core form of life upon which other forms of life depend.
44 Again, the commitments driving Shields' argument for focal homonymy are called directly into question by my argument in chapter seven. I argue there that Aristotle has resources for definitions which outstrip the synonymy, discrete homonymy, focal homonymy triad.
77
This argument fails to establish its conclusion, however, for we may work the same argument for
other forms of living.
Chapter 1 Aristotle identifies life with the possession of nutritive soul (DA ii.1 412a14, ii.2 413a31-b1, ii.4 415a23-5, PN 479a28-9, GA ii.4 741a1). or: 1''. Aristotle identifies life with the possession of the sensitive soul (NE ix.9 1170a15).
Chapter 2 Aristotle allows that some living beings, e.g., god, are alive but do not possess the nutritive soul.
Chapter 3 If god lives but has no nutritive soul, then life cannot be literally identical to the possession of the nutritive soul.
Chapter 4 Therefore, plants' way of living (i.e., the possession of nutritive soul) must be the core of life upon which other forms of life depend.
Aristotle seems to recognize three basic soul functions: nutrition/reproduction, sensation, and
thought. Further, at various places throughout the corpus he identifies life with each of these
soul functions. Given this, we cannot without further argumentation claim on the basis of an
identification of living with a particular power of the soul that Aristotle considers that power of
the soul to be the core power on the basis of the fact that other living things do not engage in
the activities associated with that power.
Now, I grant that it may be possible to think of Aristotelian reasons why we might
suppose that god's way of living constitutes the core of life rather than the life of sensation or
nutrition. What is striking, and what I wish to point to here, however, is that Aristotle himself
does not offer any such reasons. Aristotle seems to restrict himself to identifying various living
creatures with the possession of their peculiar (idion) ends, as is made clear at Nichomachean Ethics
ix.9: "life is defined in the case of animals (to de zên (h)rizontai tois zô(i)ois) by the power of perception,
in that of man (anthrôpois) by the power of perception or thought" (1170a15, emphasis added). In
these passages Aristotle makes clear that he is restricting his identity claims to particular forms or
ways of life, and we have no reason to suppose that he is making any stronger claim when he
discusses god's living in Metaphysics xii.7.
78
No commitment of Aristotle's suggests the additional claim that god's way of living
constitutes the core of a focal homonym.45 If we draw that conclusion it is because we recognize
that Aristotle considers life to be a type of homonym, but also believe that focal and discrete
homonymy exhaust the types of homonymy, and further, that it is wholly implausible to suppose
that life is a discrete homonym either in Aristotle or in fact. These are highly theoretical beliefs
concerning the understanding of Aristotle's principles, and the argument for them constitutes the
bulk of part one of Order in Multiplicity. This is not the time or place to address those arguments
directly, although in chapter seven I argue for the falsity of the theses by developing an account
of a further form of associated homonymy as a reading of certain theoretical and illustrative
passages in Aristotle's texts. At this stage I restrict my point to noting that Shields has not
established that Aristotle conceived of god's way of living as the core way of living and has, to
that extent, failed to support the claim that life is considered a focal homonym by Aristotle.
2.3.4.2 The relation between core and noncore cases of living.
On Shields' interpretation, life is a focal homonym and all focal homonyms display
relations of asymmetrical causal dependence between core and noncore cases of being F. I have
just argued that Shields does not successfully establish that life has a core sense. In this sub-
section I argue that independently of this problem, Shields has not established a plausible form
of causal dependency for core and noncore cases of living. This provides additional reason to
deny that Aristotle conceived of life as a focal homonym. If we could find a plausible relation
which had textual support in Aristotle we would be strongly inclined to grant that Aristotle
considered life a focal homonym. In the absence of such a relation and in the absence of textual
support for the claim that life is a case of focal homonymy at all, however, we have strong
45 In his argument that god's way of living constitutes the core of life Shields refers only to Metaphysics xii.7 1072b26-7 and iv.2 1003b6. The first of these passages does not support the focal homonymy of life, however, as I argue in the text. The second, quoted above on p. 65, is a general description of focal homonymy that makes no reference to the particular case of life; it can therefore offer no support at this juncture since Aristotle's general commitment to focal homonymy is nowhere in question.
79
interpretive grounds to doubt that Aristotle considered life a focal homonym. By the same
token, of course, plausible textual bases for another form of definitional relation would
constitute strong evidence in favor of Aristotle's commitment to that form of explanation.
Shields claims that noncore cases of being alive depend causally for their being alive on
formal causal ties to god's way of being alive. The argument is as follows:
If 'life' is a core-dependent homonym, all living things must stand in one of the four causal relations to the form of intentional system god is.46 Surely god is not the material cause of other living things; nor is god in Aristotle's Metaphysics an efficient cause of the life of all living things. Rather, given that all living things are kinds of intentional systems, it follows that they stand in formal causal relations to the core notion. (1999, 190)
In the previous sub-section I argued that the antecedent of this conditional has insufficient
textual support. Granting the antecedent, however, we may note that the conclusion of the
argument does not follow even if the premises are granted: from god's life not being the material
or efficient cause of living it does not follow that it must be the formal cause, for there is a fourth
causal relation � final causation � to consider.
In fact, in a footnote47 Shields considers and rejects the possibility � raised on the basis
of Metaphysics xii.7 1072b3 � that the relation might be one of final causation. A stronger case
that the tie is a final causal one might be raised, however, on the basis of de Anima ii.4 415a25-b7,
where Aristotle identifies participation in the eternal and the divine as a final cause of all living
beings. In fact, I do not believe that this text can be used to support any claim stronger than the
claim that one non-proximate cause of life in mortal living beings is their aim of keeping the
species form in existence,48 and this claim is too weak to support the claim that the causal
connection uniting core and noncore cases of being alive in Aristotle's view is a final causal tie.
46 This is a consequence of Shields' understanding of what it is that makes the definitional ties noncore cases of being F have to the core case of being F 'important' � that the tie is a (four) causal tie. See my recapitulation of Shields' views, above, or Shields (1999), part one. 47 Shields (1999, 190n21). 48 See chapter seven for my argument for this claim.
80
The point which must be emphasized here, however, is not that the textual evidence in
favor of some form of connection appropriate to focal connection other than a formal causal
connection is plausible; I do not believe that any others are plausible. Rather, the point to be
emphasized is that Shields offers no textual support at all for the claim that life's core and
noncore instances are related through asymmetrical relations of formal causal dependency. The
textual thread upon which Shields attempts to hang this thesis is, then, either extremely thin
(because it is so remote, depending as it does wholly his on general account of focal homonymy)
or nonexistent (because it cites no texts specific to the issue of the account of life).
In fact the situation is, I believe, worse than this, for if we pursue the nonproximate
textual case for a formal tie between life's core and noncore cases contained in Shields' general
discussion of homonymy, then we find that general case is extremely thin both textually and
philosophically. First, it is not the case, as Shields claims in his general account, that it is clear
"minimally that Aristotle expects some non-core homonyms to stand in formal causal relations
to core homonyms" (118). As a matter of interpretation, Shields offered only one purported
example from Aristotle to back up this claim (1003b2-3) and he admits that the interpretation of
this passage is difficult, obscure, and disputed.49 It would strain traditional standards for textual
support to conclude on the basis of this one controversial passage that Aristotle accepts focally
connected formal causation. A more sober assessment is that Aristotle's commitment to such a
tie is established, at best, as a hypothesis to be considered tentatively and with great hesitation.
This hesitation ought to be increased, moreover, by considerations of interpretive
charity, for the notion of a formal focal tie is itself unclear. Throughout his treatment of
homonymy, Shields is careful to point out that the claim that any particular disputed term is
homonymous will need to be adjudicated and defended by rigorous argumentation. However,
no such argumentation is provided in support of the only purported examples of focally
49 Indeed, the passage is so difficult that Shields himself offers two distinct and apparently unrelated accounts of what the passage in question indicates (cf. 112n10 & 118).
81
homonymous formal causation that Shields produces. The only two examples (the first of which
he claims to find in Aristotle) Shields produces are:
1. The physician is medical.
1'. The folk healer is medical.
and
2. The Buddhist monk is a teacher.
2'. The chief mechanic is a teacher.
2''. Experience is a teacher.
Neither of these cases appears upon reflection to be a genuine case of homonymy, focal or
otherwise.
Shields' explanation of (1) and (1') is that the folk healer instantiates the form inchoately or
incompletely (118). This is insufficient for establishing homonymy, however, since a balding man
instantiates being hairy incompletely or inchoately but is not therefore homonymously hairy.
The fact that one thing can incompletely instantiate a form which another thing instantiates
completely reveals that the instantiations of some properties shade off into borderline or vague
cases where it is no longer clear to us whether we should say the thing instantiates the property at
all or not. Such cases do not support homonymy claims.
No explanation at all is given of the purported focal connection in (2) through (2''), but
the example is far from being an uncontroversial case of focal homonymy. It seems highly
plausible that the monk and the mechanic are synonymously teachers although they teach
different skills, and that the sense in which experience is a teacher is metaphorical: we learn (gain
insight) from experience, but there is no implication that experience teaches at all; we can after all
learn without teachers. Shields generally shows great awareness of the disputed status of
homonymy claims and their need for rigorous support in his defense of homonymy. When it
comes to these cases, however, cases which turn out to be absolutely crucial to both his four
82
causal account of focal homonymy in Aristotle and to his account of life, Shields offers scant
defense for highly controversial claims.
Shields' claim to have uncovered a formal causal tie legitimate for establishing focal
connection is thus undermined both textually and philosophically. The existence of such
relations has only extremely slight and highly controversial basis in Aristotle's texts: there are no
texts relating to life that support the view, only one highly controversial case that might display
the view, and no texts supporting a general commitment. Further, the relation itself does not
appear to withstand critical philosophical scrutiny. Given Shields' positive claims that this is a
legitimate form of focal connection recognized by Aristotle in the case of life, the burden was on
him to establish both the textual and philosophical grounds for focal formal ties. I have argued
that this defense fails, both in its most important instance (life) and in general. I conclude that in
the absence of convincing evidence in favor of the thesis that Aristotle considered life a focal
homonym, we ought to conclude that we do not understand, thus far, Aristotle's preferred
solution to the dual problems of unity and multivocity.50
2.3.4.3 Section summary.
In the previous two sections I argued that Shields' account of Aristotle on the nature of
life fails to find the correct form for that definition to take. Shields argues that we ought to
accept a particularly controversial form of focal connection, but I argued above, first, that he did
not textually establish that Aristotle considered life a focal homonym, and second, that his
textual and philosophical defense of the form of focal connection he defends (formal focal
connection) fails. Aristotle is committed to the multivocity of life, but how we are to understand
that commitment is a matter of interpretation.
Given these failures we are without a solution to the puzzle generated by Aristotle's
apparent commitment to both the unity and multiplicity of life. Matthews and Shields have both
50 My own solution to these problems is developed in chapter seven of this dissertation.
83
emphasized, in different ways, the substantive point that life, as Aristotle conceives it, is
teleological. At least on one reading Shields deepens this point with reference to the fact that
different types of living things will be defined with reference to distinct ends or goals. Neither
Matthews nor Shields, however, gives us an adequate understanding of the form of definition of
life that Aristotle seeks to give given his commitment to life's being said in many ways, and
Matthews supplies a philosophically inadequate and textually unmotivated reductivist account of
teleology.
2.4 Conclusion: prospectus for an account.
Our survey of contemporary interpretations of Aristotle on the nature of life has proven
profitable. Matthews and Shields both make Aristotle's commitment to the teleology of life
central to their accounts, and Shields offers the further insight that the different ends possessed
by different types of living beings are potential grounds for explaining Aristotle's commitment to
the multivocity of life. Suggestive as each account is, however, neither account succeeds as an
Aristotelian account of life, and their shortcomings and strengths define the tasks for the
chapters to come.
First, I take it that Matthews and Shields are correct to place teleological commitment at
the core of Aristotle's account of life. As Aristotle says,
[F]or any living thing that has reached its normal development and which is unmutilated, and whose mode of generation is not spontaneous, the most natural act is the production of another like itself, an animal producing an animal, a plant a plant, in order that [hina], as far as its nature allows, it may partake in the eternal and divine. That is the goal towards which all things strive, that for the sake of which [ekeinou heneka] they do whatsoever their nature renders possible [hosa prattei kata phusin]. (DA ii.4 415a27-b2)
[A]ll living things (ta zô(i)a)51 both move and are moved for the sake of something, so that this is the limit of all their movement � that for the sake of which. (MA 6 700b15-6)
51 The context of the quote makes it clear that 'living things' rather than 'animals' is the correct translation here.
84
Taking our cue from these passages, we may make the provisional supposition that teleological
orientation is the key to life in Aristotle.
Thus, we may begin with a provisional analysis of life given in the following terms.
Life in Aristotle � Provisional Definition (LAPD): a is alive =df a is a teleological system.52
This basic insight is widely shared,53 and in chapter seven I argue in great textual detail that this
provisional definition is indeed largely correct.
Substantive claims concerning Aristotle's final account of life are contained in the
provisional definition, yet the definition requires a great deal of unpacking, as do any teleological
claims in a modern context. My intention is to approach this provisional definition of life both
as a serious contemporary solution to the problem of life's definition and as a promising
interpretation of Aristotle's texts. The problems besetting Matthews' account of Aristotle on the
nature of life therefore largely determine the course of chapters three through six of the
dissertation. In essence, Matthews offers us an analysis of Aristotle's commitment to teleology
when he suggests that we understand de Anima ii.4 415a27-b2 to mean that something is alive if
and only if it possesses one among a list of powers such that the possession of those powers is,
in general, necessary for the survival of beings of that thing's kind. The inadequacy of this
analysis of teleological commitment to the task was displayed above in my criticisms of
52 Obviously this definition will need unpacking. What is it to be teleological or end directed? What is it for some (set of) thing(s) to constitute a 'system'? The account is provisional as it stands and will be developed throughout the dissertation (see esp. chapter five). Nevertheless, we may profitably consider the account at this stage in company with a modern account of life offered by P. Taylor. He says, "We conceive of the organism as a teleological center of life, striving to preserve itself and realize its good in its own unique way. To say it is a teleological center of life is to say that its internal functioning as well as its external activities are all goal-oriented, having the constant tendency to maintain the organism's existence through time and to enable it successfully to perform those biological operations whereby it reproduces its kind and continually adapts to changing environmental events and conditions.... All organisms, whether conscious or not, are teleological centers of life in the sense that each is a unified, coherently ordered system of goal-oriented activities that has a constant tendency to protect and maintain the organism's existence" (1986, 121-2). 53 Among contemporary writers who associate life and teleology closely, in addition to Matthews and Shields, we may note Randall (1960), Gotthelf (1976/ 1987), Charlton (1992), Irwin (1988, pp. 283-4) and Nussbaum (1978, pp. 76-80). The view has been challenged in recent years and in different ways by
85
Matthews' account,54 but the attempt clearly indicates the deep metaphysical problems
surrounding our understanding of teleological claims. Teleology has been suspect for hundreds
of years, and reasonable and philosophically deep concerns about teleological commitment
abound both in contemporary Aristotelian scholarship and in contemporary philosophical
accounts of teleology.
Given the disputed status of teleology in both scholarly and philosophical contexts my
final teleological account of life must be pushed back to chapter seven. In the intervening four
chapters I first elucidate the nature of and grounds for Aristotle's commitment to teleology in the
living world (chapters two and three), and then defend the contemporary relevance of Aristotle's
conception of teleology in a modern context (chapters five and six). The problems that arose for
Matthews' account of Aristotle's teleological commitment act as a signpost for the deep
metaphysical and methodological problems we must overcome before we may accept the
teleological substance of the provisional definition of life.
Supposing that scholarly and philosophical problems surrounding teleology can be
overcome, however, Shields' subtle account of Aristotle's commitment to the multivocity of life
sets a further hurdle to our understanding of Aristotle's view. I argued above that we do not
have an adequate conception of the form of the definition Aristotle believes ought to be given to
life even though we have (in broad outlines) the substance of the account. One major project of
chapter seven's interpretation of Aristotle on the nature of life involves filling this gap in our
understanding of Aristotle's account. Chapter seven's other main project is to precisify and
textually defend the details of our provisional definition of life in teleological terms.
Cooper (1982), Furley (1985; 1996), Owens (1968), Kahn (1985), and Sedley (1991). These challenges are taken up in chapter three. 54 Reductive accounts of teleology such as this one form the focus of chapter five of this dissertation.
86
Chapter 3
The Extent of Teleology in Aristotle
In chapter one I argued that the question of life's nature is a deep and interesting
philosophical one and that its solution requires sustained philosophical attention to an
interconnected set of problems at the intersection of science and philosophy. I argued, further,
that Aristotle's theses hold surprising promise in this area despite his out-moded biology.
Chapter two partially vindicated both of these claims: Aristotle's approach to the nature of life
has been revealed to be deep and interesting, but also philosophically (and interpretively)
problematic. Aristotle appears to analyze being alive teleologically, but both the grounds for the
acceptance of and the extent of teleology are highly controversial in both contemporary and
scholarly contexts. This chapter therefore begins a four-chapter investigation of the scope,
nature and defensibility of Aristotle's conception of teleology. These chapters form the
philosophical foundation of the final chapter's account of Aristotle on the nature of life.
This chapter focuses on the question of the scope of and motivation for Aristotle's
commitment to teleology. If Aristotle truly analyzes being alive in teleological terms, then life
and teleology (in the relevant sense) must, minimally, be coextensive. We may formulate this
idea as a thesis.
The coextensiveness thesis (CT): According to Aristotle, the following biconditional obtains: x is alive if, and only if, x is a teleological system.
Given that Aristotle clearly endorses the existence of a certain class of nonliving things which
nonetheless possess a form of teleology (i.e., artifacts), we must understand the coextensiveness
thesis to indicate that only things which have intrinsic rather than derived teleological directedness
are alive.1
1 I argue in chapter seven that Aristotle recognizes and explicitly draws this distinction at GA ii.6 742a21-36 and employs the distinction consistently throughout his texts. The distinction is crucial to a proper
87
As is well known, however, long interpretive tradition maintains � not without textual
basis � that Aristotle rejects the coextensiveness of life and teleology. There is little debate that
being alive was, for Aristotle sufficient for the possession of intrinsic teleological directedness;2
the controversial claim expressed in the coextensiveness thesis is instead the claim that only living
things possess teleological directedness. Challenges to the thesis that only living things are
teleologically directed in the relevant sense take one of two main forms which I will call the
regularities challenge and the natures challenge respectively.
According to the regularities challenge, which bases itself largely on Physics ii.8, Aristotle
holds that the teleological is broader than the living in that all things which happen always or for
the most part (hôs epi to polu) happen teleologically. According to the natures challenge, which has
various bases,3 all things with natures (i.e., at least living things and the elements) are
teleologically directed. In this chapter I dispute each of these challenges in turn, and thereby
defend from critical attack the view that in Aristotle's philosophical account of nature all and
only living things are teleologically directed (in the relevant sense).4
After explaining the basis for the regularities challenge to the coextensiveness thesis on
the basis of a reading of Physics ii.8, I argue that those who offer this challenge find themselves
caught in a dilemma. The reading of Physics ii.8 required by the regularities challenge appears to
be in ineliminable conflict with a large number of Aristotle's commitments. If the challenger
revises the commitments of Physics ii.8 in one of many possible (and popular) ways to avoid this
understanding of any adequate teleological conception of being alive, but plays only a minor role in the argument of this chapter. 2 The claim that all living beings are teleologically directed is defended with textual support in chapter seven. 3 See On Generation and Corruption ii.10 and Generation of Animals ii.1 where Aristotle appears to commit himself to the view that all things with natures (i.e., living things and the elements) aim at what is best. Aristotle's frequent refrain that nature always strives for the best also suggests, at least prima facie, a wider reading of teleology than suggested by the coextensiveness thesis. 4 A defense of the coextensiveness thesis can, of course, only provide a necessary first step toward a full defense of the thesis that Aristotle defines life in terms of teleological directedness. I support the stronger claim concerning Aristotle's definitional commitments in chapter seven.
88
problem, however, the interpreter is left without a plausible or charitable understanding of the
argument of Physics ii.8 � the revision modifies the original interpretation in a way that is both
philosophically implausible unsignaled in the text.
Most interpreters see something like this problem, and choose to embrace the second
horn of the dilemma: Aristotle restricts his argument in a way that is both implausible and
unsignaled, and is left with no plausible argument in favor of teleology.5 I argue � by appeal to
the principle of charity and through an alternative reading of the texts � that we ought not
accept this last ditch position. There exists a textually and philosophically plausible reading of
the main argument of Physics ii.8 on which Aristotle elucidates a moderate and plausible defense
of teleology in a unique dialectical context. Once the dialectical context of the argument is made
clear Aristotle's argument in favor of teleology is seen to be on a par with his defenses of the
formal and material causes in Physics i.
Against the natures challenge, I argue that despite Aristotle's frequent rhetorical
flourishes, he is not committed to the view that all natures � i.e., even the nonliving elemental
natures � are teleologically directed. In order to undermine this thesis, I investigate in depth the
arguments of the two main contemporary expositors of the view that all natures are teleologically
directed towards ends: those by Charles Kahn and David Sedley. Their arguments in favor of
the stronger view of the natures challenge are rejected for having inadequate textual basis.
If my arguments succeed, then the challenges to the coextensiveness thesis are
undermined and the thesis that only living things are considered by Aristotle to be intrinsically
teleologically directed reveals itself as a plausible and textually sensitive reading of the corpus.
Given the claim (supported in chapter seven) that all living things are teleologically directed, we
5 Such a reading, by the way, fits in well with the modern presupposition that there simply are no good grounds for positing Aristotelian teleology in a world of efficient causes. The modern suspicion against the acceptability of Aristotelian teleology therefore lends support to a reading on which Aristotle makes a valiant effort at grounding the final cause � an effort which may even have appeared to have some plausibility in bygone days � but that his reasons inevitably fail. I examine and critically undermine the
89
will have established the coextensiveness thesis on the plausible view that we ought not attribute
to Aristotle a view stronger than that required by a charitable, sober and historically plausible
reading of his texts. In the course of this chapter I elucidate a plausible argument � Aristotle's
� in favor of accepting the final cause in nature. With these defenses in hand, we turn in
chapter four to an investigation of the ontological nature of Aristotle's final cause: is it reducible
to any other form of cause, or does it constitute a sui generis causal factor in the structure of the
physical world?
3.1 The regularities challenge.
Commentators who offer a version of the regularities challenge to the coextensiveness
thesis argue largely on the basis of Physics ii.8 198b10-199a8. As is well known, Aristotle
frequently refers to a class of things that happen not of what we might call broadly metaphysical
necessity, but which happen instead either 'always or for the most part' (hôs epi to polu).6 For
simplicity in what follows I will refer to what happens, as Aristotle says, hôs epi to polu as a
'regularity' in nature.7 In Physics ii.8, in the course of defending his teleological account of nature,
Aristotle appears to argue that things either happen always or for the most part or by chance,
apparently implying that in his view these options are mutually exclusive and exhaust the ways
things may happen.8 Aristotle then appears to argue that all things that happen regularly (i.e.,
either always or for the most part) possess teleological directedness. But the class of things
which happen always or for the most part is much broader than the class of living things, and it
would therefore appear that Aristotle rejects the coextensiveness thesis in his main theoretical
passage on the final cause.
traditional modern presumption against the acceptability of final causality in chapters five and six of this dissertation. 6 On hôs epi to polu, see Judson (1991, §iii) and Reeve (1992, §2). 7 In this I follow Curzer (1998, 362). 8 At least in the relevant cases.
90
The text of Aristotle's argument requires a closer look. Aristotle opens Physics ii.8 by
setting its goal: "We must explain then first why nature belongs to the class of causes which act
for the sake of something" (198b10-11), and notes immediately that the supposition that there is
teleology among natural things has not gone unchallenged (198b17-198b33). After noting the
basis of the challenge from those who claim that natural things can be accounted for adequately
without reference to the final cause, Aristotle claims that it is impossible that the deniers of
natural teleology can be expressing the correct view. He says,
It is impossible that this should be the true view. For (a) teeth and all other natural things either invariably or for the most part come about in a given way; but (b) of not one of the results of chance (tuxês) or spontaneity (tou automatou) is this true. (c) We do not ascribe to chance or mere coincidence (sumptômatos) the frequency of rain in winter, but frequent rain in summer we do; nor heat in summer but only if we have it in winter. (d) If then, it is agreed that things are either the result of coincidence (sumptômatos) or for the sake of something, and (e) these cannot be the result of coincidence or spontaneity, (f) it follows that they must be for the sake of something; and that such things are all due to nature even the champions of the theory which is before us would agree. Therefore action for an end is present in things which come to be and are by nature. (Phys. ii.8 198b34-199a8)
In (a) and (b) Aristotle distinguishes between natural occurrences, which come about regularly,
and chance occurrences. Then in (d) and (e) he appears to commit himself to the view that all
regular occurrences involve final causality, and concludes in (f) that there must therefore be
teleology in nature.
More formally, the argument of the passage appears to be as follows (following Curzer
1998, 368-9).
1. Natural things come to be in a regular way.
2. Things that come to be by chance are irregular.
3. Things are either caused by an end or by coincidence.
4. Natural things are not caused by coincidence (chance). [from (1) and (2)]
5. Natural things are caused by an end. [from (3) and (4)].
This argument has been subjected to numerous interpretations, but I believe that the above
formulation captures the core of what we may call the traditional reading of the argument. The
91
key commitment of all traditional readings is expressed in premise (3), which states what we may
call the 'regularities thesis'.
The regularities thesis (RT): Among natural occurrences, all things which happen regularly have final causes.
As I mentioned above the argument poses a challenge to the coextensiveness thesis: it would
appear that teleological directedness is not coextensive with living, since, according to this
reading of the argument of Physics ii.8, there will exist many regularities among nonliving things
and these regularities will all involve teleological directedness.
3.1.1 Against the coextensiveness of regularity and teleology.
This cannot be the end of the matter, however, for (as I mentioned above in the
introduction) there exist numerous texts in Aristotle's corpus where he seems, sensibly, not to
accept an unrestrained regularities thesis. I say 'sensibly' since there exist any number of regular
occurrences which appear to involve no final causality, even on the most liberal views on the
matter. Thus, to pick just one example, two things may have a co-cause and therefore occur
regularly together without either one of them being finally caused by the other or by their co-
cause. Aristotle himself provides at least two examples: in On Divination in Sleep Aristotle notes
that a lunar eclipse and a star's moving into a shadow may have a co-cause (PN 462b29-30), and
he notes in the History of Animals v.22 that large olive harvests and swarms of bees have a co-
cause (heavy rain). There is no suggestion in fact or in Aristotle that in virtue of the regular co-
occurrence of these events one is for the sake of the other.
William Charlton constructs an argument from silence to the conclusion that Aristotle
did not accept that all regularities are teleological. Charlton says,
Aristotle does not argue that everything which is due to nature is due to form and susceptible of teleological explanation.... When... as in the Meteorologica, he deals with sublunary physical phenomena, such as weather, the sea, coction, his explanations are exclusively in terms of necessity, chance, and the natures of different kinds of matter. (1992, xvii)
92
Charlton refers to a large number of texts where Aristotle seems to endorse and offer
explanations through material necessity alone.9
Clearly an argument from silence will not win the day in this context, however: so long
as Aristotle accepts the compatibility of efficient and final causal explanations � which he does
� there is no problem in his remaining silent in certain cases. Nevertheless, Howard Curzer
develops a stronger argument; "Aristotle does not seem to endorse or act upon [the regularities
thesis] elsewhere." However, "Aristotle surely would have employed this type of teleological
explanation more than once10 if he had considered it to be legitimate" (1998, 370). The
argument from silence alone is unpersuasive. Nevertheless, the prevalence of these silences in
theoretical passages where the regularities thesis could be expected to make an appearance
coupled with the lack of direct textual evidence for the regularities thesis outside of Physics ii.8
highlights the crucial importance of Physics ii.8 to the case of those who offer the regularities
challenge. Without Physics ii.8 the reading is wholly implausible and without support.11
Turning to direct textual evidence against the an interpretation on which Aristotle
accepts the regularities thesis, William Charlton (1992, 113) cites Metaphysics viii.4 1044b8-12 as
proof that Aristotle did not seek final causal explanations in all cases of regularities. Aristotle
says,
Nor does matter belong to those things which exist by nature but are not substances; their substratum is the substance. E.g. what is the cause of an eclipse? What is its matter? There is none; the moon is that which suffers eclipse. What is the moving
9 He cites APo. ii 94b37-95a3 with 200a1-5; APo. ii 94b32-3; Meteor. ii 369a24-33, iv. 378b31-4, 380b13-4, 384b15-18, 390b2-14; GA ii. 734b31-735a2, 735b16-21, 26-37, v. 778a29-b1, b10-19, and 789b2-15. 10Curzer (1998, 381n24) cites Cooper's admitted inability to find any examples other than Phys. ii.8 199a1-2 (Cooper 1982, 217-8) and notes that Cooper and Sedley (1991) only find Pol. 1256b10-22 to support teleological relationships holding between species. 11Sedley (1991) replies to this argument by suggesting that the absence of such reasoning is explained by the fact that Aristotle so infrequently discusses theology. It is only from that perspective that things are end directed according to Sedley: "It is only when the world's interactive structure is examined as a whole, from the top down, that the contribution of weather, animals and everything else to the natural hierarchy moves into focus" (195, citing Met. xii 1075a11-25 & Pol. 1253a19 ff.). This response, however, is open to the seemingly decisive objection that it would make the argument of Physics ii.8 an argument in theology rather than physics, as it surely is an argument in physics.
93
cause which extinguishes the light? The earth. The final cause perhaps does not exist (isôs ouk estin). (Met. viii.4 1044b8-12)
Aristotle here clearly considers the possibility that a regular occurrence has no final cause,
suggesting strongly that he does not believe any thesis so strong as the regularities thesis. To this
passage we may add the following.
[T]hough even the residua are occasionally used by nature for some useful purpose, yet we must not in all cases expect to find such a final cause; for granted the existence of this or that constituent, with such and such properties, many results must ensue as necessary consequences of these properties. . . . the bile is not for the sake of anything, but is a purifying excretion. (PA iv.2 677a12-31. Emphasis added. See also Pol. vii.8 1328a24-6)12
There are then two causes, namely, necessity and the final end. For many things are produced, simply as the results of necessity. (PA i.1 642a1-2)
In these passages Aristotle explicitly distinguishes between those results in an organism which
happen for the sake of something and those which happen regularly but merely of necessity.
Aristotle utilizes this distinction in practice again and again when he distinguishes
between those qualities and parts of animals which (although they occur regularly) require final
causal explanation and those for which efficient and material causes are sufficient. Thus,
[S]ince the animal body must undergo natural changes of quality, and when the parts [i.e., the heart and the penis] are so altered some must increase and others decrease, so that the body must straightaway be moved and change with the changes that nature makes dependent upon one another (the causes of the movements are heatings and chillings, both those coming from outside the body, and those taking place naturally within it) � so the irrational movements (hai para ton logon) in the aforesaid parts occur when a change of quality supervenes. (MA 11 703b10-16; slightly modified from the ROT)
12 Gabriela Carone has suggested that while the texts cited are decisive in showing that regularity is not sufficient, in Aristotle's view, for teleology, that he may nevertheless consider it to be a necessary condition for teleological directedness. I agree with this as a theoretical point, but two replies are necessary. First, the argument of the main text need not address the necessity claim, interesting as it may be in its own right. Second, I believe conclusive evidence exists which suggests Aristotle rejects even the necessity thesis, for, as Aristotle says, "The end and the means towards it may come about by chance" (Phys. ii.8 199b19, see Met. xi.8 1065a29). As confirmation of this view, spontaneously generated animals and plants have ends in Aristotle's view, although they are paradigms of beings which do not come about 'always or for the most part' through some cause (see PN 468a18-21; HA iv.7 532a5, a11; PA ii.8 653b38-654a2, 654a32, iv.5 678b11-14, 679b35-6, iv.7 683b10-11).
94
Here again Aristotle seems clearly and explicitly to deny that he himself holds the regularities
view.
To these texts we may add more indirect theoretical evidence against attributing the
regularities thesis to Aristotle. In the discussion of completeness (teleion) in his philosophical
lexicon, Aristotle says that the end (telos) is among the things which are last or ultimate (tôn
eschatôn) (1021b24-5), and goes on to distinguish two ways of being last: being last as death is last
( = the last point in a series) and being the last as that for the sake of which (Met. v.16 1021b28-
30). Aristotle here distinguishes clearly between two sorts of ends: those ends which are merely
last (i.e., death), and those ends which are teleological and normative as the attainment of form is
among animals. This distinction is no mere passing thought in Aristotle; it earns a place in his
philosophical lexicon and is repeated at Physics ii.2 194a30-3.13 The importance of Aristotle's
explicitly drawing this distinction, for our purposes, is that it shows clearly that Aristotle rejects
the regularities thesis in its bald, unmodified form.
As further indirect textual evidence we may point to the following passage. Aristotle
says,
It is a wrong assumption to suppose universally that we have an adequate first principle in virtue of the fact that something always is so or always happens so. (Physics viii.1 252a32-3)
In context Aristotle is warning inquirers not to suppose that we have discovered an essential
property of a thing from evidence which suggests merely that we have found a property that is
coextensive with it.14 This general point constitutes evidence against attributing the regularities
13 This shows the superiority of Guthrie's translation of On the Heavens i.9 279b4-5 over that of Stocks in the ROT. Guthrie has, "for the body whose motion is circular the place where it ends ((h)on teleuta(i)) is also the place where it begins." Stocks translates, "the body whose path is the circle has one and the same place for starting-point and goal ((h)on teleuta(i))." In general, both translators consistently overtranslate Aristotle's statements about the end-points of elemental motion in On the Heavens as if Aristotle considered those motions to have teleological-ends rather than mere-ends. These overtranslations keep alive the incorrect view that Aristotle considered all natures, elemental and living, teleological. 14 Here is Aristotle's definition of a proprium: "A property is something which does not indicate the essence of a thing, but yet belongs to that thing alone, and is predicated convertibly of it. Thus it is a property of man to be capable of learning grammar; for if he is a man, then he is capable of learning
95
thesis to Aristotle since Aristotle holds that things which functions or ends are defined by the
possession of those ends.15 Functions are therefore part of the essence of the things that have
them � they are not mere universal regularities. Given Aristotle's methodological commitments
concerning the care necessary in determining essential properties, therefore, we ought not
attribute the regularities thesis to him.16
These passages may be supplemented with a number of further passages where Aristotle
strongly suggests that certain regular occurrences have no final cause. Aristotle endorses the
existence of useless but necessary (i.e., regular) bodily excretions such as bile,17 the regular but
nonteleological end of death,18 of the spleen,19 the uselessness of a peacock's tail,20 the
uselessness before the seventh month of milk in human mothers21 and the necessity of there
being useless residue in females owing to their cold nature.22 The horns of deer which are said
to be not merely useless but actually detrimental to the deer.23
None of these texts meshes with a reading of Physics ii.8 on which Aristotle has a deep
theoretical commitment to the regularities thesis. What they do show, instead, is that Aristotle's
regular practice and theoretical commitments are in direct conflict with the regularities thesis.
grammar, and if he is capable of learning grammar, he is a man" (Top. i.5 102a18). Thus Aristotle's point in Physics viii.1 252a32-3 is probably to note that while getting hold of a universal regularity gives one excellent reason to believe that one has hold of a proprium of a thing, one does not thereby have additional reason to go on to say that one has an essential property of a thing. 15 See Meteor. iv.12 390a10, EE iii.7 1115b23, OH ii.3. 286a8, Pol. i.2 1253a23, DA ii.4 416b23, OH ii.14 297b21. See also chapter seven for further support of this view. 16 I am indebted to Gabriela Carone and Robert Pasnau (independently) for bringing the weakness of an earlier version of this argument to my attention. 17 See PA iv.2 677a12-31; cf. the definition of a useless residue at GA i. 725a4-6. 18 See Phys. ii.2 194a30-3, OH ii.6 288b15-18, Met. vi.3 1027b8-10, PN 465a14-5. 19 See PA 670a30. 20 See IA 10 710a3 ff. 21 See GA iv. 776a15 ff. 22 See GA ii 738a9-b3; cf. 739a8-13. 23 See PA iii.2 663a9-12, 664a3-8; cf. 663b12-4.
96
3.1.2 Restricting teleological regularities.
In the previous section I argued that we have strong Aristotelian grounds for denying
that Aristotle held that all regularities were teleological. If this is correct, however, our
interpretation of Physics ii.8 requires significant modification, for the passage's straightforward
interpretation appears committed to this thesis.
Perhaps surprisingly, I know of no one who offers the regularities challenge to the
coextensiveness thesis who would deny the broad result of the previous section: it is widely
recognized that Aristotle does not accept an unrestricted version of the thesis that all regularities
are teleological. Somehow, all commentators agree, Aristotle must have had in mind in Physics
ii.8 a narrower thesis on which only some restricted class of things which happen always or for
the most part are teleological. Even David Furley, whose essay "The Rainfall Example in Physics
II.8" is the contemporary locus classicus for the regularities challenge, believes that Aristotle
restricts the thesis somehow. As he says, "[S]ome limitation on the range of this phrase in the
argument is inevitable" (1985, 179-80).24
This fact however, has an unwelcome consequence for those endorsing the regularities
challenge. The problem is this. As Furley claims in an oft-quoted passage, on the traditional
reading of the argument of Physics ii.8 Aristotle must endorse an unrestricted version of the
coextensiveness of regularities and teleology, for the supposition that he does not "makes
nonsense, or something close to that, of [Physics ii.8 198b34-199a8]" (Furley 1985, 179).
Against all such strategies25 Furley argues that we cannot make sense of Physics ii.8
198b34-199a8 without the claim that premise (3) above26 is unrestricted in its scope. Furley
argues,
24 William Charlton (1992, 122-3) and Terrance Irwin (1988, 552n18) offer other restrictions of the thesis. 25 Including his own! Presumably Furley failed to notice this fact. 26 According to premise three of the traditional interpretation of Aristotle's argument Aristotle is committed to the claim that things either happen for an end or by coincidence. See above p. 90.
97
[I]t would be a fatal weakening of Aristotle's argument if he even suggested, at this point, that the disjunction 'either coincidence or for the sake of something' is not exclusive, that there is sometimes a third possibility. It would be necessary at least to explain what class admitted a third possibility and what class did not, and why the relevant cases belonged to the latter; and there is of course no trace of this explanation. (1985, 180-1)27
Furley thus argues on the basis of charity that Aristotle cannot have in mind any restriction on
the stated coextensiveness of regularity and teleology among natural things, for if he did the
argument in its dialectical context would be an immediate failure. Aristotle has just laid out his
opponents' case, after all, that certain regular occurrences in nature (e.g., the growth of teeth) happen
without the aid of final causality. If Aristotle himself accepts that there are exceptions to the
claim that all regularities are teleological (and he gives no indication in the text, remember), then
surely Aristotle has not even begun in this argument to address the position of his opponents.
Aristotle's silence over the supposed (but unstated) exceptions to the coextensiveness of
teleology and regularities would destroy any pretense of the argument's dialectical plausibility.
A sign of the seriousness of this problem is the fact that no interpreter has claimed �
starting from a traditional reading of the passage � to find an argument which we, from a
modern perspective, would find plausible. The strongest claims suggest instead that while we can
see glaring faults in Aristotle's arguments, these flaws would not have been apparent to Aristotle
or his contemporaries due to their different worldview and/or the poverty of the empirical
science available to them.28 This strategy, however, embodies a desperate and unsteady
compromise between a failure to find any plausible reading of the text and charity, which forces
us29 to blame ourselves rather than the texts when we cannot find a plausible reading.
27 Wardy (1993) comments thus on the power of Furley's argument, "As Furley established, there is no gainsaying the conclusion that winter rain falls within the scope of panta ta phusei (198b35). Rain falls (almost) always in the winter, thus neither from luck nor coincidence, so for a natural purpose" (20); further, he is explicit in claiming that Aristotle takes being a regularity to be sufficient for being teleological (21, 22). Against this, see Irwin (1988, 106). 28 Gotthelf (1976/ 1987, pp.225, 229) and Cooper (1982, pp.211-2) employ this strategy. 29 The desire for charitable readings is not without its methodological perils. For a sceptical tour of such perils, see Cohen and Keyt (1992).
98
3.1.3 Summing up problems for the regularities challenge.
This situation ought to motivate us to take a renewed look at the main argument of
Physics ii.8. We are in an extremely awkward interpretive situation, after all. Texts abound which
suggest that Aristotle held the regularities thesis neither in theory nor in practice. Further, on the
supposition that Aristotle's argument in Physics ii.8 relies on the regularities thesis, we are unable
to find a plausible reading of the argument of the text.
We appear caught in a dilemma. Either we deny the seemingly overwhelming textual
evidence that Aristotle rejected the regularities thesis both in selected theoretical passages and
consistently in practice or we accept that he rejected the unrestricted thesis and "make nonsense"
of the argument of the passage as we have chosen to read it. As I have said, the most common
reaction to this dilemma is to embrace the second horn; interpreters who avail themselves of this
route attempt to make finding a bad argument in Aristotle into a victory for their interpretation
� a good interpreter, after all, doesn't push charity beyond the limits of textual and historical
fidelity. The interpreter may seek further comfort in the not wholly plausible thought that while
we can see the colossal mistake Aristotle and his contemporaries were locked in a world-view that
kept them from seeing what we see.
Despite the popularity of embracing the second horn of this dilemma, I suggest that this
unattractive interpretive situation should push us to consider instead the possibility that the
dilemma is a false one: must an interpreter embrace either horn?
3.2 A more plausible reading of Physics ii.8.
David Charles offers the starting point for a more plausible reading of the argument of
Physics ii.8. The core of Charles' reading comes from noting that we need not take the argument
of Physics ii.8 198b34-199a8 to be wholly a reflection of Aristotle's own positions. Instead, portions
of the argument may be ad hominem, directed against the specific theses of Aristotle's
philosophical opponents. If this were true, the argument of Physics ii.8 might be reconceived to
constitute a plausible ad hominem argument against the eliminitivist views of select opponents
99
without constituting a general defense of teleology against eliminitivists of all stripes and without
committing Aristotle to the implausible premise (3).30
Charles argues that the troubling premise (3) of the formalized version of the argument
expresses not Aristotle's commitment to the regularities thesis but his opponents' views. One
argument in favor of adopting such a view of the texts is that it allows us to accept at face value
the variety of evidence marshaled above against attributing a regularities thesis to Aristotle.
Another, more indirect argument develops a more plausible interpretation of Physics ii.8 on the
basis of this reading than is possible on the basis of the traditional reading in which the argument
is beset with problems. I will argue below that a plausible reading of the main argument of
Physics ii.8 can be mounted from Charles' hypothesis that the regularities thesis is one endorsed
by Aristotle's opponents rather than by him. Still, more direct evidence in favor of the reading is
available.
Charles argues first that Aristotle frequently confronts materialist specific opponents
who offer explanations of the formation of living things through chance happenings.31 This fact
makes plausible the claim that the opponents Aristotle has in mind in Physics ii.8 themselves hold
the view expressed by the regularities thesis that things happen either by chance (as they hold) or
teleologically (as Aristotle holds). He argues next that his interpretation of the Physics passage is
preferable to the alternative traditional reading on three grounds: it gives full weight to
Empedocles' theory as stated in the Physics and throughout the corpus;32 it avoids attributing a
baldly bad argument to Aristotle; and it "makes better sense of Aristotle's subsequent criticism of
his predecessors' theories for being incomplete in certain fundamental respects" (23).
30 Again, premise three reads: "Things are either caused by an end or by coincidence." Aristotle's words are, "If then, it is agreed that things are either the result of coincidence or for the sake of something...". See above, p. 90. 31 Charles cites Met. 984b11-4; Phys. 198b34-6, 199b17-8; PA 640a22-4, 641a1-5, 10-15, 15-8, 20-24, 26-7, 641b5-8, 27-30, 34-6; GA 743a21-4, 764a3-7, 12-19, 765b1-7; and GC 314b1-3. 32 See the texts cited in n. 31 above.
100
The prima facie case for taking Charles' suggestion seriously may be summed up, then, by
noting that (i) we have a great deal of direct and indirect evidence against attributing a regularities
thesis to Aristotle from outside of Physics ii.8; (ii) reading Physics ii.8 as required by the regularities
challenge leaves us with no charitable understanding of the argument of the passage in question;
and (iii) there exists a plausible alternative which has abundant textual grounds. The ultimate
proof of this interpretation's merits, of course, must lie in spelling out with care and precision the
exact nature of Aristotle's argument on this reading. It is to this task that I now turn.
In what follows I develop the core insight of Charles' account of Physics ii.8 into a full
reading of the relevant passage. If the passage is an ad hominem against particular opponents, how
does the argument go and what conclusion may we take from it? I begin my reading by focusing
on a discrepancy in Physics ii.8 which is infrequently emphasized. When Aristotle presents the
case for his opponents using an example favorable to them at the beginning of Physics ii.8
Aristotle gives a materialistic explanation of rain's falling. When Aristotle argues against the
position and in favor of teleology, however, Aristotle mentions the frequency of winter rain as
something that is not explained by his opponents. In what follows I make this shift in examples
an in-road into Aristotle's argument in the passage; an understanding of this shift plus Charles'
insight will, I argue, result in a rich and plausible interpretation of the main argument of Physics
ii.8.
3.2.1 The roles of rain and frequent winter rain.
In Physics ii.8 198b16-21 Aristotle offers the following materialistic account of rain's
falling as a prelude to his opponents' account of animal and plant generation without teleology.
He says,
A difficulty presents itself: why should not nature work, not for the sake of something, nor because it is better so, but just as the sky rains, not in order to make the corn grow, but of necessity? (What is drawn up must cool, and what has been cooled must become water and descend, the result of this being that the corn grows.) (198b16-21)
101
This materialistic (i.e., non-teleological) explanation of rain's falling is endorsed by Aristotle in a
number of other texts,33 and we ought to accept the view as Aristotelian unless we have
compelling reason to reject it.
When Aristotle turns to criticizing his opponents' views he raises a similar but
importantly different example to show why their view is 'impossible'. To reveal the impossibility
of their view Aristotle notes that,
We do not ascribe to chance or mere coincidence the frequency of rain in winter, but frequent rain in summer we do; nor heat in summer but only if we have it in winter. (Physics ii.8 199a1-2)
This change in examples ought, of course, to puzzle us. What motivates the switch? What is the
importance of the change to his argument?
In my view Aristotle's use of the example of frequent winter rain/summer heat is meant
to illustrate a general warning against the overly hasty materialist conclusion that rainfall's making
crops grow is coincidental on the basis of the materialist explanation of rain given in Physics ii.8
198b16-21.34 In my view, Aristotle uses the fact that rain is frequent in the winter to make the
point that there is a larger pattern of explanation in which the materialist's given explanation
finds its place. As Aristotle says elsewhere,
[W]hen the sun in its circular course approaches, it draws up by its heat the moist evaporation: when it recedes the cold makes the vapour that had been raised condense back into water which falls and is distributed over the earth. (This explains why there is more rain in winter and more by night than by day: though the fact is not recognized because rain by night is more apt to escape observation than by day.) (Meteor. ii.4 359b34-360a5)
The point of the example of frequent winter rain is, as the above quoted passage makes clear,
that the materialistic explanation of rain's falling as applied to the growth of crops at 198b16-21
is incomplete even by their own standards; a complete explanation will need to account for a
33 See Meteor. i.11 347b12 ff.; PN 457b31 ff.; PA ii.7 653a2 ff.; and APo ii.12 96a12 ff. 34 Again, Aristotle finds this explanation to be perfectly apt as far as it goes. See again the citations above in n. 33.
102
larger pattern of natural facts. In the case of rain that pattern of explanation will involve the
sun's role in determining the seasonal weather patterns.35
If this is correct, then Aristotle's point in using the example of frequent winter rain in
Physics ii.8 198b34-199a8 is that it is no mere coincidence, even in the materialist's terms, that rain's
falling (frequently in winter) causes crops to grow. What I mean can be illustrated with an
example from Aristotle's theory of chance (Phys. ii.4-6). In his discussion of chance Aristotle
employs as an example of a chance event a man's going to market for a purpose other than
collecting a debt but running into the debtor and 'killing two birds with one stone'. Collecting on
the debt in such a situation (call it case 1) is a paradigmatic chance event (Phys. ii.5 196b33-6).
Aristotle points out, however, that we would not say that the man collected the debt by chance in
the marketplace if he always or usually went to the market for the purpose of collecting debts
(Phys. ii.5 197a3-5; call this case 2). The frequent trips to the market for the purpose of collecting
debts do not lead simply by chance to the regaining of debts � so long as debtors do not catch
on, that is.36
The contrast between chance (case 1) and non-chance (case 2) debt collections
illustrated in Physics ii.5, then, offers a very good reason for the difference between the
explanation of rain's falling at 198b16-21 and the example of frequent winter rain at 199a2-3.
Aristotle's point in Physics ii.8 is that it isn't merely by chance that the frequent winter rain makes
crops grow since, first, the movements of the sun explain the regularity of rain during one
particular season, and second the nature of the plants (not the nature of rainfall) is end-directed.37
Plants have the end of growing to maturity and reproducing, and this end is served by their
synchronizing their growing season (something that is in their power) with the timing of the
35 In addition to Meteor. ii.4 359b34-360a5, quoted above, see: Meteor. i.9 346b20-4, GC ii.10, GA v.10 777b24-778a2, and Met. xii.6 1072a11-18. 36 This reveals an irrelevant disanalogy between the cases of the debtors and the cases of rain vs. frequent winter rain.
103
rainy season (a regular event determined by the motions of the sun which is well beyond the
range of their control).38
On this reading the crops are like the man who makes a habit of going to market to
collect debts (i.e., case 2); they do what they do (going to market, growing when they do) for the
sake of the results that ensue (the collecting of debts, growing to reproductive maturity).
Aristotle explicitly makes this very point in his biological works.
The reason why the hair does not grow again in cases of baldness, although both hibernating animals recover their feathers or hair and trees that have shed their leaves grow leaves again, is this. The seasons of the year are the turning-points of their lives so that when these seasons change they change with them by growing and losing feathers, hairs, or leaves respectively. (GA v.3 784a12-16, emphasis added)
The habits of animals are all connected with either breeding and the rearing of young, or with the procuring a due supply of food; and these habits are modified so as to suit cold and heat and the variations of the seasons. (HA iii.12 596b21-3, emphasis added)
As these passages suggest, Aristotle's point in Physics ii.8 is not that frequent winter rain has any
intrinsic teleological tendency to make plants grow any more than parallel reasoning would show
that the debtor who went unsuspectingly to the market in case 2 went to market in order to pay
the debt. If Aristotle conceives frequent winter rain to have a purpose at all it is in a derived
sense in virtue of his commitment to the intrinsic ends of animals and plants.39 The plasticity of
the behavior and growth of plants and animals indicated in the quotes above is a sign of their
teleological directedness, not of the directedness of rain.40
37 See DA ii.4 415a25-b7; Pol. i.2 1252a27-9; HA viii.1 588b24-5; GA i.4 717a21, i.8 718b9-10, i.23 731a25-731b7. 38 See HA iii.12 596b21-3, quoted below. 39 While the point is tangential here, we might conceive of the teleological directedness of rain towards making plants grow along the lines of artifactual teleology. The idea would be that even plants and animals may give the things they use in the course of the activities directed towards their own intrinsic ends the derived end of serving their needs. The model here would be a human's making a rock found on a hike into a doorstop (something with the purpose of stopping doors) by using it for that purpose; the rock did not (and does not after being put to use) have the purpose intrinsically but it does have it derivatively. Rain may likewise be derivatively for the crops, given that they mold their growth such that they can make use of the winter rains. 40 See again the citations in n. 37 above.
104
To sum up: I have argued that there is a serious point to Aristotle's use of the two
examples of rain and frequent winter rain. The example of rain illustrates a materialistic (non-
teleological) explanation for rain which Aristotle accepts. The invocation of frequent winter rain,
however, points to Aristotle's contention that his opponents mis-use the materialistically
explicable case of rain in their criticism of natural teleology. The key to seeing the mistake he has
in mind lies in recognizing that even on the best scenario for his opponents it is no mere
coincidence that crops grow when they do (during the rainy winter). In supposing that they can
transfer the case of rain over to the cases of natural teleology in question (crops' benefiting from
the rain) they miss a larger pattern of occurrences (i.e., seasonal changes) which require
explanation above and beyond that given by their simple materialistic explanation. The existence
of this larger pattern undermines the explanation of the growth in terms of chance alone just as
the creditor's plan in case two undermined the idea that the debt was paid by chance.
3.2.2 The growth of teeth.
To test this interpretation of Aristotle's diagnosis of the weakness of his opponents'
position we may turn to the other example given in Physics ii.8, the growth of teeth. The case of
rain was meant as a paradigm of materialistic (non-teleological) explanation; the case of frequent
winter rain was meant (on my interpretation) to show that citing the existence of this level of
material explanation is insufficient to justify the conclusion that the result was merely by chance.
Where there exist larger patterns of events within which the case at hand fits, we must seek
further causes for the patterns, and these further causes may well contain plausible roles for a
teleological explanation of the original event consistent with the truth of the material level
explanation. We now turn to the other explicitly mentioned example of an event that is (at least)
apparently teleological but which is attributed to chance occurrences by his opponents: teeth
coming in for biting and chewing (Physics ii.8 198b24-9).
I have claimed thus far that Aristotle's point in raising the example of frequent winter
rain is to illustrate the need to look to a wider context in offering explanations than the
105
materialists are accustomed to doing. As with rain, Aristotle appears to accept the bulk of a
materialist explanation for the growth of teeth (see esp. GA v.8 788b33-789a2 and 789a11-b2),
but believes that this does not constitute the whole story. What, however, is the larger pattern
that Aristotle claims the materialist is missing in the case at hand? Here is Aristotle's account of
his opponent's position.
Why then should it not 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 coincident result; and so with all other parts in which we suppose that 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 spontaneously in a fitting way; whereas those which grew otherwise perished and continue to perish, as Empedocles says his 'man-faced oxprogeny' did. (Physics ii.8 198b26-32)
In this passage Empedocles is reported as offering an explanation � through chance
occurrences � for the origin of species fit to live. This view provides an explanation for the
existence of species with parts which have no real teleology despite the appearances.
If I am right, however, there must exist a larger pattern of facts which goes unexplained
on Empedocles' view, a pattern which, once noted, makes the invocation of teleology plausible
or even obligatory on the opposition's own principles. As a starting point toward investigating
this hypothesis, we may note that Aristotle claims that the explanation rehearsed above contains
an 'impossible' view (Phys. ii.8 198b34); what about the view is impossible?
The key to answering this question lies in recognizing what is left out of Empedocles'
explanation of the existence of fit species. There is a larger pattern of occurrences that
Empedocles has (thus far) failed to explain on the account given in Physics ii.8 198b26-32.
Howard Curzer notes that Empedocles has some work to do if he is truly to have an explanation
for the prevalence of fit animals and plants in today's world. Empedocles must explain why it is
not merely that fit species occurred, but in addition why it is that they are preserved down the
106
generations;41 this wider pattern of events is not yet explained.42 The crucial point, from
Aristotle's perspective, is that Empedocles needs an account of why it is that among living things
Fs normally beget Fs43 (a fact I shall refer to as the fact that species normally 'breed true'). There
is a larger pattern in which growing teeth finds a home, the pattern generated by the fact that
living things tend to breed true. This fact will undermine Empedocles' claim that teeth grow as
they do by chance (just as in the cases of rain and the debtor), and from this basis Aristotle may
mount a plausible case that teleology has a role to play in the growth of teeth.
3.2.3 The inference from breeding true to nonreductive nature.
If I have been right thus far, Aristotle believes that the eliminitivist account of purported
cases of natural teleology such as the growth of teeth is adequate as far as it goes. That account
� parallel to the materialist's account of rain � could account for the one-off existence of fit
species. As Aristotle says, "The end and the means towards it may come about by chance" (Phys.
41 Irwin (1988) misses this point. He says, "If surviving animals uniformly or frequently have beneficial teeth, it does not follow that the benefit of the teeth has anything to do with their efficient causation. Empedocles accounts for the uniformity by saying that some animals were coincidentally born with beneficial teeth, and the animals without them died; this account avoids all final causation" (106). Irwin takes Empedocles to be making a gross mistake about simple matters of fact in claiming "that in every single generation it is a matter of chance that" animals and plants with the beneficial features arise (107). It is highly unlikely, however, that Empedocles would be so daft as to think that in every generation "many animals are born with beneficial teeth, and many are born with useless teeth." (107. Physics ii.8 199b14 ff. is charitably taken to be Aristotle's expression of an unforeseen consequence of Empedocles' view rather than as a statement of Empedocles' considered position.) Rather, Aristotle makes clear that Empedocles has assumed without reason or argument that breeding true can be explained without reference to final causation (see the main text, above). Irwin's account of Aristotle here not only attributes a highly implausible view (uncharitably) to Empedocles, it wrongly attributes something like Larry Wright's (1976) conception of teleology to Aristotle (107). On this point see chapter four below; on Wright's conception of teleology see chapter five. 42 Curzer believes that this addition to Empedocles' theory will be trivial and unimportant to the overall structure of the argument. He says, "But a minor extrapolation of [Physics ii.8 198b26-32] enables the materialist to explain the fact that well-adapted organisms fall into species. The explanation is simply this. Organisms of a type which not only flourish, but also breed true tend to multiply, while the others die out" (1998, 366. Emphasis added). I could not disagree more strongly that this is addition constitutes a 'minor extrapolation'. See the main text for my reasoning. 43 See PA i.1 640a25-6, 640b1-4, ii.1 646a35; GA iv.3 767b35; Phys. ii.1 193b8, ii.2 194b13, ii.7 198a25; Met. vii.7 1032a22, xii.3 1070a29, xii.5 1071a20, 1071a25, xiii.10 1087a21.
107
ii.8 199b19), and "as a thing may exist, so it may be a cause, either by its own nature or by
accident" (Met. xi.8 1065a29).44
Aristotle believes, however, that there is a wider pattern of occurrences which
undermines the all-things-considered plausibility of this explanation. That pattern of
occurrences involves the fact that not only have fit species occurred but that they continue to
occur regularly � species breed true. What Aristotle denies is that where sequences are regular45
they happen by chance (Phys. ii.8 199b25-6 with 199b14-18). Aristotle's worry is not with
creatures' having arisen by chance;46 it is rather to their propagating and regularly breeding true
by chance that he objects (see GA i.1 715a23-b15). Thus, Aristotle says "coincidences do not
occur according to any universal or general law" (PN 463b11, 463a2-3). Adding a further point
relevant to our case, Aristotle says,
[I]f it is through chance that one is fortunate, the cause of his fortune is not the sort of cause that produces always or usually the same result�further, if a person succeeds or fails because he is a certain sort of man, just as a man sees badly because he is blue-eyed, then it follows that not fortune but nature is the cause. (EE vii.14 1247a33-6; emphasis added.)
Applied to the case at hand, the point of these passages is that it is not by chance that living things
breed true. Instead, the fact that species breed true can only be explained by postulating animal
and plant natures. Nature, not chance, is the cause of breeding true (EE vii.14 1247a32, 35-6;
see also Rhet. i.10 1369b1-2 and Phys. viii.1 252a10-11).
Indeed, this point is the basis for one of Aristotle's arguments in Physics ii.1 that form is
nature rather (mallon) than matter. Aristotle says,
44 At GA ii.11 762b29 ff. Aristotle considers the possibility of men and quadrupeds being 'earth-born' as some say. See also Aristotle's numerous discussions of spontaneous generation, for example GA ii.11 762a9 ff. See further n.46, below. 45As they are in the breeding true of animals, at least for the most part, see 199a33-b6. 46 Aristotle does accept the reality of spontaneous generation, and he also accepts that spontaneously generated organisms have parts with functions. On spontaneous generation, see HA v.19 550b34-551a10; also GA iii.11, Met. vii.7 1032a30-2. On the functional parts of spontaneously generated things, see HA iv.7 532a5, a11, PA ii.8 653b38-654a2, iv.5 678b11-14, 679b35-6, iv.7 683b10-11.
108
The form indeed is nature rather than the matter; for a thing is more properly said to be what it is when it exists in actuality than when it exists potentially. 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. (Physics ii.1 193b7-12; emphasis added)
In this central theoretical passage Aristotle notes that the fact that living things have activities
proper to them (their reproductive activities, their breeding true) gives us reason for postulating
unreduced natures for animals and plants. These reasons are claimed by Aristotle to be
dialectically acceptable to his opponents, for they are the same reasons that lead them to
recognize elemental natures.
Aristotle makes this point again and again in similar contexts throughout his works. At
Parts of Animals i.1 Aristotle offers the following reason for thinking that natures are teleological:
he says, "For (gar) a given seed does not give rise to any chance living being; but each springs
from a definite parent" (641b24-8). Earlier in the same chapter Aristotle points out that
"Empedocles, then, was in error when he said that many of the characters presented by animals
were merely the results of incidental occurrences during their development" for, first,
he overlooked the fact that propagation implies a creative seed endowed with certain powers. Secondly, he neglected another fact, namely, that the parent animal pre-exists, not only in account, but actually in time. For man is generated from man. (PA i.1 640a19-25)
Again, at Physics ii.4 Aristotle considers thinkers who claim that the heavens come to be
spontaneously but who also deny that chance is the cause of "the existence or generation of
animals and plants, nature or mind" (196a30). The reason he offers on their behalf for denying
that these latter things come about by chance is the now familiar one: "for it is not any chance
thing that comes from a given seed but an olive from one kind and a man from another"
(196a31-2). Again, we have the following extended discussion at On Generation and Corruption ii.6
which makes the same point. Aristotle says,
[I]t is far more difficult for him [Empedocles] to account for the coming-to-be which occurs in nature. For the things which come-to-be by natural process all do so either always or for the most part in a given way; while any exceptions � any results which
109
occur neither always nor for the most part � are products of chance and spontaneity. Then what is the cause determining that man comes-to-be from man, that wheat (instead of an olive) comes-to-be from wheat, either always or for the most part? Are we to say that bone comes-to-be if the elements be put together in such-and-such a manner? For, according to his own statements, nothing comes-to-be from their coming together as chance has it, but only from their coming together in a certain proportion. What, then, is the cause of this? Presumably not Fire or Earth. But neither is it Love and Strife; for the former is a cause of association only, and the latter only of dissociation. No: the cause in question is the substance [ousia] of each thing � not merely (to quote his words) 'a combining and a divorce of what has been combined'. And chance [tuchê], not proportion [logos], 'is the name given to these occurrences; for things can be combined as chance has it [ôs etuchen]. (GC ii.6 333b4-16; emphasis added.)
Aristotle goes on immediately to complain that for the reasons given we cannot look to
Empedocles to tell us anything about nature (phusis); Empedocles says nothing about it (333b16-
7).
The principle at work in all of these cases (including Physics ii.8) is the same. Aristotle
maintains reasonably that regularities do not occur by chance (Physics ii.8 199b25-6 with 199b14-
8). Where regularities occur some nature or natures are involved as the explanation (EE vii.14
1247a33-4; PN 463b11; PA i.1 641b24-8, 640a19-25; Phys. ii.4 196a30 and GC ii.6 333b4-17).
However, a datum of natural science that must be accommodated by any account of nature is
that living things generate after their own kind (PA i.1 640a25-6, ii.1 646a35; Phys. ii.1 193b8-9,
ii.2 194b14, ii.7 198a25-7; Met. vii.7 1032a22-5, xii.3 1070a29 and xii.5 1071a24-9, etc.).47 Nature
(phusis), not chance, is the explanation of this datum (EE vii.14 1247a32, 35-6; see also Rhet. i.10
1369b1-2 and Phys. viii.1 252a10-11). It is the regularity of growth and development within kinds
over successive generations which needs explaining on Empedocles' account. Aristotle
concludes from this train of thought that nature, not chance is the explanation of the regularities
in question in Physics ii.8. This reasoning is in keeping with his argumentative practice
throughout the corpus.
47 This datum is explained in Aristotle's own natural philosophy by subsuming it under the more general rule that the efficient cause of motion must be the same in form as the thing produced: "The mover on the other hand is already in actuality: e.g. it is that which is hot that produces heat, and in general that which produces the form possesses it" (Phys. viii.5 257b8-10; cf. Phys. i.5 188a32-4; Met. vi.3 1027a31-2, vii.9 1034a22, xii.3 1070a4-5; PA ii.1 646a31-5; GA ii.1 734a30-2 and 734b20-1)
110
3.2.4 The inference from natures to teleology.
On the account I have given Aristotle criticizes his materialist opponents for leaving
unexplained a crucial fact � the fact that species breed true. This fact points to the existence of
unreduced organism natures or forms. But, it may be asked, even if the materialist must concede
this much, why must the materialist accept the further claim that the postulated animal and plant
natures are intrinsically teleological? Given that living things have natures which account for
their breeding true, why must these natures have final causes as well as their material and
efficient causes and attendant explanations?
At this point I believe we reach rock bottom in Aristotle's defense of teleology.48
Aristotle offers us no reason why we must accept that living natures are teleological, he simply
accepts (rather plausibly) that their apparent teleological nature is a datum of our common sense
and scientific experience which ought, if possible, to be reflected in our theories of the world.
The economy of the account offered by his materialist opponents turned out, Aristotle believes,
to have been a false economy for they have not reductively explained the fact that animals and
plants breed true. This fact underlies the postulation (on principles accepted by his materialist
opponents) of natures above and beyond the four material elements, and once such natures are
postulated there are no longer compelling reasons not to accept those unreduced natures as they
appear to us in common sense and science (i.e., to accept them as teleological).
This structure fits in well with the rest of the arguments in Physics ii.8 as I understand
that chapter. Aristotle first (198b33-199a8) argues that the Empedoclean account misses a larger
fact which needs explaining � the fact that species breed true (199b14-26). Once we postulate
living natures to account for this fact Aristotle assumes that we have accepted the appearance
48 In fact I believe that this and the previous point are mutually supportive in Aristotle's thinking. Part of the reason for postulating unreduced animal and plant natures in the first place is the prima facie plausibility of such entities' being teleologically directed coupled with the fact that this fact is unaccounted for by other causes (see Met. i.7 988b6-15). The possibility of unreduced natures also provides support for the existence of emergent properties such as teleology. There is no circularity here � at least no vicious circularity. We are simply at the rock bottom level of explanation.
111
given to us in common sense and science that animals and plants and their parts have functions,
purposes, goals and ends. This assumption can be bolstered by considerations concerning
similarities between the less controversial case of mentalistic teleology and the more
controversial case of non-mentalistic/natural teleology (199a9-19, a34-b6). Further, the analogy
extends well into the non-mentalistic range to include spiders and plants (199a20-33) and,
Aristotle argues, there is no reason to deny natural teleology arising from the fact that these
things do not have minds (199b27-30).
Not only does the opponents' explanation fail but Aristotle's teleological explanation
receives as much support as we should expect; unless there are objections to the theory,49 we
should accept the phenomena as they present themselves to us (OH ii.1 283b33-284a1, NE vii.3
1146b6-7).50 Note that Aristotle's opponents have not raised any objections to postulating
teleology in nature, they have merely claimed that they can account for the facts without it.
Aristotle has shown that the alternative explanation they propose does not explain the facts, and
has, further, reasserted that when all of the facts to be explained are considered we have strong
reason to postulate unreduced teleological animal and plant natures. If his opponents wish us to
revise this common sense and scientific commitment, then they have further work to do.51
3.2.5 The dialectic of the argument: summary.
In the previous sections I have argued that Aristotle's main argument in Physics ii.8 is an
ad hominem argument against opponents who offer a particular eliminitivist account of the being
and coming to be of creatures that appear, prima facie, to possess teleological directedness. On
49 Of course there are modern objections to the postulation of Aristotelian teleology of which Aristotle was unaware. Those objections form the basis of chapter six of this dissertation. 50 The type of reason offered here for accepting final causes in nature is roughly the same as the reason expressed in the following passage: "teleological language is also used throughout biology, including physiology and ethology, and shows no sign of abatement, even as physiology merges with biochemistry. It is this continued use which has attracted attention, especially the use of teleological language in describing parts and processes of organisms and the behavior of lower animals and all plants, where minds or adequate minds cannot be safely assumed" (Nissen 1997, vii-viii). 51 See Top. viii.2 157a34-5, 157b31-3, viii.8 160a38-b13, viii.11 161a37-b1.
112
the opponents' account, apparently teleological structures and beings come to be as the chance
result of the interactions of lower level material constituents.
Aristotle's view, as elaborated here, is that this explanation is incomplete. While the
materialists may have an explanation for the one-off existence of particular living things with
only apparently functional parts, they have not explained a further datum � the fact that living
beings breed true. Accounting for this fact, Aristotle believes, removes the appeal of
materialistic reductionism in that it motivates � on dialectically accepted grounds � the
postulation of unreduced animal and plant natures. But, once we have postulated unreduced
animal and plant natures, we have every reason to accept those natures as they appear to us �
i.e., to accept them in their teleological glory. When all the facts come into view, the opponents'
claim to have accounted for the phenomena nonteleologically is revealed as implausible.
More formally, Aristotle's opponents and he both distinguish between occurrences that
happen regularly and those that happen by chance. His opponents have maintained that among
natural things (i.e., biological living things), those things that appear to be teleological really
happen simply by chance. But Aristotle points out in response that on their own view things
happen either by chance or regularly, and if regularly, then teleologically.
This last step, of course, is the controversial one. Why should the materialist opponents
accept the conditional that if things happen regularly then they happen teleologically? The
answer, I believe, requires reference to Aristotle's argument from parity for the existence of
animal and plant natures (see again Physics ii.1 193b7-12, quoted above). If animals and plants
breed true, then we have the same reason to accept the existence of unreduced animal and plant
natures that the materialists have for positing elemental natures � we observe things with
sources of motion which are characteristically their own. But if we postulate animal and plant
natures, then we ought (Aristotle believes) postulate them as they appear to us. Thus, if it turns
out that animals and plants and their parts do not occur by chance (as shown by the fact of
113
breeding true), then on the materialists' own principles we ought to accept that they occur due to
final causation.
3.2.6 What is Aristotle's argument for natural teleology?
On the traditional view of Physics ii.8 the purpose of the chapter is to provide a general
anti-eliminitivist argument for the acceptance of natural teleology. On the present reading,
Aristotle offers no general anti-eliminitivist argument, but instead addresses only a specific sort
of opponent, one who rejects teleology in favor of a particular brand of reductivist eliminitivism.
If this is correct, we might nevertheless wonder whether Aristotle offers any general reasons for
accepting natural teleology; that is, whether he offers us any reasons to accept natural teleology.
I believe that the answer to this question is that he does, although the goals of the
argument Aristotle offers in my view are likely to appear quite anemic when compared to the
ambitious (albeit wholly unconvincing) argument postulated in the traditional reading. On my
view, Aristotle's general considerations in favor of the acceptance of teleology stem from his
commitment to accepting the facts as they present themselves to us in critical common sense and
science.52 As Aristotle says,
[I]t is the business of experience to give the principles which belong to each subject. I mean for example that astronomical experience supplies the principles of astronomical science; for once the phenomena were adequately apprehended, the demonstrations of astronomy were discovered. Similarly with any other art or science. Consequently, if the attributes of the thing are apprehended, our business will then be to exhibit readily the demonstrations. For if none of the true attributes of things had been omitted in the
52 Joseph Owens emphasizes the aspect common sense's natural commitment to the final cause: "The form is the knowable content, the principle of intelligibility in the individual thing. It seems regarded without further ado by Aristotle as functioning in an intelligible manner, and therefore after the fashion of mind. . . . But no explicit account is given in the Aristotelian text. Probably if he were pressed for an answer, Aristotle would have replied as on another occasion [citing APo. ii.19 100a13-14] that such is the way natural things are, and that they have to be accepted as they manifest themselves to human cognition" (1968, 163). David Balme emphasizes that systematic scientific study of nature does not refute but confirms this commitment of common sense: "Aristotle's argument for teleology has two bases: (i) we see ends in nature; (ii) we see that nature is regular, not random. (i) We do not see only interacting elements, but composed wholes having their own properties. We see that animals grow from seed to adult; this is a movement towards a goal.... (ii) We see that animals breed true, which shows (a) that their kind is predetermined in the parent's seed, (b) that their growth is organized in a determinate direction; in neither case do we see random results" (1992, 93, commentary on 641b10-642a1).
114
survey, we should be able to discover the proof and demonstrate everything which admitted of proof, and to make that clear, whose nature does not admit of proof. (APr. i.30 46a17-26; see also GC i.2 316a5-13; DA i.1 402b21-5; OH iii.7 306a16-18; GA iii.10 760b29-32 and APo. i.18 81a38-b9)
Teleology in nature is among the phenomena of the lay person, but the point to emphasize, as
Aristotle does, is that teleology figures in the best critically developed systematic conceptions of
nature as well. Even from a regimented scientific point of view teleology is among the facts to
be explained.
At the beginning of the inquiry we must postulate the principles we are accustomed constantly to use for our scientific investigation of nature, that is we must take for granted principles of this universal character which appear in all nature's work. Of these one is that nature creates nothing without a purpose, but always the best possible in each kind of living creature by reference to its essential constitution. Accordingly if one way is better than another that is the way of nature. (IA 704b11-16; see also Phys. viii.7 260b23 and GA v.8 788b20-1)
This belief that the facts of teleology are among the facts a natural scientist must account for and
explain has held up well over the centuries, and it was not naïve for Aristotle to accept teleology
on these grounds.53
This type of argument has its limitations, of course. Against one who is willing to press
eliminitivism about teleology at the expense of the facts, Aristotle has very little to say.
About all these matters we must try to get conviction by arguments, using the phenomena as evidence and illustration. It would be best that all men should clearly concur with what we are going to say, but if that is unattainable, then that all should in some way at least concur. (EE i.6 1216b26-8)
That nature exists, it would be absurd to try to prove; for it is obvious that there are many things of this kind, and to prove what is obvious by what is not is the mark of a
53 In the introduction to a recent anthology on biological teleology, David Buller writes, "In spite of the difficulties associated with teleology, however, biologists continued to use the teleological concept of function in describing the characteristics of organisms, finding the organization of organisms and the operation of their parts virtually incomprehensible in strictly non-teleological terms" (1999b, 6). In a recent book-length treatment on the problem of teleology, Lowell Nissen writes, "Teleological language is controversial because of some apprehension that it might presuppose either reverse causation or minds.... However, teleological language is also used throughout biology.... It is this continued use which has attracted attention" (1997, vii). Karen Neander, one of the most prominent of contemporary writers on biological teleology writes, "[T]he apparent explanatory power of teleological explanations which appeal to biological functions is quite robust. That the koala's pouch has the function of protecting its young does seem to explain why koalas have pouches" (1991b, 127).
115
man who is unable to distinguish what is self-evident from what is not. (This state of mind is clearly possible. A man blind from birth might reason about colours.) Presumably therefore such persons must be talking about words without any thought to correspond. (Phys. ii. 193a4-9; see also Phys. i.2 185a14-16 and Met. iv.4 1006a5-7)
Aristotle's goal is not to address the philosophical sceptic who wishes to be given apodectically
certain premises from which the truth of Aristotle's conclusions may be reached through
reasoning which is itself infallible.
The goal, from which Aristotle never wavers, is instead the weaker one of taking a
critical examination of the apparent phenomena, saving the appearances where they can be saved
and relinquishing them when they conflict with other better-attested phenomena or the prior
results of theorizing.54 This weaker goal is apparent in Aristotle's general and specific defenses
of his postulated four causes. Aristotle offers only a historical survey of theories of cause in
Metaphysics i.3-7 (see esp. i.7 988a18-24 and 988b16-8) as his defense of the four causal theory (he
offers no argument for it in the Physics).55 In the Metaphysics Aristotle repeatedly claims that his
predecessors were 'forced by the facts themselves' to accept the existence of the causes he lists.56
The existence of the causes to which Aristotle appeals is taken by him to be a datum familiar to
both the lay person and any serious student of nature.
These points are not undermined by the arguments presented for form and matter in
Physics i.7. The general support that form and matter receive from the facts and Aristotle's survey
in Metaphysics i is supplemented in Physics i.7 by the role they play in Aristotle's theory of change.
Change itself is presumed, not argued, to exist. Just as importantly, Aristotle's theory of change is
never argued to be the only possible theory of change. Aristotle accepts the existence of form
and matter on the basis of their role in a plausible theory of change which has no competitors he
54 See GA iii.10 760b29-33; Met. xii.8 1073b10-6, 1074a15-7, xiii.9 1085b36 and PA i.1 640a12-16. 55 Ross says, "We do not know how Aristotle arrived at the doctrine of the four causes; where we find the doctrine in him, we find it not argued for but presented as self-evidence" (1936, 37). Wieland explains the fact that no "thoroughgoing passage designed to prove the doctrine of causes exists" (1962, 261/146) to be explained by the doctrine's being the produce of linguistic analysis.
116
has not considered and rejected; nothing in his form of argument rules out the in principle
possibility that there be other theories of change which would not avert to form and matter.
Such a theory, if offered, would need to be evaluated on its merits, and may undermine his case
for the existence of form and matter.57
Form and matter thus receive support for their existence from the role they play in a
particular account of change. Final causality gains the same sort of additional support by its
acceptance among the phenomena by both common sense and systematic study of nature. We
are not offered reasons for the existence of the principles in question which would satisfy all
forms of sceptics or eliminitivists when the topic is form and matter. Nor on this subject are we
offered arguments which bar the possibility of accounts that do without the principles (i.e., form
and matter) to which Aristotle appeals. Aristotle's argument for final causality is directly parallel.
Nevertheless in both cases Aristotle has provided strong and plausible grounds for the existence
of the principles in question.
Aristotle's general defense of teleological commitment is both modest and plausible.
That it has held up very well indeed is demonstrated by the vitality of current realist accounts of
biological teleology.58 Aristotle refutes the specific alternative views of his opponents, marshals
evidence in favor of taking the phenomena at face value, and gives a theoretical account of the
compatibility of the final and material causes (Physics ii.9).59 Having shown his opponent's
position not to be adequate,60 Aristotle concludes that the teleological data from which he began
56 See Met. i.3 984a17-9, 984b9-10, i.5 986b30-1; cf. PA i.1 642a19-20, 27-8; Phys. i.5 188b28-30 and PN 455b20-1. 57 Nevertheless, it is crucial to understand that the mere possibility that another theory might be offered does not itself undermine Aristotle's case for form and matter. 58 On which see chapter five of this dissertation. 59 This aspect of Aristotle's defense of final causality is beyond the scope of this dissertation. 60 Aristotle's standard methodology does not even require that one be able to accomplish this before one is justified in claiming that one's own view is correct. See Met. iv.7 1012a18-20 and Top. viii.8 160b7-8.
117
the study of nature61 are correct.62 Aristotle's ultimate argument in favor of teleological
explanation comes to little more than this: first, the world presents itself to us in both common
sense and biological science as being teleological and, second, alternative accounts that appear to
sustain a simpler ontology are shown, upon analysis, not to offer alternatives in genuine conflict
with the data of common sense and science.63 Aristotle has proven correct in his estimation of
the explanatory value and staying power of teleological explanation in biology.64 While modest,
Aristotle's reasons for postulating teleology in nature are nevertheless extremely powerful.
3.2.7 Summary of the argument thus far.
In the previous two sections I have argued that the regularities challenge to the
coextensiveness of life and teleology in Aristotle fails. The regularities challenge is based on a
particular reading of the main argument of Physics ii.8, but that argument either runs into conflict
with Aristotle's theoretical statements concerning teleology elsewhere in his corpus or else
"makes nonsense" of the argument of the passage. I have argued that instead of accepting the
second horn of this dilemma we ought to seriously consider an alternative reading of the passage
on which it offers a plausible and moderate defense of teleology against a particular kind of
opponent.
I argued that David Charles' reading of Physics ii.8 as an ad hominem argument against a
particular brand of materialist opponent is both plausible on textual grounds and can be
61 See IA 2 704b11-17, Phys. viii.7 260b22, GA v.8 788b20; cf. APo. 10 76a31-6. 62 See EE i.3 1215a6, Rhet. iii.17 1418b5, NE vii.1 1145b1-7, OH ii.1 283b33-284a1, NE vii.14 1154a23-5, EE i.3 1215a6. 63 Jacobs makes a parallel point with regard to the teleological characteristics of organisms in response to an objector who considers teleology a delusion. "If here one still insists that this is only evidence that we conceptualize and explain biological phenomena as if there are de re necessities and that there is no demonstration that there are, one may coherently persist in that insistence. But this does not relegate our claims to the status of mere proposal. I take them to be a faithful articulation of the manner in which we do regard the biological part of this world or any world. The objection to this is that our essentialist intuitions are really delusions and an articulated delusion is still a delusion. But the centrality and firmness of these intuitions do, I think, constitute a prima facie presumption in their favor" (Jacobs 1984, 31). 64 See again the quotations cited in note 53, above. See also chapters five and six of this dissertation on contemporary accounts of biological teleology.
118
developed into a full and rich interpretation of the argument of Physics ii.8. On my interpretation,
Aristotle believes that his materialist opponents fail to explain an important phenomenon in the
natural world � the fact that living species breed true. Recognition of this phenomenon ought,
on pain of inconsistency, persuade his reductivist opponents to accept the existence of
unreduced animal and plant natures. Once such natures are accepted into one's ontology,
however, one ought simply to accept them as they appear � i.e., one ought to accept them as
teleological.
If this understanding of Aristotle's reasons in favor of postulating biological teleology is
correct, then the basis for the regularities challenge to the coextensiveness thesis fails. That
challenge was based on the assumption that Aristotle himself accepted the regularities thesis,65
but on the reading I have offered Aristotle does not endorse this thesis in Physics ii.8. In addition
to providing (for the first time) a plausible and moderate Aristotelian defense of natural
teleology, this reading then has the additional advantage that it brings Aristotle's commitments in
Physics ii.8 into line with his repeated denials of the regularities thesis in both theoretical
statements and practice.
In the next section I turn to consider the next challenge to the coextensiveness thesis;
the natures challenge.
3.3 The natures challenge.
According to the natures challenge to the coextensiveness thesis Aristotle does not
believe that life and teleology (in the relevant sense) are coextensive. Instead, Aristotle holds that
all things with natures are teleologically directed. We may formulate this view as the natures
thesis.
The natures thesis: All things with natures (e.g., animals and plants and their parts, and the simple bodies: earth, air, fire and water. Phys. ii.1 192b9-11) are intrinsically teleologically directed toward some end.
65 The thesis that all regularities are teleological.
119
Given that the material elements have natures in Aristotle's view (i.e., they have principles of
motion and rest intrinsic to themselves as what they are, Phys. ii.1 192b21-3), it follows that
teleology in the relevant sense is broader than living in Aristotle and, again, Aristotle therefore
rejects the coextensiveness thesis and therefore cannot offer the account of life I offered
provisionally at the close of the previous chapter.
The strongest argument that Aristotle is committed to the natures thesis is really quite
simple. In On Generation and Corruption ii.10 Aristotle offers a final causal explanation of the
subject of that work, coming to be and passing away (GC ii.10 336b25-34). The scope of the
'coming to be and passing away' discussed appears not to be limited in the text and thus appears
to range over all the natural bodies of the lower cosmos (cf. Joachim 1926, xxxvi-xxxvii). The
mutual transformations of the elements (see GC ii.1-8), then, appear to be given final causal
explanation in On Generation and Corruption ii.10. If these appearances are correct then in at least
one text Aristotle maintains the natures thesis and denies the coextensiveness thesis.
I believe that this appearance is deceptive and that a closer reading of the text reveals the
mistake. Here is Aristotle's text:
[a] Coming-to-be and passing-away [hê genesis kai hê pthora] will, as we have said, always be continuous, and will never fail owing to the cause we stated. [b] And this continuity has a sufficient reason. For in all things, as we affirm, nature always strikes after the better [tou beltionos oregesthai...tên phusin]. [c] Now being (we have explained elsewhere that being is spoken of in many ways) is better than not-being [beltion de to einai ê to mê einai]; but not all things can possess being, since they are too far removed from the principle. [d] God therefore adopted the remaining alternative, and 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 approximation to eternal being. (GC ii.10 336b25-34; slightly revised from the ROT)
I understand this passage as follows.
120
In (a) Aristotle refers us to the explanation for the eternality of coming to be and passing
away offered earlier in the chapter through the advance and retreat of the sun.66 In (b) Aristotle
announces that it is time to give the final cause for the eternality of coming to be. In (c) Aristotle
states the principle that "being ... is better than not-being [beltion de to einai ê to mê einai]". We must
be careful in understanding the meaning here, for Aristotle explicitly refers us in (c) to the fact
that being is spoken of in many ways. Aristotle must here express the principle that being actually
is better than being potentially; he cannot mean that being ( = existing) is better than not-being
(=not-existing) since these two are noncomparable.67 In (d) Aristotle metaphorically invokes
'god' as choosing (demiurgically) on the basis of this principle to fulfill the best state. The literal
meaning here, given his impersonal and noninterventionist conception of god,68 must simply be
that the eternality of coming-to-be happens in some sense because it is better thus.
Two sections of this text require comment to deflate the notion that Aristotle here
supports the natures thesis and rejects the coextensiveness of life and teleology. First, we need
to ask whether the principle stated in (c) is a general principle that applies to all things with
natures, living and nonliving. Second, we must ask in what sense coming-to-being is said in (d)
to be finally caused by being's being better than not-being in the relevant sense.
To the first question, whether the principle stated in (c) is a general principle claiming
that all actuality is always better than potentiality, I believe we must respond in the negative.
Aristotle clearly states in a number of places that it makes no sense to attribute being well off to
things which are not alive. He says,
[I]t would surely be ridiculous to wish wine well; if one wishes anything for it, it is that it may keep, so that one may have it oneself. (NE viii.2 1155b30-1)
[N]either health nor disease can exist in lifeless things. (PN 436a18)
66 See GC ii.10 336a13-24; see also Meteor. ii.4 359b34-360a5, i.9 346b20-4; GA v.10 777b24-778a2 and Met. xii.6 1072a11-18. 67 See Phys. vii.4 248b6-10; Cat. 8 11a5-13; Top. i.15 107b17 and GC ii.6 333a22. 68 See Metaphysics xii.7.
121
[T]he fact that these eggs go bad shows that they previously participate in some way in life. (GA ii.5 741a19-20)
Indeed, Aristotle makes theoretical pronouncements which give his reason for making these
particular claims.
[A]n inanimate thing or a beast or a child cannot do anything by chance, because it is incapable of choice; nor can good fortune or ill fortune be ascribed to them, except metaphorically. (Phys. ii.6 197b7-9)
[L]ife is in the soul (and therefore well-being [eudaimonia] also; for it is a certain kind of life). (Met. ix.8 1050b1-2)
Being well off belongs only to living things, and cannot be ascribed (except metaphorically) to
nonliving things. It would seem, therefore, that Aristotle accepts a view � widely shared today69
� that it is nonsense to claim that nonliving things are better in some states than others in a
sense other than being better-for-us.70 This fact would appear to rule out claiming that (c)
expresses a general principle which applies to living and nonliving beings.
Turning to the second question: in what sense is the eternality of coming-to-be finally
caused by being-actually's being better than being-potentiality? I argue that the main plausibility
of the claim that this passage contains a commitment to the natures thesis comes from
conflating71 two distinct theses: first, the thesis that the eternality of coming to be and passing
69 Thus, for instance, Mark Bedau says, "[I]t would seem that sticks, like stones and specks of dust, are simply not the kind of thing for which anything could be good (or bad); they do not have interests that can be promoted (or thwarted)" & "survival is neither good nor bad for non-living things. Non-living things are not the kind of thing that can be beneficiaries" (1992d, pp.272 & 282). 70 Realists about aesthetic value may hold that a beautiful world without consciousness is better (aesthetically) than an ugly world without consciousness � independently of any consideration of the possibility that such worlds be observed and appreciated for their aesthetic qualities. Even if Aristotle held such a view (and I am unsure how one would go about establishing such a thesis) the texts cited in the body of this chapter count against interpreting Aristotle as holding that things with natures are intrinsically directed towards putting themselves into such better states. In other words, although there may be (in fact if not in Aristotle) coherent senses in which nonsentient beings can be better or worse (e.g., aesthetically), these senses will do nothing to bolster the case that Aristotle conceives of all natures as directed towards their better states. My thanks to Richard Geenen and Jennifer Everett for bringing this issue to my attention. 71 This reflects the history of my reading of the passage at any rate.
122
away has a proximate teleological cause and second that thesis the eternality of coming to be and
passing away has an ultimate teleological cause.
The eternality of all (sub-lunar and heavenly) coming to be and passing away may have a
teleological explanation of one or two sorts. Either the explanation is proximate, in which case
individual sub-lunar natures have an intrinsic teleological directedness upon an end, or the
explanation is ultimate and only the outermost heavens72 (which are, of course living things on
Aristotle's account) have an intrinsic directedness upon an end. Sub-lunar natural entities are not
themselves directed upon any end in a proximate sense, but in the ultimate sense, given that the
efficient causal explanation of their coming to be etc. refers to the outermost heaves, their
movements do have an (ultimate) teleological cause (see Met. xii.7 1072b3-4).
The main argument against reading this passage such that it commits Aristotle to the
stronger thesis that the motions of all things with natures are proximately rather than merely
ultimately finally caused comes from the fact that the weaker reading fits in more closely with
Aristotle's statements and practices regarding the extent of teleology in nature throughout his
corpus. While defending this point in a thoroughgoing way is impossible in this context given
Aristotle's frequent use of metaphor in speaking of nature seeking to do the best,73 I will limit
myself to a brief argument to the effect that Aristotle did not suppose nonliving things with
natures to be teleologically directed.
In the Eudemian Ethics i.8 Aristotle argues against the view that numbers 'aim at' (ephiêmi)
unity by noting first that those who hold this view do not explain in what sense numbers 'aim at'
72 The outermost heavens and their motion are clearly relevant to the understanding of On Generation and Corruption ii.10 given Aristotle's invocation of them in 336a13-22. I am indebted to Gabriela Carone for bringing to my attention the need to explain the suitability of appealing to the outermost heavens in this context. 73 See, for example, PA iv.10 687a16, ii.14, 658a18-24, iii.3 665a18-26 and iii.4 665b18-21. For an exhaustive survey and evaluation of the texts in the biological works which agrees with the conclusion presented here, see Antony Preus, who says "'Nature does nothing in vain'... means that any essential process of any species which has in itself, qua itself, a principle of movement, is a purposive process.... this is, in fact, the limit of Aristotle's commitment to the existence of a purposive nature" (1975, 246-7).
123
unity, and second, he asks, "how can one suppose that there is desire [orexis] where there is no
life? [(h)ois zôê mê (h)uparxei]" (1218a27) In the de Anima iii.10 Aristotle makes clear that orexis "is
in every form of it relative to an end [(h)eneka tou]" (433a15). At Politics i.2 Aristotle clearly states
that both animals and plants nonconsciously74 'aim at' (ephiêmi) reproduction, where Aristotle
makes clear that this is a teleological aim (1252a27-33).75 In this group of passages from
different works Aristotle expresses no hesitancy to attribute teleological 'desire' (orexis) and
teleological aims (ephiêmi) to nonconscious beings such as plants � he clearly is no mentalist
about teleology � but he does balk at attributing such teleology to a class of things he himself
marks out as being lifeless.
This passage from On Generation and Corruption provides by far the most compelling
evidence in favor of a natures thesis in Aristotle's corpus. We have seen, however, that even this
passage is best understood as expressing views consistent with the coextensiveness of life and
teleology and fails to positively support the natures thesis. I have argued, first, that the principle
that being is better than not-being is limited in scope to living things and, second, that the
eternality of coming to be is only teleologically caused in a nonproximate sense. This reading
makes good sense in the context of Aristotle's consistent pattern of employing teleological
explanations only to living things throughout the corpus in both theoretical passages and in his
work as a practicing scientist. The passage contains no compelling grounds to break with the
uniform evidence of the bulk of the corpus and attribute the natures thesis to Aristotle.
74 Aristotle says such beings aim ouk ek proaireseôs, not from deliberate purpose. See also Physics ii.8 199b27-31. 75 He makes this clear in the context by saying that the male and female join for the purpose of ((h)eneka) reproduction.
124
3.3.1 God as an End in Aristotle
Charles Kahn's essay "The Place of the Prime Mover in Aristotle's Teleology"76 frames
an argument whose goal is to persuade us that Aristotle accepts a thesis even stronger than the
natures thesis, the thesis that all natures are intrinsically finally caused to seek actuality in god, the
prime mover. This thesis is stronger than the simple natures thesis in that it specifies a particular
end as the end of all things with natures. We may thus formulate a stronger natures thesis which
Kahn endorses.
The natures thesis (first strengthening): All things with natures (i.e., animals and plants and their parts, and the simple bodies: earth, air, fire and water. Phys. ii.1 192b9-11) are intrinsically teleologically directed toward being like a specific being: god.
It is difficult to be more precise about Kahn's goals and accomplishments, however, since he
admits that his argument doesn't demonstrate the conclusion he would like us to accept.
Kahn distinguishes between three views about the role of the prime mover in Aristotle's
cosmology with reference to final causality.
V1: The prime mover's efficacy as a final cause is restricted to finally causing the outer sphere of the heavens to move (strongly suggested by Met. xii.8 1073a24-8). The prime mover is the mechanical or efficient cause of other motions (cf. GC ii.10), but no more.
V2: All things with natures have as one of the proximate causes of their motion their intrinsic teleological directedness towards the prime mover. The prime mover is conceived "as the eternal guarantee and ultimate driving (drawing) force for all patterns of teleology in nature as well as in human life and action" (184; see DA ii.4 415a24-b2, Met. xii.10 1075a18, 1076a4).
V3: The demiurgic view: All things with natures have ends which they derive from the providential ordering of the prime mover in the way that an orchestra derives its ends from a conductor. The prime mover acts as "a creator who shows providential concern for mankind and who rules the world as his kingdom" (185; receiving possible support from GC ii.10 and infrequent phrases like OH i.4 271a33, "God and nature create nothing that is pointless").
Kahn maintains that Aristotle does not endorse V3, the demiurgic view of god's teleological
influence on the world. He holds, further, that V2 is true to Aristotle, but acknowledges that his
76 Kahn (1985). All references in this sub-section are to this essay unless otherwise noted.
125
argument does not demonstrate this conclusion. What he does believe that his argument shows
is that Aristotle does not endorse V1 � Aristotle believes that god is more than simply an
efficient cause of the motions of things with natures. Since Kahn does not define the position
stronger than V1 but weaker than V2 which his argument does support, I will proceed in what
follows to show why I believe we may accept V1 despite Kahn's argument.
Kahn's argument relies on his reading of a series of texts. I will not focus on each text
in turn, but will select those texts that give the most support to his view. The first two texts
Kahn cites are either inconclusive (the first, GC ii.10 337a21) or actually undermine his position
(the second, Meteor. i.2 339a19-32) unless they are supplemented by his third text (GC ii.10
336b27-337a7 � quoted and discussed above) � I therefore ignore them and move to the third
text. I have already argued against taking On Generation and Corruption ii.10 336b27-337a7 to
support attributing the nature's thesis to Aristotle. Kahn believes that the passage supports as
stronger thesis, the natures thesis plus a commitment about the nature of the particular end of all
things with natures. He takes it to show not merely that the goal of all potentialities is actuality
but that "The goal here is fullness of 'being', in other words77 approximation to eternal existence
and activity" (189).
Utilizing the distinctions of the last section, then, Kahn reads On Generation and Corruption
ii.10 as support (i) for attributing proximate teleological ends to all natures plus (ii) his particular
account of the form which those ends take. Against (i), Kahn offers no argument in favor of
reading this text as support for the strong view that Aristotle is committed to proximate as
opposed to ultimate teleological causation for all things with natures (as distinguished in the
previous section). I have already argued that there exist strong textual grounds to resist
attributing this view to Aristotle.
77 Kahn's phrase 'in other words' is misleading. It is possible that all potentialities have their full actualization as a goal, and that one 'approximates to eternal existence and activity' by fully actualizing potentialities without the two being identical. Thus, Aristotle argues at OH ii.14 296b6 ff. that earth's
126
In defense of (ii), Kahn believes that Aristotle is committed to the stronger natures
thesis by this passage because of its talk about simple bodies imitating more perfect movements.
This theme is picked up by the next text, quoted here in Kahn's translation:
Thus the sun and stars and the whole heaven are always in act <since they have no potentiality, except for change of place>... And things involved in change, like earth and fire, imitate the imperishable things <i.e., the heavenly bodies>. For they <sc. the elements> are also always in act; for they possess motion per se and in themselves. (Met. viii.8 1050b22-30; Kahn's translation)
Kahn concludes from these two passages that, "In following their own nature, then, the elements
imitate their ontological superiors" (189). However, this result is not supported by the texts in
any sense that supports the natures thesis. Kahn takes 'imitation' to be equivalent to 'being
finally caused by' but there is a weaker reading of this phrase which stays closer to Aristotle's
texts.
The claim here that the elements imitate the imperishable things immediately follows the
point that the heavens have (almost) no potentiality (i.e., they have only potentiality for circular
movement). The 'imitation' which the elements do is, I take it, directly related to this fact: the
elements, like the heavens, have by their nature only 'one-way' potentialities. Fire is not eternal
or indestructible, yet it does lack the potentiality (insofar as it is fire) for being cold (IN 13 22b36-
23a6; cf. Cat. 12b38, Met. ix.2 1046b5-24, ix.5 1048a1-24). Likewise, while living things die, their
forms lack the potentiality for not being. Death is an accidental result of their composition; their
essence is directed towards life and reproduction (see Met. viii.5 1044b34-1045a2). Destructible
things 'imitate' the always actualized indestructibles by being continuously actualized either (a) as
having only 'one way' potentialities as fire does or (b) as being members of a species which is
always actualized due to the reproductive ends of members. Aristotle's texts neither require nor
imply that the 'imitation' destructibles do of indestructibles involves any reference to final
causality by the prime mover.
natural place is the center of the whole universe rather than the center of the earth even though the two happen to coincide.
127
Kahn calls his next text "probably as close as Aristotle ever comes to a general assertion
of the point of view I am ascribing to him" (190). The text follows in Kahn's translation:
[Actuality (energeia) is not only prior to potentiality (dunamis) in substantial being (ousia) but also in time:]78 one actuality precedes another in time right back to the actuality of the primary eternal mover. (Met. viii.8 1050b5-6)
It is difficult to see, however, how this passage invokes a final rather than an efficient cause.
Potentialities, according to Aristotle, do need something which is actually the thing their
potentiality is defined in terms of in order to activate their own potentialities,79 but there is no
suggestion here or elsewhere80 that potentialities as such are teleologically directed towards
actuality.81
A further problem with this passage considered as support for any natures thesis is that
Kahn ignores the words 'as we said' which appear immediately before this passage at Metaphysics
viii.8 1050b4. As Ross notes, the passage to which Aristotle refers when he says 'as we said',
1049b17-29, "contained no explicit reference to the primum movens" (1924, ii 264). Ross points us
to Metaphysics xii chapters six and seven for the explanation of the reference, but those texts
contain no direct suggestion of a natures thesis and Kahn supplies us with none. The passage's
connection with teleology is tenuous � a very slim thread indeed upon which to rest a dramatic
thesis.
Kahn offers three further texts in favor of attributing his particular version of a strong
natures thesis to Aristotle. On the Heavens i.9 279a22-30 and Generation of Animals ii.1 731b20-
732a9 make claims restricted to living things and therefore cannot support Kahn's natures thesis.
The remaining text, On the Heavens 292a22-b25 involves Aristotle's attribution of life to the
78 Kahn supplies the gloss, in brackets, to the context of the passage just as I have included it. 79 See Phys. iii.2 202a10, viii.5 257b8-9, OH iii.1 299b20, DA ii.5 417a18, iii.7 431a3, GA ii.1 734a30-1, 734b20, Met. vii.9 1034a22-4, ix.8 1049b24-6, xii.3 1070a4, xii.6 1071b29-31. 80 Recall our earlier discussion of On Generation and Corruption ii.10 336b25-34, above. 81 I am indebted to Gabriela Carone for emphasizing a version of this point to me in comments on an earlier draft of this chapter.
128
heavens as a contribution to understanding their movement. Since they are alive we can
understand their motions as being directed towards a good, which (to deepen the analogy to life)
they attain easily and in a single (perfect, circular, unending) motion whereas other less fortunate
beings need many motions to attain their ends. The problem with Kahn's use of this passage is
that if we read it as he wishes � i.e., as attributing final causality all the way down to the
elements � we would be committed by the same reasoning to supposing that Aristotle held the
elements to be alive since it was the reference to life which was meant to make clear the
application of final causal explanations to the heavens. But, as Kahn notes later in a different
context, "Aristotle is no panpsychist" (199).
I conclude that the strongest texts in Kahn's favor do not support his particular version
of a strong natures thesis. The texts he cites cannot support the view that Aristotle held a
natures thesis of any kind without some further argument. Kahn is correct that his argument
does not establish his particular brand of a natures thesis � i.e., V2 � but he is incorrect in
supposing that he has shown V1 or the coextensiveness thesis false. We have yet to see any
reason to reject attributing to Aristotle belief that only beings which are alive are teleological (in
the relevant sense). Before we may conclude, however, we must look at one other version of a
strong natures thesis, that of David Sedley.
3.3.2 Human well-being as an End in Aristotle
David Sedley has argued in his essay "Is Aristotle's Teleology Anthropocentric?"82 for
the thesis that in addition to being committed to Kahn's version of the strong natures thesis,
each thing with a nature has as an end the benefit of humans. We may thus formulate a second
version of a strong natures thesis.
The natures thesis (second strengthening): All things with natures (e.g., animals and plants and their parts, and the simple bodies: earth, air, fire and water. Phys. ii.1 192b9-11)
82 Sedley (1991). All page references in this section are to this article unless otherwise noted.
129
are intrinsically teleologically directed toward a specific end: the benefit of humans.
The argument of Sedley's essay is restricted to establishing this as a thesis in Aristotle's
theoretical work as a physicist.83
Sedley's argument comes in two stages. In the first stage he uses a reading of Physics ii.8
198b16-199a8 to establish a regularities thesis and uses this regularities thesis as the grounds for
going beyond that thesis to his stronger natures thesis. I argued above that the regularities
reading of Physics ii.8 is incorrect,84 and will therefore focus on the main problem for the second
half of Sedley's argument.
The brunt of Sedley's position is borne by his interpretation of the following passage
from Physics ii.8. After arguing that both in the arts and in nature, where a sequence as a whole
has an end all the members of the sequence have the end as a goal,85 Aristotle says,
And in general, art perfects some of the things which nature cannot complete, and imitates others. Therefore, if artistic things are purposive, clearly so are natural things. (Physics ii.8 199a15-8; Sedley's translation)
Sedley offers the following interpretation of this passage:
83 'As a physicist' is meant to contrast here with 'as a political philosopher' since in Politics i.8 1256b10-22 Aristotle says the only thing in his entire corpus which would seem to directly commit him to the view in question. The standard moves by those who deny that Politics ii.8 supports a nature's thesis (for example, Balme 1987; Wardy 1993, 22; Wieland 1962, pp.157-9) are to (a) deny that Aristotle is speaking theoretically as a physicist in the passage, hence the passage has no ramifications for his physics; (b) to reduce the apparent commitment to teleology in the passage to something more benign; and (c) to claim that it might just be an incautious (given how the passage might be taken with reference to his scientific works) rhetorical flourish in a popular rather than technical introduction. Sedley replies to these responses (181), but I do not believe his responses succeed. If he could make out his case in the physical works then the strong reason based on Aristotle's theoretical works on nature would be overcome and Sedley would carry the day. Barring that, the standard responses succeed; hence the importance of Sedley's main thesis concerning Aristotle's theoretical commitments as a scientist. 84 Sedley remarks against the claim that Aristotle accepts that "something can be necessary and regular without being purposive" that he knows "of no evidence that Aristotle would allow [anyone] this escape route" since Aristotle holds that mechanical and teleological explanations are compatible (183). However, we cannot argue from Aristotle's acceptance of the compatibility of the two types of explanation to his endorsement of regularities theses. Further, I have provided numerous citations from Aristotle's text and a number of independent arguments to establish that Aristotle would allow us exactly this escape route. 85 Phys. ii.8 199a9-14; see Phys. ii.3 194b36-195a2, Met. v.2 1013a36-b3; cf. PA i.1 642a12-4, i.5 645b15-20.
130
The argument must be as follows. Art functions by either imitating or improving on nature. But art acts for a purpose. Therefore art must borrow this purpose from the natural processes which it is imitating or improving on. (187)
The problem with this interpretation is that it is difficult to understand how what follows
'therefore' is supposed to connect with what came before.86 Aristotle's passage mentions no hint
of a dependency of artistic ends or functions on natural ends or functions. It would seem
instead that arts (like agriculture, to use Sedley's example) derive their functions from our ends
(i.e., gaining quantities of food efficiently), not from the ends of the plants we cultivate (i.e., are
their ends to provide humans with food? See Sedley 1991, 192).
I can see no basis in Aristotle's texts for Sedley's claim that the ends of arts are derived
from the natural things on which they work. The natural reading of Aristotle's passage is to take
the comment that arts take up what nature cannot complete (Phys. ii.8 199a15-8) with what
comes before, i.e., Physics ii.8 199a9-14. When nature cannot complete a sequence with an end
on its own, human arts may be developed to help nature along and bring the sequence to its
natural end.87 Such a reading makes perfect sense of the passage in its context and contains no
hint of the elements Sedley requires in order to support his strong natures thesis.
Unfortunately, this is the entirety of Sedley's positive textual argument, and I find no
reason in it to accept that Aristotle is committed to his version of the strengthened natures
thesis. There is one further caveat to his position which is worth dwelling on, however. In
formulating a possible objection to his own view, Sedley notes that it would be quite implausible
to say that a pig's own nature is teleologically directed towards our benefit even if it is the case that
86 Wieland is more forthcoming about how the derivation in question is supposed to run. Wieland says, "If art, which in itself already proceeds in a goal-directed manner, imitates nature, then the order in nature can only, logically, be goal-directed. So art derives from nature the goal-directed character of its products, not nature from art" (1962, 269/153). This interpretation of 'imitation' language in Aristotle is even more extravagant than Kahn's; once again we shall see that there is a less extravagant reading which stays closer to Aristotle's texts. 87 Sedley's Aristotelian example comes from fr.11 where Aristotle remarks that there are some agricultural seeds which need the art of farming in order to survive (quoted on p.188). Sedley uses fr.11 and Physics ii.2 194a33-6 as supporting evidence for his thesis, but neither passage (even taking into account Sedley's
131
we use pigs for our ends and benefit (192). But then whose nature is directed toward the benefit of
humans? Not pigs'. Sedley's answer is that "the nature which is exhibited by the anthropocentric
natural hierarchy must be not so much individual nature as global nature � the nature of the
entire ecosystem" (192).
Did Aristotle have any notion of an ecosystem, however?88 Sedley rejects the idea that
Aristotle thinks of the cosmos or ecosystem as an organism (whose parts have functions in the
system), but argues that Aristotle conceives the universe as a system whose parts have ends on
the model of a household on the basis of Metaphysics xii.10 1075a11-25.
An apparently decisive objection against taking this view, however is that unless we take
the demiurgic view of the prime mover's role in the world89 it is impossible to see how the order
in the universe described in the passage could support the view that parts of the world have
interlocking functions as a result of being a part of this sort of whole. A household is set in
order by a master who assigns functions to members of her household, but this is not the view
of the cosmos or the prime mover expressed by Aristotle in Metaphysics xii.10. How, then, are
members to get functions in a household/universe without a head of the household/universe?
The analogy Sedley employs breaks down given Aristotle's view of the prime mover.
Further, Wardy (1993) argues that the supposition that the 'whole of nature' has a nature
directed towards the human good is incompatible with central tenets of Aristotle's metaphysics.
Wardy argues that no substance (which 'the whole of nature' would be on Sedley's view) could be
made up of substances; therefore people could not be substances on this view (25; see GA i.18
722b23-4). Wardy takes such problems to highlight the fact that the passage in question is highly
metaphorical and must be interpreted with caution. In particular, Wardy reads "somehow jointly
translation of ôs plus the participle in 194a33-6) offers any support for the claim that arts derive their ends from natures they work on. 88 Nussbaum (1978, 97n.53) offers a negative response. 89 See above, p. 124.
132
arranged [suntetaktai pôs]" at 1075a16-7 as Aristotle's warning that the analogy he is working with
is really quite loose; pôs serves as "cover for an embarrassing void" (26) in Aristotle's highly
metaphorical account. Given the problems attaching to a literal reading of Metaphysics xii.10
1075a11-25, that text can lend no positive support for Sedley's version of the strong natures
thesis.90
As I noted above, Sedley restricts the argument of his essay to consideration of
Aristotle's comments as a physicist and metaphysician. I have argued that his case from
Aristotle's physical and metaphysical works cannot support the view. Nevertheless, one may
retain lingering worries concerning Aristotle's commitments given his one apparently
unequivocal statement of the view. Aristotle says,
[a] For some animals bring forth, together with their offspring, so much food as will last until they are able to supply themselves; of this the vermiparous or oviparous animals are an instance; and the viviparous animals have up to a certain time a supply of food for their young in themselves, which is called milk. [b] In like manner we may infer that, after the birth of animals, plants exist for their sake, and that the other animals exist for the sake of man, the tame for use and food, the wild, if not all, at least the greater part of them, for food, and for the provision of clothing and various instruments. [c] Now if nature makes nothing incomplete, and nothing in vain, the inference must be that she has made all animals for the sake of man. (Politics i.8 1256b11-21)
Even if it is not the case that Aristotle endorses the thesis that all natures have as ends the
benefit of humans in his physical or metaphysical works, surely nothing could be plainer than that
he is committed to it by (b) and (c) of the text above?
Despite the plausibility of the standard reading of this text, I believe that it is incorrect.
I will argue we have strong reason from within the Politics itself to suppose that Aristotle does not
commit himself even here to the thesis that other things with natures have as their end the
benefit of humans.91 The reason is this. Aristotle makes the claims which appear to support
90 My primary objection to Sedley's reading of the text is not based on Wardy's argument but on the failure of the analogy on which his reading rests. 91 This claim derives support, of course, from Aristotle's uniform silence elsewhere in the corpus concerning the apparent commitments of this passage. I am at pains to emphasize, however, that the main reason derives from the text itself properly understood.
133
Sedley's natures thesis in (b) and (c) of the text above. Section (a) of the text, however, sets the
context for those passages. In section (a) Aristotle notes that some animals provide milk for the
sake of their young, and extrapolates from this point to the conclusion in (b) and (c) that other
things are for the sake of human benefit. The key point, however, is that we are only justified in
attributing the sort of ends to plants and animals with reference to our uses that we are justified
in attributing to mothers' milk with reference to the child's use. What kind of end does mothers'
milk have in relation to young?
The natural answer to this question is that it has derived ends, ends derivative upon the
intrinsic ends of other beings. On a plain understanding, milk has derivative ends in two ways:
the milk subserves the intrinsic reproductive ends of the parents and it is used by the offspring to
subserve its own intrinsic nutritive ends. There is no suggestion either in Aristotle or in fact that
milk has as an end intrinsic to itself its serving the parents' or the offsprings' ends. This reasonable
deflationary hypothesis finds support both within the Politics itself and in Aristotle's general
biological inquiries. At Politics i.10 Aristotle says,
[A]s political science does not make men, but takes them from nature and uses them (chrêtai autois), so too nature provides them with earth or sea or the like as a source of food. At this stage begins the duty of the manager of a household, who has to order (diathenai) the things which nature supplies � he may be compared to the weaver who has not to make but to use wool, and to know, too, what sort of wool is good and serviceable or bad and unserviceable. (Pol. i.10 1258a22-7)
Aristotle clearly indicates here that it is the household manager who supplies order to the things
which nature supplies � the raw materials do not have our good as their intrinsic end but as an
end derived from our use. This is why Aristotle says at Physics ii.2 that we use everything as being
for our sake ((h)ôs (h)êmôn (h)eneka pantôn (h)uparchontôn � Phys. ii.2 194a35).92
92 Sedley translates (h)ôs as indicating the reason we use things: "we use it on the ground [i.e., (h)ôs] that everything exists for our sake" (1991, 189). I take it that this is an overtranslation; (h)ôs simply indicates that we use the things as being for our sake, and this statement is neutral concerning whether the things themselves are intrinsically or extrinsically and derivatively for our sake. Sedley requires the stronger veridical reading, but the text itself is at the very least open to the weaker reading and cannot decisively point in favor of Sedley's reading. My thanks to Gabriela Carone for discussion of this difficult passage.
134
Further, we have strong evidence from Aristotle's biological works to suppose that
Aristotle himself conceived of such ends as derivative in exactly this way.93 The attainment of
soul constitutes the intrinsic end of living beings.94 Parts of animals serve this intrinsic end in
their various ways as tools, and like tools, they derive their ends from the ends of their user.
[A]ll natural bodies are organs of the soul. This is true of those that enter into the constitution of plants as well as of those which enter into that of animals. This shows that that for the sake of which they are is soul. (DA ii.1 415b17-18)
As every instrument and every bodily member is for the sake of something, viz. some action, so the whole body must evidently be for the sake of some complex action. Thus the saw is made for sawing, for sawing is a function, and not sawing for the saw. Similarly, the body too must somehow or other be made for the soul, and each part of it for some subordinate function, to which it is adapted. (PA i.5 645b15-20; see also PA i.1 642a12-4; GA i.2 716a23-5, ii.1 734b27-30; EE vii.9 1241b18-24 and vii.10 1242a13-7)
The relevant sense in which Politics i.8 is committed to milk's having an end is the same sense
that a tool or instrument has an end: not intrinsically, but derivatively upon its user's independent
possession of intrinsic ends.
If this is correct, then the sense in which (b) and (c) above attribute to animals and
plants the end of serving the benefit of humans is derivative upon the nutritive and reproductive
ends humans have and uses to which we put animals and plants in order (ourselves) to meet
those ends. Given that humans, as composite substances of form and matter, have nutritive and
reproductive ends, they use other things such as plants, animals, tools, and servants to serve
those ends. All of these things will, in a derivative sense, have ends in virtue of this use: their
derivative end will be to serve the intrinsic ends of the human(s) who use(s) them.
I conclude that Politics i.8 � even considered on its own terms � contains no hint of an
Aristotelian commitment to the strong thesis that all things with natures have as one of their
intrinsic ends the benefit of humans. Further, there is no textual basis for attributing to Aristotle
93 I emphasize and elaborate upon this point further in chapter seven. 94 I explain and elaborate upon this point in chapter seven.
135
Sedley's version of the strong natures thesis that all natures are intrinsically teleologically directed
toward the benefit of humans.
3.4 Conclusion
In this chapter I have focused on various challenges to life's provisional definition in
terms of teleology, challenges which arise out of worries concerning the scope of teleological
causation and explanation in Aristotle's metaphysics. If Aristotle defines life in terms of
teleology, then life and teleology must, minimally, be coextensive. I defended this thesis on a
rigorous textual basis against the two main sources of opposition, the regularities challenge and
the natures challenge. I argued that the regularities challenge fails to offer a sound interpretation
of its key textual basis, Physics ii.8, and that a proper understanding of this text lends no support
to the challenge. I argued further that Aristotle does not believe that all natures (including
elemental natures) are intrinsically teleologically directed, and defended this claim against the
arguments of its two main contemporary proponents.
The two main results from this chapter have been the exegesis of a plausible, powerful
and textually well grounded epistemic basis for Aristotle's invocation of final causality in nature
and a defense of the thesis that life and teleology (in the relevant sense) are coextensive.
Achieving this later result required a great deal of textually sensitive argumentation against
competing views, but the results have uniformly supported the limited view required for our
provisional definition of life. The next stage of my argument involves a deepening of our
understanding of the philosophical commitments involved in defining life teleologically. In
chapter four we take up the task of determining the ontic status of the final cause as it appears in
Aristotle's corpus. Chapters five and six engage in an extended contemporary evaluation of that
conception of the final cause.
136
Chapter 4
The Analysis of Aristotelian Teleology
The topics of the previous chapter included the scope of and the epistemological
grounds for Aristotle's natural teleology. I argued that the scope of Aristotle's commitment to
teleology is no broader than living beings1 and that Aristotle's epistemological grounds for
appealing to teleology in nature were moderate, appealing and powerful. Aristotle's commitment
to teleology is grounded in his commitment to preserving the data of common sense and science
in finished accounts of the structure of the world2 within the context of a sustained critical
defense of the commitments generated by the view.
This chapter focuses on the question of the ontological status of the final cause. Are we
to give an ontologically eliminitivist, reductivist, or nonreductivist account of Aristotle's
commitment to teleology in nature? Each of these interpretations has been forwarded in the
philosophical literature, although I hope to show in this chapter that focused attention on the
question of ontic status (i.e., separating the ontic task out from elucidating epistemological
grounds and scope issues) makes one answer compelling: Aristotle is committed to the final
cause's being a sui generis irreducible form of real causal factor in the world. Accounts of the
ontological status of the final cause are currently too intertwined with questions of scope and
justification for this point to be perspicuous. Once ontic issues take the foreground as they do in
this chapter, however, the philosophical and textual poverty of alternative interpretations
becomes clear.
If I am correct that Aristotle conceives of the final cause as a sui generis sort of causal
factor, then serious problems for the thesis of this dissertation arise. My thesis is that an
1 In chapter seven I argue for two stronger claims: first, that life and teleology are coextensive in Aristotle, and second, that Aristotle defines life teleologically.
137
Aristotelian analysis of life in terms of teleology constitutes a viable contemporary account of the
nature of life. While teleology is currently undergoing a rehabilitation in the modern literature on
the philosophy of biology,3 this resurgence of confidence in the scientific credentials of realist
approaches to teleology leaves behind accounts of teleology that are not grounded in the
reductionistic approaches currently in favor.4 In the next chapter I critically survey
contemporary reductive accounts of teleology as a first step toward evaluating Aristotle's
commitments. Chapter six completes our foundational work on teleology by arguing for a
physical ontology that is scientifically acceptable and in which Aristotelian teleology finds a
natural home. Chapter seven returns to Aristotle's account of life, deepening our understanding
of life's teleological nature and providing the provisional account with rich textual support.
4.1 Ontological issues in contemporary scholarship.
Given the current state of the secondary literature on Aristotle's teleology it is imperative
that we clearly distinguish ontological, epistemological and scope issues concerning Aristotle's
teleology. Modern worries about teleology pervade the secondary literature on teleology,5 and
the desire to find a philosophically defensible conception of teleology in Aristotle appears to
2 See Joseph Owens (1968, 163) and David Balme (1992, 93, commentary on 641b10-642a1). See also chapter three of this dissertation. 3 See especially Allen, Bekoff and Lauder (1998b), Buller (1999a), and Nissen (1997). 4 Part of the intent of Buller's recent anthology on biological teleology is to chart the "near-consensus .... [concerning] certain fundamental commitments" which has emerged in the literature on biological teleology (Buller 1999b, 1). This consensus centers around agreement "that the biological concept of function is to be analyzed in terms of the theory of evolution by natural selection", and "represents as great a consensus as has been achieved in philosophy" ( and Allen and Bekoff1995; 1995, who label this view 'the standard line'; 20, cf.26; see also Godfrey-Smith 1993). The view that I attribute to Aristotle, unlike the interpretations favored by Furley (1996), Irwin (1988), Lewis (1988), Balme (1987), Matthen (1989) or Bradie and Miller (1984; see also Miller 1995), goes against this biological and philosophical consensus. All of the previously mentioned commentators on Aristotle gesture in the direction of modern accounts (different ones, mind you) for which they claim to find a precursor in Aristotle; none looks with a critical eye at any of the various competing accounts on offer, however. Gotthelf's (1988) gesture towards literature criticizing Woodfield's (1976) analysis is the near exception that proves this rule. 5 For general statements of modern concerns about teleology, see especially Woodfield (1976, 3) and Mayr (1988, 40). The invocation of these problems as posing unavoidable interpretive challenges to understanding Aristotle on teleology displays itself clearly in Gotthelf's formulation of the problem in
138
drive many commentators to find links between Aristotle's uses of teleology and modern
reductive analyses. The issue of the interpretive and philosophical adequacy of these accounts of
the ontic status of Aristotle's commitment to final causality, however, is obscured by the
entanglement of such discussions of ontic status with elucidations of epistemic grounds for
accepting teleology and questions concerning the scope of Aristotelian teleology. My goal in this
chapter is to disentangle these threads, focus solely on the question of the ontic status of the
final cause (is it eliminable, reducible, or a sui generis kind of causal factor in the world?), and to
argue for the view that Aristotle is committed to sui generis teleology independently of any qualms
we may feel towards such a position.
The fact that epistemological, ontological and scope issues are insufficiently
distinguished in the literature is readily apparent in Gotthelf's taxonomy (1997) of the relevant
contemporary secondary literature. Gotthelf determines to distinguish between different
accounts of Aristotle's teleology on the basis of interpreters' distinct answers to the question,
"What fact or facts grounds or licenses Aristotle's use of teleological explanation?" (74). But this
question is ambiguous between epistemological and ontological readings. On an epistemological
reading, the question asks what reasons Aristotle offered for accepting a commitment to final
causality. On an ontological reading, the question asks about the ontic status of the final cause: a
reductive account will seek to ground Aristotle's apparent commitment to sui generis final causality
in his (more acceptable) commitment to material, efficient, or formal causes. Eliminitivist views
claim there is no ontic ground, but claim that teleology has an important pragmatic or heuristic
role in our theories nonetheless. Nonreductive accounts ground Aristotle's final causality in the
existence of a sui generis real causal factor in the world.
Gotthelf uses both of these questions (without distinguishing them) to classify
interpretations of Aristotle's teleology. Thus, Gotthelf distinguishes a group of ontologically
terms of the problem of 'biological reducibility' (1976/ 1987, pp.208-12). See also M. Bradie and F. Miller (1984, 133), Balme (1987), Wieland (1962, 141), Matthen (1989, 159), and Furley (1996, 59).
139
eliminitivist accounts from realist accounts, employing a clearly ontological criterion. But
Gotthelf distinguishes another group according to the authors' shared rejection of the thesis that
one of Aristotle's (epistemic) reasons for accepting natural teleology was the unavailability of
complete material causal explanations for natural phenomena. David Charles' account of
teleology satisfies this epistemological criterion, but his account is not included in the grouping;
his account instead merits its own category on the basis of his unique answer to the question of
ontic status. While John Cooper does not share Charles' view on the ontological question, his
view receives a mention with Charles' on the basis of the fact that Cooper believes that some of
the considerations Charles raises constitute Aristotle's epistemic grounds for accepting teleology.
Mohan Matten's answer to the ontic question is identical to that of Balme, Bradie and Miller, but
Matthen is placed in a different category because of his distinct answer to the question of
Aristotle's reasons for postulating final causes.
Gotthelf's taxonomy is far from apt, but my point in making these cross classifications
explicit is not specifically to criticize Gotthelf's taxonomy. His taxonomy in fact accurately
reflects the way in which ontological and epistemological questions concerning Aristotle's
commitment to teleology have been approached in the literature. If we are to give a satisfactory
account of the ontic status of the final cause in Aristotle, we need to separate out the
epistemological and scope issues along with modern concerns about the acceptability of
Aristotle's final answer and simply confront Aristotle's view. I argue in what follows that such
focused attention yields a clear answer: Aristotle's notion of the final cause is the notion of an
irreducibly distinct sui generis real causal factor in the structure of the world. Justifying the
acceptance of teleology is another task, as is assessing the contemporary merits or demerits of
Aristotle's answer to the question of its ontic status.
4.2 Eliminitivism, reductionism, or irreducibility?
Answers to the question of ontic status may be either eliminitivist or realist. Among
realists, one may be a realist of a reductive or nonreductive sort. Among reductive realists, it is
140
possible to hold that commitment to final causality in Aristotle reduces to commitment to the
mental or to material, formal or efficient causality.6 Reductions to each of these types of cause
may take a number of forms; as it happens each of the main types of reductive interpretations
has at least one contemporary proponent, and there exist two distinct types of accounts which
reduce Aristotle's commitment to teleology to the material cause.
In what follows I survey and critically assess the extant accounts of the ontic basis of
Aristotle's commitment to final causality.
4.2.1 Eliminitivist accounts.
Contemporary attention to the ontological question concerning Aristotle's final cause
has been least entangled with extraneous issues as it has focused on the distinction between
realist and eliminitivist interpretations of Aristotle's teleology. Not surprisingly, scholarship has
also been most decisive and convincing in precisely this area.7 A number of commentators8 have
proposed that Aristotle thought there to be no ontological ground for the final cause. According
to such views there is nothing in nature corresponding to Aristotle's notion of a final cause;
commentators incorrectly reify Aristotle's talk about final causes on the basis of shallow linguistic
evidence. Such a view appears to run afoul of Aristotle's consistent use of final causality in
explanatory contexts, but the position holds that Aristotle's use is justified on heuristic or
pragmatic grounds compatible with the claim that there is nothing in nature corresponding to a
final cause.
As I mentioned at the outset, the distinction between realist and eliminitivist accounts of
Aristotle's teleology has received the most focused attention in the scholarly literature, and I
endorse what appears to be the consensus view that the eliminitivist view radically distorts
6 Or, I suppose, some combination of these. I know of no one who offers such a view. 7 With one possible exception; the case against a mentalistic reduction of teleology in Aristotle is so overwhelming that it is also widely accepted. See below.
141
Aristotle's position.9 In brief, we may note that Aristotle consistently conceives of causes as real
things or events in the world, and there is no principled reason to suppose that he abandons that
view with regard to teleological causation. The view that Aristotle held causal factors of some
(or all) sort(s) to be mere explanatory devises appears to be grounded in contemporary fears that
more realistic interpretations of the final cause in Aristotle must run afoul of sensible constraints
on adequate scientific explanations. Whether or not this was indeed the main motive for the
view, commentators on Aristotle have of late � and in my view rightly � turned their attention
to accounts of Aristotle's teleology that find some basis in reality for the final cause. I turn now
to such accounts.
4.2.2 Realist accounts 1: reductive accounts.
An intuitive conception of reduction is all that is necessary for understanding the sense
in which the accounts to follow are reductive.10 Our interest is in metaphysical or ontological
claims concerning reduction rather than linguistic or explanatory reductions. The core claim of
an ontological reduction is the claim that our apparent ontic commitment to entities of type F in
certain modes of expression really commits us to no more than the ontology of other more
ontically perspicuous modes of expression G.11 Thus, a reductive physicalist claims that
mentalistic expressions such as 'I believe in Santa' or 'I see an orange after-image' which might
seem, prima facie, to commit one to a sui generis mentalistic ontology of beliefs and after-images in
fact express no ontic commitment to anything beyond purely physical entities and properties (or
predicates). Idealists believe that expressions such as 'There is a table here now' when properly
8 This view was argued by Wieland (1962), Nussbaum (1978), and Sorabji (1980). Nussbaum has since abandoned the view; see Nussbaum (1980). 9 For detailed arguments, see for instance Gotthelf (1980), Nussbaum (1980), and Balme (1987). 10 For a formal treatment of reduction, see the appendix to chapter six. 11 Thus, Paul Teller (1995) says that reduction is "the replacement of one expression by a second expression that differs from the first in prima facie reference" (679). For further preanalytic characterizations of ontological reductionism, see Kim (1999, 15) and Klee (1997, 83).
142
understood express no commitment to the existence of physical things such as tables, but in fact
reveal commitment only to ideas.
Let us define a generic notion of reducibility as follows. Supposing that expressions of
type E1 commit one, prima facie, to entities of type F, where F things seem prima facie to be
distinct in kind from entities of type G, then:
Generic reduction (roughly): F properties, relations and entities are reduced to G properties, relations and entities iff, for each entity F there are properties or entities G1...Gn such that x's being F just is x's having G1...Gn in relation R; or, x's being F is nothing ontologically over and above x's having G1...Gn in relation R; or, being F need not figure as an independent existent in a complete ontological inventory of the world.
This generic notion of reduction admits of a plurality of understandings depending on how one
understands what it is for one thing to 'just be' or 'be nothing over and above' another. These
distinct understandings form distinctive conceptions of reduction.
We need not pursue these more specific forms of reduction in great depth here;12 the
intuitive notion explained above suffices to understand and evaluate the reductive credentials of
the views which follow. Our questions are twofold: is there a viable Aristotelian reduction of
final causality to some other (hopefully more philosophically acceptable) form of causality? And,
did Aristotle himself actually conceive of final causality along any of the reductivist lines?
4.2.2.1 Reductions to the mental.
Although the view is currently � and in my view correctly � out of favor, some
commentators13 have taken Aristotle's notion of teleology to be mentalistic. On a mentalistic
conception of teleology, all end-directedness depends for its existence on the mental properties
� specifically the beliefs and desires � of some agent. It is my use (or the designer's
12 Again, for greater detail, see the appendix to chapter six. 13 See the citations on the demiurgic view of teleology in Kahn (1985). For a contemporary advocate of a mentalistic conception of teleology, see Nissen (1997).
143
intention)14 that gives my coffee cup its function. In contrast, I have goals that derive not from
an external agent but primitively from my being an agent. On the mentalistic view, all
nonagentive teleology (i.e., in particular biological teleology, the functions and purposes of plants
or organs) is understood on the model of the account of the function of my coffee cup. All
nonagentive teleology depends on or is derivative upon the ends, purposes, or designs of agents.
Despite some potentially misleading linguistic cues in Aristotle's corpus, there is strong
theoretical reason to reject a mentalistic reduction of Aristotle's final cause.15 The reason is,
quite simply, that in Aristotle's world there is no mind on which all the directedness can depend.
Aristotle accepts that plants have goals but do not themselves have minds and Aristotle's god is
too impersonal to provide the mind a mentalistic reductionist would need. There seems simply
no way to reconcile Aristotle's views on final causation with his theoretical commitments in
theology and biology.
4.2.2.2 Reductions to the formal cause.
Cooper (1982; 1987)16 argues that Aristotle's "belief in the goal-directedness of nature
derives from" his belief that living things have both formal and material natures (198).
According to Cooper, formal and material explanations are "both of them basic to the
understanding of natural phenomena generally, and equally so � they cannot be dispensed with
in favour of anything more basic than they, nor can either be discarded in favour of the other"
(199). Formal and material causes form the bedrock level of ontic structure in Aristotle's system.
Cooper goes on to explain how the other two types of causes in Aristotle's system, the
final and efficient cause, fit in. The form of a living thing is its mature form; but "its mature
14 The issue is more complicated than this. See chapter five for the details; they ought not obtrude here. 15 Such an account would be reductionistic in this sense: it would account for prima facie nonmentalistic teleology (i.e., biological teleology) in other, independent terms � i.e., mentalistic terms. Contrary to appearance, then, nonmentalistic teleology would turn out to be something else � mentalistic teleology. 16 My account will refer solely to Cooper (1982) since the later article recapitulates the earlier (see Cooper 1987, 243n*).
144
natural condition, is a good for it, [so] reference to the form here is reference also to the goal of
the process of growth by relation to which, therefore, it is to be understood" (200). Invoking the
mature form in an explanation of a seedling's growth or structure is "at the same time to invoke a
goal" (200).17 The efficient cause, Cooper claims, sometimes plays a formal role in Aristotle's
work and sometimes plays a material role.
On this account we have two basic types of causes in the world: formal and material. As
Cooper says, "There are then two distinct and independent levels of facts and correspondingly
two levels of principles that Aristotle holds are responsible for what happens in the course of
nature" (202). Corresponding to these types of causal factors, we have two types of explanation,
formal and material, which explain by reflecting or exhibiting the causal structure of the world.
Final causal explanations are formal explanations,18 and sometimes efficient causal explanations
play a formal role (indicate formal causal relations in the world) while at other times they play a
material role (indicate material causal relations in the world).
Is this 'derivation' of the final cause from commitment to the formal cause apt? Is it
Aristotle's? In fact, the reduction is philosophically suspect. That an organism's form or mature
state is a good does not entail that that creature's growth is teleologically directed towards the
attainment of that good, for the goodness of the goal may be an accidental concomitant of a
causal process that independently results in that form.19 Insofar as we understand the process in
terms of formal causality, moreover, the immature organism's potential for form, which explains
its growth into a mature state, need not be teleologically directed towards that end but may only
17 Cooper cites Aristotle's identification of the formal and final causes at Phys. ii.7 198a25-6; DA ii.4 415b10-12; GA i.1 715a4-6 and 8-9. See also Phys. ii.8 199a30-33 and GC ii.9 335b6. 18 See the citations above in n. 17. 19 See the statement of Empedocles' view in Physics ii.8 and the account I offered of Aristotle's surrounding argument in chapter three.
145
be definitionally posterior to that end � the potential is the potential it is in virtue of being the
potential (nonteleologically) for that form.20
I of course grant that Aristotle does in fact understand the process of an organism's
growth to maturity as teleologically directed; the point here is the narrow one that the processes'
being teleologically directed is not nothing 'over and above' its being formally caused by the
mature state. Cooper's invocation of a correlation between the teleological goal and the form
does not support the claim that the final cause is understood in terms of commitment to a more
basic form of causal tie (i.e., the formal); it indicates no more than their coextensiveness in
normal cases. The formal causation of the growth of an organism is compatible with the
absence of final causal influence; formal causation is therefore not sufficient for teleological
causation. But if it is not sufficient for it, then teleological causation is something 'over and
above' formal causation and cannot be accounted for in purely formal causal terms.
That this account of the ontic basis for the final cause is inadequate becomes clear when
Cooper comes to what he calls the "crucial point" for understanding the positing of formal (and
final) explanations. According to Cooper,
the living things in question are so structured that each one's organs and other parts work together to make it possible for it to achieve to a rather high degree its own specific good, the full and active life characteristic of its kind, including the leaving of offspring behind; the actual plant and animal life that is preserved is all of it good.... [Further,] because the regular outcome of each such process is something good, one is also entitled to interpret the process itself as directed at that outcome as its goal. If it is a fundamental fact about the world, not derivable from other natural principles, that it maintains forever these good life-forms, then the processes by which it does so, being processes by which something good is achieved, are for the sake of the outcomes. (213, emphasis added)
I have here underlined where Cooper's 'derivation' or reduction of the final cause to the formal
derails. The problem is that in the crucial passage above (and in a number of other places, see
20 See DA ii.4 415a16-21, NE ix.9 1170a15-19, Met. ix.8 1049b4-16. For more on this point, see my discussion of Gotthelf's reduction of the final cause to a sort of (higher level) material cause, below.
146
pp. 200, 205, 207, and 208) Cooper reasons invalidly. The mere fact that a good is forever
associated with a process is no indication that the process occurs for the sake of the good.
We may illustrate the point by modifying one of Cooper's own examples (see 208n7).
The two-coins case: Suppose one tosses a fair coin to determine whether all one's future coin tosses will employ a fair or biased coin. Tails lands up, and so as a result of a chance happening not teleologically directed towards an infinitely long winning streak, just such a streak occurs.
Just as we may not infer from the fact that a person regularly benefits from the untrue coin that
the process occurs because it benefits that person, so we may not infer from the eternal existence
of active lives characteristic of living kinds (which also happen to be good) that the parts which
make that eternal fact possible happen because the result is good: perhaps it was the result of a
cosmic lottery as the Empedoclean objector seems to suggest in Physics ii.8.
There are in fact hints in Cooper's article which suggest that, perhaps, he takes the final
cause to be a primitive ontological causal factor alongside and related to the formal cause. In
discussing the goals, Cooper makes clear that goals are good things made possible by other
things, "where this other thing exists or happens (at least in part) because of that good" (197).
Perhaps we are to understand by this and similar claims a commitment to a sui generis sort of
causal factor which acts qua good, and are to understand the talk of 'derivation' epistemically as
part of Aristotle's reason for believing in the existence of final causes.
Which interpretation of Cooper is the right one is unimportant to the task of this
chapter; if Cooper intended an ontologically reductivist account then he has failed, for he has
failed to show how the formal cause can account for a cause's being a cause qua good. If his
account is meant merely to make clear Aristotle's reasons for accepting final causes, and he
interprets Aristotle as committed to the existence of a sui generis sort of causal factor in the final
cause, then I am happy to accept his account as allied with the thesis of this chapter. Either way,
however, it is clear that Cooper's mixed discussion of the ontological status of and epistemic
grounds for the final cause obscures the philosophical and interpretive significance of his
147
position. Aristotle's final cause is not ontically grounded in the formal cause; whether or not
Cooper thinks it to be is neither here nor there.
4.2.2.3 Reductions to the efficient cause.
Furley (1996), Irwin (1988) and (perhaps) Matthen (1989) take Aristotle's commitment
to final causality to be understood reductively in terms of his commitment to a special sort or
aspect of efficient causality.21 As Furley says, "we can understand the material, formal, and final
causes as being different aspects of the efficient cause, or perhaps different kinds of efficient
cause" (62).22 Efficient causality is ontologically basic; final, material and formal causes are
aspects or kinds of efficient causality.
The reductive account itself is expressed most completely when Furley completes
Aristotle's unexpressed thoughts with the following:
The words I am offering him [Aristotle] are something like the following. This kind of animal has such-and-such a manner of life: it is defined by having the capacity for living this kind of life.... Now this kind of physical part is necessary, or good, or at least better than anything else, for leading this kind of life. Hence the possession of this part enables its possessor to survive and reproduce. Through the process of heredity ... each of its (normal) offspring ... is equipped with this part. The cause of this individual's
21 In favor of reading Matthen along these lines, he says, "Powers are efficient causes in Aristotle's scheme of things. Final causes are not additional influences over and above powers, but are built right into the specification of powers" (161). Against this view, see below at n. 33. I will largely ignore this reading of Matthen in this section, but my comments on Irwin and Furley will apply to Matthen should anyone wish to press the case. Gotthelf includes Sauvé Meyer in this group in his taxonomy; I do not because I do not believe she actually takes a stand on the ontological issue in her paper (see especially 1992, pp. 811-12). 22 This same thesis is expressed by Irwin: "To mention the formal, material, final and efficient causes of the statue is not really to maintain four distinct causes of the same thing. When the effect is specified more clearly, reference to the first three causes turns out to be attribution of formal, final, and material properties to the efficient cause" (1988, 96). Note, however, that Irwin too makes clear that "This is a restatement, rather than a statement, of Aristotle's doctrine" (96). Taking this to be even a restatement of Aristotle's doctrines rests, I believe, on highly questionable philosophical assumptions and has a very thin textual basis. It should not pass without comment that the reductivist theses proposed by these authors are extremely ambitious both philosophically and interpretively; we ought to expect them to meet high standards of proof if we are to accept any such 'restatement' of Aristotle's main theses given Aristotle's consistent use of the four causes as four separate causal factors (see on this point Balme 1987, 281, quoted below).
148
possession of this part, then, is the fact that this part is good for this kind of animal and therefore was a part of the form inherited from the parent. (73)23
The idea � although Aristotle never brings himself to express it � is supposed to be that an
aspect of the efficient causality of a part or process grounds the final causal explanation for the
presence of the thing or process (see pp. 71 and 77). The relevant aspect of the efficient
causality is the benefit the organ's doing F has for the organism; it is the benefit provided by doing
F which explains why the organ is present in organisms of later generations. The thing or
process is present in this instance because of the benefit it brings to things of that kind; it is the
part's being good for leading a kind of life that enables its possessors to survive, reproduce, and pass
the part (and benefit) along.
This account of the ontic basis of Aristotle's final cause cannot succeed. First, Furley
gives no reason to believe that Aristotle himself conceived of the final cause as an 'aspect' or
'kind' of efficient cause. We might suppose that the fact that two interpreters arrived at the same
interpretation of Aristotle is itself evidence that there is some basis in Aristotle's text for the
account. Unfortunately, however, it is clear where commentators get the idea for interpreting
Aristotle along roughly these lines; there exists a popular modern reductive analysis of final
causality which is strikingly similar.24 Both accounts derive from a modern analysis of biological
function, and this explains their similarity; neither interpreter suggests convincing reason to
believe that Aristotle conceived of biological functions along the lines suggested by this analysis
23 Those familiar with Wright's analysis of functions (1973; 1976) and the literature which has grown up around it (see especially the collection of papers in Buller 1999a) will find Furley's account here very familiar. Irwin (1988, pp. 107, 523-4n.22) endorses an almost identical view. 24 After giving a description of the efficient causal explanation of final causality that is very similar to Furley's, Irwin (1988, 523-4n22) cites Wright's analysis of biological function (presented in his 1973; and 1976). Furley himself only claims that 'evolutionary biologists' give this sort of reading of biological function; this is true of course, and the philosophical basis for the contemporary evolutionary account of functions derives from Wright's work (see Furley 1996, 67-8). For more on Wright's analysis of biological functions and the burgeoning field of reductive analyses of biological functions, see chapter five of this dissertation.
149
however.25 Since Aristotle discusses biological functions, we may (of course) read this reductive
contemporary account of final causality back into a great many of the things Aristotle says, but
finding Aristotle himself explaining or even indicating this grounding for the final cause is
another matter altogether.
This problem has a second aspect: this account of teleology cannot account for all of the
uses Aristotle makes of the final cause in his work. The modern account from which this
interpretation derives is designed to account for biological functions, which belong primarily to
organs. While Aristotle does believe that organs do have functions, he also believes that
organisms have other ends and goals, and these ends and goals are not explicable in the terms of
this reductive analysis. Thus, on Aristotle's view reproduction itself is one of the goals of the living
thing (see Pol. i.2 1252a28-30, DA ii.4 416b23-5, GA ii.23 731a25-b7, HA viii.1 588b25-6), but
reproduction cannot be given an account along the lines indicated above.26 On Aristotle's view,
organisms have goals of individual flourishing (Phys. ii.2 194a28-33, ii.3 195a23-5, ii.7 198b8-9;
Pol. i.2 1252b34-5; EE i.8 1218b9-11, ii.1 1219a9-11; Met. i.3 983a31-2), but again, this goal
cannot be accounted for along the lines of the analysis indicated.27 The problem here is simply
that the supposed reductive analysis of teleology is not broad enough to cover the range of cases
of teleological directedness that Aristotle recognizes. Further, Aristotle does not draw
distinctions between types of teleology such that Furley and Irwin could claim only to capture
25 Irwin comes closest on this matter. He appears to suggest that Aristotle's commitment to this sort of account of final causality can be seen in Physics ii.8. I do not believe that his account of Physics ii.8 can be correct, however, for it rests on attributing to Aristotle's opposition the highly implausible thesis that in each generation of animals and plants it is a matter of random chance what parts they are born with. Other problems are apparent; see chapter three for my account of Physics ii.8 and further criticisms of Irwin's view. 26 Such an account would go along these lines: reproduction exists because it was necessary or good or better than anything else for enabling things which possessed it to survive and reproduce. This is no explanation at all. 27 Aristotle recognizes still further types of goals or ends for organisms that cannot be accounted for along the lines of the analysis. Thus, he recognizes goals of self-maintenance (PN 479a28-9, DA ii.4 416b17-18) and growth into a mature state (GA ii.3 736b4-5; see further citations in chapter seven). Chapter seven of this dissertation surveys the range of uses to which Aristotle puts his teleological notions.
150
some restricted range of Aristotle's uses. Aristotle himself does not recognize such a restricted
range; his use across cases is consistent and uniform. The reduction to efficient causality cannot,
then, be Aristotle's view of final causality because it is too narrow.
This last point focuses our attention on the fact that the account of teleology offered is
really an account of biological functions (as its modern models are) and as such cannot provide
an account of Aristotle's other uses of teleology. But once we see that the account is really an
account of biological function, it becomes apparent that it fails to render an adequate account
even of that notion. Recall that on the account, an aspect of an organ's efficient causal action �
its producing benefit to the organism � is supposed to warrant the teleological explanation in
addition to the standard efficient causal explanation of the part's behavior and existence.
Because the part benefits its possessor it gets passed on down the generations, and in later
generations we come to recognize the function of the part to be the benefit it produces for
organisms which possess it. We may formalize the account as follows.
The function of a part, x, of an organism, O, is to F just in case (a) x does F; (b) doing F is good for Os; and (c) as a result, O passes x and its ability to do F on to its offspring.
This is not an adequate notion of functions in Aristotle or in fact, for it fails to provide a
necessary condition for being a function, and further, the account either fails to provide a
reductive analysis or fails to generate a real causal role for the teleological aspect of the efficient
causality of the events and processes.
The analysis fails to provide a necessary condition for possessing functions because
spontaneously generated things may have functions.28 Indeed, Aristotle himself happens to
accept all of the following propositions: there exist beings which are spontaneously generated;
such organisms cannot reproduce after their kind (GA i.1 715a23-5, b4-15); and such organisms
have parts with functions (see HA iv.7 532a5, a11; PA ii.8 653b38-654a2, iv.5 678b11-14,
151
679b35-6, iv.7 683b10-11). It cannot be a necessary condition for the existence of functions,
then, either in Aristotle's view or in fact,29 that the organisms which have parts with functions
gain those functions through ancestral efficient causality.
Further, the definition either fails to live up to its billing of showing the final cause to be
an aspect or kind of efficient causality or else the definition fails to provide any role to the
teleological aspect of its analysans. Condition (b) of the analysis � that x's Fing be good for some
O � is meant to show how goodness enters into the efficient causation of the event, thus
explaining the role of teleology in explaining30 the presence of the organs (i.e., xs are present
because they do good). But just how does the goodness contribute to the causal story? Either the
goodness provides a sui generis causal contribution to the event, or else the goodness appears to
be an epiphenomenal rider which does no work. In the former case, the teleological aspect of
the causation appears to play a sui generis role in the causation of the event, and it is difficult to
see how this can be considered a mere 'aspect' or 'kind' of standard efficient causality. In the
latter case the teleological aspect of the causation appears epiphenomenal, all the causing is
efficient and done in clause (a) of the analysans.31 This point may be made in another way by
noting that the 'as a result' clause in (c) is ambiguous. Either the perpetuation happens as a result
28 This problem arises for the most popular contemporary reductive account of functions. See chapter five for details. 29 It cannot be correct in fact because spontaneous generation is possible even if it does not actually happen. Given that it is possible that there exist beings which do not derive the functions of their organs from the efficient causal story leading up to their appearance, the account does not provide a necessary condition. 30 We must keep in mind that insofar as we are realists about teleology and not eliminitivists, these explanations must explain in virtue of capturing some specifically teleological feature of the causation of the event. Irwin and Furley are not offering accounts merely of teleological explanation, but of teleological causation. See further n. 31 below. 31 It is crucial to keep in mind that we are here dealing with realist rather than eliminitivist accounts of teleology. The realist must find a causal role in the world for teleology. Eliminitivists might be happy to accept the argument I make in the text and accept that there is no causal role for the good to play in the course of events. Of course there isn't, the eliminitivist would say, there is after all no ontic ground for teleological explanations. Nevertheless, she might continue, the analysis presented here provides a good explanation for why we find it useful and natural to explain things teleologically � the good is, after all, tightly correlated with the causal processes generating the organs. This eliminitivist response is not open to those, like Furley and Irwin, who accept that Aristotle's account is broadly realistic.
152
of x's doing F, in which case we have a standard efficient causal story but no causal role for the
good (a good is correlated with the causation of the process, but not itself involved),32 or the
perpetuation happens as a result of doing F's being good for O, in which case its being good
provides a sui generis sort of contribution to the course of events. Either way the account fails: it
fails to provide a causal role for the good or it fails to make teleology an aspect or kind of
efficient causality.
Given these problems with the proposed reductive account, it is remarkable that neither
of its proponents has provided a more thorough exegesis and defense of the position. The fact
that so little is said in defense of these accounts is, I believe, a further symptom of the traditional
intertwining of questions concerning distinct sorts of 'bases' � epistemological and ontological
� for Aristotle's final causality. When the ontological thread of these accounts is focused upon,
very little of substance remains; the proposed reduction to efficient causality is both textually and
philosophically problematic.
4.2.2.4 Reductions to the material cause 1: the program view.
A number of interpreters (Balme 1987; Bradie and Miller 1984; Matthen 1989; Miller
1995) find a reductive ground for selected portions of Aristotle's use of final causality in a very
specific sort of material cause. The feature of these views which I wish to emphasize33 involves
the idea that at least some portion of Aristotle's use of teleology is grounded in his acceptance of
32 See again the point made in n. 31. 33 Difficulties obtrude in making a definite classification of views concerning Aristotle's teleology. Matthen (1989, 161) explicitly takes Aristotle's teleological causes to be aspects of efficient causes, suggesting that he is mis-categorized here. However, he is inconsistent with this view; the analysis of teleology he offers at the end of his essay (178) abandons the point emphasized at p. 161 that all of Aristotle's material and efficient causes are teleological. I emphasize the analysis, therefore, over the textually unsupported claim that the extent of Aristotle's teleology is universal. Balme's (1987) cybernetic account of teleology is supposed to answer the question of how teleology and necessity are related in Aristotle, but he is quite hesitant and more than a bit unclear concerning his final position on this issue making it hard to know quite where to classify his view � if it should be classified as a settled view at all.
153
the existence in biological organisms of a highly structured low level34 material potentiality that,
like a molecular 'program' or DNA molecule, controls organic growth and development.35 Each
account distinguishes this use of final causality from Aristotle's invocation for final causality as
the end of the process of development � as the mature form of the organisms (see, for example
Bradie and Miller 1984, pp. 136-7) � and none of the accounts tells us anything about the ontic
basis for this latter type of final causality in Aristotle.
This approach therefore attempts a partial defense of Aristotelian teleology. Some set of
his uses of teleological explanation are well-founded and defensible; others are undefended at
best and are at worst indefensible. One clear problem with such an approach will be the fact that
Aristotle does not himself distinguish between these two types of teleology just insofar as they
involve teleology. In his treatment throughout the works the two types of teleological cause (if
indeed they are two distinct types) are treated uniformly. As an interpretation of Aristotle's
views, therefore, this reading succumbs to the obvious problem that while there may be a basis
for reading into Aristotle's text such a distinction between types of teleological cause,36 Aristotle
himself gives no indication of wishing to approach the issue in this way. We may conclude on
this basis that there is insufficient textual support for supposing that this is Aristotle's view of final
causality; we may, of course, be interested in the much weaker thesis that this is a view of final
causality for which we can find antecedents in Aristotle's works.37
34 This phrase is intended to indicate that the sort of material cause we have on these accounts is a micro-cause. These accounts tend to be 'reductionistic' in at least the sense that they ground the existence of macro-phenomena such as teleology in behavior and properties of things' micro-parts and processes. 35 Bradie and Miller (1984) explicitly link Aristotle's material basis for this type of teleology with a DNA molecule: "The type of movement required on Aristotle's account for a potential for form is the type of movement exemplified by the DNA molecule. The genetic 'program' contained in the molecule's structure directs and limits the organism's growth in the manner set forth in Aristotle's biological writings" (143). 36 The clear motivation these authors have for distinguishing types of teleological commitment derives from the belief that the modern account they provide for one of Aristotle's uses of teleology vindicates (at least partially) Aristotle's usage. Aristotle himself gives no indication of feeling that one sort of use is differently grounded than the other. 37 Gotthelf (1997, 80) correctly criticizes the Bradie/Miller interpretation for advancing this interpretation as Aristotle's without any textual basis. Matthen (1989) is most forthcoming in recognizing the point in his
154
Further problems for such an account arise from the fact that such a view of teleology is
currently widely considered to be unacceptable on philosophical grounds. The main problem for
all such views is that for any sort of material structure that is claimed to ground teleology in
biological systems, a materially identical counterpart to that structure can be found in systems
which appear not to be teleological (see Bedau 1992a; 1992d; Nissen 1997, chapter one).38
Given the fact all these theorists abandon vast swaths of Aristotle's use of final causality to
concentrate on this apparently more adequate and modern notion the failure of the modern
account to provide a proper reductive analysis of teleological causation further undermines the
plausibility of our reading it back into Aristotle.39
I have argued in this section that the program approach to reducing Aristotle's final
cause to material causes fails in a number of ways. The interpretation rests on a distinction
between types of teleology among cases where Aristotle himself gives no hint of seeing or
desiring to draw such distinctions. Further, the account is unsatisfactory on philosophical
grounds and therefore cannot rescue even a portion of Aristotle's use of final causality. Again,
when we focus exclusively on the ontological claims made for purported reductive bases for
Aristotle's final causality we see glaring inadequacies in the accounts offered. We have no reason
own work. On his view Aristotle's overall teleology is "fatally flawed" but nevertheless "incorporates a remarkably subtle and prescient use of the four causes" (160). 38 Matthen (1989) is most explicit concerning the analysis of teleology that he claims to find in Aristotle, so the relevance of this point to these Aristotelian accounts is most obvious in his case. On his account, "a series of events is directed towards a certain end if a) it normally culminates in that end, b) it occurs within a substratum so fashioned that a particular action, performed by a particular sort of agent, will ensure that the series of events occurs in its entirety (unless some subsequent event interferes) and c) this series of motion is natural (not fortuitous or forced). The cause of the ordered sequence is form" (178). However, on this definition death would have a teleological explanation, as would a rock rolling down a hill towards the bottom. 39 Oddly, Balme (1987) appears both to see this point and to offer the interpretation nonetheless. He makes the rather surprising claim at the start of his essay that the 'cybernetic model' which he claims Aristotle was 'moving towards' offers no solution to the basic problem of teleology since the cybernetic model shows only that some "apparently teleological processes may in fact be necessary outcomes" (275), showing them thus to be merely (?) apparently teleological.
155
to accept, and many to reject, the program view of final causality as an interpretation of
Aristotle's ontic commitments vis-à-vis teleology.
4.2.2.5 Reductions to the material cause 2: irreducible potentials.
Allan Gotthelf calls his influential interpretation of Aristotle's teleology the "'irreducible
potential' interpretation" (Gotthelf 1976/ 1987).40 On this view to be teleologically directed (in
the basic sense)41 is to be a potentiality (dunamis) � irreducible to lower level material
potentialities � for a developed form. Officially, Gotthelf says:
Aristotle's concept of coming-to-be for the sake of may be defined42 as follows:
A stage in development, A, comes to be for the sake of the mature, functioning organism which results from the development, B, if and only if: (1) A is a necessary (or 'best possible') stage in a continuous change resulting in B, and (2) this change is (in part) the actualization of a potential for B which is not reducible to a sum of actualizations of element-potentials whose identification does not mention the form of B. (213)
According to Gotthelf, what makes this an explanation of teleological directedness is the fact
that "the identity of a nature or potential is given in part by its object or end (i.e. by what it is
irreducibly for)", and reference to a potential for a form that is irreducible to lower level material
40 Gotthelf's view undergoes essentially no development from its original form in his repeated statements of the view, and so my account concentrates on the original formulation. See also Gotthelf (1988; 1989; 1997). 41 Gotthelf believes that 'being for the sake of' is to be analyzed in terms of a more basic form of teleology, 'coming to be for the sake of', and it is that notion that he is concerned to explicate (see Gotthelf 1997, pp.72-3). Gotthelf's account, unlike the accounts of Bradie, Miller, Matthen and Balme (discussed above) claims to cover all the cases of Aristotelian teleology. For criticism of Gotthelf's attempt to analyze all Aristotelian teleology in terms of this one basic notion, see Charles (1991). 42 It will be crucial for us to realize that there is ample and consistent evidence from Gotthelf's writings that he does indeed take this statement to be definitional rather than to be the weaker claim about correlation which is suggested at least once in his writings. Gotthelf writes elsewhere that "I have argued that final causality is operative in nature ... only when there is being actualized a potential for a complex organic outcome which is not ontologically reducible to a sum of the actualizations of potentials of the organism's elemental constituents" (1988, 113). This quick formulation of his view might suggest that he takes his account to provide only a necessary condition for (the basic form of) teleology, or that he means to establish (more strongly) a necessary co-extensiveness between teleology and irreducible potentials for form. There is ample evidence throughout his various writings, however, that he is making the stronger definitional claim with his account. This fact plays an important role in the argument which follows.
156
potentials "puts into the explanans an irreducible reference to an outcome for which the
explanandum is antecedently necessary" (232, emphasis in the original).
The account is clearly reductive, for the analysans appeals to no teleological notions.
The central notions in the analysans, being a potentiality and being irreducible, are both
nonteleological notions and have broader application than to cases of teleology.43 Further,
despite Gotthelf's express wish to disavow teleological reduction to lowest level material causes, his
analysis does count as a form of reduction to the material cause in Aristotle's scheme in virtue of
the fact that the analysans is presented in terms of potentiality, and potentiality is clearly a
material notion on Aristotle's view.44 There can be lower and higher level material potentialities
given Aristotle's hylemorphic view of matter according to which what is formal at one level of
analysis may be material at another.45
This account suffers from both of the two problems with which we are by now familiar:
it has insufficient textual support to qualify as an adequate interpretation of Aristotle and it fails
as a reductive account of final causality. Gotthelf provides far more extensive textual support for
his view than any of the views canvassed thus far. Nevertheless, in evaluating the support he
does provide, it is important to grasp the extremely bold nature of his reductive hypothesis.
According to Gotthelf, "Aristotle's central explanatory concepts are 'nature' (phusis) and
'potential' (dunamis)" (209). The claim that these two principles are ontologically basic46 in
Aristotle's framework is extremely bold, and it is clear that the details of this reductive claim have
43 Gotthelf acknowledges this point when he notes in the analysans that it is only irreducible potentials that are teleological. I argue for this claim concerning the scope of Aristotle's commitment to teleology in chapter three. 44 See Phys. i.9 192a25-33, ii.1 193b7-8; DA i.1 412a9, 412a21, ii.2 414a15; Met. ix.8 1050a15-16, 1050b2, Met. xii.5 1071a8-9, and xiv.4 1092a3-4. 45 See Phys. ii.2 194b9-10; PA i.1 646a12-b10, ii.1 646b11; Meteor. iv.12, 389b27-9 and GA i.1 715a9-11. 46 Gotthelf's claim may be a purely heuristic one: we can understand Aristotle best by taking these notions as the basic notions. If this is what he means, then the thesis is an expository thesis irrelevant to the ontological investigations of this chapter. I presuppose above that it is relevant to the task outlined in the chapter.
157
not been adequately worked out in Gotthelf's work.47 Even concerning the one notion on which
Gotthelf does focus in his reductive scheme � the final cause � there are prima facie puzzles
which Gotthelf's analysis does not even begin to address. Consider the following.
As an interpretation of Aristotle, it is quite odd to suggest that final causality is to be
analyzed in terms of the possession of a potentiality or dunamis � this seems to get the priority
relations between Aristotle's notions exactly backwards. In Aristotle's thought, final causality is
frequently associated with formal causality (see Phys. ii.7 198a25, ii.8 199a30-33; GC ii.9 335b6;
DA ii.4 415b10-12; GA i.1 715a4-6, and 715a8-9), form and matter are frequently contrasted
(DA i.1 412a9, ii.2 414a15; Met. ix.8 1050b2; Phys. ii.1 193b7 and iv.5 213a2), and matter and
potentiality are frequently linked (see Met. ix.8 1050a15, xii.2 1069b19, xii.5 1071a8, xiv.4 1092a3,
OH iv.4 312a16-21). It is unclear how Gotthelf's view is meant to accommodate these facts. My
point here is not that Gotthelf could not reply to these concerns and deepen his account; my
point is precisely that before we could accept Gotthelf's bold reductive thesis we would need to
see such an account worked out in plausible detail.48
Further, Gotthelf's reductive analysis is an inadequate analysis of teleological
commitment. I have mentioned already that the notion of a potentiality is not a teleological
notion, and that the account is therefore a reductive account. Further, adding that a potentiality
is irreducible to some material level also fails to elucidate teleology, for irreducible potentials may
47 It is worthwhile to draw attention to the fact that Cooper, Furley, Irwin and now Gotthelf have all proposed extremely bold reductive theses concerning the relations between Aristotle's central metaphysical concepts. Cooper holds form and matter to be basic, Irwin and Furley derive these qua causes from the efficient cause, and Gotthelf brings them to a supposedly more basic level with reference to nature and potentiality. None of these bold reductive accounts has received sufficient textual defense or close critical scrutiny within the authors' own work. 48 More problems arise. Two strands in a thing's nature in particular require distinguishing. Nature is a cause of movement (see Phys. ii.1 192b21-3, iii.1 200b11, viii.3 253b9; OH i.2 268b15, iii.2 301b18; Met. iv.1 1015a13-9, and ix.8 1049b9). Nature is also the form, structure, or order of a thing (see OH iii.2 301a5, DA ii.4 416a15-8; Phys. vi.10 241a34, viii.1 252a10; GA iii.10 760a32 and iv.2 767a17). The natural move in interpreting these strands of Aristotle's thought is presumably to suppose that nature as cause and nature as form are not identical except numerically � i.e., they are distinct in being. See citations in n. 55 below.
158
clearly be teleological or nonteleological. How, then, does Gotthelf bridge the reductive gap and
explain teleology in these nonteleological terms?
Gotthelf's explanation warrants repetition and analysis. He points to the fact that "the
identity of a nature or potential is given in part by its object or end (i.e. by what it is irreducibly
for)", and reference to a potential for a form which is irreducible to lower level material potentials
"puts into the explanans an irreducible reference to an outcome for which the explanandum is
antecedently necessary" (1976/ 1987, 232, emphasis in the original). Thus, Gotthelf explains
teleology in terms of an aspect of potentialities: potentialities are posterior in definition to
actualities49 and therefore the definition of a potentiality will make reference to what the
potentiality is a potentiality for. But here we have a slip: every potentiality is a potentiality for
some actuality, but the 'for' here carries no teleological connotations; 'for' is in fact ambiguous
between its teleological and nonteleological uses.50 From the fact that a potentiality is the
potentiality it is with reference to some specific actualization we can draw no conclusions about
teleology. Gotthelf's abbreviated and oft-repeated slogan for his view is that to be teleologically
directed is to have an 'irreducible potential for form', and each occurrence of this slogan contains
the seeds of its own refutation. If we read the 'for' in his account teleologically, then we have
neither a reductive account of teleology nor any account at all � sui generis unreduced teleology
of the sort claimed to be ontically grounded in something more familiar has crept into the
analysans. If we read the 'for' in his account nonteleologically, then we have no analysis of
teleology.
Scaltsas has already diagnosed this main problem for Gotthelf's reductive account:
The difficulty I find with this analysis of teleological explanation is that the notion of potentiality is different from, and does not capture or explicate, the notion of something
49 See Met. vii.1 1028a35-6, ix.8 1049b4-16, 1050b3, 1051a3, xii.6 1072a9; DA ii.4 415a18 and NE ix.9 1170a15-19. 50 The sentences 'I went to the store for groceries', 'The chewing is for digestion' and the like are all teleological uses of the word 'for'. Nonteleological uses include: 'The message bears bad tidings for Clinton' and 'Gore's personality is a potential disaster for the Democrats'.
159
being for the sake of something else. Both relations are forward looking, but a potentiality is not generated for the sake of its actuality. Further, the notion of potentiality is too broad for the notion of final causality.... The difficulty that arises is that there is no distinction between the way that element potentials are for their end from the way that what Allan considers teleological potentialities are for their end. The potentiality for being hot is defined in terms of the form of being hot, just as the potentiality for being terrestrial is defined in terms of the form of being terrestrial. It follows that Allan's limitation of the teleological relation to only some of the potential-actual relations--to the non-elemental ones--lacks justification. (Scaltsas 1988, 141)
Scaltsas is correct; whether a potentiality is reducible or irreducible to the potentialities of the
material elements is irrelevant in the following sense to the potentiality's being teleologically
directed towards its fulfillment. All potentials (reducible and irreducible alike) are definitionally
tied with their fulfillment, but this explains nothing one way or another concerning their possible
teleological directedness. Irreducible potentials may be nonteleological,51 and reducible
potentials may be teleological.52
Gotthelf has published a response to this criticism; I present it here in full.
As for Dory's [Scaltsas'] second argument, it is true that elemental natures are specified in terms of outcomes rather than (e.g., as for Newton) in terms of mathematical relations of earlier to late stages. But it remains the case that for Aristotle there is directiveness over and above that. It [is] for this latter directiveness�directiveness over and above that of the element potentials�that Aristotle reserves his teleological language: that much is pre-interpretative fact (Gotthelf 1987a, p.214n19). What distinguishes potentials for form from elemental potentials on my view is just that actualizations of the latter but not actualizations of the former are irreducible to elemental potentials! (1988, appendix, p.138)
Gotthelf is, I believe, correct that Aristotle reserves teleological language for the level of
irreducible potentials,53 but noting this weak claim about correlation is insufficient when our
problem concerns whether or not teleology can be reductively analyzed in terms of irreducible
51 Silberstein and McGeever (1999) argue that quantum mechanics provides instances of emergent properties (roughly, properties which are irreducible to the sum of the lower level components of the thing). There is absolutely no reason such properties must be teleological. 52 It is possible that the elements themselves be teleologically directed. I have argued in chapter three that Aristotle's view is that they are not, but the point here is that the possibility that they be teleological constitutes a counterexample to Gotthelf's proposed analysis. 53 The coextensiveness claim holds, at least, for mortal beings.
160
potentials;54 the point about correlation, even if true, is not strong enough to establish the
definitional claim. This coextensiveness is insufficient to establish that final cause and irreducible
potentiality are the same in being rather than being, more weakly, one in number.55
Gotthelf misses Scaltsas' point entirely in exclaiming that the irreducibility of the
potentials is what makes them teleologically directed. Scaltsas was not unaware of the obvious
point that irreducible potentials differ from reducible potentials in their reducibility; he was
instead making the important point that the correlation between irreducible potentialities and
final causality is insufficient to provide a reductive analysis of final causality given the
nonteleological character of the notion of a potentiality. Given that a thing may, just insofar as it
is a potentiality, be either teleologically directed or not, Gotthelf's attempted reduction fails.56
I conclude that Gotthelf's interpretation of Aristotle's conception of final causality in
terms of irreducible potentials for form cannot work. Not only does the analysis appear
implausible given Aristotle's frequent associations of form and final cause and disassociation of
form and material cause, but Scaltsas has conclusively shown that the notion of a potential is of
the wrong categorial sort to explicate teleology. Irreducible potentials may be associated with an
important kind of teleology in Aristotle, but Aristotle's teleology cannot be analyzed in terms of
potentiality, reducible or irreducible.
54 This is true on at least some conceptions of reduction. Perhaps Gotthelf accepts a weaker view of the requirements for reduction. Even if this is the case, there will remain a perfectly coherent sense in which the point I make in the text � i.e., that his analysis fails � is defensible. On notions of reduction, see the appendix to chapter six of this dissertation. 55 For the distinction between spatial or other coincidence and difference in being or definition, see Top. i.5 133b31-6; Phys. i.3 186a30-1, iii.3 202b9-10, iii.3 202b20-1, iv.2 219a20-1, 219b18-9, v.5 229a19, 229a28, vii.8 262a20, vii.8 263b7-8, 263b12-4; OH ii.14 296b8 ff., iv.4 312a19-21, iv.5 312a30-2; GC i.5 320b13-4, 320b23-5; DA iii.4 429a10-1, iii.7 431a13-4, 431a28, iii.9 432a20, iii.10 433b24; PN 449a14-9, 467b26-7; Met. vii.8 1033b31, viii.6 1045b18-9, xii.10 1075b4-5 and xiii.2 1077b1-4. See also APr. 67b12, Top. 133b33, DA 431a9. 56 Ironically, Gotthelf makes this same sort of criticism of the Bradie/ Miller view: "If there is a regular direction upon an end, guaranteed by an intrinsic efficient cause of a material sort, this seems only to mimic teleological causation; what we need is that the intrinsic cause's activity that leads to the end occurs because it leads to that end" (Gotthelf 1997, 81). Gotthelf misses the application of this criticism to his own view: intrinsic efficient/material causes, whether reducible or irreducible to element potentials, only mimic teleology. They cannot analyze it.
161
4.2.3 Realist accounts 2: nonreductive accounts.
[T]he notion of 'for the sake of' seems to be taken as basic throughout and not further explained. (Charles 1991, 116)
For [Aristotle] nature as such is not endowed with intelligence, nor is there any outside demiurge or world soul or creator to do the directing. The all-pervading direction of natural activities towards a purposive goal seems accepted merely as a fact, on the ground of immediate observation and deduction. (Owens 1968)57
I have argued negatively in the previous sections that none of the proposed ontic bases
for teleological causation in the secondary literature presents a view that is faithful to Aristotle's
texts, and further that none appears to offer a defensible reduction of final causality to a more
basic (and hopefully more acceptable) sort of causal factor. In this section I argue that we are
most faithful to Aristotle's texts if we take him to invoke a sui generis form of causation in his
appeals to final causes; on Aristotle's view our apparent commitment in teleological explanations
to a unique kind of cause is not merely apparent, it is real.
4.2.3.1 A caveat on nonreductive 'accounts'.
A note about accounts is perhaps in order before we begin. On the interpretation I offer,
Aristotle holds that final causality is a real thing, irreducibly distinct from other kinds of real
causal factors in the world. The ontological 'account', then, will simply state the following claim:
final causality is not 'something else'; it is a primitive sui generis form of causal factor in the world.
If we were to take a complete inventory of Aristotle's commitments to basic causal factors in the
world, then on my interpretation we would find final causality on the final list as a basic notion
57 Owens here expresses a thesis concerning the scope of Aristotelian teleology against which I argue in chapter three. The quote further refers to the epistemic grounds Aristotle had for accepting the final cause; as it stands I accept roughly these same grounds (see Owens 1968, 163 and chapter three of this dissertation). The quote is included above because it also expresses precisely the ontological claim that this chapter endorses.
162
along with efficient, final, and material causality.58 Now, problems may appear to arise for this
view given that such 'accounts' are often derided as 'uninformative' or claimed simply to 'label a
mystery' or 'multiply entities beyond necessity'. It is important to see why these charges are both
unfounded and unfair.
The charge that such accounts must be uninformative is unfair because this ontological
claim (that F things constitute a primitive, irreducible kind or way of being) is fully compatible
with the existence of further informative nonreductive analyses of the notion at stake.59 Perhaps
the most familiar instance of this sort of informative nonreductive analysis is the interdefinability
of possibility and necessity; modal notions are, as is frequently noted, not explicable 'outside the
circle of modal terms', but this does not entail that they are not explicable simpliciter. They are
explained by examples plus analyses displaying essential relations between the properties that
make up the 'circle'.60 The same point holds concerning physics itself.
Physical theories do not derive the existence of these features [i.e., matter, motion, space, and time] from anything more basic, but they still give substantial, detailed accounts of these features and of how they interrelate. (Chalmers 1996, 213)
There is no principled objection to nonreductive accounts just insofar as they are nonreductive.
We have a satisfactory ontological account of the status of the final cause insofar as our
questions concern reduction and elimination; we ought not fault our answer to this question for
not answering other questions about the analysis of final causality.
58 I have not, of course, argued that Aristotle takes all of these notions to indicate sui generis real causal factors in the world. Nothing much hangs on the claim here; my point is simply that whatever else appears on the list, final causality will. 59 Alvin Plantinga offers one such nonreductive account of function statements in his (1993). I offer another below in chapter seven. 60 Arguing on empiricist assumptions, Quine famously claimed that synonymy and analyticity cannot be noncircularly defined and are for this reason bogus. Jerrold Katz has argued in a number of papers and books that Quine's argument fails to account for the possibility that synonymy and analyticity could informatively be defined within a circle of semantic terms such as antinomy, synonymy, redundancy, ambiguity, meaningfulness, meaninglessness and analyticity. See his (1982; 1990; 1992).
163
The charges that in claiming that the final cause is sui generis we are simply labeling a
mystery or multiplying entities beyond necessity are similarly unfair. There is nothing a priori
more or less coherent about a view of the world on which it turns out to be monistic or
pluralistic in the basic kinds of entities and relations it posses. Those who seek monistic
ontological accounts of the world have no principled ontological61 basis for objecting to there
being more basic kinds of things in the world than they suppose: an ontology two basic kinds is
no more mysterious than one with one kind, and so on.
Whether we are multiplying entities beyond necessity is another issue, one not dealt with
in this chapter. The answer to this question will depend on the strength of our justificatory
reasons for postulating such sui generis entities. In chapter three I argued that Aristotle has
reasonable and moderate grounds for his postulation of final causality. In chapters five and six
below I argue that reasons of the same basic sort are defensible in a contemporary context. But
these questions have nothing to do, per se, with the question we have been addressing in this
chapter. We have been concerned narrowly to investigate the ontic status of the final cause in
Aristotle. Our investigation seems thus far to point to Aristotle's accepting the final cause as a
sui generis real causal factor in the world. This position is unobjectionable just insofar as it is a
type of ontic pluralism as opposed to monism. Of course, there are special scientific and
philosophical concerns with this specific type of pluralism. These concerns constitute the focus
of chapters five and six of this dissertation.
4.2.3.2 Aristotle's commitment to sui generis final causality.
With these worries behind us, one premise in the argument in favor of attributing to
Aristotle the view that the final cause is a sui generis real causal factor in the world comes from the
61 One might suppose that there were principled methodological or epistemological reasons for accepting monism, but this is incorrect. The goal of science is truth, and if there exists more than one kind of basic thing, then we want to recognize that fact in our final theories. We might seek the simplest theory consistent with the facts, but it will be the facts themselves which determine which theory is simplest relative to them. For an especially deep and forthright discussion of these issues, see C.D. Broad (1918-19).
164
poverty of the reductive accounts we have surveyed. None received adequate textual support
(some were positively impoverished in this regard), and none indicated a philosophically
defensible analysis.
Another premise is supplied by the fact that Aristotle's practice seems positively to treat
each of the causes on a par with the others; none seems on educated and scholarly surveys to be
held by Aristotle to be more basic than any of the others.62 As David Balme says,
Aristotle always presents the four causes as four separate factors in a causal situation (Phys. ii.3). They are not one factor plus three alternative descriptions or views of it.... [I]n Aristotle's usage ... causes are objective things and events. Therefore if the efficient cause is one objective feature in nature, so too is the final cause another one. (1987, 281)
While it is epistemically possible63 that this claim could be proven false, this appearance raises a
strong prima facie presumption against any reductive analysis. If Balme's claim is correct, then
there is a good reason that, to paraphrase Furley, these interpreters need to offer Aristotle some
words before we can understand his view.
Neither of the above two considerations can, of course, be conclusive given the
epistemic possibility that some interpreter may develop an adequate reductive account of
Aristotle's commitment to final causality, whether in response to arguments of this chapter or
independently. I encourage such focused work on the ontological question; it has been one of
my theses in this chapter, after all, that the unsettled state of debate over the ontological status of
Aristotle's final cause is in large part due to the lack of focused attention to the ontological
question.
Nevertheless, there is a further premise in the argument for taking Aristotle's notion of a
final cause to be that of a sui generis real causal factor in the world. The reason is simply that
62 Of course many other scholars claim to be able to find and support bold reductivist claims. Nevertheless, the plurality of reductive schemes on offer and the relative poverty of the textual support offered for them must make such attempts to reduce Aristotle's pluralism about basic factors to a more acceptable number dubious. See n. 47, above. 63 Epistemic possibility as I use it indicates roughly 'imaginability'. Imaginability is neither necessary nor sufficient for genuine metaphysical possibility; see Yablo (1993).
165
Aristotle expresses his own view of the matter quite clearly in criticizing his opponents. Aristotle
says,
That for the sake of which actions and changes and movements take place, they assert to be a cause in a way, but not in this way, i.e. not in the way in which it is its nature to be a cause. For those who speak of reason or friendship class these causes as goods; they do not speak, however, as if anything that exists either existed or came into being for the sake of these [ou mên ôs heneka ge toutôn], but as if movements started from these. In the same way those who say the One or the existent is the good, say that it is the cause of substance, but not that substance either is or comes to be for the sake of this. Therefore it turns out that in a sense they both say and do not say the good is a cause; for they do not call it a cause qua good but only incidentally [ou gar haplôs alla kata sumbebêkos legousin]. (Met. i.7 988b6-15, emphasis added to the ROT)
[I]n all things the good is in the highest degree a principle. The school we first mentioned is right in saying that it is a principle, but how the good is a principle they do not say � whether as end or as mover or as form. Empedocles also has a paradoxical view; for he identifies the good with love. But this is a principle both as mover (for it brings things together) and as matter (for it is part of the mixture). Now even if it happens that the same thing is a principle both as matter and as mover, still being them is not the same. In which respect then is love a principle? (Met. xii.10 1075a37-b8)
Past students of nature, however, took the opposite view. The reason for this is that they did not see that the causes were numerous, but only saw the material and efficient and did not distinguish even these, while they made no inquiry at all into the formal and final causes. (GA v.1 778b7-10)
Democritus, however, neglecting the final cause, reduces to necessity all the operations of nature. Now they are necessary, it is true, but yet they are for a final cause and for the sake of what is best in each case. Thus nothing prevents the teeth from being formed and being shed in this way; but it is not on account of these causes but on account of the end; these are causes in the sense of being the moving and efficient instruments and the material. (GA v.8 789b1-15, in part.)
In each of these passages Aristotle expresses in unambiguous terms that on his conception, but
not on the conception of his predecessors, there is a peculiar form of causality, final causality,
which cannot be accounted for in independent terms. His predecessors fail to talk about the
cause 'qua good'; the forms of causation they appeal to 'differ in being' from the final cause;64
they failed to notice that the causes were numerous; it is not on account of the types of causes they
cite, but on account of the end that certain things happen. It is unsurprising that Aristotle's
theoretical commitments should thus show themselves in his discussions of the failures of his
166
predecessor's views, and the existence of such passages constitutes extremely strong positive
evidence for attributing to Aristotle a belief in the sui generis irreducible ontic status of the final
cause.
4.3 Conclusion
In this chapter I have argued that Aristotle's conception of the final cause is the notion
of a sui generis, irreducible real factor in the causal structure of the world. Aristotle is committed
to teleology, and he is committed to the view that teleology is not 'something else', something
more palatable, but simply is what it is � primitive directedness upon an object.
My argument has invoked positive theoretical claims concerning the relationship
between the causes Aristotle accepts (Met. i.7 988b6-15, xii.10 1075a37-b8 and GA v.8 789b1-15,
quoted above), reference to Aristotle's uniform nonreductivist practice, and objections to such
reductivist accounts as exist. I have also argued that the conclusion of this chapter is more
readily apparent than it may appear from a casual survey of the variety of reductivist views on
offer; it is more readily apparent, that is, once the ontological issue is separated off from
questions concerning Aristotle's justification for accepting teleology and questions concerning
the scope of Aristotelian teleology. When the views surveyed above stand alone as accounts of
the ontic relations between the various causes in Aristotle, they reveal themselves to be both
philosophically implausible and textually unsupported.
This result may be unwelcome to those who wish to find in Aristotle a philosophically
defensible view of the ontic status of the final cause.65 The acceptance of sui generis teleology in
64 See the citations in n. 55. 65 This need not be the final word on the worth of Aristotle's invocation of teleology, however, even if one accepts that sui generis teleology is dead in the water. Aristotle's other doctrines concerning teleology, including the justification he offers for accepting teleology as a part of the natural world, his theses on the scope of teleology, and his discussions of the ends of particular creatures and activities may all have great value even if his idea of the ontology of final causality is hopeless. I argue in chapters five and six that Aristotle's conception of the final cause as ontically sui generis and irreducible has surprising philosophical plausibility and defensibility given the status of contemporary realist reductivist accounts of teleology and the independent support available for an ontology of the physical world in which sui generis teleology finds a natural home. Even those who fail to accept these arguments may take heart in noting that the ontic
167
nature coincided historically with centuries of scientific stagnation, and the outright rejection of
sui generis teleology ushered in an as yet unfinished period of astounding success in the sciences.66
These facts may appear too glaring to be coincidences, and most have certainly drawn the
conclusion that it was sui generis teleology which held science back.67
In the next two chapters I call this modern consensus of opposition sui generis
Aristotelian teleology into question. In chapter five I note first that there are strong
contemporary scientific grounds for being realist concerning teleology. I then survey
contemporary reductivist accounts of this realistic commitment and find them wanting. I argue
in chapter six that the standard modern objections to teleology can all be overcome, and that our
grounds for realism about teleology thus survive the failure of reductivist accounts. We have
strong scientific evidence for the actuality of a physical ontology in which sui generis teleology has
a natural home. If these arguments are correct, then Aristotle's sui generis conception of the final
cause has much more promise than the consensus view allows.
status of Aristotle's commitment to teleology does not eviscerate his other interesting theses concerning teleology. 66 It is important to note that this bold claim is, as it stands, untrue. One highly successful science � biology � has never abandoned teleology except, occasionally, in name. Biology has, of course, thrived despite this seeming handicap. Most scientists and philosophers would claim that this is because biological teleology can be given (must be able to be given) a reductive analysis, but this claim about the availability of adequate reductive analyses is an independent philosophical issue from the bare fact that biologists have continued to find the invocation of teleology essential to their work. My point is the narrow one that biology has both thrived and accepted teleology. The success of reductive analyses of teleology is in question in chapter five. 67 Joseph Owens expresses the sentiment nicely. "In his [Aristotle's] wake the project of calling on an internal teleology directed towards substantial form, in order to explain the specific traits and functions of natural things, was tried long enough through the centuries, and without success. It should serve as a sufficient warning against expecting philosophy to do any of the work of the experimental sciences." (1968, 172)
168
Chapter 5
Teleological Reductionism: Is there a need for Aristotelian Teleology?
Contemporary analyses of teleological explanation generally attempt to 'sanitize' it, usually by trying to assimilate it to some uncontroversial descriptive form of explanation. This trend is misguided. (Bedau 1992d)
I argued in the previous chapter that Aristotle's notion of a final cause is the notion of a
sui generis real causal factor in the world; teleology (as Aristotle conceives it) is not reducible to or
fully explicable in terms of any other cause � material, formal, or efficient. Stepping back from
this scholarly project, one main thesis of this dissertation is that Aristotle's account of life (which
is, by hypothesis, teleological) is of contemporary interest. Despite teleology's rehabilitation in
contemporary philosophy of biology1 however, the philosophical presuppositions of Aristotle's
account will be unacceptable to modern philosophers and scientists if it employs � as I have
claimed it does � irreducible sui generis final causes. Given my acceptance of the findings of the
best modern biological accounts of living things and my argument thus far, the following
dilemma for this interpretation appears sound: either Aristotle was right about the teleological
character of life but wrong about the nature of teleology, or right about the nature of teleology
but wrong about the nature of life.
My goal in this chapter and the next is to argue that the foregoing dilemma is a false one.
Despite the longstanding and well known objections to the scientific credentials of robust
teleology, I intend to argue in this chapter and the next that Aristotle's conception of the final
cause as a sui generis real causal factor in the world is the correct understanding of teleological
causation. I will argue further that sui generis teleological causation is both philosophically and
1 See especially Buller (1999b), Nissen (1997) and the papers collected in Buller (1999a) and Allen, Bekoff, and Lauder (1998b).
169
scientifically acceptable given a proper understanding of the methodological and ontological
commitments of modern science.
The existence of widely accepted naturalistic and reductionistic accounts of teleology
counts strongly against my claims, however, for if teleology can be accounted for without the
radical revisioning of our ontology that Aristotelianism would involve, we may feel strongly
inclined toward the ontically more parsimonious worldview. In this chapter I argue that the
most promising reductivist accounts of teleology fail. The next chapter, then, is devoted to
picking up the pieces, for if reductivistic accounts fail longstanding philosophical presuppositions
about the ontology of the physical world are likely to kick in and push us to conclude that the
cost of accepting teleology is simply too high. I intend to argue in chapter six that, properly
understood, the acceptance of sui generis Aristotelian final causes is scientifically and ontologically
respectable in light of the achievements of modern science.
5.1 The modern problem of teleology.
As David Buller writes, "one of philosophy's foremost metaphysical problems has been
whether, and if so how, teleology is possible in nature" (1999b, 3). The project of the next
chapter is largely devoted to offering a solution to this metaphysical puzzle; this chapter attempts
to motivate the investigation offered there through a focus on the recent history of philosophical
and biological accounts of teleology.
Contemporary work has focused on the most controversial form of supposed
teleological causation, biological teleology. It may seem easy enough to understand how agents
act teleologically: their belief-desire psychology explains their goal directed behavior. It may also
seem obvious how artifacts and found objects gain their functions derivatively through agents'
intentions: the key's function is to open the lock because that is what it was designed (by an
170
agent) to do, etc.2 If all cases of teleology were of these two types teleology would perhaps be
no more (and no less) problematic than other issues in the philosophy of mind.
Biological functions, however, present new difficulties. Biological functions such as the
heart's function of pumping blood, the eye's function of sight and innumerable other cases from
the biological world are enigmatic. How can we explain single celled organism's seeking food
(something which they certainly seem to do) without resorting either to a divine mind designing
such behaviors or to an implausible panpsychism that attributes to such creatures minds and
their attendant beliefs and desires?3
The central role of teleological notions in biology has led to a tendency � which I will
follow � to focus the general metaphysical puzzle concerning the possibility of teleological
explanation into a question in the philosophy of science such that "What we principally want to
understand", in Karen Neander's words, "is how the biological notion of 'proper function' can be
both teleological and scientifically respectable" (1991b, 124).
5.1.1 Eliminitivism or realism?
In light of such puzzles, one might consider resorting to an eliminitivism about
biological teleology, explaining our use of teleological talk in the biological sciences as a
reflection of our tendency to take a selective interest in the multitudinous effects organs have.
However, such a solution to our problems is fraught with difficulties stemming from the
patterns, persistence, and apparent explanatory power of our functional attributions.4 Bigelow
and Pargetter emphasize that patterns within function attributions do not map interests; we are
interested in things which have no functions (in physics and chemistry), and find functions in
things which fail to interest us. Further, we believe that things would have had functions even if
2 For accounts of the difficulties involved in this seemingly straightforward idea, see Bigelow and Pargetter (1987, 101-2), Griffiths (1993, §§7-8). 3 See Bedau (1990) for a convincing case against mentalism in teleology. See Nissen (1997) for a defense.
171
there had been no one around to take interest in the structures that have them (1987, §II).
Various authors appeal to the persistence of functional attribution: despite hundreds of years of
philosophical antagonism to teleology, biological science has made little effort to remove
teleological language from its explanations and reflective biologists are positively resistant to the
idea that teleology should be eliminated from their field.5 Finally, it is the explanatory power of
teleological explanations which seems to account for this persistence. Karen Neander is perhaps
most emphatic on this point:
the apparent explanatory power of teleological explanations which appeal to biological functions is quite robust.... That the bee's dance is for directing other bees to pollen does seem to explain why bees dance. I suppose it is just barely possible, perhaps, that this apparent explanatory power is illusory.... However the thesis that we are persistently irrational in this respect is psychologically implausible in contrast to a theory of functions that shows such explanations to be legitimate. (1991b, 127)
Neander calls teleology the 'conceptual glue' of biology (1991b, 137) and notes that it would be
"hard to exaggerate" the concept's importance to biology (1995, 227).6
Recent history in the philosophy of biology has taken this data seriously, and has
undertaken the challenge of understanding biological teleology without resort to mentalism,
panpsychism or eliminitivism.7 We may thus note a condition of adequacy that the main
contemporary approaches to biological teleology have attempted to satisfy: an adequate account
4 See Bedau (1990): "even if a theory of teleology does not take natural teleology at face value, it should provide some sort of explanation of why organic nature appears teleological" (66). 5 David Buller says, "In spite of the difficulties associated with teleology, however, biologists continued to use the teleological concept of function in describing the characteristics of organisms, finding the organization of organisms and the operation of their parts virtually incomprehensible in strictly non-teleological terms" (1999b, 6. Emphasis added). As David Hull puts it, "Teleological systems do seem to force themselves on us" (1974, 120). Andrew Woodfield says, "Although some biological function-statements may seem metaphorical, some seem to be literally false (e.g. 'Noses exist in order to support spectacles'), and some seem literally true.... Whatever the correct analysis of these TDs [teleological descriptions] may be, there is a strong presumption that they make at least some objective claim on reality" (1976, 32, see also p.1). See also Nissen (1997, vii), Allen and Bekoff (1995, 244) and the quotes to follow by Karen Neander and Fred Dretske. 6 See also Bigelow and Pargetter (1987, 100), Allen, Bekoff, and Lauder (1998a, 1-2), Buller (1999b, 6). 7 For exceptions to this trend, see Nissen (1997), who argues that only a mentalistic account of teleology respects the data, and Cummins (1975/1984/1998), who claims (but does not argue) that teleology is
172
of biological teleology should reflect realism about teleology; teleology is not something we
'spread' onto the world, but is a real feature of the world that we discover through investigation.
Condition one: the realist stance. It is a condition of adequacy on an account of biological functions that the account respect the apparent commitment in common sense and science's to teleology as an objective mind- and language-independent feature of the world. This commitment reveals itself in the patterns, persistence, and explanatory power of teleological explanations.8
The greater part of this chapter will be occupied with addressing theorists who accept this
condition of adequacy on roughly these grounds.
5.1.2 Aristotelianism?
One obvious solution to the problem of biological teleology would be to claim that we
live in an Aristotelian world where final causes play genuine causal roles. As Mark Bedau writes,
there would be non-mental biological teleology in a hypothetical non-mental world in which, as a matter of some sort of natural law, creatures typically have features because of the goodness they confer. (1992d, 285)9
In such a world, biological teleology would be scientifically respectable in a straightforward
manner. In broad outlines, this is the solution I offer to the problem of teleology as stated
above;10 however, there is nothing straightforward about making sense of this view given
prevailing views about the ontology of the physical world and methodological concerns
incoherent and offers an eliminitivist account of function ascriptions. Cummins' view is described in more detail below. 8 Fred Dretske says, "We are accustomed to hearing about biological functions for various bodily organs. The heart, the kidneys, and the pituitary gland, we are told, have functions�things they are, in this sense, supposed to do. The fact that these organs are supposed to do these things, the fact that they have these functions, is quite independent of what we think they are supposed to do. Biologists discovered these functions, they didn't invent or assign them. We cannot, by agreeing among ourselves, change the functions of these organs" (1988, 63). See also the citations and quotes above in n.5. 9 There is nothing incoherent in the notion of such a world. As Richard Spencer-Smith says, "It is empirically possible that, for instance, in the atoms of all and only living things, force #5 [≠ any of the four fundamental forces postulated by physics] is discovered.... The discovery would be that there are forces which do not unify with the forces of physics, and in this the natural world is greater, in ontology, than the physical world.... [T]here is no more of a mystery [to this] than that of the brute fact of nature's containing N distinct fundamental forces." (1994, 118-119). 10 The details of the account, developed in chapter six, attempt to make this solution plausible. The goal of this chapter is to make room in contemporary philosophical discussion for such an heterodox view.
173
surrounding its investigation. Acknowledging this fact sufficiently will occupy both this chapter
and the next.
As is well known, longstanding objections to teleology in modern science and
philosophy11 have made the Aristotelian solution to our problem unattractive. Thus, the
problem of understanding the possibility of teleology � already narrowed to the problem of
determining how biological teleology can be scientifically respectable � is focused even further
in contemporary discussion:
The reductivists project: "How can the biological concept of function, which is prima facie infected with final causation, be analyzed so as to make it compatible with a scientific world view that countenances only efficient12 causation?" (Buller 1999b, 6)
It is in the guise of the reductivist's project that the contemporary literature on biological
teleology has approached the problem of teleology, and it is this narrow formulation of the
problem that has motivated the effort (noted in an epigraph to this chapter) to 'sanitize'
biological teleology of Aristotelian taint. In this sense, the common reductivist goal of each of
the popular modern accounts of teleology is well stated by Ernest Nagel: the goal is "to analyze
the concepts of being a goal and being goal-directed so that the analytic versions of these
concepts... contain no undefined teleological notions" (1979, 291).13
Since modern science accepts only efficient causation all our teleological talk must
ultimately commit us only to an ontology of efficient causality. We can therefore formulate a
condition of adequacy under which contemporary analyses have labored.
11 These objections are surveyed and evaluated in chapter six. 12 I will not be concerned in this essay to address the vexed issue of the precise nature of 'efficient' causality. The contemporary usage (which is not equivalent to its use in Aristotle) seems to equate it roughly with Humean 'billiard ball' causation and, by extension, to whatever other sorts of causes the sciences of physics and chemistry (at least) eventually admit. 13 Eliott Sober says, "An interesting feature of all extant philosophical accounts of what the concept of function means is that they are naturalistic" (2000, 87); he endorses this trend: "Function is a concept", he says, "that should not be taken at face value" (2000, 87). Mark Bedau (1992d), Alvin Plantinga (1993) and Lowell Nissen (1997) offer accounts which falsify Sober's generalization, but Sober correctly identifies the main trend in modern accounts.
174
Condition two: efficient causality. An adequate account of teleology may make an appeal to no factors in the world incompatible with the modern scientific view of the world as governed solely by efficient causality.14
My main intention in this section has been to remind the reader that the reductivist's project,
accepting as it does the efficient causality condition, is not the only obvious � or even the most
obvious � solution to the problem of teleology; sui generis approaches are not a priori incoherent.
Whether the reductivist approach can succeed is under investigation in this chapter; whether the
Aristotelian approach can succeed is the topic of chapter six.
5.1.3 Precursors to the standard line.
Excellent surveys of the contemporary debates concerning teleology already exist and
would at any rate be beyond the scope of the current project.15 Here I will present only the
barest outlines of the contemporary controversy, hoping to reveal the broad contours of the
debate; I engage with the analytical details only where this is unavoidable.
The 'cybernetic' or 'systems' approach to teleology attempts to explain teleology through
the internal structures of complex systems.16 Some mechanical and biological systems contain
'feedback' mechanisms that give them a tendency to maintain a particular state in the face of a
limited range of environmental changes. Systems analysts take such states to be the 'goals' of the
systems, and attempt to specify the exact requirements systems must meet in order to be genuine
goal-directed feedback mechanisms.
14 See also Godfrey-Smith: "The concept of function was bequeathed to post-Darwinian science from an earlier conceptual scheme.... But the categories we recognize now should be determined, of course, by our own worldview" (1993, 194). Karen Neander says, "Any theory which delivers a fully-fledged teleological notion of biological function, consistent with the tenets of modern science, has a plus in its favor" (1991b, 127). 15 For book-length historically-oriented treatments of the issue, see especially Woodfield (1976) as well as the more up-to-date Nissen (1997). The classic papers by Wright (1973) and Cummins (1975/1984/1998) both offer excellent treatments of earlier views. Buller (1999b) is a very good, current and compact historical overview, and the series of papers on teleology by Mark Bedau (1990; 1991; 1992a; 1992b; 1992d) is extremely helpful. 16 This approach to teleology was first developed by Rosenblueth, Wiener, and Bigelow (1943). See Bedau (1992a, 49n3) for further references.
175
Such analyses become extremely complex very quickly because counterexamples to the
analyses are easy to formulate. A bowl with a marble in it forms a complex system which
maintains a 'goal' state (in the cyberneticist's sense) in the face of perturbations, but it is as clear
as can be that 'staying at the base of the bowl' is not the marble's goal (in the sense we are
interested in); its staying at the base is a mere result of the physics of the situation. As the
systems analyses become more complex, so (inevitably) must the counterexamples; nevertheless,
we need not pursue the details here for Mark Bedau has argued convincingly that all such
analyses are in principle subject to counterexamples of the bowl and marble sort (1992a; 1992b,
30). The moral of the story is that "a system's [intrinsic] causal dynamics do not by themselves
determine whether the system is genuinely goal directed" (Bedau 1992d, 265).
Given this problem, the most popular response in the contemporary literature has been
to follow Larry Wright's lead and offer an etiological analysis of biological functions. On Wright's
view, it is not the intrinsic causal dynamics of a system which gives its function, it is rather the
details of the item's causal ancestry which determines its function. On his analysis,
The function of X is Z means
(2) (a) X is there because it does Z,
(b) Z is a consequence (or result) of X's being there.
According to this analysis, a thing's function explains why it is present � its being present is an
efficient causal effect of its doing whatever is its function. Thus, a heart's function (i.e., to pump
blood) explains why hearts are present in humans (i.e., humans have hearts because in their
evolutionary history ancestral hearts were selected for their ability to pump blood), and pumping
blood is a result of hearts' being in humans. The analysis seems, therefore, to add a feature �
i.e., reference to causal ancestry � which eliminates counterexamples to analyses offered in
terms only of intrinsic features of things.
Wright's analysis itself succumbs to counterexamples, however. The first of many
versions of such counterexamples was developed by Christopher Boorse (1976). Boorse argued
176
that nonfunctional accidental features of systems could meet the conditions of Wright's analysis.
Thus, a break in a hose17 which allows chlorine gas into a room is there because it lets gas into
the room (i.e., the gas knocked out the person who would have repaired the hole had she not
breathed in the gas), and letting gas into the room is a consequence of the hole's being there.
But, the function of this accidental break is not to let gas into the room, despite the fact that it
meets the conditions of Wright's analysis.
5.2 The etiological account of functions.
The contemporary response to the liberality of Wright's account has been to modify the
original analysis such that an item must have a particular kind of causal ancestry in order to have
a genuine ('proper') function; roughly, items with proper functions must have a history of selection.
In common usage it is this later development of Wright's view which has taken over the name
'the etiological account of functions', and I will follow this usage here. As Karen Neander says,
"Roughly speaking... the proper function of a trait is to do whatever it was selected for" (1991b,
124). More precisely, Paul Griffiths notes that "the central element of the etiological approach,
[is] the idea that the effects of past tokens of a type provide an explanation for the existence of
current tokens of that type" (1993, 148).
In Neander's formal account:
It is a/the proper function of an item (X) of an organism (O) to do that which items of X's type did to contribute to the inclusive fitness of O's ancestors and which caused the genotype, of which X is the phenotypic expression (or which may be X itself where X is the genotype) to increase proportionally in the gene pool. (1991a, 319)
Thus,
17 It is important to note that on Wright's analysis the origin of the thing with a function may be perfectly accidental. The hose accidentally has a leak, and its doing the things it does leads it to satisfying the requirements of his analysis. A gene on an organism randomly mutates; its doing what it does then leads to the phenotypic effect satisfying the requirements of his analysis. The origin of the item with the function can be, and usually is in biological cases, purely accidental and Wright's analysis allows this.
177
the function of your opposable thumb is to assist in grasping objects, because it is this effect which opposable thumbs contributed to the inclusive fitness of your ancestors, which caused the underlying genotype, of which opposable thumbs are the phenotypic expression, to be selected. In brief, grasping objects was what the trait, the opposable thumb, was selected for, and that is why it is the function of your thumb to help you grasp objects. (Neander 1991b, 130)
Accounts of functions along these lines (with variations in detail) have become the 'standard line'
in the philosophy of biology,18 and analyses of these sorts have been hailed by enthusiasts as
being of great moment in the history of philosophy.
Despite the disagreement ... there is a common core of agreement that unites etiological theorists ... and this core of agreement represents as great a consensus as has been achieved in philosophy. (Buller 1999b, 20)
[T]he achievement of a near consensus concerning how to understand the biological concept of function certainly represents progress in a field notorious for failing to produce any. (Buller 1999b, 26)
[S]ince 1973 there has been a convergence towards a view of functions which has Wright's idea at its core. I think of this trend as an example of real progress in philosophy. (Godfrey-Smith 1993, 185)
The details of the basic analysis are still being debated adherents,19 but most contemporary
philosophers and biologists20 appear to agree that the etiological account developed beyond
Wright's original formulation to include reference to a history of natural selection contains � at
last � a workable, modern, scientific account of teleology. This account is intended to respect
both science's commitment to explanation exclusively in terms of efficient causality and the
condition imposed by the realist stance.21
18 Allen and Bekoff use the phrase 'the standard line' to describe such accounts (1995; 1995). 19 See Buller (1999b) for a nice summary of issues remaining to be hammered out, and the articles collected in Buller (1999a) and Allen, Bekoff, and Lauder (1998b) for the details of such disagreements. 20 This appears to be true, at least, of those philosophers and biologists contributing to the literature on the subject. There are voices of dissent: see Prior (1985), Nissen (1997), Plantinga(1993), Searle (1995) and, more ambiguously, Bedau (1992d). 21 Buller (1999b, 18-9) offers a concise summary of how the account addresses traditional objections to teleology.
178
5.3 Objections to the etiological analysis.
Despite the enthusiasm of proponents, the etiological theory of functions suffers from
serious � I believe fatal � flaws. The most general symptom of the account's failure is the fact
that the account generates a disunity in our pretheoretic � even our tutored and scientific
pretheoretic � understanding of the phenomena of biological functions; this disunity seems to
be an artifact of theory rather than a reflection of genuine types of functions. This basic
symptom displays itself along a number of dimensions, surveyed below. The cause of this
failure, I argue, is that the etiological account fails to obey categorial constraints on the analysis
of biological functions. Given that theorists are engaged in the project of solving the traditional
problem of biological teleology and do not take themselves only to be explicating a purely
technical notion labeled 'proper function', the etiological account fails because it analyses a
nonhistorical notion (i.e., biological function) in essentially historical terms (i.e., contributions to
natural selection).
In the sections that follow I first lay out the general form of the argument that the
etiological account generates an artificial disunity in the phenomena of biological functions. I
then explain a number of particular dimensions along which this argument may be developed.
As the objection has been presented in the literature to date it fails to be decisive against the
etiological account; etiological theorists have correctly responded to weak versions of the general
argument that they seek to give a theoretical definition � not a conceptual analyses � of
biological function. This response fails, however, against a strengthened version of the argument
from disunity which invokes legitimate constraints placed on theoretical definitions (at least
where those theoretical definitions claim to match up in significant ways with historical
philosophical and biological problems). According to this stronger version of the argument the
etiological account fails to provide either a necessary or a sufficient condition for being a
biological function and the theoretical definitions response is disarmed.
179
I close the chapter by noting one final objection to the etiological account. Etiological
theorists are realists about biological functions; they accept the data of biological functions
largely as they appear in common sense and science. Given this, etiological theorists have felt
constrained to acknowledge and naturalistically explain the normativity of biological functions:
hearts are supposed to pump blood and they malfunction if they fail to do so properly. I argue that the
etiological account fails to naturalistically explain this feature of biological functions.
The overall conclusion of the argument, then, is this. Insofar as the etiological theorists
(and others) are correct that the persistence, patterns and apparent explanatory power of
attributions of biological function demand the realist stance the failure of the etiological account
(and all other modern reductive accounts) leaves us in a position to reconsider commitment to
sui generis teleology. In this situation, where our reductive analyses continue to fall short but all
our motivations for realism survive, it is time to take a closer look at the philosophical and
scientific presuppositions driving the search for a reductive account. It is to this project that I
turn in chapter six.
5.3.1 The general form of the uniformity objection.
We may formulate a number of distinct objections to the etiological account of
biological functions, all of which take the same basic form. According to all the objections, there
is a unity in the phenomena of biological functions that is distored by the etiological account.
Depending on the type of unity we have in mind, the distortion may be quite mild, consisting
merely in the narrowing the extension of biological functions, or it may be more severe � such
as postulating distinct types of functions solely on the grounds of theory or ruling common
conceptions of biological functions analytically false. Nevertheless, all versions of the objection
seek to undermine the etiological account on the basis of its necessitating revisions (whether
drastic or minor) in our practices of attributing biological functions, which revisions are artifacts
of theory rather than features of the phenomenon.
180
5.3.1.1 Three weak uniformity objections.
In this section I develop the general objection just discussed along three dimensions of
uniformity in the phenomena of biological functions. Biological function attributions are
uniform across the following three dimensions.22
Uniformity across examples of functions. There is no difference qua function attribution between claiming that the function of the heart is to pump blood and claiming that a function of the heart is to contribute to survival and reproduction.
Diachronic unity in the understanding of function attributions. There is no difference qua function attribution between a pre-Darwinian's claim that the heart's function is to pump blood and a post-Darwinian's claim that the heart's function is to pump blood.
Synchronic unity in the understanding of function attributions. There is no difference qua function attributions between a 'scientific' creationist's claiming that the function of the heart is to pump blood and an evolutionist's claiming that the heart's function is to pump blood.
Uniformity objections may be raised against the etiological account of functions along each of
these dimensions.
First, while the etiological account shines in its application to a number of standard
examples of functions such as the heart's function of pumping blood, there exists a type of
function attribution that pervades our functional talk about the biological world but with
reference to which the etiological account appears vacuous. It is standard to suppose that some
of the functions of organs generally involve their promoting survival and reproduction but the
etiological account does not informatively apply to these functions.23 Thus, choosing a single
example, we might try to account for the thumb's function of promoting survival and
reproduction along the following lines:
* A nonproximate function of the opposable thumb is the promotion of survival and reproduction because it was this effect that the trait (nonproximately) contributed to the inclusive fitness (survival?) of our ancestors and which
22 In the next section I raise a 'fourth dimension' of uniformity which to some extent cross categorizes with types (1) and (3) listed here. The examples serve to illustrate the cases. 23 The objection is raised by Cummins (1975/1984/1998, 72).
181
caused the underlying genotype of which the thumb is the phenotypic expression to proportionally increase in the population (reproduction?).
The problem with (*) is not that it is false, but that it appears very nearly vacuous. The
etiological account is persuasive along wide ranges of cases precisely because it has explanatory
and theoretically interesting content across those cases; in the pervasive cases of the functions of
survival and reproduction, however, the account comes up stale.
Second, the etiological account does not capture uniformity of historical use. Lowell
Nissen remarks on this fact in a number of places. As he says,
[F]or any account of functions to be of general interest and not a mere private excursion, it must connect with existing subject matter and existing terminology, in this case, the subject matter and terminology that have been around for a very long time. (1997, 186. See also pp.57 & 110-11.)
Nissen quotes five other theorists remarking on this history of consistent usage (110-11); the
point of citing such a number of authorities is, presumably, not as the basis for an argument
from authority but to reveal the strong presumption in favor of this widely accepted insight's
being correct. Intelligent inquirers looking into the history of functional attributions find no
discernable shift in import over the centuries. Nevertheless, if the etiological account of
functions is correct, pre-Darwinians were employing a concept whose analysis involved reference
to Darwin's theory of natural selection. This claim may appear highly implausible.
Finally, function attributions are uniform in their import synchronically between groups
with very different commitments concerning the natural world. Thus, functional attributions
seem to have the same import as they are used by evolutionists and 'scientific' creationists.
Contemporary creationists who invoke function in the service of design arguments for the
existence of God mean to indicate the same phenomena indicated by function attributions as
employed by contemporary evolutionary biologists. Functional attributions, then, appear to have
uniform purport across the divide between scientists (i.e., evolutionary biologists) and lay people
or pseudoscientists (i.e., creation 'scientists'). Again, on the etiological account creationists are
employing function attributions to organisms conceived not to have evolved, but this is (on the
182
account) incoherent. Again, while no disunity in import appears in the data, the etiological
account generates disunity.
5.3.1.2 The etiological theorist's response.
The obvious reply to each of these uniformity objections is the same, and it is employed
(in slightly different guises) in two important essays on the etiological account of functions.24
The etiological theorists may reply to each of the above objections by distinguishing between
two possible goals of giving an account of functions. If one's goal is conceptual analysis, then
one's analysis must accommodate what it is that people who assign functions have in mind in
attributing functions. If one's goal, however, is not conceptual analysis but to seek a 'theoretical
definition', then one needs to give an account of the real nature of the phenomenon picked out
by our language and concepts.
With this distinction in hand,25 the etiological theorist may claim that uniformity of use
along each of these dimensions is accounted for by a continuity in the notion of a function at the
level of conceptual analysis. Nevertheless, the etiological theorist maintains that the theoretical
definition of biological function reveals the underlying real nature of the phenomena in question
to be evolutionary at heart. That pre-Darwinians or creation scientists fail to understand their
uses along the lines indicated in the etiological account is no failure in the account; one need not,
after all, know the theoretically perspicuous account of one's data to speak intelligibly. Similarly,
the etiological theorist may simply be willing to acknowledge that her account does not map
perfectly onto the extension of our preanalytic uses of function talk, but claim that it
nevertheless captures the core cases magnificently. The etiological theorist may therefore
recommend that we accept this minor revision in extension in light of the gain we make in a
theoretical understanding of the central cases.
24 Millikan (1989) and Neander (1991a) offer versions of this reply which have important differences. 25 See also chapter one for another discussion of theoretical definitions and conceptual analysis.
183
I believe that this response adequately addresses the worries raised thus far.26 If there
were no further objections to the etiological account of functions then we might rest content
that modern science and philosophy had finally achieved a lasting and satisfying solution to the
problem of biological teleology � a solution that meets both the realist and the efficient
causality constraints on an adequate account.
Unfortunately, a more troubling disunity in the phenomenon is produced as an artifact
of accepting the etiological account, a sort of disunity that cannot be addressed along the lines of
the theoretical definitions defense; I address this more troubling disunity in the next section.
Further, the general uniformity objection may be strengthened, and I argue below that the
uniformity objection as elucidated in response to the theoretical definitions reply reveals the fatal
flaw in the etiological account. With this strengthened objection in hand, it will become
apparent that the etiological account of functions provides neither a necessary nor a sufficient
condition for possessing a biological function.
5.3.1.3 Uniformity among the experts.
The key to the theoretical definition response to the uniformity objection lies in
distinguishing between the concepts people in general (whether scientists or lay people) employ
when they think about functions and what the experts have discovered about the natural
26 We might suppose that the argument as stated fails for the following reason. If the etiological account's definition of biological function is correct, then it turns out that creation 'scientists' are not merely wrong in making their claims, but in fact their claims are analytically false since the analysis of those terms makes essential use of terms whose meaning is evolutionary despite the fact that the surface meaning of the creationist's sentences denies evolution. The objection, then, is that the etiological account incorrectly rules that 'scientific' creationism is analytically false; but however lopsided the scientific issue is in favor of evolutionists, it cannot turn out that creationism is analytically false. To suppose that it was would be to suppose that a substantive disagreement over scientific theory could be settled by analytic investigation of the meanings of our terms.
This response fails, however. The etiological theorist need neither deny that there is a sense in which the creationist's statements are analytically false nor deny that in the only important sense the issue between creationists and evolutionists is an empirical scientific issue. The theoretical definition of functions reveals the statements to be analytically false, but the scientific evidence in favor of adopting that theoretical definition as correct is the scientific evidence on which the debate between evolutionists and creationists hangs. The scientific evidence still plays the pivotal role in determining which side is correct; there is no 'easy' analytical victory for the evolutionists.
184
phenomena picked out by our concept. We may have an adequate conceptual analysis of
phlogiston; this analysis will describe the various ideas scientists associated with the concept.
But we will never have a theoretical definition because the phenomenon phlogiston was
postulated to account for turned out to require no such thing in its account: there is no
phlogiston to be described in our finished theory of the phenomenon.
This distinction cannot help the etiological theorist, however, if there exists a case where
function attributions (1) are made by experts in different subfields in biology in the course of their
work as experts, (2) such attributions carry the same import, and (3) there is a strong presumption
that one group of scientists' use makes no tacit reference to evolutionary histories. In such a case,
we would have highly successful scientific uses of function revealing no distinction in their
import except when we understand one use along etiological lines � something we cannot do for
the other use. This possibility is in fact actual.
The case I have in mind is famous in the contemporary literature on biological functions
and was raised against the etiological account most prominently by Elizabeth Prior in her (1985).
One of Prior's objections to the etiological account of functions is that it emphasizes only the
attribution of functions by evolutionary theorists. But other, older, sciences discover and appeal
to functions without any reference to evolutionary theory. Prior objects,
[A] statement like 'The function of the heart is to pump blood' is accepted by us as true because physiology tells us that it is true. It is not based simply on unschooled intuition. Rather, it has standing behind it in the sciences of physiology and anatomy. If we take the etiological account seriously then it follows that any claim of function supplied by physiology will be subject to overthrowal by the appropriate evolutionary evidence. (317)
Prior argues that this result is unacceptable.27 Physiology discovers functions and has
autonomous authority in its field of expertise.28 Further, physiological attributions of function
27 In fact Prior argues something stronger, namely that evolutionary accounts of functions themselves depend on physiology's prior discovery of functions. See pp.320-1.
185
make no reference to evolutionary history. When this uniformity objection is raised against the
etiological account the defender cannot reply with the theoretical definitions response for both
users of functional talk in this instance are respected scientific experts but one's use carries no
evolutionary import even at the level of theoretical definition.
Instead of offering the theoretical definitions response to this challenge, then, etiological
theorists have tended to accept the authority of physiology and anatomy29 and have instead
declared that what this shows is that we ought to be pluralists about kinds of functions: there are
etiological functions analyzed in terms of evolutionary analysis and there are physiological
functions.30 It is essential to note that it is respect for the independent results of other branches
of biological science � physiology and anatomy � that motivates this pluralism about
functions.
The first thing to notice concerning this strategy is that the plurality of kinds of
functions we have here is entirely an artifact of the etiological account of functions. There is no
preanalytic distinction to be drawn between the expert attributions of function as practiced by
evolutionists and by physiologists; the motive for postulating a plurality of types of teleology
here is parochial to etiological theorists and entirely theory-driven. This fact ought to count
heavily against the etiological account of functions; insofar as the etiological account set out to
tackle the traditional problem of biological teleology it set out to handle a phenomenon
uniformly invoked by physiology and in broader biological theory both before and after Darwin.
One natural response to noting that the etiological account fails to capture the notion as
28 Richard Dawkins is blunt about this; he says, "any engineer can recognize an object that has been designed, even poorly designed, for a purpose, and he can usually work out what that purpose is just by looking at the structure of the object" (1986/1996, 21). 29 The statement that etiological theorists respect the authority of physiologists is not quite accurate, as I argue below. On the standard etiological development of pluralism, physiological functions are eliminated ontologically and accounted for as artifacts of our contingent interest in particular features of the world. This is decidedly not the respect paid to the evolutionist's employment of functions, an employment which generates a realist account. See below. 30 See Millikan (1989), Godfrey-Smith (1993), and Allen and Bekoff (1995; 1995).
186
employed by physiologists would be to conclude that the etiological account cannot for that very
reason be the correct account of functions. To instead declare a plurality of types of functions
appears an ad hoc move taken to defend a parochial theory of functions rather than a reasonable
response to the phenomena under investigation.31
A second point worth considering about the pluralists' response has to do with the
motive for moving to this pluralism. Etiological theorists claim to be respecting the data of
another science � physiology � in postulating a plurality of types of functions; indeed, this
motive appears to express the etiological theorist's consistent desire to respect the persistence,
patterns, and apparent explanatory value of function attributions. The question arises, however:
if the physiologist's employment of biological functions is not accounted for along etiological
lines, then how are we to account for that use (consistent of course with the reductivist project)?
The pluralist's standard answer has been to account for the second type of functions
along the lines of the analysis given by Robert Cummins. My last objection to the etiological
approach in this section will be the ad hominem point that Cummins' analysis is entirely unsuited
to the etiological theorist's own commitment to the realist stance and its underlying motivation
� respect for the persistence, patterns and apparent explanatory power of biological functions.
The reason that Cummins' analysis is unsuited to the etiological theorist's needs is that
Cummins is an eliminitivist about biological teleology. He says,
Function-ascribing statements do occur in explanations having a teleological form, and, when they do, their interest is vitiated by the incoherence of that form of explanation. It is the legitimate use of function-ascribing statements that needs examination, i.e., their contribution to nonteleological theories such as the theory of evolution. (1975/1984/1998, 81n7)
31 This is not to imply that there are no precedents for a move such as the etiological theorist makes. Thus, jade, once thought to form a natural kind, is now distinguished into jade and jadeite. This plurality of types mapped no preanalytic distinction but is instead recognized on the basis of sound theoretical work. The etiologist may, of course, invoke such parallels as a response to this objection. Nevertheless, the disunity deserves our attention given the plurality of symptoms of analytic failure we are accumulating; only a healthy and thriving theory can ignore problems such as this, and the brunt of this chapter argues that the etiological account is nowhere near so well founded as its proponents would perhaps like.
187
Genuinely teleological explanations are incoherent in Cummins' view;32 scientifically legitimate
function statements have no teleological taint. On the Cummins account functions in their
legitimate employment are capacities had by parts of systems to contribute to the workings of
the wholes of which they are parts. Thus, my heart's function is whatever it contributes to the
working of the larger system(s) which contain(s) it (the cardiovascular system, the entire body, or
the body and the computer I'm typing on, etc.). Following what has become standard usage, I
will call the 'functions' discovered through such functional analyses of the causal contributions of
sub-systems to containing systems 'Cummins functions'.
In addition to the eliminitivism of Cummins' approach, the second central feature of his
account for my purposes � a feature noted by Cummins (1975/1984/1998,.§4) and numerous
commentators33 � is its wild liberality. Cummins functions can be found for anything for which
one can dream up (a) a system of which it forms a part and (b) a role � no matter how
recondite that role may be � for the part to play in that system.34 Stones in riverbeds have the
Cummins function of contributing turbulence to the waters. The dust on my side table has the
Cummins function of cushioning my flower pots, etc., etc. The key point is that Cummins
functions are hopeless as specifications of the sense of 'function' we set out to investigate under
the heading of the problem of teleology. We certainly engage in functional analyses of the sort
Cummins describes,35 but such analyses do not explain the use of 'function' we are interested in
when we speak of biological functions.36 In the sense we are interested in, hearts and eyes have
functions, but dust and stones in riverbeds do not. Cummins deals with the liberality of his
32 Whether or not he (and many other moderns) are correct in this assessment is the subject of chapter six. 33 See Buller (1999b, 15), Millikan (1989, 120), Griffiths (1993, 145-6), Kitcher (1993, 173), Godfrey-Smith (1993, 195-6), Neander (1995, 229-30n3), Walsh and Ariew (1996, 259). 34 Elliott Sober says, "There is an extremely minimal interpretation of function, explication by Cummins (1975), according to which everything has a function" (1985). 35 For an eloquent (as usual) statement of their use, see Dawkins (1986/1996, pp.11-13). 36 See my discussion of the patterns, persistence and explanatory power of talk concerning biological functions, above p.170, for the defense of this claim.
188
account by claiming that biological functions constitute just that subset of the Cummins
functions of organs in which we happen to take an interest.
The important point for this ad hominem argument against pluralists, then, is that their
'pluralism' injects an unheralded element of eliminitivism and antirealism into what was until this
point a steadfastly realist account. Indeed, it is highly misleading to call the account 'pluralist'
about functions at all;37 better to say that the etiologists have accounted for some of the uses of
function realistically and have chosen to eliminate all the rest. In substance this approach to
saving the phenomenon of biological teleology falls short of the respect its rhetoric promises to
pay to function attributions from independent biological subdisciplines.
I have argued in this section that at least one form of the uniformity objection against
the etiological account of functions has real teeth. Faced with the fact that physiologists employ
functions without even a hidden reference to evolutionary theory most etiological theorists feel
driven to a 'pluralism' about functions that at best injects an artificial disunity into an apparently
unified phenomenon and at worst abandons the project of seeking a realist account of a vast
swath of theoretically important function attributions.
5.3.2 The core problem with the etiological account.
I argued above that the etiological account of functions has the resources to successfully
reply to a first round of objections stemming from uniformities in uses of functional language
not captured by the etiological account. At the core of this reply is the distinction between
offering a theoretical definition and a conceptual analysis of biological function: if our goal were
conceptual analysis, then the threat from historical discontinuity (for instance) would be great,
but given that we seek to give a theoretical definition � a scientific account of the real
37 The etiological theorist's pluralism is no more a pluralism about functions than I am a pluralist about types of humans when I endorse the claim that some people are nonmagical but that others are witches when what I actually believe is that there really are no witches but that we are sometimes interested in calling nonwitches 'witches'.
189
phenomenon in the world corresponding to our use of functional concepts � the supposed
problem is dissolved.
In this section I argue that the theoretical definitions response fails. We may pose the
problem as a dilemma. We may conceive of the project of giving theoretical definitions for key
terms in science in one of two ways. If we understand the etiological account to provide a type
one or stipulative theoretical definition, then the account is irrelevant to the traditional problem of
teleology that the etiological theorist set out to solve.38 If we understand the etiological account
to provide a type two or minimally constrained theoretical definition, then the account is relevant
but fails as a definition of functions.
Merely stipulative theoretical definitions are proposed as an account of entities and
relations previously unheard of under the sun; such theoretical definitions earn their keep solely
by their usefulness to the science in question. With this type of account, theorists have total
freedom to propose any concept for our acceptance; there are no constraints on such definitions
qua definitions, for they are entirely stipulative. A second type of theoretical definition is
proposed as an explication of entities, properties or relations with common but problematic or
interesting employments. In these cases such definitions are judged for their adequacy not only
by their usefulness to science, but also by certain elements of their faithfulness to traditional use;
the traditional employment of such terms puts constraints on proposed theoretical definitions of
the second sort.
This last point bears clarification. The point is not that theoretical definitions that are
meant to have ties to past usage must capture everything in past usage precisely; the newly
proposed theoretical definition may not capture the exact extension of the original notion; it may
38 Alvin Plantinga makes this point against Millikan's assertion that she is not trying to give a conceptual analysis but merely giving a theoretical definition of 'Proper Function'. Plantinga says, "Strictly speaking, of course, if her aim is not to give an account of the common notion of function, then her project is not as it stands directly relevant to the question at issue: the question whether there is available a naturalistic understanding or analysis of proper function" (1993, 201). See also Lowell Nissen (1997, 186), quoted above.
190
be more precise than the original conception; it may be different in many ways. The point is that
when a theoretical definition purports to tie into past usage and to solve familiar problems it is
constrained in some ways by previous use.
We may dramatize this point by offering up the following definition of biological
function:
X is the biological function of Y =df X is to the left of Y.
On the first conception of theoretical definition, there is no objection to my proposing this
theoretical definition; if the relation being to the left of does useful work, then so will the new
technical notion of biological function. If I propose this as a theoretical definition of biological
function, however, and at the same time claim that this definition gives us the long sought after
solution to the problem of biological teleology (the problem that has exercised scientists and
philosophers since Aristotle), then my definition is inadequate; it strays too far39 from traditional
use to provide a helpful resolution to the problem of teleology.
My claim is that the etiological account of functions fails because it purports to be a
theoretical definition of the second sort, of the sort that ties in with traditional usage and
attempts to solve traditional problems, but nevertheless strays 'too far' from traditional use to
provide the correct analysis. With this background distinction drawn, the argument for the
second horn of the dilemma proceeds as follows.40
The paradigm case of a theoretical definition tied to past use is the discovery that water
is H2O. This definition is connected properly with past use in that while before the discovery of
the molecular nature of water people did not know that water's chemical composition was H2O
they nevertheless believed it to be a stuff amenable to having its nature discovered through
39 What it is for an account to 'stray too far' from traditional usage will in some cases be unclear. Below I spell out one precisification of this notion as it applies to the etiological account of biological function. 40 See George Bealer (1987) for a detailed version of this argument in relation to scientific essentialism. Richard Spencer-Smith argues along similar lines that qualia cannot be reduced to any physical property; see (1994, 128).
191
scientific investigation.41 Water can have a type two theoretical definition in terms of the kind of
stuff it is as revealed by scientific investigation because it has always been conceived as the kind
of thing amenable to analysis into composing materials. We could not do the same kind of
scientific analysis and come to discover that food is XYZ, however. The reason is that 'food'
does not pick out a stuff kind as 'water' does; instead, 'food' seems to be a functional kind � x is
food iff x nourishes.42 Food cannot be given a type two theoretical definition as XYZ because it
is simply the wrong kind of thing to turn out upon investigation to be defined as a physical
natural kind.43
These examples illustrate the nature of at least one of the limits beyond which a
theoretical definition that purports to tie in with past use may not stray; type two theoretical
definitions must satisfy categorial constraints on the notions they attempt to analyze.44 In
chapter one I introduced this idea with the categorial constraint on proper definitions.
The categorial constraint on proper definitions: Roughly, all definitions of F must minimally capture features of F which place it in fundamental ontological categories.
41 David Chalmers says, "At a rough approximation, we might say that the primary intension [of 'water'] picks out the dominant clear, drinkable liquid in the oceans and lakes; or more briefly, that it picks out the watery stuff in a world" (1996, 57). Primary intensions are intended by Chalmers to specify "how reference depends on the way the external world turns out" (57). The citation functions in this context as independent evidence that what Chalmers calls our 'a priori' conception of water, the conception we have of it before empirical investigation, specifies a stuff open to physical analysis. 42 This argument does not hang on the details of this analysis of food. The point is that food is not a stuff kind whose chemical nature can be discovered through scientific investigation; the details of the positive analysis are not important here. 43 We might of course come to discover that there is a contingent correlation between things which serve as food and a particular natural kind XYZ � this would be surprising, but possible. This possibility does not undermine the claim that food is not identical to the natural kind; food would be composed of or correlated with XYZ, but these two relations are weaker than identity and insufficient to give us the identity relation given in the paradigm type two theoretical definition where water is identical to H2O. 44 Another sort of limit on theoretical definitions, even less easy to formulate with precision, may be that where X proposes to be the reductive analysis of Y, it cannot be the case that X applies clearly and straightforwardly to paradigm cases of things we would ordinarily consider non-Y. This appears to get at the kind of failure indicated by the earlier counterexamples to the systems approach to teleology and Wright's original etiological account of functions. The counterexamples are accepted as defeating the proposed analyses presumably because the extension of the analysans so clearly extends to paradigmatic nonteleological systems. The etiological account by contrast was allowed to have vacuous extension to some attributions of teleology because those uses were not considered 'central' or 'paradigm'. The metaphilosophical principles which guide these decisions are, however, difficult to elucidate clearly.
192
While we may allow that theoretical definitions of everyday notions may fail to include every aspect of our preanalytic notion of an F, there are limits to what even a theoretical definition may revise out of our ordinary conception.
The point applies here. Water is the appropriate kind of thing � it belongs to the right category
(i.e., stuff) � to have a chemical kind as its type two theoretical definition. Food is not of the
appropriate kind to turn out to have a chemical kind such as XYZ as its type two theoretical
definition because it is a functionally defined kind rather than a stuff kind. Categorial features of
properties constrain possible theoretical accounts of those properties that purport to tie to past
use. This why the relation biological function of could not turn out to be analyzed by to the left of �
one is categorially a spatial relation, the other is not.
Applied to the etiological account of functions the point is this. On the etiological
account, the function of x is (roughly) that effect of x in the organism for which x was favored
in the process of natural selection. The analysans here is essentially historical. Let us say that a
concept is an historical concept just in case a proper analytic account of the concept makes
ineliminable reference to past events as necessary for the current applicability of the concept.45
The etiological account makes the relation biological function of essentially historical: the analysans
makes ineliminable reference to past history. But, the analysandum is not and never was
categorially an historical relation in the way that 'ancestor', and 'predecessor' are. Given this, it
can no more turn out to be the case upon analysis that the traditional notion of function is an
historical notion than it can turn out after analysis that food is identical to XYZ or that the
number four weighs eight pounds. Nor, of course, can it turn out to be the case upon analysis
that my grandfather's being an ancestor of mine can be revealed to have nothing to do with our
relative places in a temporal sequence.
Karen Neander points to the nub of this particular issue. She notes that conceptual
analysis is necessary to "delimit the scope" of theoretical definitions (1991a, 318), and she is
193
absolutely correct. The point is that theoretical definitions of commonplace terms (as in the
cases of 'function', 'water', and 'food') must be tied to past use; specifically, past and present use
reveal categorial constraints on adequate analyses.46
Examples of such categorial constraints on accounts can easily be multiplied. Thus,
Berkeley was just wrong (and brazenly so) in claiming that his idealism captured common sense
beliefs about the world; the common conception of a physical object simply has no room for
physical objects to turn out to be ideas.47 The number two could not turn out to be taller than
Michael Jordan. Ideas do not sleep, furiously or otherwise. And the notion of a function is not a
historical notion. Past use constrains possible contemporary analyses which seek to account for
that use and the problems it has generated; this task is more ambitious than offering stipulative
type one theoretical definitions which eliminate or ignore old concepts, replacing them with
newer, better notions. Intuitions about the categorial status of concepts place firm constraints
upon adequate theoretical definitions of the second sort (i.e., the sort that attempts to tie in with
past use).
Various arguments may persuade one that biological function of is not an historical relation.
That function is not an historical notion was pointed out by David Hull twenty-five years ago.
(Hull's comment comes temporally prior to more recent enthusiasm for historical accounts. His
judgment is extremely revealing � partially for this very reason.) Hull said,
45 Richard Healey defines being an historical concept as follows: "A historical property of an object is a property that object has at a time simply in virtue of what happens to it at some other time. Having been driven 30,000 miles is a historical property of my car." (1991) 46 Alvin Plantinga offers a version of this criticism to David Lewis' reductionistic account of modality. Lewis accepts (from common sense) that there are propositions, but claims that their nature is not revealed except under analysis. Under analysis these propositions turn out, he claims, to be sets. Plantinga responds, "Clearly this is partly right: there is much about the nature of propositions we don't pretheoretically know.... But we do know something about the nature of propositions, prior to theory.... Conceivably they could turn out to be idealized sentences or divine thoughts; but they couldn't turn out to be just anything�donkeys or fleas, or tables, for example" (1987, 207-8). 47 Of course that does not rule out there being no physical things and all the things of our acquaintance which we take to be physical turning out to actually be ideas. To learn this would be to have eliminated physical objects from our ontology, not (as Berkeley claimed) to preserve them as ideas.
194
Wimsatt's analysis is confronted by. . . . the need to justify his interpretation of 'teleological systems' as an historical concept. At the very least, the adoption of his analysis requires a modification of current usage, because currently teleological systems are identified without any necessary reference to their origins. (1974, 113)
Hull here notes that the (then) rising trend of giving historical analyses of functions is a proposed
shift in use, a shift that requires justification. No such justification is provided, and Hull
diagnoses the root mistake being made by etiological theorists once this purported shift is
brought to light.
Hull notes that there is a reason that we might be confused into thinking that being the
function of could be analyzed through reference to natural selection; he says, "In point of fact, all
those systems we view as being teleological arise, directly or indirectly, through selection
processes" (113). This correlation claim is correct: having a biological function is indeed
sufficient for having a history of selection, at least in this world.48 This claim is insufficient to
ground the strong claim of the etiological theorist, however, that biological functions are to be
identified with selected effects. As another early critic notes, "We know why evolutionary
biologists are interested in effects contributing to an organism's capacity to maintain its species,
but why call them functions?" (Cummins 1975/1984/1998, 72) The etiological account confuses
a weak correlation between function and selection for the basis of an analysis of being a
function, but this proposed analysans fails; the correlation is too weak and the analysans imports
reference to history into an analysandum that is not historical.
Another sign that the identification fails displays itself when we return to the types of
uniformity functional attributions display across time and between divergent groups.49 There is
strong reason to believe that the notion of a biological function is not an historical notion arising
from the fact that for the greater part of Western history working biologists have employed the
48 If creationism is possible, as it surely is, then this correlation does not extend across all possible worlds. 49 As I mentioned above, this central criticism can be seen as a strengthening of the original uniformities objection, generated in response to the etiological theorist's theoretical definitions reply to those objections.
195
notion of a biological function but have been creationists or have at least thought creationism
possible. If one believes that organisms are created with functions, then the notion one is
employing cannot be one whose theoretical definition involves essential reference to history of
selection; to suppose otherwise is to suppose that all working biologists before Darwin were as
confused about functions as they would have been had they thought being a function of could be
analyzed by to the left of. We must suppose wild categorial errors on an enormous scale to begin
to grasp the failure of comprehension envisioned for the greater part of the history of human
attributions of biological function.
Further, one need not be a creationist or even appeal to history to have the intuition that
biological function is not an historical notion.
[W]hile it might even be a law of nature that all teleological systems and processes result from some species of 'blind variation and selective retention', we at least conceive it to be logically possible that teleological explanations are appropriate to the behaviour of, and purposes are possessed by beings or systems which have never felt the molding force of, these two processes. (Wimsatt 1972, 15. Quoted in Nissen 1997, 93)
As Larry Wright says,
[T]his seems to suggest that it is impossible by the very nature of the concepts � logically impossible � that organismic structures and processes get their functions by conscious intervention (design) of a Divine Creator. This, I think, is analytical arrogance. I am, personally, certain that the evolutionary account is the correct one. But I do not think this can be determined by conceptual analysis: it is not a matter of logic. (Wright 1976, 96-7. Quoted in Nissen 1997, 140)50
These authors point to a modal intuition that is widely shared; history of selection is not necessary
for the possession of biological function. This is because the notion of a biological function
which has grounded the historical problem of teleology is not an historical notion.
It is important to keep the dialectic of my argument in mind at this stage for the
etiological theorist may want to reassert the point that her definition is not intended to offer a
50 See also Bigelow and Pargetter (1987, 248-9) and Plantinga (1993, 208).
196
conceptual analysis of ordinary uses of terms51 but a theoretical definition of the reality behind
the uses.52 This reply is not available to the etiological theorist at this stage in the argument.
I have argued as follows. Given that the etiological theorist intends her account to be
tied to the past usage of biological function that has generated the problem of biological
teleology the etiological account cannot provide the correct account of functions because its
analysans makes essential reference to history of selection whereas the analysandum is not an
historical relation. The core of the problem with the etiological account lies in a mistaken
analogy between paradigm cases of theoretical definition (i.e., water =df H2O) and the case
currently in question (i.e., function =df (roughly) naturally selected effects). While water's having
always been conceived to be (categorially) a stuff explains how its can turn out upon analysis to
be identical to H2O, the notion of a biological function was never (categorially) an historical
notion and so it cannot turn out upon analysis to be historical any more than it could turn out be
identical to to the left of, or than food can turn out to be identical to a stuff, or than physical
objects could turn out to be identical to ideas, or... 53 The epistemological basis of this error lies
in taking a factual correlation between (a) possessing a biological function and (b) having an
evolutionary history for an analysis of biological function. Despite the correlation, biological
function is not analyzed in terms of evolutionary history. Further, I have supported the claim
that the notion of a biological function is not an historical notion through appeal to uniformities
in usage and intuitions concerning possibilities which give strong warrant to the claim that
having a history is not necessary for having a function.
51 Robert Brandon says, "I can't see the interest in an account indifferent between divine creation and Darwinian evolution" (1981, 79). 52 Or, a conceptual analysis of a particular subgroup's (evolutionary biologist's) use of the term. See Neander (1991a, 323). 53 Again, if what we standardly think of as physical objects turned out to be ideas then, pace Berkeley, physical objects would have been eliminated from our ontology and replaced with ideas; physical objects would not (could not) be ideas.
197
Given this argument, the etiological theorist cannot respond by simply reiterating that
she is seeking to offer a theoretical definition. Either her theoretical definition is of type one
(i.e., stipulative) and irrelevant, or it is of type two and it fails because it fails to obey categorial
constraints on proper type two definitions. Insofar as the etiological account of biological
functions purports to be a reductivist solution to the traditional problem of biological teleology it
fails.
5.3.3 Replies: direct and indirect.
While no etiological theorist has explicitly replied to the above line of argumentation,54
two types of responses appear available from a survey of the contemporary literature. First, the
direct response argues that in cases where history is absent we are at a loss concerning what to
say about function. Second, the indirect reply points to similarities between paradigm cases of
functions: both artifact functions and biological functions are 'selected'. This gives rise to an
argument that it is history of selection which in both cases gives rise to functions. I argue in this
section that neither of these replies can succeed.
5.3.3.1 The direct response: no history, no function.
On the direct response to my argument, etiological theorists may argue that it is not at
all clear that organisms without evolutionary histories have parts with functions. Karen Neander
asks us to imagine a world where there are no lions until suddenly, "half a dozen lions pop into
existence" (1991a, 325). These turn out to be strange lions (relative to our experience), for we are
told that they have "wing-like protuberances" on their flanks about which we ask, 'What is their
function?' The lions cannot fly, but we know that other organisms have malfunctioning parts or
vestigial organs, and so we cannot be sure that their function is not or was not flight. Neander
concludes on the basis of this case that,
54 To the best of my knowledge this argument has never been formulated with reference to the etiological account of functions.
198
we could not reliably place them in any category until we knew or could infer the lions' history. And if we were to somehow discover that the lions had no history, and were the result of an accidental and freak collision of atoms, they would definitely not belong in any of our familiar functional categories. . . . Without history the usual biological/functional norms do not apply. (1991a, 326)
Despite the prima facie intuitive plausibility of the idea that instant organisms would have parts
with functions, Neander thus argues that evolutionary history is a necessary condition for the
possession of "the usual biological/functional norms."
Ruth Millikan makes a similar claim about an instantly occurring physical duplicate of
you.
[Y]our randomly created double exhibits no purposive behaviors and has no purposive parts because there is no way that any of his/her states or parts could be defective or might fail. That creature of accident ... falls under no norms. (1989, 304)
Thus, each etiological theorist argues that, despite initial appearances to the contrary, deep
reflection on cases reveals that function really is an historical notion. Contrary to our untutored
modal intuitions (appealed to above), our considered intuitions about cases involving instantly
occurring organisms reveal that that such organisms lack functions because they lack
evolutionary history.
The problem with the etiological theorists' assessments of these cases is that they
depend for their plausibility either on highly selective choices of examples or solely on bravado.
Neander's strategy is to employ a highly selective choice of examples: wings on instant lions are
certainly 'hard cases'; we are mystified as to what such parts' functions might be.55 And she is
clearly right in at least one sense (although wrong in the sense that is at stake) that "biological
norms are determined by the history of traits" (1991a, 327, see also 330). This is correct at least
to the following extent: biological functions may in some cases be epistemically available to us only
55 Indeed, the first time I sat down to seriously read the paper where this example is used, it took me a number of re-readings of the passage to convince myself that I was not simply reading a typographical error or misunderstanding her intent.
199
through investigation of evolutionary histories.56 Nevertheless, even granting this
epistemological point modal intuition tells us that biological function does not depend on
evolutionary history for its existence (as opposed to discovery). For as soon as we turn from
Neander's highly peculiar case to the lions' more 'standard' parts, we judge (again) that the parts
of the instant lions would have functions. Thus, the lions' eyes would have a function which we
could determine with great assurance: sight. The lions' ears would have a function which, again,
we could determine with great assurance: the apprehension of sound. The point generalizes.
Alvin Plantinga notes:
Whether or not God did create Adam and Eve instantaneously out of the dust of the earth, he could have; and if he had, wouldn't Adam's heart have had a function � the same function served by your heart and mine?.... [T]he proposed necessary and sufficient condition is clearly not necessary. (1993, 203)
The list goes on and on;57 for innumerable parts, the functions of those parts is available to us
quite clearly even in the case of the 'instant duplicate', and this reveals, pace Neander, that the
notion of function is not a historical notion.58
The upshot is that, although following out evolutionary history may be an epistemically
indispensable tool in our search for functions in certain 'hard cases', nevertheless, only selective
attention to those examples most favorable to the standard line's approach could make one think
that this is true across the board. Any tendency to support such a claim ought to be discouraged
sufficiently by reflection on physiology and anatomy before the advent of evolutionary theory.
Neander's evaluation of the problem of instant duplicates relies, then, on a selective choice of
56 We ought not forget, of course, that evolutionary history is not our only, or even always our best, way of determining functions in hard cases. See Elizabeth Prior (1985, 317), Richard Dawkins (1986/1996, 21), and pp. 183-188. 57 This is a point made clearly in Elizabeth Prior's criticisms of the etiological theory of functions (1985, see esp. p. 317). As Richard Dawkins says, "any engineer can recognize an object that has been designed, even poorly designed, for a purpose, and he can usually work out what that purpose is just by looking at the structure of the object" (1986/1996, 21). This fact constitutes a firmly entrenched datum about physiology and anatomy.
200
examples and a confusion of an epistemological means for answering questions about functions
in certain hard cases with the ontological analysis functions.
Millikan's strategy, on the other hand, is to employ sheer bravado.59 A duplicate of me
(with no evolutionary history) would "exhibit no purposive behaviors", she claims. But this is
claim is both wholly implausible and presented as true without a shred of supporting evidence or
argumentation.60 It would fail to have functions if functions were just selected effects, of course,
but whether or not this is a proper account of the notion of function that has given rise to the
problem of biological teleology is precisely what is at issue; the account's rulings cannot be
invoked at this point without begging the question.
58 Of course it could turn out that we were wrong about the functions, even of eyes and other 'easy cases', in the case of the instant duplicate lions. But this skeptical worry arises for any realist account � including the etiological account � and does not count as any special problem for the view presented in the text. 59 The argument of this paragraph misrepresents Millikan's view slightly. Millikan holds that the program of conceptual analysis, of using cases such as these to test our 'intuitions', is "a confused program, a philosophical chimera, a squaring of the circle, the misconceived child of a mistaken view of the nature of language and thought" ( see for further defense1984, chapters 6, 8, and 9; 1989, 297). Given this, Millikan is unconcerned about the fact that our intuitions don't match hers in the case of the duplicates: she is making a claim about reality, not about how we use our concepts, and in reality there are no functions in the cases at hand despite what our conceptual intuitions tell us. This radical methodological move on Millikan's part cuts all ties to the traditional notion of a biological function which would determine adequacy conditions for her analysis. If we follow this strand in her work rather than strands which attempt to tie her notions with notions which have generated traditional problems of teleology, then her analysis is only misleadingly thought of as analyzing the concept of a function; it would cause much less terminological and analytic confusion if she would simply develop a neologism to label the new-under-the-sun concept she proposes for our use. The dilemma her radical strategy faces is that it either cuts all ties to traditional notions of biological functions, in which case it cannot solve problems of teleology, or it does not cut such ties, in which case it is open to decisive counterexamples. In fact I do not believe that Millikan has a stable position between her strong denials that she is engaged in conceptual analysis (such denials push us to read her work as offering merely stipulative definitions) and her insistence that her notion of a Proper Function solves traditional problems of teleology. 60 George Bealer notes in a parallel context: "For example, by a fantastically improbable but nevertheless causally possible coincidence, a being physically indistinguishable from me could arise spontaneously without any relevant causal interaction with the things in its environment. According to the etiological analysis, such a being would in that case have no intentional contents. But this is absurd. Although the being would not (let us assume) have familiar natural kind intentional contents [because of twin-earth arguments for externalism about meaning], the being could have at least some intentional contents. For example, the being could be aware that it is in pain..." (Bealer 1993a, 107-8). Bealer's point is the one I have been making above; the bare assertion in the face of plausible counterexamples that the etiological account is correct or that it solves the problem of intentionality is so plainly flimsy as to be absurd. It would be extreme analytical arrogance to suppose otherwise.
201
While I suppose that it is just barely conceivable that someone's intuitions about this
case could be so infected by his or her attachment to the etiological theory of functions that she
would deny the above intuition, such denials should carry little weight with one deciding which
theory of functions to accept. To the informed but not dogmatically committed, such denials
must have the air of extreme implausibility. We have strong reason from our ordinary and our
scientifically and historically tutored grasp of the phenomena involved � the phenomena that
standardly constrain our search for theoretical definitions � to believe that Neander and
Millikan are simply wrong in their analyses of these cases.
Our ability to mount convincing responses to the critical challenge of supporters of the
etiological account supports my main contention that the concept of a function is not an
historical concept. There are clear cases of functions for organs and behaviors of 'instant
organisms' despite the fact that there might also be cases which are epistemically hidden from us
without the possibility of reference to evolutionary history, and despite the analytical arrogance,
daring, or sheer bravado of some of the theory's proponents.
5.3.3.2 The indirect response.
Given the failure of the direct response to my argument, we might search for an indirect
solution the etiologist's problem. On one indirect approach to establishing the historical content
of the notion of a function, the etiological theorist points to uses of the word 'selection' in cases
of artifact teleology and natural selection.61 Agents select or design artifacts that will be useful,
thereby conferring functions on the artifacts, and natural selection (metaphorically) 'selects' items
that contribute to overall fitness, thereby (over time) conferring functions on them (according to
the etiological account). Neander holds that "Natural selection is one type of selection process,
which counts literally as such, through the death of a rewarding metaphor" (1991b, 132). The
metaphor has died because there is an important likeness between the teleology and selection in
61 In addition to the case discussed in the text, Paul Griffiths attempts this strategy in his (1993).
202
artifactual and natural cases: in each case the effect of the trait being explained 'plays a role' in the
causal explanation of the trait's existence. In natural selection, the relevant role that traits play is
contributing to increased replication of the type; in agent selection "it may be that a mental
representation of the effect is what plays the causal role" (1991b, 132). Since selection 'plays a
role' in both processes, and since agent selection confers function in the artifactual case, natural
selection may confer function in the other case as well.
Unfortunately, it is only the looseness of Neander's conception of what it is to 'play a
role' that yields whatever plausibility her comparison enjoys. What plays the decisive role in
conferring functions in the case of agent selection is and always has been an agent's intentions with
regards to the artifact, not her having merely selected or represented it mentally. Items that have
no function are represented mentally and then created (e.g., artworks) and items that are never
represented or selected mentally end up having functions even in certain artifactual cases.62
Whatever it is about agents' behaviors that gives rise to artifactual functions, we know
that it is not their selecting or mentally representing them. Further, the fact that problems arise
for naive attempts to cash artifactual functions out in terms of intentions gives etiological
theorists no license to claim that mere selecting or representing does the work; mere selection is
less adequate to account for teleology than intentions, not more. In light of this, the superficial
similarity between the two kinds of selection breaks down, and we no longer have an informative
parallel between how it is that things in general come to have functions. 63 Since there are no
intentions in natural selection, and since past 'selection' and 'representation' are not what
determine functions in the artifact case, the fact that we use the same word, 'selection', to cover
both cases does nothing to bolster the etiological theorist's case.
Neither the direct nor the indirect response to my strengthened continuity objection
succeeds. The weaknesses of the case that can be marshaled in favor of the etiological account is
62 See Bigelow and Pargetter (1987, 101) and Griffiths (1993, §7) for arguments and examples.
203
further evidence that biological function is not an historical notion. This is revealed in standard
ways in our categorial intuitions about the concept and is supported by critical reflection on
tough cases. The standard line on biological functions cannot be correct any more than the
number three can weigh seven pounds.
5.3.4 Necessary and sufficient conditions.
Above I noted that etiological theorists may point to a genuine correlation between
possessing a biological function and possessing an evolutionary history: possessing the one is
sufficient for possessing the other (in our world). This correlation does not form the basis for an
adequate analysis of being a function, however.64 Indeed, in this section I argue that the
correlation between biological function and history of selection is in fact quite weak, far weaker
than identity. The possession of a history of selection is neither necessary nor sufficient for the
possession of a biological function. The examples I cite are (for the most part) well known in
the contemporary literature, but they bear repeating here for the counterexamples can be seen in
a new light given their grounding in this chapter's diagnosis of the etiological account's main
failing: the account analyzes a notion that is not historical in terms of evolutionary history.
The argument against Neander above has already established that the standard line fails
to provide a necessary condition for the possession of functions. An instant lion would have
parts with biological functions (eyes, ears, heart, etc.) which we could discover with as much
confidence as we do in ordinary cases. This would be true even if it were also true that the lions
had parts ('wing like protuberances', for instance) whose functions would be inaccessible to us
without the possibility of investigating relations of evolutionary descent. My treatment of the
direct response to my categorial argument against the etiological account above shows how this
evaluation of the cases can be supported in a dialectical context; that discussion will not be
63 In addition to the case to be discussed in the text, see Paul Griffiths (1993, §8). 64 See the appendix to chapter six for an account of the difference between mere correlations and 'genuine' (or at least stronger) reductive analysis.
204
repeated here. Categorial intuitions about the irrelevancy of history to the notion of a function65
and long histories of consistent usage of the term66 enter into the argument at this stage;
categorial intuitions and historical employment both indicate clearly that the possession of an
history is not a necessary condition for possessing a function. The etiological analysis fails to
provide a necessary condition for the possession of biological functions.
Mark Bedau and Alvin Plantinga argue convincingly that the etiological analysis also fails
to provide a sufficient condition for function possession. Plantinga's argument rests on two
hypothetical examples which invoke processes of artificial selection. In the first,67 supposing
that a Hitler-like tyrant bred into non-Aryans a mutation with a painful result associated with
sight and made their environment such that those without the mutation died and only those with
the mutation bred, then, after some number of generations the mutation would have come to
dominate the population because of its capacity to cause pain in the non-Aryans' selection
history, thereby gaining, on the etiological account, the function of causing the pain. "But
wouldn't it be wrong", Plantinga asks, "to say that m's visual system is functioning properly"
(1993, 204).68 This case may be objected to on the grounds that it employs artificial rather than
natural selection. While pursuing the issue would, I believe, be fruitful,69 it is ultimately
65 See the quotes, above, by David Hull (1974, 113), Robert Cummins (1975/1984/1998, 72), Larry Wright (1976, 96-7) and Wimsatt (quoted from Nissen 1997, 93). 66 See Nissen (1997, pp.186, 57, and 110-111). 67 The second is in all respects similar and will not be described here. 68 The fact probably has no ability to shock us anymore, but it seems that we have actual cases like Plantinga's from the realm of genetic engineering. Peter Singer refers to the development, and patenting, of 'onco mice' for the purpose of scientific research (onco mice are bred for their propensity to develop cancers); see Singer (1993, 35) and his source, Mooney (1990). The proper thing to say about this case (from our perspective) seems to be that the mice are bred to malfunction; their being selected for their propensity to malfunction does not, however, turn the malfunction into a function. 69 'Creation' scientists argue that Darwinist comparisons between natural and artifical selection are plainly rubish. Contemporary advocate Philip Johnson says, "The analogy to artificial selection is misleading. . . . The point of Darwin's theory ... was to establish that purposeless natural processes can substitue for intelligent design. That he made that point by citing the accomplishments of intelligent designers proves only that the receptive audience for his theory was highly uncritical" (1993). The proper evolutionary response � given many times before the publication of Johnson's book to the wild adulation of a truly highly uncritical audience � is that the intentions of agents in artificial selection play a role for all intents
205
unnecessary given the fact that other counterexamples exist which employ only natural selection
among objects in the actual world.
Bedau explains that there are actual cases of entities with complex structures that
undergo processes of natural selection70 but concerning which we feel no temptation to ascribe
teleology. As he says,
natural selection can take place in an entirely inorganic setting. This is possible because of the nature of crystals, specifically, the microscopic crystallites of inorganic materials out of which clays and muds are composed. ( see also 1991; 1992b, 37)
The results of processes of natural selection among such crystallites are structures that we have
no tendency to suppose exist for the sake of anything. They have the history of selection required
by the etiological account, but the clay crystals have no function.71 The etiological account
therefore fails to provide either a necessary or a sufficient condition for the possession of
functions.
The fact that the etiological account provides neither a necessary nor a sufficient
condition for the possession of a function reveals that the correlation between possessing an
evolutionary history and having a function upon which the etiological account relies is actually
quite weak and contingent. It is weak because the correlation does not even hold without
and purposes equivalent to a consistent selectional environment. Agent's intentions play an inessential role in the selection pressures on species such as dogs and cattle undergo under artificial selection. The essential fact of natural selection � that selectional pressures (in this case human generated) nonrandomly sort between competing traits leaving a preponderance of certain types of traits in future generations � is present in both artificial and natural selection. (For an example of this type of response, see Gould 1983, p. 143.) Given this, I believe that anyone who wished to object to Plantinga's case on the grounds that it employed artificial rather than natural selection has a burden of proof to meet before we accept the objection as stated. 70 Bedau gives an extended and detailed defense of the claim that such structures undergo natural selection just as biological organisms do. 71 In a response to Bedau (1991), Mohan Matthen appears to bite the bullet and accept that such crystals have functions since they fit the etiological account (1991, 656). Such moves are, of course, always possible, but their plausibility depends in large part on (a) strength of the grounds offered in favor of the theory overriding one's intuition in such a case and (b) the lack of alternative theories that account better for wider ranges of our central pretheoretical intuitions. The task of the next chapter is to provide such an alternative account. Throughout this chapter it is the strength of the case in favor of the etiological account that is on the line.
206
exception in this world (as Bedau's case of clay crystals shows). It is contingent because
counterexamples to both necessity and sufficiency are readily available through employment of
our modal intuitions concerning possibilities. It would be a mistake to take even the strongest
form of correlation (necessary correlation) to prove identity; it is a more serious error to take a
correlation as weak and contingent as this as the basis of an analysis.72
5.3.5 Normativity.
To this point I have offered two independent reasons to suppose that the etiological
account fails. First, it postulates a disunity in the phenomenon of function attribution among
experts in biology that is entirely an artifact of theory and no part of the data independent of the
etiological account. Second, I have argued that the etiological account fails because it analyses a
notion which is not historical � i.e., biological function � in terms that would transform it into
an historical notion. With this second argument in place, I have argued further (what could not
be argued plausibly independent of the forgoing argument) that the etiological account provides
neither a necessary nor a sufficient condition for the possession of biological functions.
In this section I offer a third independent reason to reject the etiological account of
functions. Etiological theorists attempt to capture the data of biological functions in realistic
terms. A datum about biological functions is that functions introduce biological norms.73 In
other words, functions tell us what organs are supposed to do, rather than simply what they
actually or even normally (in the statistical sense) do. Organs can malfunction; they may not work
properly. On the basis of these observations we may formulate another condition of adequacy on
a proper analysis.
Condition three: normativity. An adequate account of teleology must account for the normativity of biological functions. Hearts are supposed to pump blood, and they
72 Again, see the appendix to chapter six for a more focused discussion of the requirements for reductive analysis worthy of the name. 73 See Neander (1991b, 123, 131, 137); Neander (1995, 222); Walsh and Ariew (1996, 259, 260, 262); Millikan (1984, 5, 17; 1989, 304), and Dretske (1988, 63).
207
malfunction when they fail to do so, or fail to do so as efficiently as they are supposed to.
Ruth Millikan ties normativity in with the reductionist's project: "The task of the theory of
functions is to define this sense of 'designed to' or 'supposed to' in naturalist, nonnormative, and
nonmysterious terms" (1984, 17). The etiological account of functions (see above, p. 176) is
intended to play exactly this role; it is meant to offer a nonnormative analysis of the sense in
which, for instance, hearts malfunction during heart attacks. All of this is uncontroversial in the
literature on biological function � a part of the data to be explained. Nevertheless the
normativity of biological functions is left entirely unexplained despite claims to the contrary by
etiological theorists.74
In attempting to account for this feature of teleology, Karen Neander claims that
"biological norms are determined by the history of traits" (1991a, 327, emphasis added. See also
330), but why this should be the case in the only relevant sense is left completely unexplained.
An ambiguity in Neander's statement needs revealing. Biological functions may (in some cases)75
be determined by history in the epistemic sense that they are (sometimes) discovered by us through
historical (i.e., evolutionary) investigations. This would not suggest � much less entail �
however, that biological norms are determined by history in the ontological sense that history of
selection provides the basis for a reductive identification of biological normativity with aspects of
selection history.76 Neander has a strong point that our identification of functions often
depends epistemically upon our discovering histories of selection in at least in some 'hard'
cases.77 But we are given no reason to suppose that biological norms require evolutionary
histories for their existence. The counterexamples to the necessity claim, discussed above, reveal
74 John Searle (1995, 18) criticizes the etiological view for leaving the normative character of functions unexplained. 75 It is crucial to keep in mind that physiology was in the business of ferreting out biological functions long before the advent of evolutionary theory. See Prior(1985). 76 See the appendix to chapter six for close discussion of issues surrounding reducibility.
208
this fact: an instant lion's heart could malfunction. Etiological theorists will want to appeal to
their understanding of cases of 'instant duplicates' to bolster their claim here, but I have already
argued that their responses fail because of their reliance either on the highly selective choices of
examples or on sheer bravado (see above, pp. 197-201).
To reiterate the points already made in this new context: instant duplicates with parts
such as eyes and hearts that have clearly discoverable functions may suffer malfunction in those
parts, showing that having a history of selection is not necessary for the existence of biological
norms.78 Bedau's discussion of selection among clay crystallites shows that the proper history of
selection is not sufficient for normativity either, since there is neither a function that the
structures of clay crystals performs nor (obviously) any thing they are supposed to do.
Of course, this is just what we should expect preanalytically: there is a strong prima facie
case to be made that the way things have worked in the past gives rise to no norms about how things
should work now or in the future. The parallel case with moral norms is particularly revealing:
that our ancestors enslaved other of our ancestors gives rise to no norms for our behavior; they
should not have done it, and we should not do it. Any theorist who claims otherwise in either
the biological or the moral case has a strong burden to overcome, and it is clear that etiological
theorist's brief remarks on the subject have not met this burden. They are therefore unentitled
to the claim that the etiological account captures the normativity of functions; it simply does not.
77 Her discussion involves the case of the instant lions, discussed above at p. 197. 78 Again, epistemic difficulties may obtrude in the case of instant organisms. In case one, suppose an instant lion appears, and this lion has what we would normally (with lions with which we are familiar) call a malfunctioning heart. We may not be sure with such a creature whether the organ really is malfunctioning simply because we do not know the context in which such a creature's heart normally functions. In case two, however, suppose an instant lion appears and has what we would normally call a functioning heart, and this lion lives a long and healthy life in environments just like the lions with which we are familiar until developing late in life what we would call a heart problem. In this case, despite the lack of evolutionary history, we would be confident that the lion's heart was (now, late in life) malfunctioning. Epistemological problems with instant duplicates abound, but I do not believe they fatally undermine the point in the text. The point holds so long as plausible cases of norms exist in the absence of evolutionary history, and such cases are abundant.
209
It should be noted, again, that this objection to the etiological theory has readings on
which it is and on which it is not ad hominem. Etiological theorists are vocal in their endorsement
of the claim that biological functions are normative and support normative claims; the argument
above is therefore in one of its guises an ad hominem argument addressed directly to those
etiologists who maintain this view. On the other hand, we ought to recall that this commitment
to the normativity of biological functions arises out of contemporary theorists' strong
commitment to the realist stance on teleology. In affirming the normativity of biological
functions such theorists are consistently following out their commitment to seek an account of
biological functions that accepts the data from common sense and the science of life at face
value. It is typical in biology to speak not only of functioning and proper functioning, but also
of malfunctioning and of what an organ is supposed to do. Respect for this feature of biological
functions is not a merely idiosyncratic feature of some etiologists' views, it arises naturally out of
the strong commitment to preserving the phenomena of biological functions. In this sense the
argument of this section is not merely ad hominem, but is directed towards any theorist who takes
a commitment to the realist's stance seriously.
I have labored this last point because it is clear that if reductivist accounts of biological
functions fail, as I argue they do, then if we maintain our desire to preserve the phenomena
associated with biological teleology we are saddled with serious philosophical questions
surrounding the ontology and scientific methodology necessary for a sui generis account of
teleology which accepts the phenomenon. These issues are addressed in chapter six.
5.3.6 Taking stock.
We have strong reasons for being realists about teleology79 and (it is presupposed) we
have strong reasons for being reductivists about teleology.80 I have argued in this chapter that
79 See above, pp. 170-172. 80 See above, pp. 172-174.
210
these two sets of reasons are in tension, for there exists no adequate reductivist account of
teleology. The most promising reductivist account of teleology � the etiological account
deriving from Wright's analysis � fails for three reasons. First, it injects an artificial disunity into
the phenomena of expert attributions of biological functions. Second, it fails to obey categorial
constraints on proper theoretical definitions of problematic notions with histories of use, a
failure that also reveals why it provides neither necessary nor sufficient conditions for the
possession of biological functions. Finally, the account fails to handle a central aspect of the data
of biological teleology: the normativity of functioning. The standard contemporary line on
functions appears, despite the near consensus it enjoys, to fall far short of its advertised
advantages. The most promising attempt so far to analyze biological functions under the
efficient causality constraint of the reductivist's project fails.
5.4 The value of the etiological account.
The focus of this chapter has been on the failings of the etiological approach considered
as offering a reductive account of biological functions. I believe the account has serious failings
as an account of functions, and have argued that point at length. This fact may leave the reader
with the impression that my evaluation of the etiological approach is wholly negative. This
would be a false impression, however, for despite the fact that etiological theorists are wrong that
they have in their analysans a proper account of biological functions, they are emphatically not
mistaken that that analysans plays an important role in our theoretical understanding of living
things.
One way of making this point is to note the striking similarities between the definition
of a biological function on the etiological account and a common definition of the notion of a
biological adaptation.81 On the etiological account of functions:
81 There is debate in theoretical biology and the philosophy of biology over the correct analysis of the notion of biological adaptation. I cite the definition above not to settle decisively such debates, but simply to point out one way in which the notions etiological theorists appeal to have genuine scientific value in helping us understand living beings.
211
It is a/the proper function of an item (X) of an organism (O) to do that which items of X's type did to contribute to the inclusive fitness of O's ancestors and which caused the genotype, of which X is the phenotypic expression (or which may be X itself where X is the genotype) to increase proportionally in the gene pool. (Neander 1991a, 319)
On Elliot Sober's conception of a biological adaptation:
Characteristic c is an adaptation for doing task t in a population if and only if members of the population now have c because, ancestrally, there was selection for having c and c conferred a fitness advantage because it performed task t. (2000, 85)
This account is strikingly similar to the account of functions given by etiological theorists. One
way of understanding the appeal of the etiological account of functions, then, is by noting that its
analysans picks out a central explanatory concept in evolutionary biology: adaptation. Etiological
theorists by no means underestimate the explanatory role played by their analysans in modern
biology.
Nevertheless, it remains clear that this definition of adaptation (or 'function') does not
itself provide a solution to the problem of biological teleology. Sober points out that, "If
function is understood to mean adaptation, then it is clear enough what the concept means"
(2000, 87), but if this is what 'function' means, then we have dropped all pretense of tying our
use of words to the historical uses which gave rise to the problem of teleology. As Sober notes,
"adaptation is a historical concept" (86, see also 84), but function is not an historical notion.
Functions give rise to norms, but adaptations do not. One way of understanding the explanatory
power and appeal of the etiological account of functions, then, is to note that the analysans does
have a real explanatory role in biological theory; nevertheless, this explanatory role is simply not
identical to the explanatory role of biological functions.
Another way of highlighting the explanatory value of the etiological account's analysans
is to note that the evolutionary explanation of the origin of parts on the basis of what they do for
organisms supplies us with at least a partial understanding of how it is that biological functions
212
are instanced in this world.82 There is a correlation between biological functions and histories of
selection,83 and this correlation leads to genuine insight into the material origin and maintenance
of functions in biological populations. Further, the correlation grounds the epistemic role
evolutionary analysis has in the discovery of functions.84 The etiological account therefore
indicates a sound methodology for discovering functions without providing a reductive analysis
of what it is to be a biological function.
We are not entitled, therefore, to conclude that the analysans of the etiological account
lacks theoretical explanatory value simply on the basis of our denial that it supplies a solution to
the problem of teleology. The analysans is important to contemporary biology in a number of
ways, but not as a solution to the problem of biological teleology.
5.5 Conclusion.
[S]uppose... you are convinced (as most of us are) that there really... are (for natural organisms) such things as proper function, damage, design, dysfunction, and all the rest. You think there really are these things and are unwilling to take the functionalist stance: then if you also think there is no naturalistic analysis of these notions, what you have is a powerful argument against naturalism. (Plantinga 1993, 214)85
I have argued that the etiological theory of functions fails as a reductive analysis of what
it is to be a biological function. The etiological account yields an artificial disunity among the
phenomenon of expert attributions of biological function; its analysans makes essential reference
82 My thanks to Robert Pasnau for emphasizing this point with reference to an earlier draft of this chapter. 83 We ought to remember, of course, that this correlation is not as tight as we might hope. Possessing a biological function is sufficient for having a history of selection in this world, but possessing a history of selection is neither necessary nor sufficient for having a biological function. 84 This role was mentioned above in discussion of Karen Neander's case of instant lions. 85 I argue in chapter six that Plantinga's conclusion here is too strong. One has not an argument against naturalism but against reductionism. The solution to the problem of biological teleology sketched in chapter six (and developed in the appendix to chapter six) constitutes a naturalistic nonreductivist account.
213
to history, but the analysandum is not an historical notion; it provides neither a necessary nor a
sufficient condition for the possession of biological functions; and finally, it does not explain the
normativity of biological functions. This is no trivial set of problems; these are systematic and
revealing failures all of which point to the current unavailability of a plausible account of
biological functions that respects the realist and reductivist constraints on analyses of biological
function that generated the traditional problem of biological teleology.
Given the problems with naturalistic approaches to teleology surveyed briefly here, it is
reasonable to try a different approach. The history of analytic attempts to achieve Nagel's goal
(see above, p. 173) of analyzing teleology such that the analysans contains no teleological notions
has come up short. This is not a conclusion we should shrink from; as Lowell Nissen says, "If
certain consequences of teleological language are, as a matter of brute fact, present, it is only
being forthright and realistic rather than medieval or obscurantist to acknowledge them" (1997,
211). If our options have appeared to be eliminitivism, mentalism, panpsychism, or
reductionism and none seem satisfactory, it may be time to reevaluate the metaphysical
commitments and methodological worries that underlie the belief that these options exhaust the
field; it may be time, that is, to take an extended look at unreduced Aristotelian final causality.
We now turn to this task.
214
Chapter 6
An Ontology for Aristotelian Teleology.
Mightn't it still be open to us to embrace a systematically teleological world-view? Is there any rational reason why we should not switch back to an Aristotelian conceptual framework? (Woodfield 1976, 14)
In light of the previous chapter's analysis of state of the art attempts to do away with sui
generis Aristotelian teleology in accounting for biological functions, the goal of this chapter is to
develop an ontology in which sui generis Aristotelian teleology can find a natural home, and which
meets the following three nontrivial constraints.1
C1: The ontology should be independently motivated.
C2: The ontology should make clear how unreduced Aristotelian teleology makes a difference in the causal structure of the physical world.
C3: The ontology should respect the legitimate claims and achievements � both methodological and ontological � of modern science.
The crux of the problem of teleology is contained in a combination of the second and third
constraints. Teleology appears to be a causal factor in the life cycles of living things generally,
and an account of teleology should explain not just the heuristic usefulness of teleological
language in highlighting relevant causal features of organisms, but also teleology's peculiar causal
contribution to the life cycles of organisms.2
The goal of this chapter is resolve the tension between these constraints through the
development of a scientifically respectable ontology that is independently motivated and in
1 Unless otherwise specified I use 'teleology' and its cognates to indicate a sui generis irreducible real causal factor in the structure of the world. 2 David Charles (1988, esp. pp. 38-9) explains how teleology can be metaphysically robust in a 'static' sense. On his view, teleology is explanatory because it picks things out by their essential features. This, while perhaps true, is insufficient to solve the traditional problem of how biological teleology can be explanatory; the traditional problem seeks a metaphysical basis or understanding of how it is that final causality makes a causal contribution to say, the growth and behavior of plants. How is it that 'seeking the
215
which teleology finds a natural home. If this goal can be achieved, my account of teleology
should have strong appeal given the persistence and apparent explanatory fruitfulness of
teleological explanations in the life sciences and the failure of modern naturalistic accounts of
teleology.3
In this chapter I will approach the problem of teleology by first surveying historical
objections to positing Aristotelian teleology in the modern world. This survey will eliminate
from consideration irrelevant (but nonetheless common) objections to Aristotelian teleology, and
focus our attention on the core of the explanatory task facing any theorist looking to make a
place for Aristotelian teleology in a physical ontology adequate to modern science. The main
objection to sui generis Aristotelian teleology is that it appears to have no home in the ontology of
the physical world as revealed by science in the modern era. I respond that it is not modern
science but certain philosophical presuppositions of modern science that has seemed to bar
teleological causation from inclusion in our physical ontology. In fact, there is mounting
independent evidence from the science of quantum mechanics that we must abandon the
presuppositions that exclude teleology in order to make sense of certain physical phenomena.
This strong evidence is sufficient to show the substantive and methodological compatibility of
science with an ontology in which teleology fits naturally, and I develop and defend the ontology
necessary in what follows.
First, however, we turn to a critical survey of the main contemporary objections to
biological teleology.
6.1 Objections to Aristotelian teleology.
Given that teleological explanations have been enormously controversial for the whole
of the modern period, we must survey the reasons for this controversy before we can enter into
sun' can explain the leaves' turning? Charles' explanation while valuable in itself cannot help us to answer this question, and it is the latter sort of problem that I attempt to address in this chapter. 3 Both of these claims are defended in the previous chapter of this dissertation.
216
a productive discussion of purported solutions to the problem Aristotelian teleology. I here
present a compilation of supposed problems with Aristotelian teleology taken from
contemporary biological and philosophical literature on the subject.
O1. Aristotelian teleology is inextricably bound up with X, and a commitment to X is unacceptable; X may take the form of one or more of the following:
a. the belief that the world is a nice place.4
b. the belief that evolution is progressive; i.e., that evolution is itself teleologically directed toward some end.5
c. the existence of a creator God.6
d. animism or panpsychism; the existence of minds in things which do not have minds.7
e. backwards causation.8
f. vitalism or certain ad hoc entities.9
O2. Aristotelian teleology is incompatible with the methodological constraints of modern science:
a. The progress of science has shown reference to teleology to be irrelevant to scientific explanation.10
b. Teleological explanation is inevitably vacuous; it is committed to phony (and often parodied) 'natural tendency' explanations which offer a redescription of the explanandum as if that constituted a genuine explanation.11
4 See Bedau (1992d, 283). 5 See Bedau (1992d, 283). 6 See Bedau (1992d, 283), Mayr (1988, 40), Nissen (1997, vii, 128, 162) , and Woodfield (1976, 3). 7 See Buller (1999b, 6), Jacobs(1986, 390), Mayr(1988, 40), Nissen(1997, vii, 128, 162), and Woodfield(1976, 3). 8 See Buller (1999b, 6), Jacobs(1986, 392), Mayr (1988, 40), Nissen (1997, 96, 105, 134), and Woodfield (1976, 34). 9 See Bedau (1992d, 283), Mayr (1988, 40). 10 See Woodfield (1976, 8-9). 11 See Woodfield (1976, 7).
217
c. Teleological claims are not (but must be) open to empirical test.12
O3. Aristotelian teleology is ontologically unacceptable because it is incompatible with the ontology of the physical world as revealed by modern science.13
Aristotelian teleology is often rejected on the contemporary scene for a combination of the
above reasons. The adequacy of the account offered in this chapter, then, will depend upon my
having surveyed the best and strongest objections to teleology and on my ability to mount an
adequate defense against the objections. Let us take the objections in turn.
6.1.1 Guilt by association.
The problems under (O1) have a common form: Aristotelian teleology is claimed to be
bound inextricably to unacceptable commitments and suffers guilt by association. It is no doubt
true, of course, that teleology has been thought by some of its supporters at some times to be
bound up with each of (O1a-f). Nevertheless, it is equally clear that Aristotelian teleology need
be committed to none of these unacceptable views. Aristotelian teleology may find its home in a
world which isn't 'nice' (O1a), and where evolution isn't progressive in the sense that the
evolutionary process itself (rather than some of its products) is directed towards goals (O1b).14
Less clearly perhaps, but true nonetheless, teleological directedness may be immanent in
natural things and hence independent of and not derivative upon a creator God (O1c) and free
from panpsychical worries (O1d). This is a result of Hume's subtle reflections on the nature and
observability of the causal relation. We cannot know a priori what sorts of causal connections
there are in the world, and thus the supposition that Aristotelian teleology must be either
modeled on mentalistic teleology or be non-existent (e.g., because it cannot be modeled on
'billiard ball' causation or some other paradigm case of efficient causality) rests on the false
presupposition that sui generis Aristotelian teleology cannot exist precisely because it cannot be sui
12 See Allen and Bekoff (1995, 244; 1995, 9). 13 See Mayr (1988, 40).
218
generis. But whether or not some particular sort of causal relation is instantiated is not something
we can know a priori. The causal relations that exist and deserve our acceptance need not 'pass
muster' in front of our intuitions as to whether they match up to models of causation that are
supposed to be 'intuitive'. The existence (or nonexistence) of a kind of causal relation reveals
itself to us (if at all) through experience, and it is our duty to accept the kinds we find with
'natural piety'. We have an especially strong duty to accept those our best sciences appear unable
to do without.
The backwards causation worry has always been a red herring inspired by
misunderstanding teleology as if it were efficient causality running backwards in time.15 On this
supposition, teleological causation just is efficient causation, but working backward rather than
forward in time; the goal of a plant's growth (i.e., the mature structure) is somehow thought to
reach back in time and pull the plant's juvenile form toward itself. The proper response to this
challenge is twofold: we may first grant that if teleological causation were conceived along these
lines then it would be incoherent. But second, we may simply note that no one who postulates
sui generis Aristotelian teleology is likely to endorse the view that teleology is correctly understood
as efficient causality of any sort.16 That teleological causation appears incoherent when it is
modeled on backwards efficient causality is no threat to teleology as such; it is a threat to the
claim that we ought to conceive of teleological causation along these rather unpromising lines.17
Fear of vitalistic commitment (O1f) is a bit more troublesome. The charge that the
invocation of teleology is vitalistic and hence proven false would be warranted and correct if (as I
argued in chapter one) commitment to sui generis teleology committed one to the view that
14 Good discussion of the three aspects of evolution's 'blindness' to progress can be found in Campbell (1974) and Falk (1981). See also the articles collected in Part IX: Progress, of Hull and Ruse (1998). 15 For an example of this mistake, see Reichenbach (1957, pp.192-5). 16 See Woodfield (1976, 34). 17 Of course, if this model of teleology fails one may rightly ask for some explanation of how teleology does work in the world; this chapter is devoted to providing just such an explanation. The narrow point here is that this particular challenge to sui generis teleology is misdirected.
219
teleology 'floated free' of physical properties either in the sense that it did not depend for its
existence on physical properties or in the sense that its action in the world broke physical laws.
However, as I argued in chapter one, there exists a perfectly coherent middle path between
sophisticated mechanism and vitalism that allows room for the causal efficacy of irreducible
properties with novel causal powers so long as those properties depend for their existence on the
instantiation of further physical properties and those causal powers do not break any physical
laws. Science has not refuted the view that there are such properties; indeed one main goal of
this chapter is to bring to the fore the fact that one science (quantum mechanics) seems to
provide excellent grounds independent of our narrow concerns with biological teleology for
believing that such emergent properties and causal powers exist. There is room, therefore, for sui
generis teleology in a nonvitalistic conception of the world; a large part of the burden of this
chapter is to elucidate just such a view.
I conclude, then, that none of the supposed associations thought to cause trouble for
teleology (O1a-f) is fatal to a commitment to Aristotelian teleology. While some versions of the
theory may have made such commitments, they are not inevitable counterparts of Aristotelian
theses generally, nor in particular of the theory to be offered in this chapter.
6.1.2 The methodological acceptability of Aristotelian teleology.
Problems (O2a-c), the supposed problems facing Aristotelian teleology from the realm
of scientific methodology, are no more worrisome than problems under (O1). It is simply not
the case that the advancement of science during the modern period has shown teleological
explanation to be irrelevant as (O2a) charges. Biologists have never abandoned teleological
explanation; it performs explanatory work in their work qua biologists, and there is no indication
that theoretical biology will give it up in the near future.18 If my arguments were sound, then the
18 David Buller says, "In spite of the difficulties associated with teleology, however, biologists continued to use the teleological concept of function in describing the characteristics of organisms, finding the organization of organisms and the operation of their parts virtually incomprehensible in strictly non-teleological terms" (1999b, 6. Emphasis added). As David Hull puts it, "Teleological systems do seem to force
220
claim that teleology (under any analysis) has been shown by science to be irrelevant must rely on
either (what amounts at this point to) a question-begging assumption that teleology is reducible
to some other form of causation19 or a narrow identification of science with physics and
chemistry. Neither of these assumptions is justified, however. The persistence, patterns, and
apparent explanatory value of our functional attributions in everyday life and science give ample
justification to a realist stance on teleology, and these reasons to be realists are not undermined
by the failure of reductive accounts of teleology unless there are independent problems with the
supposition that teleology is a sui generis causal factor in the world. Determining whether or not such
independent problems exist is the motivation for this survey and for this chapter, but we need
not fear that (O2a) � the claim that teleology has been shown by science to be irrelevant � will
undermine sui generis teleology for (O2a) is false and, as such, cannot undermine anything.
The worry that explanation in terms of Aristotelian teleology must be vacuous, (O2b),
has no special bearing on the problem of teleology for the same purported problem arises with
other forms of causation. The vacuity problem is typically illustrated with the explanation of a
pill's power to put one to sleep through mentioning a directedness in the pill toward putting one
to sleep. But the existence of this type of vacuous teleological explanation cannot undermine
teleological explanation generally since cases of such vacuity arise in the case of efficient causality
as well; thus we may claim not that the pill has a directedness but simply a power to put to sleep.20
This explanation is as vacuous as the first but makes no reference to teleological directedness.
We are right in this instance not to find this to be a problem with efficient causal explanation in
themselves on us" (1974, 120). Andrew Woodfield says, "Although some biological function-statements may seem metaphorical, some seem to be literally false (e.g. 'Noses exist in order to support spectacles'), and some seem literally true.... Whatever the correct analysis of these TDs [teleological descriptions] may be, there is a strong presumption that they make at least some objective claim on reality" (1976, 32, see also p. 1). See also Nissen (1997, vii), Allen and Bekoff (1995, 244). 19 I argued in chapter five that extant reductivist accounts of biological teleology fail. 20 See Woodfield (1976, 7-8).
221
general, but instead diagnose it as a local failure concerning a particularly lazy invocation of an
otherwise useful form of causal explanation.
Just as clearly, however, there are plenty of nonvacuous teleological explanations, as
evidenced by biology's continued devotion to seeking and offering such explanations. As we saw
in chapter five, biologists find teleological explanation quite powerful. Far from being the
vacuous and trivial affairs invoked in the vacuity objection biological functions as invoked by
scientists may take a great deal of empirical work to discover and can be quite surprising to their
discoverers. Only selective attention on the worst abuses of teleological explanation could
convince one that teleology is undermined by the existence of vacuous teleological explanations.
There may of course be local failures of teleological explanation of the sort cited in the vacuity
objection, but such failures do nothing to undermine teleology generally.
There is a further worry associated with (O2b), however, which needs to be aired. In
explaining this supposed problem for teleological explanations, Andrew Woodfield claims that:
The nub of the criticism is that appeals to natural tendencies are nonexplanatory unless there is or could be evidence for them which is independent of their alleged manifestations. (1976, 7)
Woodfield here places a constraint on genuinely explanatory invocations of teleology. To be
genuinely explanatory invocations of teleology must be such that there is or could be evidence
for the teleological directedness independent of its alleged manifestation. This requirement would
imply that many function attributions are not genuinely explanatory. Fortunately, however, the
requirement that generations this implication must be rejected, for it is far too strong.
That this is so becomes obvious when we generalize the application of the same
requirement to nonteleological cases. Physical explanations could not meet the requirement
stated since they cannot be vindicated by evidence independent of the manifestation of physical
properties, entities and relations. But surely any requirement that rules out physical explanation
in the sciences is too strong. Taking further examples, we neither do, nor should we, require
evidence of the existence of the external world independent of the world's alleged
222
manifestations.21 Nor should we require evidence for the existence of moral norms independent
of their alleged manifestations in intuitions concerning wrongness, rightness, and the like.
The requirement, then, is far too strong; no autonomous (i.e., sui generis, nonreductive)
area of inquiry could meet the challenge. It is not the case that legitimate explanations of sort F
must 'back themselves up' with explanations of sort G where F ≠ G, and it will not be the case
that such explanations are available for autonomous domains of inquiry � whatever those
happen to be. Singling teleological explanation out by placing this requirement on it alone may
exhibit a tacit presupposition that teleology must be reducible to some other (independent and
independently specifiable) relation between things.22 But in the context of the vacuity problem
this reductionist demand is quite simply out of place: nonreductive explanations need not be
vacuous. Physical explanations are nonvacuous but nonreductive,23 explanations of modal and
ethical terms may be both nonreductive and nonvacuous, and teleological explanations may be
both nonreductive and nonvacuous for all the vacuity objection has shown. There is no reason
offered here to doubt sui generis teleological explanation which is not also a reason to doubt any
autonomous type of explanation.24
Aristotelian teleology cannot be saddled with explanatory vacuity, therefore, unless we
unfairly restrict our attention to the very worst cases of teleological explanation or place
21 Descartes (1967) disagreed. He attempted to find grounds for belief in the physical world through invoking intuitions about the nature and existence of God, requiring that we have such independent grounds for our belief before we are justified in claiming to know that there is an external world. Thomas Reid objects that Descartes' demand, taken up unwittingly by the empiricists, is unwarranted (1983). For expositions of Reid's arguments, see Alston (1985) and Wolterstorff (1983). 22 Given the ambiguities in the notion of 'reduction', and corresponding ambiguities in what Woodfield may mean by 'independent' in the statement of his requirement, this point may require revision. If 'reduction' requires only correlation, then it is unobjectionable as I note below. Stronger readings of Woodfield's requirement presuppose stronger forms of reduction, and it is those stronger forms that are objectionable in this context. See my discussion of reduction in the appendix to this chapter. My thanks to Chris Shields for bringing this point to light in comments on an earlier draft. 23 As David Chalmers expresses the point, "Physical theories do not derive the existence of these features [i.e., matter, motion, space, and time] from anything more basic, but they still give substantial, detailed accounts of these features and of how they interrelate" (1996, 213).
223
unmotivated requirements on teleological explanations, requirements stronger than any
purported explanatorily autonomous domain could hope to meet. If we set the requirements for
adequate teleological explanations sufficiently high we can, of course, trivially guarantee that no
nonreductive account of teleology will meet those standards. The interesting question is whether
teleology can pass the reasonable and motivated standards for explanatoriness that are employed
elsewhere in everyday life and successful sciences � both formal and physical � and the vacuity
objection has supplied us with no grounds to suppose that it cannot.
Nevertheless, the worry concerning the supposed vacuity of sui generis teleological
explanations is probably only heightened by (O2c), the problem that such explanations seem not
to be open to empirical test. This claim is false, however.25 26 So long as there exist reliable
causal correlations between teleological properties and physical phenomena (which their should
be, if the two27 forms of causal relation are compatible and are at play in orderly phenomena),
teleological explanations will be open to empirical testing through standard scientific methods
(both physiological and evolutionary). So long as we are careful not to confuse the correlation
between the Darwinian origin of a trait's function and an analysis of its being a function, we are
free to employ all the sophisticated tools of modern biology (including nonevolutionary tools
24 Alvin Plantinga makes this point clearly and convincingly in chapter one of his Warrant and Proper Function (1993, see esp. p. 4). 25 Above I cited Allen and Bekoff (1995, 244; 1995, 9) as raising this worry. To be fair to those authors I must make clear at this point that they do not phrase the worry in the terms that I do, nor do they commit themselves to the falsity of what I say in the body text. They do, however, raise a problem for teleological explanation which I believe is fairly represented by (O2c). 26 Bedau addresses this issue convincingly. He illustrates the possibility of such a science with an uncontroversial example � the materialistic study of physical primes. 'Physical primes' refers, roughly, to prime numbered groups of objects some distance from each other. The point of the example is that one could study such groups of objects in a materialistic science despite the fact that being a physical prime isn't an entirely naturalistic property since it has a mathematical component. He likens this uncontroversial case to the cases of the study of conscious beings (even if consciousness is sui generis) and life. He says, "Thus, by analogy, even though biologists study living entities naturalistically, it does not follow that the notion of life is itself purely naturalistic. The naturalistic practice of the science of life is consistent with the hypothesis that the notion of life contains a non-naturalistic, evaluative component" (Bedau 1992b, 46). 27 There is, of course, no reason to suppose that there are only two sorts of sui generis types of causal relatedness, efficient and teleological, but nothing hangs on the simplification expressed.
224
from physiology and anatomy) in our investigation of biological functions. Investigation will
continue to turn up functions that were unavailable even to tutored common sense before
empirical investigation, and further empirical investigation may even overturn the assignment of
functions given by tutored common sense and science in particular cases. Thus, there need be
no problem concerning the testability of statements involving genuine Aristotelian teleology. If
our teleology were strongly vitalistic in the sense that it was claimed to float free of or be
independent the interactions of physical objects then the charge of being unopen to empirical
test would have real bite. As it stands, so long as teleological directedness is causally enmeshed
in and compatible with standard efficient causal relations we have no reason to suppose that they
will not be strongly correlated with other salient properties and hence open to standard empirical
tests and investigations.
To make this point more clear, note that the worry as it applies to teleology may rest on
the assumption, stated by Woodfield (above p. 221), that causal relations are not legitimate unless
they can be confirmed to exist independently of their manifestations. That requirement was too
strong; it would not rule any causal relations (or any other sui generis phenomena) as legitimate. If
the requirement of testability raises a skeptical worry such that we need a special independent test
for the existence of teleological causes, then it too places too strong a requirement on genuine
teleological explanation and should be rejected for that reason. If, however, all we have is the
weaker non-skeptical demand that teleological claims be open to scientific test according to the
standards and practices of our best scientific theories, then there is no worry that sui generis
Aristotelian teleology cannot rise to meet the demand.
6.1.3 Sui generis teleology and physical ontology.
I have argued that the bulk of the objections to Aristotelian teleology are easily met. We
arrive at last, however, at the greatest objection to sui generis final causality. What is needed is an
account of the possibility of sui generis teleological interaction in the causal structure of the world
that coheres with the ontology of the physical world as revealed by modern science. The most
225
serious problem is that sui generis Aristotelian teleology seems incompatible with the prevailing
conception of the ontology of the physical world (O3). According to the standard picture, there
is no place for teleology in a world that runs according to physical causes and laws.
The best argument against the supposition that efficient and teleological causation are
incompatible is to develop a strong case for their compatibility. Science is often claimed to have
shown us that there is no room in the world's physical ontology for teleological causes, but I
shall argue that it is not science but certain philosophical presuppositions of science that have
generated the apparent incompatibility. Indeed, I argue that the science of quantum mechanics
provides evidence that the ontology of the physical world as revealed by science is not only not
incompatible with sui generis teleology but is in fact downright hospitable to it. It has been the
philosophical presuppositions of modern science, not the science itself or any method definitive
of science, that has made teleology seem incompatible with science, but the quantum mechanical
(and possibly biological and psychological) evidence gives us strong reason to abandon those
presuppositions.
In what follows I give a brief sketch of an ontology for the physical world in which sui
generis teleology may find a natural home, namely, an emergentist ontology.28 I then argue on the
basis of contemporary work on the philosophy of quantum mechanics that our best scientific
theories indicate that our world instantiates just such an ontology. Thus far from it being the
case that modern science has revealed a physical ontology at odds with the postulation of sui
generis teleology, if we take our conclusions from science rather than from a priori presuppositions
concerning the way the world must be causally structured, then we have good reason to suppose
28 I have included a an appendix to this chapter in which I discuss and defend in more detail the ontological commitments of the ontology invoked here. I refer readers to that discussion for more details concerning emergentist ontology.
226
that there is room for both science and sui generis Aristotelian teleology.29 The chapter proper
closes with defenses against common objections to such an ontology.
6.2 Emergentism and teleology's compatibility with contemporary science.
The thesis of this section is that there is nothing incoherent or contradictory about the
supposition that the ontology of the physical world includes sui generis teleology. Strong forms of
emergentist ontologies such as the one I sketch below (and develop in detail in the appendix to
this chapter) are intended to elucidate 'middle ways' between strong and unacceptable dualisms
and equally strong and unacceptable materialisms.30 Such ontologies are designed to handle the
dual tasks of accommodating the results of science, both methodological and substantive, while
at the same time resisting the strong methodologically and ontically monist or reductivist
tendencies of contemporary science and philosophy of science. The goal of emergentist
ontologies is to illustrate that there exist coherent conceptions of the ontology of the physical
world that neither challenge nor undermine our best science, but which also accommodate the
possibility that the world is ontically more robust than strict reductivist and mechanist
conceptions would allow.31
The notion of property emergence was originally explicated and employed in attempts to
develop a tertium quid between vitalism and mechanism in biology and also between materialism
29 If Aristotle were telling the narrative, he might say, "When these men and the principles of this kind had had their day, as the latter were found inadequate to generate the nature of things, men were again forced by the truth itself, as we said, to inquire into the next kind of cause" (Met. i.3 984b9-11; see also Met. i.3 984a18, PA i.1 642a19-20, 642a27-8, Phys. i.4 188b28-9). Richard Healey (1991) draws our attention to the moral that our philosophical presuppositions about the world need to stay in touch with what our science is telling us about the world in exactly the context I have in mind. Hilary Korblith (1994, pp.40-2, quoted below) opts for an even stronger commitment to the ontic authority of the sciences. See also Crane (1994, 480) and Rudder Baker (2000, 23). 30 See chapter one's discussion of vitalism for examples of such positions. 31 To say that the ontology is coherent is not to say that it has gone unchallenged. See Bedau (1997), Kim (1992; 1993; 1999) and Klee (1984) for challenges to the view. John Searle (1992) supports a weak form of the ontology while challenging a stronger form. Defenders of the ontology include Broad (1918-19; 1925), Lovejoy (1927), Hasker (1982; 1999), Healey (1991), Humphreys (1995; 1996; 1997a; 1997b), O'Connor (1994), Silberstein and McGeever (1999), and Teller (1986).
227
and substance dualism in the philosophy of mind.32 Emergentists attempt to keep the best of
both camps while avoiding their excesses. From the vitalistic and dualist camps, they took the
notion that there are properties of physical things that are irreducible to the physical but
nevertheless (and least possibly) causally efficacious.33 From the materialistic and mechanistic
camps the emergentists took both a respect for the material continuity of physical things and a
firm commitment to the causal dependence of such emergent properties as there may be on
physical realization bases. Microphysical explanation and causation have a firm place in
emergentist thinking about the physical world despite the fact that they enjoy neither the
explanatory nor the causal hegemony they have in traditional monistic and reductionistic
conceptions of the world.34
Emergence is associated with the vague claim that a whole may be greater than the 'sum'
of its parts, and contrasted with the denial of that claim.35 There are a number of ways of
cashing this claim out, and they will lead to various conceptions of emergence that need
distinguishing.36 Clearly, however, all doctrines of emergence will be dealing with properties of
complex physical systems composed of parts, systems which are such that at least some of their parts
may exist outside of those complex systems. We may deepen our positive conception of
emergentism by bringing to the fore some claims that emergentists deny.
32 See especially Broad (1918-19; 1925), Goudge (1967), Lovejoy (1927), and McLaughlin (1992). 33 I argued in the last chapter that teleology was one such property. See also Plantinga (1993) and Bedau (1992d). In the philosophy of mind there exist arguments for the sui generis status of intentionality (see, for example Bealer 1993a; Searle 1980; 1983) and especially qualia (see, for example Bealer 1994; Jackson 1986). 34 See Philip Pettit (1993) for an exposition of the traditional view. I discuss features of Pettit's account below. 35 Emergentism is thus closely associated with holism. Healey (1991) and Teller (1986) draw this connection. 36 The detailed work distinguishing varieties of emergence has been removed to the appendix.
228
According to Arthur Lovejoy, underlying the opposition to emergentism lies an
adherence to a medieval assumption about causality which he calls the 'preformationist
assumption'.
The preformationist assumption: "there cannot be more in the effect than there is in the cause." Alternately, the doctrine that an effect is not understood until and unless "the eye of reason could somehow discern it in the cause" (1927, 20).37
The preformationist assumption is a 'medieval' doctrine concerning causality in the sense that it
was refuted decisively in the modern era by Hume.
Hume argued that we cannot know a priori which things will or can be causally related to
which other things; there are no nontrivial a priori constraints on what two things may enter into
causal relations. Emergentists take this Humean point firmly to heart and affirm that there are
no a priori bars on what types of events, properties, or entities might emerge from the causal
interactions of complex groupings of micro-entities.38 We cannot know a priori that the
interactions of strictly physical properties and entities will not causally give rise to the existence
of sui generis mental properties such as consciousness or sui generis teleology, for instance. Further,
according to emergentists Hume's insight concerning causality reveals that there is no a priori bar
on whether these emergent properties may enter into primitive and novel causal relations with
other entities whether the causal relations are on the 'same level' (i.e., between two emergent
novelties) or 'downward' (i.e., between a cause which is an emergent novelty and an effect at the
level of its base properties and entities) or 'upward' (i.e., between a cause which is an emergent
novelty at level L1 and an effect which is an emergent novelty that itself emerges only from the
ordering of L1 properties into complex wholes).
37 Lovejoy is not alone among emergentists in drawing these connections. See also Goudge (1967) and Popper (1977). 38 Hume's arguments are in his (1975, §IV) and (1990, Book I.iii). As Hume says, we are in a "natural state of ignorance with regard to the powers and influence of all objects" when we consider them a priori (1777, §IV.ii.32).
229
From this brief overview, we may take the following core commitments of the
emergentist's ontology.
1. Emergentists accept an 'ultimate physical ontology'. "[Ultimate Physical Ontology] There are basic, nonemergent entities and properties, and these are material entities and their fundamental properties." (Kim 1992, 122)
2. The denial of preformationism (1): the affirmation of emergent novelty. Emergentists accept that the causal relations micro-entities enter into in certain complex wholes may generate emergent novelties in the sense that some properties of wholes may be irreducibly different in kind from the properties and entities which generate them and upon whose interactions they depend for their instantiation.
3. The denial of preformationism (2): the affirmation of 'downward' causation. Emergentists hold that novel emergent properties may enter into novel causal relations; specifically, while novel properties depend for their existence on the interactions of their base properties, they may nevertheless influence the course of lower-level events in ways compatible with their continuing to be instantiated.
These basic positions lead us to a definition of an emergentist ontology strong enough for our
purposes in this chapter.
Emergentist ontology: The ontological doctrine that there exist or may exist properties P of a structure X with components a1...an such that, (i) P properties depend causally for their existence on the interactions of a1...an in X; (ii) P properties augment the ontology of the world in the sense that they are not reductively identifiable with any of the properties or relations of a1...an in their interactions outside of structures of the same type as X; and (iii) P properties may have novel causal powers; the causal relations P enters into are not wholly identifiable with any of the causal powers and relations of a1...an or their aggregates outside of structures of the same type as X.
A few comments on this ontology are in order.
First, the emergentist ontology respects the findings of modern science. Emergentism
does not deny the fact or value of micro explanation in modern science. Emergentism accepts
that there exist no natural systems concerning which it will fail to pay dividends to seek a
reductive micro-level understanding of the behavior and being of those natural systems.
Emergentists do, however, deny the hegemony of mirco-level explanations at least in principle;
autonomous higher-level properties and causal relations are not incompatible with the pervasive
dependency of higher-level entities and structures on lower-level entities and structures.
230
Second, the ontology is controversial; the standard view of what science has revealed
about the physical ontology of the world is not an emergentist picture. This point is illustrated
amply by the acceptance of the reductionist's project by the overwhelming majority of theorists
working on the problem of teleology. As Elliot Sober says, "Function is a concept that should not
be taken at face value" (2000, 87).39 The assumption at play in such statements and
commitments is that since the micro-parts of entities with biological functions are composed
display no sui generis teleological directedness, the complex wholes made up of those parts must
not either. The emergentist's reply to this inference is that Hume's arguments concerning the
nonexistence of substantive a priori constraints on causality show the inference to be
ungrounded.
The emergentist picture goes against the grain of the standard view of the ontology of
the physical world in another way. On the standard view, micro-to-macro causation rules to the
exclusion of macro-to-micro causation. As Robert Klee says,
We find micro-explanation to be a powerful and impressive form of explanation. . . . [But w]e really have no established model of what a macro-determinative connection would be like. Direct determination from higher-levels to lower-levels seems somewhat mysterious when one attempts to construct a relatively precise scenario of the 'how' and the 'why' of it. (1984, 59-60)40
Contemporary views of the ontology of the physical world not only attempt as a matter of
principle to do without entities and properties that cannot be reductively identified with entities
and properties at lower physical levels, they also have a bar on the existence emergent causal
relations from higher levels back 'down' to lower levels. But again, the emergentist reasserts that
Hume has taught us that there is no a priori bar on the existence of such causal relations. As a
39 See also Buller (1999b, 6), Nagel (1979, 291), Godfrey-Smith (1993, 194) and Neander (1991b, 127). 40 See also Mark Bedau (1997, 377), quoted below.
231
number of emergentists emphasize, the world may quite possibly be more causally 'intertwined'
than the neat micro-to-macro exclusively picture maintains.41
Even opponents of this ontology tend to grant that it is a coherent description of one
possible way the world might be.42 The importance of this point for our purposes is that sui
generis teleology may find a natural and scientifically respectable home in an emergentist ontology
of roughly the stripe indicated above. On such a picture, sui generis teleological directedness is
taken to emerge from the causal interactions of micro-entities and states within organisms at and
above some level of organic complexity.43 Further, such directedness may influence the course
of events in the world at both the macro- and the micro-levels. The plant's directedness towards
flourishing may be causally efficacious in the turning of its leaves towards the sun given that
there can be no a priori bar on which kinds of things can causally interact with which other kinds
of things.44 The findings of science do not debar teleology from the world; certain philosophical
presuppositions of modern science do. That there is a coherent ontology for the physical world
which respects the substantive and methodological findings of contemporary science and has a
place for sui generis teleology establishes this claim.
Elucidating an ontology that is rich enough to accommodate Aristotelian teleology but
not so strong as to fail to respect the findings and methodologies of modern science45 does not,
41 For authors who stress this point see especially Broad (1918-19), Silberstein and McGeever (1999) and Humphreys (1995; 1996; 1997a; 1997b). 42 For example, Mark Bedau, who argues that we have no need for such a strong conception of emergence, notes that it is "logically possible" that some such ontology obtain (1997, 377). Robert Klee argues not that the ontology is incoherent, but that we lack any models for understanding its detailed workings and thus ought not accept it (1984, pp. 60-1 see below for a response to this criticism.). Jaegwon Kim also grants that the ontology is coherent; he simply argues that it is incompatible with certain things we know as a matter of contingent fact to be true of the world ( see also 1992; see especially 1993; 1999. I respond to Kim's arguments below.). 43 Of course it is possible that teleology could arise causally from complex interactions of inorganic materials, but I will stick to the case with which we are familiar. 44 Whether this sort of causal interaction would be utterly 'mysterious' and hence to be avoided, or whether we need some 'model' for how it might work are issues which are taken up below. 45 Again, this ontology is treated in depth on its own terms in the appendix to this chapter.
232
of course, establish that the actual world has such an ontology. Nevertheless, the theoretical
point has value, for the modern presumption against unreduced teleology is so strong that
establishing even this theoretical point is quite significant. In the next section I argue that the
ontology outlined here is not a mere theoretical fantasy but describes the ontology of the actual
world.
6.3 The ontology of the physical world.
I argued in the last section that certain modern philosophical presuppositions
concerning physical ontology generate the seeming incompatibility of teleology and modern
science; the results and methodologies of science are compatible with an ontology in which sui
generis Aristotelian teleology finds a natural home. In this section I argue for the stronger claim
that we have strong scientific evidence from quantum mechanics for the actuality of the ontology
necessary for the acceptance of sui generis teleology.
This is a crucial dialectical point in my argument for the acceptability of sui generis
teleology. There are a number of philosophical46 reasons to suppose that the ontology of the
natural world is emergentist: the continuing failure of reductivist accounts of teleology combined
with biology's thriving as an explanatory discipline that invokes functions argues strongly for a sui
generis realist conception of teleology and an ontology in which such teleology makes sense. In
the philosophy of mind, and in particular with regard to problems of mental causation and
freedom of the will the purported irreducibility of certain mental states and properties to physical
states and properties along with the robustness and explanatory value of our belief that mental
states are causally efficacious constitutes another philosophical argument for the actuality of an
46 As I note below, the reasons which follow need not be construed as 'philosophical' rather than scientific (i.e., conclusions from the sciences of biology and psychology). There is a strong tendency still alive in philosophy and contemporary philosophy of science, however, to venerate physics; in discussions of ontic emergence in psychology and in biology this tendency reveals itself in invocations of the incompleteness of our knowledge in these sciences in response to claims that such sciences may reveal an ontology richer than the ontology presupposed by physics and chemistry alone.
233
ontology where such facts make sense.47 Philosophical presuppositions concerning what science
has revealed about the nature of the world, however, stand in the way of wide dialectical
acceptance of these arguments.
In such a dialectical standoff, then, it is imperative to note that many contemporary
philosophers accept a broad thesis of the ontic authority of science; this commitment will prove
relevant to the dialectical acceptability of sui generis teleology and other purported emergents.
Hilary Kornblith expresses the general sentiment well.
The ontic authority of science: In metaphysics, I believe, we should take our cue from the best available scientific theories. As Wilfred Sellars so nicely put it, '... science is the measure of all things, of what is that it is, and of what is not that it is not.' . . .The task of the naturalistic metaphysician, as I see it, is simply to draw out the metaphysical implications of contemporary science. A metaphysics which goes beyond the commitments of science is simply unsupported by the best available evidence. A metaphysics which does not make commitments as rich as those of our best current scientific theories asks us to narrow the scope of our ontology in ways which will not stand scrutiny. (quoted from Kornblith 1994)48
Because of the wide acceptance of the ontic authority thesis by most of emergentism's
detractors, this thesis plays a central role in forming a dialectically sound argument for the
acceptability of sui generis teleology.
Of course, we might suppose that biology and psychology (not the philosophy of biology or
philosophy of psychology) provide arguments for ontic emergence on the basis of the facts
referred to above, and that given the ontic authority of science we therefore have reason already
to revise the philosophical commitments of our ontology of the physical world. As C.D. Broad
noted long ago, however, scientific arguments on the basis of biology and psychology may be
dialectically ineffective given the strength with which those presuppositions are held.49 Due to
the relative complexity and uncertainty pertaining to issues in biology and psychology arguments
47 See William Hasker (1999) for a development and defense of an emergentist solution to the problems of mental causation and freedom of the will. 48 See also Tim Crane (1994, 480), Alex Rosenberg (1996), Lynne Rudder Baker (2000, 23), and Richard Healey (1991).
234
based on findings in these fields are open to the charge that motives for postulating ontic
emergence are artifacts of our relative ignorance of the subject matter rather than a genuine
feature of the world.
The strategic best hope for an argument that our world's ontology is an emergentist
ontology roughly along the lines indicated above, then, comes from our most established science:
physics. Fortunately, there are good arguments that quantum mechanics reveals a world with an
emergentist physical ontology. If these arguments prove to be correct, then we have strong
dialectical ground to claim that the ontology sketched above is not merely theoretically
interesting but describes the ontology of the physical world. In that case, the philosophical
presuppositions that generate opposition to sui generis teleology require revisioning on the basis of
arguments wholly independent of concerns centering on biological teleology. There is strong
independent motivation for postulating the ontology teleology requires.
The strongest case to date for the existence of strong ontological emergence comes
from the science of quantum mechanics. Although admittedly I lack the expertise to evaluate
arguments for the existence of emergence from this area, a number of authors50 concur that the
best interpretations of our best theories point to the existence of instances of ontological
emergence along the lines indicated above. Here I will summarize only one of these arguments.
Silberstein and McGeever argue that certain quantum phenomenon51 � EPR Bohm
systems � have only three possible explanations, two of which are to be rejected on empirical
grounds, and that the remaining explanation postulates ontic emergence. In one such system,
quantum mechanics theorizes that should a single zero-spin particle decay into two spin-half
particles moving outwards in opposite directions, then conservation of angular momentum
49 Broad (1918-19, 87; 1925, 44); see also Humphreys (1997b, pp.4-5). 50 See Humphreys (1995; 1996; 1997a; 1997b), Teller (1986), Stapp (1993), Penrose (1995), Healey (1991), Silberstein and McGeever (1999), and Stairs (1990). 51 Paul Teller says that cases like these "inundate quantum mechanics" (1986, 76), and that emergent phenomena appear to be "all pervasive" in the field (83).
235
implies that the two particles' angular momentum must always equal that of the original particle
(zero), and so their spins must always be opposite.
Silberstein and McGeever argue that it is implausible in the extreme to suppose that the
particles could 'agree in advance' about their spin states given all possible future observation
conditions and that it is empirically implausible given general relativity to think that the particles
could exert nonlocal influence on each other at later times. They conclude that the quantum
system of the two particles displays emergent properties; "These systems exhibit correlation
properties that cannot be accounted for by local properties and dispositions possessed
individually by their parts.... The world cannot be governed solely by local [rather than
holistic/emergent] interaction" (1999, 188-9). The idea is that the 'higher level' EPR Bohm
system itself, the system that is micro-composed of the two half-spin particles, is a higher-level
entity that determines the lower level states of its component parts. Rather than the state of the
system as a whole being wholly determined by the intrinsic local features of the half-spin
particles that go to compose the system, the system itself is a whole which determines some of
the states of its component parts.
If such cases present sound interpretations of good scientific theory, then we have
strong evidence for the actuality of ontological emergence. If quantum mechanical arguments
such as Silberstein and McGeever's are correct, then
The best interpretation of our best science tells us that the properties of things in the world may not be fixed absolutely with respect to some unchanging space-time background, but rather that these properties arise from their interactions with and relationships among the other things in the world. . . . It is perhaps therefore necessary to seek non-reductive explanations... of some phenomena, not merely as a function of ignorance, but because the phenomena in question truly are ontologically emergent. (1999, 199)
The upshot is that not only is an emergentist ontology coherent and logically possible but we
have strong evidence that this ontology is instantiated in the actual world.
Given the commitment of dialectical opponents of emergence to the ontic authority of
our most developed science (i.e., physics), this argument should exert strong dialectical pressure
236
to revise the presuppositions concerning the ontology of the natural world such that we may
accommodate the arrival of new states and systems from the interactions of lower-level
component parts and, further, the possibility that these higher-level systems might exert causal
influence 'downward', effecting the states of their component parts. There are strong grounds52
to suppose that the causal structure of the world is not neatly organized along the lines of the
micro-to-macro only model, but is much more 'entangled'.
One important feature exemplified by this example of ontological emergence is the entangled nature of parts and wholes. The kind of emergence found in QM [quantum mechanics] and quantum field theory completely explodes the ontological picture of reality as divided into a 'discrete hierarchy of levels'. . . . It would seem that the world is much more complex and intertwined than the current cartography of scientific disciplines leads us to believe. (Silberstein and McGeever1999, 189)53
Insofar as we want to reflect in our ontology the commitments of our best science, it appears
that we ought to accept that our world's ontology is emergentist in at least some of its respects.
This means, of course, that the ontology of our world has room in it for sui generis Aristotelian
teleology, as I argued in the previous section. We have evidence, then, from our best science
that not only is the ontology of the physical world as revealed to us by modern science not in
conflict with the postulation of sui generis Aristotelian teleology but that it in fact has a natural
place for teleology.54
52 Richard Healey estimates the strength of evidence in favor of accepting ontic emergence on the basis of quantum mechanics as follows. (Healey's use of 'holism' in what follows is very similar to the broad conception of emergence I describe above.) Healey says, "the metaphysician should not use controversy about the correct interpretation of quantum mechanics as an excuse for not investigating the implications of holism and nonseparability; for quantum mechanics comes about as close to demonstrating holism and nonseparability as it does to establishing indeterminism" (1991, 394). 53 See also Humphreys, esp. (1995). 54 It is worth noting one possible response to this argument. Philip Pettit has given a definition of physicalism according to which physicalism is committed (since physics itself is supposed to be committed) to there being no irreducible high level properties with irreducible high level causal powers. All the primitive entities, properties and causal relations are micro-physical on this view. However, Pettit does believe that physicalism is consistent with the claim that below a certain size level, "there is no single level of grain at which microphysical laws obtain hegemonically. Thus the fundamental microphysical laws may include some laws at a relatively smaller level of grain, some at a relatively higher" compatible with his definition of physicalism and its commitments (1993, 221n8). Thus, Pettit makes room in physicalism for the expression of novel causal powers at higher levels so long as all such entanglements and downward causation is isolated within the micro-physical realm. Thus, Pettit's definition would 'screen off' the results of the quantum
237
That there are quantum mechanical cases of strong ontological emergence does not, of
course, establish that there is also sui generis Aristotelian teleology, but we have already seen that
there is strong and persistent biological evidence for real teleology in the world and that
reductive explanations of the phenomena fail. Given this state of affairs, it seems reasonable to
put forward the hypothesis that we should indeed take teleological phenomena seriously, and we
should rethink the main objection to robust interpretations of these phenomena � the objection
that sui generis Aristotelian teleology is ontologically suspect in the context of the ontology
revealed by modern science. It has been assumed that the ontology of science can allow only
efficient causality, but the quantum mechanical cases of strong ontic emergence (if successful)
show that this ontology needs supplementing in exactly the way required to make sense of the
causal influence of sui generis Aristotelian teleology. Far from it being the case that modern
science stands in the way of an ontology of strong ontological emergence, it would appear that
our best, most fundamental and precise science indicates that the ontology emergentism requires
is the actual one.
If the case for emergence among quantum phenomena is sound, then we have strong
reason to believe that the automatic and unthinking modern rejection of sui generis teleology is an
artifact of philosophical presuppositions about the ontic structure of the physical world which
simply do not obtain. This is strong reason to rethink and re-evaluate the place of sui generis
teleology in the modern world. Given the failure of reductivist accounts of teleology (argued in
chapter five) and the persistence of strong reasons to be realists concerning biological teleology,
mechanical argument given above, preventing them from being applied in 'macro' cases such as teleology or the mind. (Jaegwon Kim likewise attempts to allow the emergence of genuine and causally efficatious novelty � everything the emergentist desires � only at the micro-level. See Kim (1998, 116-8).)
The thing to note about this strategy, however, is not its illegitimacy but its arbitrariness. We could postulate that emergentist ontology held only below a certain size level, but the claim must appear ad hoc and unprincipled. Indeed, Pettit recognizes as much when he notes that an emergentist challenge to physicalism in his sense would be "much more challenging" than standard objections, and would "go to matters of deeper metaphysics" (217). What these deeper matters are, or how we might evaluate them, are issues Pettit does not address. We are thus left to our own devises in evaluating the plausibility and
238
evidence from physics for the actuality of an ontology in which sui generis teleology finds a natural
home is a boon indeed.
6.4 Naturalism.
Since the thesis of this chapter � that there is a place in the contemporary scientific
worldview for sui generis Aristotelian teleology � is, I recognize, heterodox, it is worth pausing to
re-emphasize the firm commitment of this position to the findings of modern biological science.
Indeed this is one respect in which I follow the reductivist tradition concerning biological
teleology completely. The main motive for popular realist but reductivist accounts of biological
teleology is the acceptance and centrality of biological teleology in tremendously successful
contemporary evolutionary biology.55 There are further substantive senses, however, in which
this account is firmly naturalistic in its ontology and methodology.
The ontological story told here is one which meets the negative ontological constraint
on naturalistic accounts that they not be committed to super-natural entities or forces, or
nonphysical substances such as 'entelechies'.56 Sui generis teleology is, on the picture sketched
above, just the emergent product of causal relations between natural things and it is hence, itself
just another natural feature of the actual world. The account also satisfies the strong
methodological constraint on naturalistic theories that they endorse only the methods of
knowledge acquisition employed in the sciences, for my account is, again, firmly based on the
findings of quantum mechanics and the supposition that biologists need to be realists about
coherence of an emergentist ontology suitable to teleology. I survey a number of possible objections to the account below. 55 See Buller (1999b, 6), Hull (1974, 120), Woodfield (1976, 32, see also p.1), Nissen (1997, vii), Allen and Bekoff (1995, 244), Neander (1991b, 127; 1995, 227), Bigelow and Pargetter (1987, 100), Allen, Bekoff, and Lauder (1998a, 1-2) and Dretske (1988, 63). 56 See McGinn (1991, 87), Katz (1998, 12) and Post (1995).
239
teleology combined with the fact that reductive accounts of their commitments fail.57 To invoke
sui generis teleology, then, is not to invoke, in David Chalmers' terms, "the forces of darkness"
(1996, 128).58
6.5 Objections to strong ontological emergence.
The picture I have presented thus far has been extremely optimistic. Not surprisingly,
however, ontological emergence of the sort I have been explaining and advocating is not without
its critics. The body of this chapter concludes with a consideration of a number of the most
common objections to ontic emergence; these objections consist of off-hand dismissals,
objections to the 'mystery' of downward determination, objections stemming from the supposed
causal closure of the physical world, and finally, the supposed incompatibility between scientific
methodology and ontic emergence.
6.5.1 Off-hand dismissals.
[T]here might be extra, irreducible external relations, besides the spatiotemporal ones; there might be emergent natural properties of more than point-sized things. . . . But if there is suchlike rubbish, say I, then there would have to be extra natural properties or relations that are altogether alien to this world. ( quoted in Humphreys 1996, 63; Lewis 1986b, x)
In a certain style of contemporary philosophy, the bold and dogmatic statement of a
particular form of orthodoxy � like the one quoted in the epigraph � has become distressingly
57 For this constraint see BonJour (1998, 69) and Katz (1998, 12). Of course, that the account satisfies the constraint need not constrain others � nonnaturalists in epistemology � from accepting the result; the account is not the exclusive possession of naturalists in epistemology. 58 Chalmers invokes the 'forces of darkness' in defending his property dualism from the charge that fails to be naturalistic. His naturalism is stronger than the version formulated here in the (dubious) sense that it presupposes the ontology of the physical world this chapter seeks to call into question. It is interesting to note that the naturalism with which the view advocated here is compatible is stronger than Lynne Rudder Baker's 'broad naturalism'. On that view, science is not the only arbiter of ontic authority � Rudder Baker argues that we ought to take the ontic commitments of common sense seriously (if critically) as well.
240
common. Charitably interpreted, such bold statements are read as nothing more than strong
statements of commitment to a particular view's truth, and taken in this way they constitute no
objection or impediment to the view presented in this chapter � whatever objections are to be
offered must be offered elsewhere than even the gruff and seemingly dogmatic statement of a
position.
In practice, however, statements such as these often seem to function as 'argument
stoppers'59 which mark out territory that will not be questioned by another, and � the gruff
realism of the rhetorical tone implies � ought not be questioned by those who wish to deal with
the world as it is rather than as they wish it to be.60 When used or read in this way, such
statements are themselves taken to offer reasons against the views which are rhetorically
undermined.61 Nevertheless, there is no argument contained in these denials, however gruff. If
there is to be a real discussion of objections to strong ontic emergence, it must move beyond
these highly rhetorical and prejudicial opening moves.
6.5.2 Is strong ontic emergence 'too mysterious'?
Although strong emergence is logically possible, it is uncomfortably like magic. . . . such causal powers should be quite unlike anything within our scientific ken. (Bedau 1997, 377)
While I find this aspect of her 'broad naturalism' congenial, my point here has been to stress that the ontology and methodology required to arrive at sui generis teleology naturalistic in quite stringent senses. 59 The phrase is John Mullens (1995, 10). He defines an 'argument stopper' as "a verbal response to argumentation that is intended to, and has the effect of, ending rational debate." 60 As William Hasker comments, "such dogmatism is hardly rendered benign by the fact that it is fairly widespread in the philosophy-of-mind community" (1999, x). 61 Richard Spencer-Smith (1994, 119) and Mark Bedau (1997, 377) employ �arguments� of this sort as reasons to reject emergence. Each grants the coherence of the view, but either declares it refuted by empirical data which are nowhere cited (Spencer-Smith) or declares it irrelevant (Bedau). Such proclamations can carry no weight in serious debates over the acceptability of emergence; some substantive, publicly available and evaluable reason must be offered if progress is to be made.
241
As I have argued above, the claim that ontological emergence is unlike anything in our
scientific ken is false of quantum mechanics, and � more controversially but more to the point
for this dissertation � false of biology as well. Nevertheless, the fear may remain that strong
ontic emergence is 'uncomfortably like magic.' One response to such fears (sufficient on its
own) would be simply to point out that in the face of the reputable scientific evidence in favor of
ontic emergence these fears should go the way of our intuitions about absolute space in a post-
Einsteinean world. We need better reasons than this to oppose the conclusions and explanatory
tools of scientific work in quantum mechanics and biology. Despite the adequacy of this
response, it is possible to move the issue further forward before employing it.
In the epigraph Bedau has particularly in mind purported cases of �downward� or
�macro-to-micro� causation; it is supposed causal relations of this sort which are said to be
�uncomfortably like magic.� The impression that downward causation is somehow illegitimate
may be fostered by the idea that we understand upward (micro-to-macro) determination better
than the downward variety of determination. Here I quote an extended and informative passage
by Robert Klee on this matter.
We find micro-explanation to be a powerful and impressive form of explanation. Micro-explanation is powerful in virtue of the fact that when a level of organization within a system can be explained in terms of lower-levels of organization this must be because the lower-levels (i.e. micro-properties) determine the higher-levels (i.e. the macro-properties). This is why micro-explanation makes sense�the direction of explanation recapitulates the direction of determination. The intuition behind this is that when we have something explained to us we understand it, and a large part of understanding something is knowing how it is determined. What this comes to, among other things, is that a micro-explanation (if it is a good one) provides us with the effective means or mechanism by which the higher-level is determined or produced out of the lower levels. . . . How might this notion of a mechanism, or effective means, help us to evaluate macro-determinative emergence? Most mechanisms of determination that current scientific theories acknowledge indicate that the direction of determination is micro-to-macro. . . . We really have no established model of what a macro-determinative connection would be like. Direct determination from higher-levels to lower-levels seems somewhat mysterious when one attempts to construct a relatively precise scenario of the 'how' and the 'why' of it. (Klee 1984, 59-60. Italics added)
242
This rich passage well reflects the worry that 'downward causation', as I have been calling it, is
too vague or mysterious to engage our scientific interests. What is the substance of Klee's
charge?
Klee argues that explanation proceeds (in large part) through elucidating determinative
or causal relations between things and events. We want not just formal understanding in the
sciences, but also an understanding of the causal structure of the world. Micro-explanation
satisfies in this regard. We already understand, Klee claims, how the micro determines the
macro, whereas with purported cases of macro-determination we have no model of the
mechanism by which the determination is supposed to work, and the lack of a model makes the
determinative relation look "somewhat mysterious."62
Note, however, that no model or mechanism for understanding micro-causal explanation
was offered by Klee. Micro-explanation is accepted without a model; insofar as a model is
employed by Klee it consists simply in making explicit various links in the micro-to-macro causal
story. The individual upward links, however, are accepted as basic and explanatory because they
are so well attested as causal determination relations for which no model or mechanism can in
62 The accusation that macro-to-micro causation is �too mysterious� to be countenanced in our ontology has obvious parallels to the standard objection to dualistic interaction. Mental/physical interaction is taken to be �too mysterious� and in need of some sort of explanation before we can accept it. But, as William Hasker points out, "This argument may well hold the all-time record for overrated objections to major philosophical positions" (1999, 150). Crane and Mellor diagnose the problem with the objection thus, "It is indeed an old thought that mental causation is hard to make sense of, and especially causation linking the mental to the non-mental, because they seem to be so different. But why should that impress anyone who has learned from Hume that causation never 'makes sense': that it is always a matter of fact, not of reason? Nothing in either Humean or other modern analyses of causation forces causes to be like their effects; nor does anything in them stop causes and effects being mental" (1990. 192). See also E.J. Lowe (1996, 52), Arthur Lovejoy (1927, 20) and C.J. Ducasse (1961). This response transfers without loss as a response to the mysteriousness objection offered against macro-to-micro causation.
243
principle be offered.63 Being a determinative relation accepted in well founded scientific theory
itself justifies accepting this sort of determination.64
Of course, advocates of 'downward causation' maintain that it is a well attested form of
determinative relation and that in the absence of principled objections to accepting such relations
as genuine, we ought to accept them on the strength of their role in scientific theory. But of
course, claiming to find a causal relation �too mysterious� does not itself constitute a principled
objection; if we are to be given reason to reject downward determination opponents must do
more work than this.
In sum, Klee places a condition on the legitimacy of macro-to-micro causation that the
micro-to-macro case was not required to meet. This condition is illegitimate, however, as Hume
has taught us. No empirically discovered causal relation must pass before the bar of a priori
reason�s demand that it be made �transparent� to the mind before we may accept it into our
ontology.65 No causal relation, upwards, downwards, or horizontal is anything but opaque to
our reasoning. The invocation of mystery by Bedau and Klee is wholly out of place (or, rather,
time) in a post-Humean context.
63 C.J. Ducasse says, �[T]he objection that we cannot understand how a mental event could cause a physical one (or vice versa) has no basis other than blindness to the fact that the 'how' of causation is incapable at all of being either mysterious or understood only in cases of remote causation, never in cases of proximate causation. For the question as to the 'how' of causation of a given event by a given event never has any other sense than through what intermediary causal steps does one cause the other.� (1961) 64 Perhaps there is simply an ambiguity at play here in �determination�. As the emergentists use the word, macro-to-micro determination is a causal relation; macro events are related as causes to micro events as effects. Often, however, micro-to-macro determination may be thought of not as a causal relation at all, but as a logical or supervenience relation, as when we say that the dots in a dot-matrix picture determine the picture�s qualities. This sort of aggregative determination is, perhaps, well understood in a way that the emergentist�s downward causal efficacy is not, but if this is what Klee has in mind then he has simply failed to address the emergentist�s position. Emergentists do not conceive of downward causation on the model of supervenience, they understand it as an instance of ordinary causal relations between things and events in the world. 65 As John Foster says, "why should any explanation be demanded?. . . . Why should he [i.e., the dualist in this case] be called on to offer a deeper mode of explanation than that which is available to physical science?" (1991, 160-1)
244
6.5.3 The causal closure of the physical.
One might object to the emergentist ontology elucidated here on the grounds that it
involves a violation of the causal closure of the physical. The idea behind causal closure is the
supposition that it should be possible to give a complete account of the causal structure of the
world in the terms of physical causes. Thus, the causal closure of the physical maintains
(roughly) that for any event E that has a cause we can cite a physical cause, P, for its happening,
and that citing P explains why E happened. Further, for any event, E, once we have cited all of
E's physical causes, P1...Pn, we have cited all of E's causes simpliciter; there are no nonphysical
causes of E.66
Doctrines that embrace strong ontological emergence may appear to violate the
principle of the causal closure of the physical, but it is not clear that all versions of such a theory
will do so. Thus, if quantum mechanical phenomena exhibit strong ontic emergence then it
seems plausible to suppose that both the properties and entities in the emergence base and the
properties and entities which emerge from the interactions of the base properties and entities are
physical, and their causes and effects are all physical. It would appear, then, that there is no
incompatibility between the causal closure of the physical world and ontological emergence.
Nevertheless, while ontic emergence and causal closure may be compatible, it is surely
the case that if sui generis Aristotelian teleology emerges, there is some reason to believe that the
principle of the causal closure of the physical has be violated. Surely teleological causes are not
physical causes, and thus cannot be accommodated into our ontology on the supposition that the
physical world forms a causally closed system with no �outside� nonphysical influences.67
66 Jaegwon Kim says, "One way of stating the principle of physical causal closure is this: If you pick any physical event and trace out its causal ancestry or posterity, that will never take you outside the physical domain. That is, no causal chain will ever cross the boundary between the physical and the nonphysical." (1998, 40). 67 Note of course, that on the emergentism presented here teleological causes can only be �outside� the physical in an extremely weak sense if at all. On this view they are not supernatural properties or entities added into the causal matrix from without for free, like supernatural beings or entelechies. Rather, the teleological properties of things in virtue of which they direct aspects of their behavior are themselves the
245
I have two responses to this worry, developed in the next two sub-sections. First, I
question the evidence in favor of causal closure; how strong are our reasons for accepting causal
closure? Second, the principle of the causal closure of the physical can be no more substantive
than our independent conception of what it is to be physical. There can be no principled
objection to purported nonphysical causes unless we have a principled (i.e., not merely stipulative
or question begging) distinction between the physical and the nonphysical.
6.5.3.1 Evidence for causal closure.
Evidence cited in favor of the causal closure of the physical consists, generally, in a
vague reference to the enormous success of the physical sciences in the modern era and what is
claimed to be a sound inductive inference from that success to causal closure. How strong is this
argument?
Not terribly. Inductive inferences which extrapolate our experience in one domain X
into another domain Y are only as good as our reasons to believe that domains X and Y are
relevantly similar. Thus, famously, medical science once generalized the results of studies
conducted solely on men to women � we are all human after all. But this was a hasty,
dangerous, and as it turns out ill-founded generalization in many crucial cases. Bearing this sort
of general issue concerning inductive inferences in mind, the case for causal closure is
extraordinarily weak.
I grant of course that physics and chemistry have been enormously successful using
reductive models of micro-to-macro causation invoking only efficient causes. Nevertheless, to
suppose that this success transfers straightforwardly into the extremely controversial domains of
biology and psychology would be rash. I have argued in the last two chapters that there are no
acceptable reductive accounts of biological teleology and that there does exist an ontology that
we have strong reason to believe is instantiated in this world within which sui generis teleology
causal result of interactions among physical things; they are a part of nature, not outside of it. See pp. 238-239.
246
finds a natural home. But if this is correct, then the widely shared grounds for being realists
about teleology themselves weigh against the supposed inference to causal closure: biology
appears to require sui generis teleology and the success physics and chemistry have enjoyed
without it is neither here nor there given the principled differences between the two domains.68
My point is the simple one that we cannot suppose at this stage in the argument that reference to
causal closure will refute sui generis teleology, for the grounds for the inference to causal closure
are themselves undermined by the arguments thus far offered in favor of accepting sui generis
teleology.
It is important at this stage of the argument to keep in mind that this conclusion is not
as heterodox as it might seem. On the view presented in this chapter, sui generis teleology
constitutes a natural part of the physical world and not a supernatural or mysterious and
independent new entity added into the world �from without�. Sui generis teleology is the causal
result of the interactions of physical things as they behave in complex organized systems; their
appearance is no more mysterious or unnatural than any effect of any cause given the opacity to
reason of causal relations generally. To undermine the grounds for causal closure, then, is not (in
at least this instance) to invoke supernatural causal factors independent of or in conflict with
physical causes, nor is it to undermine the results of science. In fact the view attempts to fully
respect the grounds for being realists about teleology that arise out of reflection on
contemporary biology.
6.5.3.2 What system is causally closed when the physical is causally closed?
Thus far I have argued that the inference to causal closure is only as strong as the claim
that the success of physics and chemistry transfers without hitch to biology and psychology. The
argument of the last two chapters, however, undermines this claim, and causal closure therefore
cannot be invoked against sui generis teleology at this stage without some substantive objections
68 Similar arguments for the sui generis status of causally efficacious mental properties likewise weigh in
247
independent of the claim of causal closure itself. We may see that causal closure fails to provide
a principled independent argument against emergent teleology in another way, however, for the
principle of the causal closure of the physical is itself only as substantive as the notion of the
physical that it invokes.
A priori definitions of the physical are out of vogue;69 contemporary definitions of the
physical are almost exclusively formulated with reference to the principle of the ontic authority
of science. According to the thesis of ontic authority (see above, p. 233), we ought to accept
into our ontology whatever completed science requires. Of course, on this view of the physical
sui generis teleology poses no threat to the causal closure of the physical for I have argued that
biology needs sui generis teleology. If �physical� is understood to apply thus without any a priori
constraints on what science might one day determine to be necessary, then this position is not in
conflict with the causal closure of the physical. Again, we require substantive objections to the
ontology independent of bare claims concerning causal closure if sui generis teleology is to be
threatened.
Nevertheless, there is a widely shared intuition that something has gone terribly wrong
with any conception of the physical on which sui generis teleology or mental properties count as
physical properties. Recognizing this, many theorists attempt not to give a priori positive
descriptions of the physical, but rather to place a priori constraints on what is excluded from the
physical: if there is sui generis teleology, then it is a nonphysical property, for instance. Of course,
if we simply stipulate the teleological out of the physical in this sense, then once again the
principle of the causal closure of the physical cannot be invoked against the emergentist
teleology of this chapter, for no such stipulative terminological move may carry substantive
against the inference to causal closure. 69 We may think, for example, of definitions of the physical in terms of what possesses extension, has mass, solidity, excludes other objects, etc.
248
weight. If this conception of the physical constitutes our whole grounds for objecting to
teleology, we have been given no substantive reason to give it up.
Given this, we might try to define the physical such that it applies primarily to things
which are very small (i.e., atoms and smaller, perhaps) and those things which are composed out
of small things. Philip Pettit (1993) has formulated a definition of physicalism (the claim that
everything that exists is physical) on this model. Pettit claims that his definition gives substantive
content to the notion of being physical; it would be false, for instance, if there were sui generis
teleology. Perhaps if we understand the physical in Pettit's way, then, we can generate a
substantive case from causal closure against emergent teleology.70
This strategy cannot succeed, however, for while Pettit�s definition does rule out sui
generis teleology, it does so (again) by stipulation. The problem for our purposes comes to the
fact that Pettit must specify what it is for a thing to be composed out of physical things and still
to be physical in his restrictive sense. If Pettit's constraints were too weak, then teleology would
count (again) as a physical cause in virtue of its constitution base in the physical. Pettit excludes
teleology, then, by stipulating that on his version of physicalism composition relations must be
�noncreative�, and by further stipulating that such macro-level laws as might exist all be wholly
derivative upon micro-level causal laws � Pettit excludes emergent novelties and emergent
causal influence from physicalist ontology by stipulation. While this is a perfectly legitimate move
to make in defining the physical, it cannot form the basis for a principled objection to
emergentist teleology for that sort of teleology is ruled out of the physical (again) merely by
definitional stipulation. Such terminological manuevering may give substance to the thesis of
physicalism, but it cannot be invoked as the basis for a principled objection to sui generis teleology
on the basis of the causal closure of the physical.
70 Ignoring, of course, the objections of the previous section.
249
6.5.3.3 Section summary.
The principle of causal closure, then, cannot form a principled objection to sui generis
teleology as conceived here. The inductive inference to causal closure is weak in the absence of
independent objections to the grounds for sui generis teleology offered in the last two chapters
and there does not exist a substantive conception of the physical which could form the ground
for a principled objection to positing a sui generis teleological cause. Instead, the primary move in
excluding teleology from the realm of the physical is to invoke a priori constraints on the nature
of the physical and to define causal closure on the basis of those a priori constraints. But, given
Hume's argument that all causal relations are opaque to us we cannot infer on the basis of
supposed a priori constraints on what can cause what that teleology is not a real causal factor in
the structure of the world. If we respect the ontic authority of what science has revealed to us a
posteriori then there is no objection to accepting sui generis teleology as a physical cause. If we
impose an a priori constraint on what may count as a 'physical' cause then we lose the basis for
any substantive objection to postulating sui generis teleology.
6.5.4 Methodological worries about emergence.
Perhaps, however there are valid worries from the domain of scientific methodology
which require us to forego emergentist explanations. Macro-micro and micro-macro
explanations are not on a par not because one is inherently less well understood or legitimate
than the other, but because one is suspect on methodological grounds.
Thus, one possible objection to an emergentist view is that it will lead to a complacency
about scientific explanation on the part of those who accept the ontology. One of the chief
reasons, it will be claimed, for the stunning success of science in the modern period is the fact
that scientists have sworn themselves off from �easy answers� to complex problems about the
nature and operation of the physical world and have continually pushed themselves to
understand the world strictly in terms of physical causes. An ontology of strong ontological
emergence that accepts with equanimity such suspect entities as sui generis teleology and/or
250
mental properties can only stand in the way of genuine scientific progress by encouraging lax
pseudo-explanations in sciences where genuine (but difficult to discover) micro-explanation has
much to tell us.
I believe that there are two proper responses to this objection. The first is simply to
reiterate the point made above (see p. 220) that it is simply not the case that allowing emergent
properties into ones scientific ontology requires abandoning genuine and systematic scientific
inquiry. The assumption that it does can only rest on selective attention to the facts or the
unfounded assumption that all emergentist explanation must be vacuous. But invocations of
biological teleology are far from vacuous. The history of biological investigation of functions
reveals unquestionably that serious investigation is needed for the discovery of numerous
biological functions, and none of this work is undermined in either substance or methodology if
we recognize teleology as strongly ontologically emergent. Similar remarks may be made about
the neurophysiological investigation of the mind and its relation to our first-person experience of
the world. Further examples of good scientific work yielding emergentist results are available if
the argument above concerning quantum mechanical examples of strongly emergent phenomena
are sound. Indeed, this last example, if defensible in the long run,71 ought to provide conclusive
reason to dismiss this objection as unfounded. Good science, good physics, is compatible with
ontic emergence; quantum mechanics and biology prove it.
There simply is then no incompatibility between good science and ontological
emergence. To claim that accepting an emergentist ontology must mean the abandonment of
serious scientific inquiry is either to ignore these stunningly successful sciences or to assume in a
question-begging fashion (in this context) that those sciences investigate and uncover only
properties strictly reducible to physical properties and events. To suppose that the methods and
successes of these scientific investigations must be abandoned because of the adoption of a
71 See again Richard Healey's estimation of the strength of evidence quantum mechanics provides for ontic emergence, quoted above in n. 52.
251
particular ontological view of the underlying facts of the matter is utterly unfounded and fails to
take seriously the ontological presuppositions of the standard view of scientific method. The
presuppositions of emergentist ontology are not stronger nor more extravagant than the
presuppositions of the standard view � they are simply different.
Second, the objection appears to assume a doctrine of the unity of scientific method and
investigation which is highly questionable at best. Certain scientific disciplines are committed by
their subject matter to a continual effort to seek further micro-physical causes for macro-events,
and there is no reason for a strong emergentist to suggest any changes in the methodologies of
such inquiries. Nevertheless, there is no reason to believe that all genuine science must proceed
under the same (microphysical) explanatory constraints, and the doctrine of strong emergence
articulates one way in which such explanatory and scientific pluralism makes sense.
In a world where causal interactions between higher and lower levels are �entangled�72
rather than being strictly uni-directional the door is open to a realist causal basis for different
methods of investigation. While microphysical investigations have been extraordinarily fruitful
and it would be foolish to suggest that we discontinue seeking them, that rash proposal is not on
offer. One of the advantages of the emergentist position is that it honestly seeks to acknowledge
the strong basis we have for our beliefs about the fundamental levels of physical reality and our
ability to understand macro events through micro causes. The emergentist simply denies that the
success of microexplanation either entails or even strongly suggests the falsity of views on which
there exists greater ontic and causal heterogeneity in the structure of the world than a simple
micro-to-macro reductivist model of science would suggest.
I conclude that the objection that a strong emergentist ontology is suspect on grounds
of scientific methodology is either emperically falsified or question-begging. It is empirically
false if the objector disavows the evidence of quantum mechanics and biology. It is question-
72 The expression is Silberstein and McGeever�s (1999, 189, quoted above).
252
begging if the objector assumes that the explanatory successes of biology and quantum
mechanics can only be understood on the reductive micro-to-macro model. We all grant the
methodological power of the scientific method in biology and quantum mechanics, but we
cannot assume that the correct ontology underling this method can only be the reductivist one
traditionally presupposed. Sound scientific practice is not wedded to the idea that the world
must be ordered only according to micro-to-macro determination rather than being more
heterogeneous.73
6.5.5 Conclusion.
In this chapter I have argued that there exists an ontology of the physical world that
enjoys independent scientific and philosophical support and that sui generis Aristotelian teleology
both makes sense within this ontology and may play a robust role in the causal structure of the
world. I have attempted to defend both this conception of teleology and the ontology that
underwrites it through positive arguments in favor of adopting such an ontology and through
defenses against traditional and contemporary objections both to teleology and to emergentist
ontology. I conclude that there is a scientifically and philosophically respectable place for sui
generis Aristotelian teleology in the modern world.74
In the next chapter we return to the question of Aristotle on the nature of life. In
chapter two I hinted that Aristotle may define life in terms of teleology. In the intervening
chapters I have investigated both Aristotelian teleology and the prospects for teleology in
contemporary science in a preliminary investigation of the viability of such a conception both as
an interpretation of Aristotle and as a contribution to contemporary concerns about the nature
of life. Having defended the theses that the scope of Aristotle's teleology presents no problem
for an analysis of life in teleological terms, that Aristotle's teleology is a sui generis causal factor in
73 See Broad (1918-19) and Humphreys (1995). 74 This conclusion is buttressed further in the appendix which follows. In the appendix I deepen the picture of emergentist ontology necessary to support the claims of this chapter.
253
the world, and that such a conception of teleology has a place in a modern conception of the
world, I now return to the task of elucidating and defending Aristotle's conception of life in
teleological terms.
254
Appendix to Chapter Six: The ontology of emergent properties and laws.
This appendix constitutes an extended defense of my claim that emergentist ontology is
coherent. Below, I survey the history of debates concerning emergentism, formalize a definition
of emergence strong enough both to do the work required in this chapter and to make clear what
the view is � and is not � committed to.
Some terminological confusion is bound to arise given contemporary uses of related
terms, and I will attempt to clarify and distinguish my view in what follows. Richard Spencer-
Smith points to the fact that 'emergence', as it is invoked in the philosophical literature, can
express both a reductionist and an anti-reductionist sense1 and compares such a situation to
there being a linguistic community which uses the same word "as a synonym both for 'black' and
for 'white'" (1994, 113).2 As a number of authors point out, the term 'emergence' also seems to
be used in both epistemological and ontological senses.3 All of these senses need distinguishing
and clarifying before we can move on.
6.6 The core sense of emergence.
As with so many concepts, emergence is best introduced through an explanation of what
it denies. Emergence is associated with the vague claim that a whole may be greater than the
1 For deflationary or reductionistic accounts of emergence, see Spencer-Smith (1994), Bedau (1997), Klee (1984), and Newman (1996). Jaegwon Kim joins the reductionists in arguing against the coherence of stronger forms of emergence in a number of important papers; see his (1992; 1993; 1999). For anti-reductionists, see Broad (1918-19; 1925), Lovejoy (1927), Humphreys (1995; 1996; 1997a; 1997b), O'Connor (1994), Silberstein and McGeever (1999), Hasker (1982; 1999), Lowe (1996) and Searle (1992). 2 Unfortunately, Spencer-Smith's attempt to disambiguate and find the common ground between these two uses falls flat. On his view, "What entitles us to see the radical and interactional conceptions as aiming at the same kind of phenomenon is the idea of a property emerging at a higher level as a result of the behaviour of the lower level constituents" (1994, 122). This solution to the problem, however, merely equates the common core of the concept of emergence with the use employed by one camp: the reductionists. 3 See, for example, Silberstein and McGeever (1999).
255
'sum' of its parts, and contrasted with the denial of that claim.4 There are a number of ways of
cashing this claim out, and they will lead to various conceptions of emergence which need
distinguishing. Clearly, however, all doctrines of emergence will be dealing with properties of
complex physical systems composed of parts, systems which are such that at least some of their parts
may exist outside of those complex systems.5
According to Arthur Lovejoy, underlying the opposition to emergentism lies an
adherence to a medieval assumption about causality which he calls the 'preformationist
assumption'.
The preformationist assumption: "there cannot be more in the effect than there is in the cause." Alternately, the doctrine that an effect is not understood until and unless "the eye of reason could somehow discern it in the cause" (1927, 20).6
The preformationist assumption is a 'medieval' doctrine concerning causality in the sense that it
was refuted decisively in the modern era by Hume.
Hume argued that we cannot know a priori which things will or can be causally related to
which other things; there are no nontrivial a priori constraints on what two things may enter into
causal relations. Emergentists take this Humean point firmly to heart and affirm that there are
no a priori bars on what types of events, properties, or entities might emerge from the causal
interactions of complex groupings of micro-entities.7 We cannot know a priori that the
interactions of strictly physical properties and entities will not causally give rise to the existence
of sui generis mental properties such as consciousness or sui generis teleology, for instance. Further,
according to emergentists Hume's insight concerning causality reveals that there is no a priori bar
4 Emergentism is thus closely associated with holism. Healey (1991) and Teller (1986) draw this connection. 5 The next two paragraphs borrow introductory material from the discussion in chapter six. 6 Lovejoy is not alone among emergentists in drawing these connections. See also Goudge (1967) and Popper (1977). 7 Hume's arguments are in his (1777, §IV) and (1990, Book I.iii). As Hume says, we are in a "natural state of ignorance with regard to the powers and influence of all objects" when we consider them a priori (1777, §IV.ii.32).
256
on whether these emergent properties may enter into primitive and novel causal relations with
other entities whether the causal relations are on the 'same level' (i.e., between two emergent
novelties) or 'downward' (i.e., between a cause which is an emergent novelty and an effect at the
level of its base properties and entities) or 'upward' (i.e., between a cause which is an emergent
novelty at level L1 and an effect which is an emergent novelty that itself emerges only from the
ordering of L1 properties into complex wholes).
Emergence is also opposed to the thesis (�Humean supervenience�) that all the
properties of wholes are determined by the local intrinsic properties of micro parts.
Emergentists thus reject (what have been variously labeled) materialistic, mechanistic,
reductionistic, physicalistic or particularistic conceptions of complex systems. Again, quoting
Lovejoy:
The doctrine of particularism:8 (a) [T]he conception ... of natural events as combinations or rearrangements9 of relatively simple, preëxistent entities, of which the total number or quantity remains invariant, and of each of which the qualities and laws of action remain the same through all the combinations into which it may enter. . . . (b) in this second form10 the preformationist assumption implied a program of reduction or simplification; it. . . . means learning to see in the complex nothing but its beggarly elements�the meager qualities and limited repertoire of the simple, merely multiplied a certain number of times. (1927, 20-1)11
8 Lovejoy calls what follows the doctrine of physicalism. I have changed the name because 'particularism' is an apt phrase to describe the metaphysical claim indicated in part (a). As we will see below, it is an open question whether or not emergentist doctrines are 'physicalist' or 'materialist', and so these descriptions of the opposition would be inapt. 'Reductionist' may seem to indicate an explanatory tendency (although it also has a central metaphysical sense) and may for that reason be more misleading than 'particularism'. 9 A proper understanding of the phrase 'combinations or rearrangements' in this definition is crucial to marking off Lovejoy's intended meaning. Emergentists distinguish between properties of wholes that are 'merely resultant' and those properties that are genuinely ontologically emergent. Lovejoy's indication that the physicalist (or, as I am calling her, the particularist) accepts all natural events as 'combinations or rearrangements' of lower-level entities indicates, for him, that while the particularist need not accept that all properties of wholes are properties of parts, all properties of wholes are 'merely resultant' upon the combinations and rearrangements of those parts. I make this distinction clear below. 10 The first form of the preformationist assumption was a theological or supernaturalistic version which stated that there must be in the First Cause (i.e., God) of the universe all the perfections of the universe. 11 For a contemporary formulation of the metaphysical (rather that methodological) doctrine expressed here, see David Lewis' conception of 'Humean supervenience' in his (1986b, pp.ix-x). Richard Healey's definition of particularism (1991) is a stronger notion; the weakest forms of ontological emergentism are compatible with Healey's particularism.
257
In part (a) of this quote Lovejoy explicates the metaphysical view that will form the particularist's
core metaphysical thesis; in part (b) he points to a reductive explanatory methodology which
naturally accompanies that metaphysical view.12 Emergence is tied with the denial specifically of
the metaphysical view first explained, but will (on realist conceptions of explanation) find natural
ties with the denial of a universalistic reductionist methodology as well.
With this notion of the opposition to emergence, we may define a core conception of
metaphysical emergence � following Lovejoy � as the denial of the particularist's metaphysical
thesis. Quoting again:
The core sense of emergence: [The general notion of emergence] may be taken loosely to signify any augmentative or transmutative event, any process in which there appear effects that, in some one or more of several ways yet to be specified, fail to conform to the maxim that 'there cannot be in the consequence anything more than, or different in nature from, that which was in the antecedent'. (1927, 20-1, 22)13 14
This general notion of emergence can get cashed out in a number of different ways.
6.7 Epistemological emergence.
As originally formulated15 emergence was often explained by claiming that emergent
properties were 'not deducible' from complete knowledge of base properties in structures unlike
those in which the alleged emergent properties were found. Largely as a result of this
formulation of the doctrine of emergence, there has arisen a history of treating doctrines of
emergence as making epistemological claims about our epistemic limitations rather than
12 This explanatory methodology is well described as 'extreme analytical reductionism' by Mayr (1982, 61). See also Richard Healey's definitions of explanatory reductionism (1991, 398-9). 13 Jaegwon Kim says, "At the core of these ideas was the thought that as systems acquire increasingly higher degrees of organization complexity they begin to exhibit novel properties that in some sense transcend the properties of their constituent parts, and behave in ways that cannot be predicted on the basis of the laws governing simpler systems." (1999, 3) 14 I provide a more formal definition of this core sense of emergence below, see p. 278. 15 See especially C.D. Broad (1925).
258
principled ontological claims. Here is the definition of 'epistemological emergence' given by
Silberstein and McGeever:
A property of an object or system is epistemologically emergent if the property is reducible to or determined by the intrinsic properties of the ultimate constituents of the object or system, while at the same time it is very difficult for us16 to explain, predict or derive the property on the basis of the ultimate constituents. (1999, 186, see also 182)
The idea behind so called 'epistemological emergence' is that from the fact that we cannot
deduce the existence of the emergent property from its base properties it does not follow that it
cannot be so deduced by a being (perhaps God) with superior knowledge or capacities. That we
cannot discover the deduction does not entail that the deduction does not exist or even that it
cannot be found. Nevertheless, Silberstein and McGeever's definition is ambiguous in at least
three ways.
First, it needs to be made clear that an epistemological emergentist denies the stronger
claims of the ontological emergentist and wants to deflate those ontological claims. According
to the epistemological emergentist, our inability to reduce or explain emergent properties in
terms of their basal conditions indicates a failure in us and not a failure of reducibility or
explicability in principle. Second, there is a sense in which ontological emergentists are emphatic
not to deny that emergent properties are 'determined by' their basal conditions; claims of
ontological emergence are even compatible with the claim that the behavior of wholes with
emergent properties is wholly determined by the laws governing the behavior of the elements in
the basal conditions. This phrase is therefore both insufficient to distinguish ontological from
epistemological emergence and gives the misleading impression that 'upward determination' is
something ontological emergentists wish to deny. This is simply false. Finally, there are multiple
notions of reduction at play in the philosophical literature, and on some of them ontological
emergentists will accept that emergent properties are reducible while on other (stronger) notions
16 Mark Bedau makes more precise what it means for it to be 'very difficult for us' to explain such events in his endorsement of a weak (epistemological) form of emergentism in his (1997).
259
they will deny this claim. In our finished definition of epistemological emergence, then, we must
indicate that there is some suitably strong sense in which emergent properties are not reducible.
We may therefore offer the following definition of the doctrine of epistemological
emergence.
Epistemological emergence: The doctrine that the claims of metaphysical particularism are true, combined with the claim that there exist (or may exist) complex systems for which it is very difficult � perhaps even impossible � for us to discover how the properties of those systems can simply be rearrangements or recombinations of lower-level properties and entities.
This definition captures the core claims of the epistemological emergentist: the positive claim
that the particularist's metaphysics is correct, and the negative claim that for merely
epistemological reasons we may not be able to see that it is correct in particular instances.
While I introduced this section by noting that doctrines of epistemological emergence
can find historical roots in Broad's claims that emergent properties are not 'predictable' or
'deducible' from base conditions, I believe that there are legitimate historical grounds to take
such doctrines not to reflect the spirit of emergentism as originally formulated in the works of
Broad, Lovejoy and Morgan, and I believe that much terminological confusion would be avoided
in philosophical discussion if such a doctrine were not labeled a doctrine of 'emergence' at all.17
Genuine doctrines of emergence have always been linked to ontological claims about the
existence of fundamentally novel kinds of entities and powers in the world,18 and it can only
breed unnecessary confusion to label a deflationary physicalistic view a doctrine of 'emergence' at
all.19 The doctrine of emergence advocated in this chapter is a stronger, ontological version of
the doctrine that is more faithful to the intent � if not some understandings of the words � of
its originators.
17 For an example of such a doctrine, see Bedau (1997). 18 Kim (1999) is particularly clear on this very point.
260
6.8 Ontological emergence.
According to the weakest version of the ontological doctrine of emergence, physical
systems give rise to novel properties because of the causal powers of the subcomponents of those
systems as they express themselves in complex systems. Views committed to emergent novelty
are neutral over whether or not these novel properties have novel causal powers. Stronger
versions claim that the properties have new causal powers, and even stronger versions may claim
that these novel causal powers give rise to regularities in the behavior of systems which may be
captured in laws. Weaker versions, however, are compatible with an 'epiphenomenalism' about
emergent properties on which genuinely emergent properties exist but do no work in the causal
structure of the world.
Clearly if we are to understand these doctrines of emergence we need to understand
both what emergentists claim in saying that novel properties arise, and how novel properties
could have causal influence. I take these issues up in turn before pulling together the results of
our investigation of emergence.
6.8.1 Novelty.
The novelty of emergent properties is directly related to Lovejoy's claim that emergent
properties confute the preformationist assumption (which he associates with reductionistic
physicalism) that 'there is no more in the effect than in the cause'. Emergent properties add to
the ontological richness of the world. Taking the idea that emergentism is defined by the view
that a whole is more than the sum of its parts, we may suppose that we can understand emergent
novelty along these lines.
19 Nothing in this chapter rests on the truth of this claim about the historical origins and proper use of the term 'emergence'. I raise it here only to explain (not to justify or defend) my restriction of uses of the term to doctrines of ontological emergence.
261
Additivity: A property F is an additive property if it is a property of a complex system S composed of parts and properties P1...Pn, and none of P1...Pn has F.20
Thus, my weighing 205 lbs. is additive: it is a property of a complex physical system composed of
parts none of which weigh 205 lbs. Likewise, I am 6' tall despite the fact that none of my
component parts has a length of 6'. My height and weight, then, are paradigm additive
properties.
Additivity is not strong enough to capture the ontological emergentist's intended notion
of novelty however. These paradigm properties are fully explicable in the strongest reductivist
sense: my being 205 lbs. is my being composed of parts the sum of whose weights is identical to
205 lbs. My height and weight are not identical to the length or weight of may parts, but they are
identical to the lengths and weights taken aggregatively. There is a clear sense, then, in which
these properties of mine are nothing new, ontologically speaking, in the inventory of the world;
they are not genuine novelties but merely 'resultant' properties. Our task is to distinguish between
two types of additive properties: those that are merely resultant and those which are genuine
novelties.
This task brings us to foundational and hotly disputed issues in metaphysics. Since
novel properties in the emergentist's sense are properties of complex physical systems which
arise from and are dependent upon the actions and interactions of component parts and
properties, we need a distinction (roughly) between the properties and relations which
characterize components in themselves and those that characterize the system as a whole. Let us
20 My use of the term 'additive' has historical links to early emergentists explications of what it was for a property to be 'merely resultant' rather than truly emergent. As numerous commentators have pointed out, however, it is unclear that additivity can adequately explain the emergentist's notion of a merely resultant property. I have therefore here defined a property as 'additive' not in terms of its being the 'sum' of properties of parts (although the examples I have chosen are particularly well suited to this) but as a general explication of the motto that wholes are 'more than the sum of their parts' � something is added when we bring the parts together. This general notion is refined below in my definition of novelty, which in turn yields an adequate definition of a property which is merely resultant. In my terms, both novel and merely resultant properties are types of additive properties.
262
therefore intuitively explain the 'basal conditions' of a complex whole which may (or may not)
give rise to novel properties following Richard Healey.
Basal conditions:21 The basal conditions of a complex thing "include just the qualitative, intrinsic properties and relations of the parts, i.e., the properties and relations that these bear in and of themselves, without regard to any other objects, and irrespective of any further consequences of their bearing these properties for the properties of any wholes they might compose." (1991, 401)
As Healey notes, it is surprisingly difficult to give precise formulation to this characterization of
the 'intrinsic' properties and relations of parts � those they bear 'in and of themselves'.22
Nevertheless, our preanalytic understanding of the notion is strong and fundamental enough to
warrant our employing it.
With this notion of basal conditions in hand, then, we may define a genuinely novel
property as follows.
Novelty: A property P of a complex physical entity E of type T, where entities of type T possess basal conditions B composed of parts, properties and relations b1...bn, is novel if and only if (i) P depends causally for its existence on the interactions of b1...bn ,23 and (ii) P is irreducibly different in kind from the kinds of properties and relations had by the component parts b1...bn of the basal conditions B as they appear independent of their composing entities of type T.24
This definition requires a good deal of unpacking.
21 Healey does not call this an explication of basal conditions but of the 'supervenience basis' of a complex thing. His terminology has the potential to mislead, however, since he desires to remain neutral concerning the question of whether all properties of systems with 'supervenience bases' as explained here actually supervene on this base. I have therefore employed the more general term, basal conditions. 22 On this problem, see David Lewis (1983; 1986a). 23 This qualification is added to Spencer-Smith's definition of novelty. Spencer-Smith defines novelty as follows: "a property P is novel in x if x has P, and there are no determinates P' of the same determinable as P, such that any constituents of x have P'." (1994, 117). Spencer-Smith notes that such a definition allows that being a word of English is a novel property of the letters 'n', 'u', and 't', but that these are not the types of things which were meant to be emergent properties. Restricting the class of novel properties to those which arise through the causal interactions of sub-components of the systems to which they belong is intended to eliminate this problem. 24 cf. Jaegwon Kim's explication of the notion of novelty: "[The Irreducibility of Emergents] Emergent properties are 'novel' in that they are not reductively explainable in terms of the conditions out of which they emerge" (Kim 1992, 124). Kim's definition is ambiguous in that one might have stronger or weaker conceptions of reductive explanation in mind. On weak conceptions ontological emergentists will accept reductive explanations; on stronger versions they may reject it.
263
The complex description leading up to the definition of novelty indicates (roughly) that
emergent properties are properties of complex physical systems which may fall into types
identified by the types of interactions sub-components of the systems enter into. All emergent
properties and entities, then, are dependent for their existence on the uniform lawlike natural
behavior of the component parts of systems of that type.25
The first clause of this definition states that novel properties in the sense we have in
mind are related to their constituent properties through contingent causal relations rather than
through necessary logical ties.26 Ontological doctrines of emergence as applied in particular
instances are intended to express a posteriori empirical discoveries about the causal structure of
the world rather than logical, analytical, or strongly metaphysical claims about necessary relations
between properties or predicates.27
The basic idea of the second clause is easy enough to understand by pointing to
examples, but requires the most unpacking. There are a number of salient examples of debates
in philosophy between pluralists about the kinds of things there are in the world and monists
who reject ontic plurality. On realistic conceptions of properties, relations and propositions
these entities are irreducibly different in kind from physical objects: the former belong to the
category of abstract objects whereas the later belong to the category of concrete objects, and neither
kind can be reduced to the other. Conceptualists reject the appearance of ontic diversity here,
25 In fact emeregentists need not be committed to the existence of such lawlike uniformities among basal conditions for emergent properties � they may accept a widely heterogeneous dependency relations between different individual basal conditions and emergent properties which cannot be formulated into lawlike regularities. Despite this possibility, however, the chances seem so remote that emergentists would choose to endorse this as a claim about the actual world that I have formulated the definition as it appears. 26 See Searle (1992) for an expression of this aspect of emergence in the philosophy of mind. 27 As Lovejoy says, "both assertors and deniers of any such hypotheses must address themselves to the analysis of definite empirical data. The assertor must... point out some type of observable entity, even, or quality � call it E �... which does not appear adequately describable in the same terms as would describe any entity, even, etc., which we can with probability suppose to have existed in [its constituents outside of structures of the type it currently belongs to]. The denier must attempt to show that everything in E is describable in the same terms as some class of entities, events, or qualities in Ph.A. [the emergence base]" (Lovejoy 1927, 29).
264
reducing properties, relations and propositions to conceptions in the mind (which may be
claimed to be further reducible to physical properties and relations). On dualistic conceptions of
the mind, mental properties (i.e., qualia and/or intentionality) are irreducibly different in kind
from brains or brain states. Reductivist physicalists attempt to reduce the prima facie ontic
plurality displayed by mental and physical phenomena to a monistic physicalistic ontology.
Idealists reduce the appearance of ontic plurality to a monistic mentalistic ontology. In the last
chapter I argued that teleology was an irreducibly different kind of causal tie than any physical
property indicated by contemporary reductivists. Each of these examples illustrates the robust
character of widely shared intuitions concerning important ontological differences in kind.
Despite the fact that we have a strong working conception of the distinction between
reductivistic and antireductionist theories the issue of when one thing is reduced to another is a
vexed one. Given the plurality of different conceptions of reduction which readers may have in
mind as they understand this central notion in the emergentist's arsenal, it is imperative that we
disambiguate senses of reducibility and clarify in exactly what sense emergentists claim that novel
properties are irreducible to basal conditions. The emergentist's claim is distinctive and well
defined, but as we will see there are widely shared conceptions of reduction according to which
emergentists will be the first to affirm that even properties that are novel in the intended sense
are reducible. Much hangs, therefore, on disambiguating and keeping in mind the precise sort of
irreducibility emergentists intend.
We are interested here in metaphysical or ontological claims concerning reduction rather
than linguistic or explanatory reductions. The core claim of an ontological reduction is the claim
that the prima facie ontic commitment to sui generis entities of type F expressed in one type of
expression, E1, reveals itself upon analysis to commit one to nothing more ontologically than the
type G ontic commitments of expressions of type E2; expressions of type E2 are perspicuous in
revealing ontic commitments whereas E1 claims mislead about the fundamental ontic inventory
265
of the world.28 Thus, for example, a reductive physicalist claims that mentalistic expressions
such as 'I believe F' or 'I see an orange after-image' which might seem, prima facie, to commit one
to a sui generis mentalistic ontology of beliefs and after-images in fact express no ontic
commitment to anything beyond purely physical entities and properties (or predicates). Idealists
believe that expressions such as 'There is a table here now' when properly understood express no
commitment to the existence of physical things such as tables but in fact reveal commitment
only to ideas.
We may on the basis of this understanding of reduction define a generic notion of
reducibility as follows. Supposing that expressions of type E1 commit one, prima facie, to entities
of type F, where F things are prima facie distinct in kind from entities of type G, then:
Generic reduction (roughly): F properties, relations and entities are reduced to G properties, relations and entities iff, for each entity F there are properties or entities G1...Gn such that x's being F just is x's having G1...Gn in relation R; or, x's being F is nothing ontologically over and above x's having G1...Gn in relation R; or, being F need not figure as an independent existent in a complete ontological inventory of the world.
This generic notion of reduction admits of a plurality of understandings depending on how one
understands what it is for one thing to 'just be' or 'be nothing over and above' another. These
distinct understandings form distinctive conceptions of reduction.
As I mentioned above, it is crucial that we distinguish these forms of reduction because
the sense in which the emergentist is committed to novel (i.e., roughly irreducible) properties
hangs entirely on this. In fact, on all but the strongest conceptions of reduction it turns out that
the emergentist's conception of ontological novelty is compatible with reductionism. Here I list
a number of conceptions of reduction, roughly in ascending order of contemporary use, all of
which are compatible with the emergentist's claim that there exist genuinely novel properties in
at least some complex physical systems.
28 Thus, Paul Teller (1995), says that reduction is "the replacement of one expression by a second expression that differs from the first in prima facie reference" (679). For further preanalytic characterizations of ontological reductionism, see Kim (1999, 15) and Klee (1997, 83).
266
Reduction by coextension: Properties or entities of type F are reduced to properties or entities of type G iff, for each property or entity F there are G properties or entities G1...Gn such that x is F if and only if x has G1...Gn.
Reduction by necessary coextension: Properties or entities of type F are reduced to properties or entities of type G iff, for each property or entity F there are G properties or entities G1...Gn such that, necessarily, x is F if and only if x has G1...Gn.
Nagelian theory reduction: Properties and entities of type F, whose behavior is described by theory T1 are reduced to properties and entities of type G, whose behavior is described by theory T2 iff, for each property or entity F (a) there exists a bridge law such that x is F if (and possibly only if) x has G1...Gn, and (b) the laws relating F properties and entities expressed in T1 are derivable from the laws relating G properties and entities expressed in T2 plus the bridge laws formed in (a).29
Compositional reduction: Properties or entities of type F are reduced to properties or entities of type G iff, for each F property or entity there are physical properties and entities G1...Gn such that x has F if and only if x is composed of G1...Gn.30
Metaphysical version of Kemeny-Oppenheim reduction:31 Properties or entities of type F are reduced to properties or entities of type G iff, for each property or entity F there is a theory T of the behavior of G properties and entities G1...Gn such that T describes the microcausal mechanism for the implementation of F in entities such as x.32
I will discuss these notions as applied to emergent novelty in turn.
Reduction by coextension and necessary coextension are infrequently employed in
discussions of ontological reduction, and have found their natural home in discussions of
29 This formulation of a reductive thesis is derived from Nagel (1961, chapter 11). 30 As metaphysically weak as this notion appears, it seems to be a notion of reduction with wide acceptance. Thus, Paul Teller says, "In most and quite possibly all cases of putative theory reduction by strict identities, we have instead a relation of physical realization, constitution, or instantiation, nicely illustrated by the property of being a calculator. . . .[Which can be realized] by indefinitely many ... physical arrangements. Perhaps many who have used 'reduction', particularly in the sciences, have intended the term in this sense of physical realization rather than one of strict identity" (1995, 680). This is consistent with Mayr's notion of constitutive reduction, see (1982, 60). 31 Kemeny and Oppenheim introduce an epistemological account of reduction in their (1956). The account is roughly along the following lines:
Kemeny-Oppenheim explanatory reduction: Properties or entities of type F are reduced to properties or entities of type G iff, for each property or entity F there are G properties or entities G1...Gn such that if x is F then x's possession of G1...Gn explains x's being F.
267
mathematics, logic, and language. The notions are frequently thought to be too weak to effect
true reductions; it seems that there is a clear sense in which even cross-world correlation between
entities does not entail that one entity 'just is' the other in any robust sense. 1+1=2 and 2+2=4
make quite distinct claims despite the fact that they are true of the same set of possible worlds
(all of them). Kidney bearing and heart bearing animals may be coextensive across nomologically
possible worlds, but it is even clearer in this case that being a creature with a kidney is something
different from being a creature with a heart. Likewise, emergentists may acknowledge and even
emphasize that novel emergent properties are coextensive with basal conditions across
nomologically possible worlds without giving up the claim that the purported novel properties
constitute a distinct ontic type from the basal conditions and properties.
Emergent novelty is similarly compatible with Nagel's account of theory reduction. As
Jaegwon Kim (1992, 125-6; 1998, 25-7 and 95-7; 1999, 12) observes, emergentists need not deny
that Nagelian bridge laws between basal conditions and emergent novelties exist,33 and neither
(for that matter) must Cartesian dualists.34 Bridge laws express correlations between properties
or entities, but just as coextension is too weak to give us reduction so the existence of bridge
laws from one domain to another is too weak to indicate that purported emergent novelties are
nothing ontically over and above the properties and entities with which they are correlated.
Correlation will not yield meaningful reduction, but of course if Nagelian reduction is what a
32 This metaphysical version of Kemeny-Oppenheim reduction derives from reflection on Kim's metaphysical explanation of the realization relation in response to an explanatory definition of that relation in his (1993). 33 Kim says, "If, however, all that reduction had to accomplish were this sort of nomological derivation, the emergentists would not necessarily have resisted the idea of reducing consciousness to neurobiological processes, or biological phenomena to physicochemical phenomena. . . . The emergentist could cheerfully admit all this [i.e., all that is required for Nagelian reduction], but at the same time deny that a mere logical exercise of this sort, based on brute and unexplained correlations between emergent and basal properties, affords us an understanding of the nature of the emergents." (1992, 125-6). 34 To point this out is not, of course, to equate the doctrines or commitments of emergentists and Cartesian dualists. I mean simply to point out how little about one's ontology follows from accepting a Nagelian reduction � if Nagelian reduction is compatible with Cartesian dualism, then it appears a rather weak and benign form of reduction.
268
philosopher has in mind by 'reduction' then the emergentist need not deny and may positively
affirm that (in this sense, as in others) emergent novelties in her sense are reducible to basal
conditions.
It is apparent from the literature that compositional reduction is an extremely popular
notion of reduction35 even though the correct understanding of composition is (as with so many
issues at this level of metaphysical scrutiny) a vexed issue in itself.36 Intuitively, if CB composes
E at t then the identity conditions of E are distinct from those of CB. E may survive the loss at
t+∆t of certain parts of CB while CB will not survive that loss. Further, if E is composed of CB,
then CB is not composed of E, although if E is composed of CB and CB is composed of CB',
then E is composed of CB'. This is to say that our intuitive notion of composition is the notion
of a relation which is asymmetric and transitive, and less strong than identity understood in terms
of Leibniz' law.37
But this is enough to see that emergentists can accept and even affirm the claim that
emergent novelties (in their sense) are compositionally reduced to their basal conditions. Indeed,
much of the motivation for emergentism as a 'middle way' between physicalism and other
dualisms such as the 'pre-established harmony' theory38 is to capture a strong sense in which
emergent novelties are dependent upon the existence of complex physical objects, and the
affirmation of compositional reducibility plays an important role in emphasizing this aspect of
the emergentist's position. Still, given that composition is weaker than identity clear sense can be
35 See, for example Teller (1995, 680), Mayr (1982, 60-1), and Dobzhansky, Ayala, Stebbins, and Valentine (1977, 488-9). 36 On this issue see Wiggins (1980), Lowe (1989) and Doepke (1982). 37 See Lowe (1989, 81). 38 The pre-established harmony theory is defined by the thesis that mind and body are distinct substances or have distinct types of properties, but that these distinct kinds do not causally interact. Instead, they are like two perfect timepieces set to march in step at one initial moment and marching on in harmony for eternity. Leibniz held a view of this sort.
269
made of the emergentist's claim that the purported emergent novelty is nevertheless ontically
distinct from its basal conditions.
The metaphysical version of Kemeny-Oppenheim reduction is an enrichment of
compositional reduction to the effect that we have not effectively compositionally reduced one
kind of entity to another until we have a well-confirmed microcausal theory of base entities that
explains their composition of the higher level properties. Thus, on this conception we have not
properly reduced water to H2O until we have a theory explaining the microcausal means by
which groups of H2O molecules give rise to the properties of water (its wetness, its boiling point,
etc.).
Again, however, emergentists need not fear and may positively endorse the search for
this type of a reductive account of emergent novelties. They may affirm the existence of causal
correlations between basal conditions and emergent properties strong enough to support the
formulation of laws and theories that microcausally explain the emergence of emergent
novelities.39 Nevertheless, there remains clear sense to the emergentist's claim that having a well
confirmed explanatory theory of how Xs give rise to Ys does not entail that Ys are 'nothing over
and above' Xs. Ys may still constitute a genuine � and in a sense still to be defined an irreducible
� addition to the ontology of the world conceived only in terms of the Xs.
We come finally to the sense in which emergentists are committed to the claim that
emergent novelties are irreducibly different in kind from their basal conditions.
Reduction by property identity: Properties or entities of type F are reduced to properties or entities of type G iff, for each property or entity F there are physical properties or entities G1...Gn in relation R such that being F (for xs) is identical to being G1...Gn in relation R (for xs).
That this is the 'proper' sense of reduction � the inevitable limit on all reductivist theories �
makes sense because we have finally arrived at the strict sense (identity) in which one thing can
just be another as required by the preanalytic or generic notion of reduction. Temperature (in
270
ideal gases) has in this sense been reduced to and is thus identical to the mean kinetic energy of
constituent molecules. Likewise, temperature (in a vacuum) is the blackbody distribution of the
vacuum's transient radiation.40
Jaegwon Kim (1999) offers one analysis of necessary and sufficient conditions for
obtaining this type of strong reduction. According to Kim, it is necessary and sufficient (as a
first 'philosophical' step, the rest of the work coming in the construction of scientific theories,
see p. 18) to effect a reduction by property identity that the property to be reduced be given a
functionalized definition in terms of causal/nomic relations between basal conditions.41 If E is
to be reduced to B, then
E must be functionalized � that is, E must be construed, or reconstrued, as a property defined by its causal/nomic relations to other properties, specifically properties in the reduction base B. (10)
Once being E has been identified with a set of causal/nomic relations between entities and
properties in the reduction base B, further scientific work along the lines indicated in the
metaphysical version of Kemeney-Oppenheim reduction is necessary to explain how it is that B
gives rise to E. Nevertheless, on Kim's account the viability of the reductive analysis stands or
falls on the property identification made at the functionalization stage.42
A candidate emergent property qualifies as a genuine emergent novelty if and only if it is
not identical in kind to a kind of property which can be had by the component parts of the
system from which it emerges in isolation from structures that type. Clearly emergentists will
have 'knock down' arguments against the possibility of such reduction in only extremely rare
39 This fact is widely noted in the literature on emergentism. See Kim (1999, 8). 40 These examples are from Churchland (1988, 41-2). See also Teller (1995). 41 This idea has obvious affinities to other conceptions of reduction. Nagel's bridge laws and definitions play a similar role in his account of theory reduction. When we understand coextensiveness claims as claims about sameness of meaning, we can see that reductivist claims concerning the coextension are also closely tied with the step Kim explains here.
271
cases.43 More frequently emergentists will argue for the novel status of properties on the basis of
systematic patterns of failure in historical attempts to give reductive analyses and by reference to
distinctive properties of the purported emergent (such as normativity, intentionality, or first-
person availability) which seem to drive our preanalytic intuition that properties of type F are
distinct in kind from properties of type G.
I argued for the sui generis irreducible status of teleology using this method in chapter
five. That investigation relied on arguments concerning the categorial properties of teleology
and brought to light features such as teleology's normativity which appear to resist satisfactory
reductive analysis. Such arguments for the sui generis irreducible status of purported emergent
novelties cannot rule out or address future reductive analyses, but they can make the claim of
irreducible novelty plausible and intelligible. This section's analysis of the notion of irreducibility
invoked by the emergentist shows, further, that the notion of irreducibility involved is perfectly
understandable and well motivated in light of other, weaker, notions of reducibility.
With this understanding of reduction, we may reformulate our definition of novelty,
specifying the sort of reduction emergentist's deny of genuinely emergent properties. The
definition which results from our discussion of forms of reduction will also yield a definition of
being a merely resultant property. Our foray, then, has clarified the distinction ontological
emergentists draw between those additive properties which are merely resultant and those which
are genuinely emergent.
Novelty*: A property P of a complex physical entity E of type T, where entities of type T possess basal conditions B composed of parts, properties and relations b1...bn, is novel if and only if (i) P depends causally for its existence on the interactions of b1...bn , and (ii) P is irreducibly different in kind from the kinds of properties and relations had by the component parts b1...bn of the basal conditions B as they
42 Kim also discusses this notion of reduction in his (1998): see especially pp. 24-5 and 98. David Chalmers invokes the same claims for what he calls 'explanatory reduction' in his (1996): see esp. pp.43-5 and 51. 43 George Bealer's argument for ante rem universals in (1993b) may be one such case. Of course, even 'knock down' arguments such as this one have presuppositions which may be denied by committed opponents.
272
appear independent of their composing entities of type T; that is, (ii') P cannot be reductively identified with any of the kinds of properties and relations had by the component parts b1...bn of the basal conditions B as they appear independent of their composing entities of type T.
Mere resultants: A property P of a complex physical entity E of type T, where entities of type T possess basal conditions B composed of parts, properties and relations b1...bn, is merely resultant if and only if it is additive but not novel.
These definitions make clear sense of the emergentist's claim that there is a distinction between
those additive properties such as weight and height which are merely resultant and those
properties which are genuine additions to the ontology of the world as it would be restricting our
view to the ontology sufficient to account for entities of the same type as those in the basal
conditions outside of entities of some complex types. My weight is reductively identifiable with
(and not merely composed of, or realistically explicable in terms of, or associated by bridge laws
with) properties of my parts, and therefore while my weight is an additive property (it is not
possessed by any of my parts), it is merely resultant. A plausible candidate for an intuitive property
of me which is additive and novel is the mental property of seeing red. If certain arguments for
the irreducibility of qualia in the philosophy of mind are correct, then we have here a property
which is, by the analysis given, an emergent novelty. This example will be acceptable to those
who consider themselves property dualists or nonreductive physicalists in the philosophy of
mind, and intelligible even to opponents.
So much for ontology. Methodologically, we discover differences between kinds of
properties through engaging in analysis that will � typically � involve investigating
superordination, subordination, exclusion and inclusion relationships between the concepts we
use and their corresponding properties. I have no theory of categories to guide us in
determining in every case when we have properties of irreducibly different kinds and when we
do not. Nevertheless, categorial and identity intuitions about properties are quite robust; to pick
one particularly clear example physicalists, dualists and idealists alike are emphatic and consistent
in the view that if mental properties turn out not to be reducible or eliminable, then physicalism
273
is either false or vacuous.44 This firm bar of the irreducibly mental from physical ontology
displays the sort of categorial intuition I have in mind to discover as we seek reductive analyses
for purported emergent novelties. Again at a highly theoretical level, Gettier examples are used
widely and widely taken to show that knowledge is not identical to justified true belief.45 Further
examples are legion in philosophy.
In chapter five I appealed to intuitions about such relations in claiming that ideas are of
such a kind/category that they cannot sleep; numbers are of such a kind that they cannot be tall;
teleology is of such a kind that it cannot be a historical property; etc. By contrast, triangles and
squares are of the same kind, they are both determinates of a determinable, geometrical object; red
and blue are of the same kind, color; mass and extension are of the same kind, physical properties;
etc. In typical cases such judgements of sameness and difference in kind will not be supported
by categorial theory but instead by a rich web of interconnected intuitions concerning
subsumption, superordination, exclusion and inclusion relations between the concepts used to
pick out the entities in question.
Obviously we will only be able to determine whether putative novel properties are
genuinely novel in the intended sense through investigation of particular cases. Still, I have
gestured in the preceding paragraphs toward a typical and traditional method that such
investigations employ and gave an extended example of such methods at work in chapter five.
Allegedly emergent properties will be novel if and only if they turn out after analysis to be
irreducibly different in kind from the kind of properties belonging (independently) to their
emergence base, or equivalently, if and only if they turn out to be identical to no property or
range of properties of the type found in the emergence base.
44 The doctrine is false if (as seems right) sui generis mental properties are not physical properties, vacuous if we simply expand our conception of the physical to embrace sui generis mental properties. One need not read far in the literature on physicalism to find the negative constraint on possible physicalist theories. For a theoretical discussion, see Crane and Mellor (1990). 45 This example is employed to much the same effect by Bealer (1992; 1999).
274
Our second task, to which we now turn, is developing an understanding of how it is that
emergent properties can have genuine causal influence.
6.8.2 Emergent causal influence.
Weak versions of ontological emergence require only that properties that are novel in
the sense described above emerge from the interactions of entities at lower ontological 'levels' in
certain circumstances. Such versions of emergence are compatible with an epiphenomenalism
about emergent properties such that emergent properties have no causal influence in the
proceedings of the world. However, these versions of ontological emergence will be too weak
for the purposes of chapter six; the goal of that chapter was to gain an understanding of how sui
generis Aristotelian final causality can be causally efficacious in the physical world. We cannot,
therefore, rely on weak versions of ontological emergence that are consistent with
epiphenomenalism, but must seek a stronger version of ontological emergence such that physical
structures may have novel properties with emergent causal powers.
There exist a number of models for understanding the emergence of 'downward causal
influence'. Let us say that a level of types of properties and relations L1 is a 'higher' level of
properties and relations than a level L2 just in case properties and relations in L1 are had by
entities composed of properties and relations at level L2.46 The claim to be understood, then, is
that at least in some cases novel and genuinely emergent properties � properties existing at a
higher level than their basal conditions � exert a causal influence on the course of events at the
level of their compositional parts.47 48
46 See Humphreys (1995, 108). 47 Emergent properties may presumably also exert 'same level' and 'upward' determinative influence, but neither of these causal relations is controversial; the only substantive question involved is the question of the possibility of downward causal influence. 48 According to Paul Humphreys, whose primary concern is not to argue that particular properties such as consciousness or teleology are emergent and have downward causality but rather that such causal influence is coherent, one coherent form of downward causal influence would be exerted if entities at higher levels were physical (not logical) fusions of lower level entities and processes. In his view, physical processes may cause properties which previously existed independently to become fused, losing their independent identity
275
Timothy O'Connor takes the metaphor of 'downward' causation very seriously in his
discussion of the issue. Here is his description of the process.
[The phrase 'novel causal influence'] is intended to capture a very strong sense in which an emergent's causal influence is irreducible to that of the micro-properties on which it supervenes: it bears its influence in a direct 'downward' fashion, in contrast to the operation of a simple structural macro-property, whose causal influence occurs via the activity of the microproperties that constitute it. Of course, if we take emergents to be a species of supervenient properties, as I have suggested, then the continuing instantiation of the emergent property is completely dependent upon some set of properties or disjunctive ranges of properties in the object's microstructure. Nonetheless, it exerts a causal influence on the micro-level pattern of events that is not reducible to the immediate causal potentialities of the subvening properties. (1994, 97-8. Emphasis in the original)
On this view, higher order emergent properties exert causal influence which is all their own by
changing micro-level relations in their emergence base. These changes are not causally explicable
in terms of the causal story applicable to the base properties and entities in isolation from their
membership in systems of the type of which they are currently members.49 Nevertheless, as
O�Connor�s account makes clear, the emergent properties all along retain their causal
dependence upon their emergent base. We could think of a horse (emergence base) and its rider
(emergent downward causal influence) as a rough analogy. Although the rider depends for her
remaining upright on the actions of the horse, the rider is also able to effect the course of events
at the level of the horse.
William Hasker takes issue with this sort of picture of downward causation. On his
view, O'Connor's type of account takes the metaphorical 'downward' of 'downward causation'
and forming a new, higher-level entity in which the component parts lose their independent identity. In such a situation, the higher-level entity would be composed of entities that no longer maintain an independent existence and the higher-level entity would have causal powers all its own that would not be reducible to the causal influence of its component parts. (See especially Humphreys (1996, 59-62) and (1997b, §5); see also Silberstein and McGeever (1999, 189).) While this model for understanding downward causal influence may show the notion to be coherent, it does not appear readily applicable in the context of a modern scientific ontology to the case I am interested in � the emergence and downward causal influence of sui generis teleology � and so I will move on to other models. 49 This, at any rate, is how I understand O'Connor's claim that emergent causal influence is not mediated 'via' the influence of lower level properties and entities, and that it is not the 'immediate' causal properties of those lower level entities that explain the emergent's causal powers. Epistemic emergents (to continue to employ the term) have causal powers that are explicable via the immediate causal powers of microcomponents; these explanations are simply unavailable to us given our cognitive limitations.
276
too literally, and the notion must be replaced by a more moderate conception of emergent
causation. In Hasker's view,
to say that the higher levels causally affect the lower levels can only mean that, because of the specific sort of organization and integration involved, the behavior of the fundamental elements is different than it would be if those elements continued to be governed by the old, non-emergent laws. (1982, 184. See also pp. 185 and 175-6) 50
It is less than clear, however, on what basis Hasker objects to the view O'Connor holds. Hasker
clearly believes that the view is in danger of giving the wrong impression that downward
influence will be exerted without change in lower level laws or behaviors, thus encouraging the
view that the two sorts of causal powers are involved in competition such that they must �push
each other around� (the metaphor is Sperry's, see Hasker 1982, 184). While it seems appropriate
to warn against viewing the relations in this way, I do not see that O'Connor's view is committed
to the competitive view in any way.
I prefer, therefore, to retain O'Connor's description of the action of downward causal
influence, which is at least compatible with Hasker's. In downward causation, emergent
properties with novel powers exert a causal influence over the course of events in an emergence
base which is not explicable in terms of the laws governing those base entities and properties in
other contexts. I prefer this way of speaking of downward causation over Hasker's for the
simple reason that it emphasizes the causal role of the higher-level property rather than the
causal interactions and powers of the emergence base.51 On a correct understanding, both sorts
of causal influences are at play, but it is, after all, the causal powers of the emergent properties
which constitute our focus.
50 Cf. Jürgen Schröder's definition of downward causation: "Downward causation is the influence the relatedness of the parts of a system has on the behaviour of the parts. It is not the influence of a macro-property itself, but of that which gives rise to the macro-property, viz., the new relatedness of the parts" (1998). 51 This conception of downward causal influence also seems faithful to the notion developed historically by Broad; see his (1918-19; 1925).
277
It may be objected, of course, that all we have described in a case of purported
downward causation is a situation such that
the physical particles or structures involved [in the complex structure] must have possessed beforehand what we may call the 'disposition' or 'possibility' or 'potentiality' or 'capacity' for producing the new properties [and their causal powers], under appropriate conditions. (Popper and Eccles 1977, 23)
As C.D. Broad expresses the point, we may surely say that the emergence base had the latent
potential for these powers (1918-19, 114). The point of the objection is to question whether the
causal power in question is best described as the exertion of a genuinely novel power by a higher
level property; it is not to question the coherence of there being new forms causal dependencies
which come into play only in certain types of highly complex physical systems.
Broad's response to this objection is the appropriate one.
There is no objection to this mode of expression so long as we remember that it is purely verbal, and that it does not alter the fact that some part of the behaviour of the second order complex could be neither deduced nor suspected from a knowledge of the behaviour of its parts in other surroundings. (1918-19, 114)
Ignoring the misleading epistemological verbal trappings of Broad's response,52 the response
reminds us that emergent causal influence is a result of two factors. There can be no objection to
claiming that emergent properties in the sense intended are 'latent' in the emergent base so long
as we bear in mind that the emergent properties are of an irreducibly distinct kind from the
properties possessed by the physical base in other circumstances � the properties which are
causally responsible are novel in the sense explained in the previous sub-section.
Noting that emergent powers were 'latent' in the base, then, is a purely verbal move
which establishes no substantive reductionistic theses, and it is harmless unless it makes us forget
that the new causal powers are supposed to be had by properties of a genuinely novel ontological
sort. Indeed, so long as we avoid this mistake, speaking of emergent causal powers as powers
which are latent in an emergent base is positively welcome in that it highlights a portion of the
52 On this point see above, p. 257 ff.
278
emergentist's claims which may be overshadowed in discussions of novelty and downward
causation. It is a fundamental claim of emergentists, recall, that emergent properties and their
powers are causally dependent upon the interactions of base properties and entities, and talk of
the 'latent powers' of these base properties and entities highlights exactly this aspect of the
emergentist's position.
6.9 Conclusion, summary, and definitions.
In the preceding sections I have developed an ontology for the physical world in which
sui generis Aristotelian teleology may find a home. On the ontological view I have explained,
there is room for sui generis Aristotelian teleology in a scientifically acceptable view of the world
so long as our ontology allows for emergent properties with novel causal powers. To explain
this view, I have distinguished the following notions, the last three of which appear here for the
first time.
The core sense of emergence: A property P of a structure X with components a1...an is emergent if and only if (i) P depends causally for its existence on the interactions of a1...an in X and (ii) P is augments the ontology of the world � P is not 'contained in' the properties of a1...an in their interactions outside of structures of the same type as X.
Metaphysical particularism: [T]he conception ... of natural events as combinations or rearrangements of relatively simple, preëxistent entities, of which the total number or quantity remains invariant, and of each of which the qualities and laws of action remain the same through all the combinations into which it may enter. (Lovejoy 1927, 20-1)
Epistemological emergence: The doctrine that the claims of metaphysical particularism are true, combined with the claim that there exist (or may exist) complex systems for which it is very difficult � perhaps even impossible � for us to discover how the properties of those systems result from the rearrangement or recombination of lower-level properties and entities.
Basal conditions: The basal conditions of a complex thing "include just the qualitative, intrinsic properties and relations of the parts, i.e., the properties and relations that these bear in and of themselves, without regard to any other objects, and irrespective of any further consequences of their bearing these properties for the properties of any wholes they might compose." (Healey 1991, 401)
Additivity: A property F is an additive property if it is a property of a complex system S composed of parts and properties P1...Pn, and none of P1...Pn has F.
279
Novelty*: A property P of a complex physical entity E of type T, where entities of type T possess basal conditions B composed of parts, properties and relations b1...bn, is novel if and only if (i) P depends causally for its existence on the interactions of b1...bn , and (ii) P is irreducibly different in kind from the kinds of properties and relations had by the component parts b1...bn of the basal conditions B as they appear independent of their composing entities of type T; that is, (ii') P cannot be reductively identified with any of the kinds of properties and relations had by the component parts b1...bn of the basal conditions B as they appear independent of their composing entities of type T.
Mere resultants: A property P of a complex physical entity E of type T, where entities of type T possess basal conditions B composed of parts, properties and relations b1...bn, is merely resultant if and only if it is additive but not novel.
Weak ontological emergence: A property P of a structure X with components a1...an is weakly ontologically emergent if, and only if P is novel*.
Strong ontological emergence: A property P of a structure X with components a1...an is strongly ontologically emergent if and only if (i) P is novel* and (ii) P has causal powers which are absent from a1...an and their interactions independently of entities of the same type as X.
Nomic strong ontological emergence: A property P of a structure X with components a1...an has strong nomic ontological emergence if and only if (i) P is novel*; (ii) P has causal powers which are absent from a1...an and their interactions independently of entities of the same type as X; and (iii) P's causal powers are formulable into laws of nature.
On the view I hold, sui generis Aristotelian teleology needs to be as strong as, but no stronger
than, strong ontological emergence to have an explicable and scientifically acceptable role in the causal
structure of the physical world.
I have argued that there is a perfectly general ontological framework � the framework
of emergent properties � in which sui generis Aristotelian teleology can find a natural home even
in the modern world. Unreduced sui generis Aristotelian teleology may be a scientifically
respectable part of our ontology of the physical world so long as it is a strongly ontologically emergent
property in our world.
280
Chapter 7
An Aristotelian Account of Life
At the close of chapter two I offered an Aristotelian analysis of life in terms of
teleological directedness. Since making that suggestion the focus of the dissertation has been on
teleology, Aristotelian and otherwise. I have argued that the scope of teleology in Aristotle is
compatible with his analyzing life in teleological terms; that Aristotle's conception of teleology is
the notion of a sui generis causal factor in the structure of the world; and that this conception of
teleology remains surprisingly relevant given the respective failures of contemporary reductivist
accounts of teleology and the traditional substantive and methodological objections to sui generis
Aristotelian teleology from philosophy and science. I have argued in addition that there is
substantive scientific evidence suggesting the actuality of the sort of emergentist ontology
necessary to understand Aristotelian teleology's place in the modern world.
With foundational work on the core notion of the provisional definition thus behind us,
this chapter provides close textual support for the earlier hypothesis that Aristotle analyzes life in
terms of teleological directedness. This support is strengthened through an extended discussion
of cases and problems deriving from the problem of the unity of life's definition in Aristotelian
terms. The hypothesis of a teleological account of life gains considerable support from its ability
to answer on firm textual grounds tough questions concerning Aristotle's substantive and formal
commitments with regard to the definition of life. The chapter concludes with a formal
Aristotelian account of life. The end result is an account of what it is to be alive that is both
Aristotle's and an important contribution to the contemporary debate concerning the nature of
life.
7.1 Teleology and life in Aristotle.
The idea that a teleological account of life is an Aristotelian account is one with a long
history, and I believe that it is fundamentally correct. The novelty of this dissertation's approach
281
to life in Aristotle lies not in the core substance of the account (i.e., teleology) but in the
systematic elucidation and defense of this core insight.
My task in this section is to make explicit some of the various textual bases for thinking
that Aristotle analyzes being alive in terms of possessing teleological directedness. The textual
case for this thesis comes in two parts. I argue first that there is enormous plausibility to the
hypothesis that Aristotle defines life teleologically arising from Aristotle's persistent and
theoretically connected ways of associating teleology and living beings. Second, I argue that this
strong prima facie case for an analysis of life in terms of teleology is vindicated by Aristotle's
explicit remarks on the definition of life; Aristotle's account of life is an application of his
theoretical commitments concerning kind individuation.
In the next section, this thesis is supported further by displaying its ability to solve
various puzzles arising out of the proposed interpretation. The strength of the interpretation
displays itself in our ability to mount textually grounded and philosophically plausible responses
to such puzzles as arise.
7.1.1 The (close) association of teleology and life in Aristotle.
In interpreting Aristotle's views on life teleologically we side with contemporary
interpreters who agree that his account is teleological at heart.1 While we found reason to reject
the formal analyses of these authors in chapter two, the teleological core of their accounts
appeared sound and justified our extended discussion of teleology, Aristotelian and otherwise.
1 Recall that Matthews (1992) took DA ii.4 415a27-b2, a passage expressing Aristotle's commitment to the teleological nature of life as the starting point for his account. I argued in chapter two that Matthews' account failed in the end not because of its teleological character, but (on the contrary) because it analyzed genuine teleological commitment out of Aristotle's view. Shields (1999) followed Matthews in offering a teleological account of life focused around the notion of an 'intentional system'. I argued in chapter two that that account failed, again, not because of its teleological base but because of its development within the context of focal homonymy.
282
Turning to textual evidence, we may take our cue concerning Aristotle's commitment to
a teleological account of life from the following passage (cited by Matthews 1992). Aristotle
says,
[F]or any living thing that has reached its normal development and which is unmutilated, and whose mode of generation is not spontaneous, the most natural act is the production of another like itself, an animal producing an animal, a plant a plant, in order that [hina], as far as its nature allows, it may partake in the eternal and divine. That is the goal towards which all things strive, that for the sake of which [ekeinou heneka] they do whatsoever their nature renders possible [hosa prattei kata phusin]. (DA ii.4 415a27-b2)
The commitment is equally obvious in the following passage:
[A]ll the movements of ensouled things (tôn empsuchôn) and of everything else have limit (peras exousin). For all living things (ta zô(i)a)2 both move and are moved for the sake of something, so that this is the limit of all their movement � that for the sake of which. (MA 6 700b15-6; first sentence my translation. See also DA ii.4 415b15-20, quoted below.)
These passages strongly suggest that Aristotle commits himself to the teleological character of
life. If Aristotle is committed to a teleological view of life, however, then we ought to be able to
answer the question: what ends are the activities of living things directed towards?
Perhaps the most striking feature of de Anima ii.4 415a27-b2 is Aristotle's specification
of one end that all living things seek: all living things are said to seek participation in the eternal
and the divine.3 This answer to our question may appear extravagant to contemporary readers,4
but is, perhaps, not one we should shy away from attributing to Aristotle. Nevertheless, oddness
of this commitment may be mitigated by drawing a distinction which has eminent Aristotelian
credentials.
2 The previous sentence's reference to all ensouled beings makes it clear that 'living things' rather than 'animals' is the correct translation here, pace Nussbaum. 3 At Generation of Animals ii.1 731b24-32 and on Generation and Corruption ii.10 336b25-35 Aristotle also seems to suggest that the goal of living things is to be eternally. 4 The answer may seem extravagant in the sense that while we may be happy to admit that organisms have reproductive ends and to admit that if these continue they will lead naturally to long lines of descent, we do not tend to think that one of the individual's aims in reproduction is contributing to a long line of descent. That result will follow (in the right conditions) given the reproductive ends, but it is not itself a goal. Nothing in my argument hangs on this point, however.
283
Aristotle suggests a resolution to modern worries about the ends of living things at
Politics i.2. He says there that the union of a male and female (among those things which require
two sexes to procreate),
is a union which is formed, not of choice, but because, in common with other animals and with plants, humans have a natural aim (phusikon to ephiesthai) to leave behind an image of themselves. (Pol. i.2 1252a28-30; slightly modified from the ROT.)
The first thing to note about the passage is its reassurance that Aristotle's conception of
teleology is not mentalistic: the 'natural aim' to procreate is present in all animals and plants, and
there is no suggestion that this aim entails mindedness; the passage suggests teleological
directedness, no more, no less.5
But secondly � and more importantly for our purposes here � Politics i.2 1252a28-30
suggests a distinction between a proximate and a nonproximate end or goal for procreative
activity. The proximate end of life activities is here claimed to be the much more palatable end
of producing another member of one's species. The existence of a nonproximate end such as
the one stated in de Anima ii.4, Generation of Animals ii.1 and On Generation and Corruption ii.10 is
fully compatible with the existence of this proximate end or goal.6 Thus, even if Aristotle is
committed to living creatures having as one of their (nonproximate) ends participation in the
eternal and divine, this does not overshadow the existence of more proximate, less extravagant
ends.
Quite the contrary, in fact. In terms of the importance of their theoretical role in
Aristotle's biology, proximate ends dominate Aristotle's attention. The proximate goal of
5 See Phys. ii.5 196b18-20, ii.8 199b27-31, Pol. i.2 1252a28-30; cf. NE iii.3 1112a35-b2. Scott MacDonald (1989, 165) suggests that the reason we have a hard time recognizing teleology in things without minds but Aristotle did not is that we associate teleology and intentions (which are mental) whereas Aristotle associates teleology and completeness (teleion, which is nonmental). MacDonald cites NE i.6 1097a25-b21 and Met. v.16. See also Preus (1975). 6 See Metaphysics viii.4: "When one inquires what is the cause, one should, as causes are spoken of in several senses, state all the possible causes... � We must state the proximate causes. What is the material cause? Not fire or earth, but the matter peculiar to the thing" (1044a33-1044b2). See also Phys. ii.3 195b22-5, Met. viii.4 1044b15-21, GA iv.1 765a35-b6.
284
reproduction is itself widely attested to in the texts. Aristotle says, "since it is right to call things
after the ends they realize,7 and the end (telos) of this soul is to generate another being such as
itself, the first soul ought to be named the reproductive soul" (DA ii.4 416b23-5; slightly
modified from the ROT). Here Aristotle clearly announces a proximate end of one type of soul
to be that of reproducing. At Generation of Animals ii.23 Aristotle claims that "to the essence of
plants (tôn phutôn ousias) belongs no other function or business than the production of seed", and
continues that animals have this generative function and one connected to knowing and sensing,
but that "when it is necessary for it [the animal] to accomplish the function of that which has life
(to tou zôntos ergon), it unites and copulates, becoming like a plant" (GA ii.23 731a25-b7; see also
HA viii.1 588b25-6).
To summarize our results thus far, then, Aristotle accepts a teleological conception of
being alive (DA ii.4 415a27-b2, b15-20; MA 6 700b15-6). The end or goal of living things is,
further, divisible into nonproximate and proximate goals (Met. viii.4 1044a33-1044b2 and Phys.
ii.3 195b22). The nonproximate goal is to partake in the eternal and divine where this means
keeping the species in existence eternally (DA ii.4 415a27-b2, GA ii.1 731b24-32, and GC ii.10
336b25-35), whereas the proximate goal is biological reproduction, the production of another
instance of the species (Pol. i.2 1252a28-30, DA ii.4 416b23-5, GA ii.23 731a25-b7, and HA viii.1
588b25-6).8 The suggestion that life is defined teleologically in Aristotle thus receives support
from the fact that Aristotle answers the obvious question: if to be alive is to be end directed,
then what are the ends of living things?
7 This thesis, often called the 'functional determination thesis', has wide textual support: see also Meteor. iv.12 390a10, EE iii.7 1115b23, OH ii.3. 286a8, Pol. i.2 1253a23, PA ii.9 655b21. The thesis is tied to teleology at OH ii.14 297b21, "it is right to call anything that which nature intends [bouletai] it to be, and which belongs to it, rather than that which it is by constraint and contrary to nature." On the functional determination thesis, see Shields (1999, pp.31-5). 8 Reproductive goals are very widely (although not universally) shared on Aristotle's view. It will pay to keep in mind, however, that not all ends that a peculiar type of creature has will be � on Aristotle's view � definitive of that type of thing. Ends peculiar to kinds have a special role in Aristotle's natural ontology, but they exist side by side with other ends in many many creatures. This point receives greater attention in the main text below.
285
This result is further confirmed and enriched by a famous passage in the de Anima ii.4
where Aristotle notes that the soul is the cause or source of the living body in three senses
(415b9-11): as the body's formal (415b12-14), efficient (415b22-7) and final (415b15-21) cause.
The soul is the final cause or goal of the body in the sense that,
[N]ature, like thought, always does whatever it does for the sake of something, which something is its end (telos). To that something corresponds in the case of animals (tôn zô(i)ôn) the soul, and in this it follows the order of nature; all natural bodies are organs of the soul. This is true of those that enter into the constitution of plants as well as of those which enter into that of animals. (DA ii.4 415b15-20)
There are four causes: first, the final cause, that for the sake of which; secondly, the definition of essence (and these two we may regard pretty much as one and the same (ôs hen ti schedon)); thirdly, the material; and fourthly, that from which the source of movement comes. We have then already discussed the other three causes, for the definition and the final cause are the same, and the material of animals is their parts. (GA i.1 715a4-8)
The attainment of the form of the living thing � its soul � is an end or goal of the living thing.
This goal may be understood in two ways, one of which confirms our previous conclusions, the
other of which deepens our conception of Aristotle's teleological view of life.
The attainment of form is the goal of the living thing in one (by now familiar) sense in
that one proximate end of a living thing is the production of another living thing � a
numerically distinct but specifically identical soul (GC ii.11 338b12-19, GA ii.1 731b24-732a11,
DA ii.4 415a27-30; see also OH ii.12 292b17-18 and NE viii.12 1161b28). When a living thing
reproduces, it actualizes the end of bringing the form into being in a new instantiation.9
In another sense, however, the soul may be the final cause of living things in the sense
that each individual living thing has the end of maturing and actualizing its own form or nature
as a member of the species to which it belongs. In this sense soul is the final cause of living in
that the full actualization of the species form in the individual is the end of the growth and
nutritive activities of a living thing. Thus, Aristotle says, "the end is developed last ((h)ustaton gar
9 Recall of course that the nonproximate end may nevertheless be specified as the attainment not of the specific goal of a new, numerically distinct soul, but of keeping the species form in existence.
286
ginetai to telos), and the peculiar character of the species is the end of the generation in each
individual" (GA ii.3 736b4-5). Two aspects of this commitment to the peculiar nature's being
the end of growth and development deserve attention.
The first point simply has to do with how widely this commitment is reflected in
Aristotle's corpus. Thus, e.g.,
We must explain the 'why' in all the senses of the term, namely,...; and because it is better thus (not without qualification, but with reference to the substance in each case (to pros tên (h)ekastou ousian)). (Phys. ii.7 198b5-9)
[S]ince nature is twofold, the matter and the form, of which the latter is the end, and since all the rest is for the sake of the end, the form must be the cause in the sense of that for the sake of which. (Phys. ii.8 199a30-3)
These texts make clear that in one sense the teleological goal or end of living things is the
individual's growth to a mature state.
But second and just as importantly, Aristotle's commitment to the thesis that things'
natures are their ends expresses itself in a range of theses which display the normative character
of the end. In saying that the 'last stage' of a living thing is its end or goal Aristotle does not
claim that death (the last stage of mortal life) is the teleological goal of living things. Rather,
Aristotle has a normative conception of the last stage to be developed in living things; the
ultimate stage, the form to be attained and the best state for the being are identical.
[T]he nature is the end or that for the sake of which. For if a thing undergoes a continuous change toward some end, that last stage is actually that for the sake of which. (That is why the poet was carried away into making an absurd statement when he said 'he has the end for the sake of which he was born'. For not every stage that is last claims to be an end, but only that which is best). (Phys. ii.2 194a29-33)
[N]ature creates nothing without a purpose, but always the best possible in each kind of living creature by reference to its essential [tê(i) ousia(i)] constitution.10 Accordingly if one way is better than another that is the way of nature. (IA 2 704b15-6)
[N]ature makes nothing without purpose, but always regards what is the best possible for each individual, preserving the peculiar substance [tên idian ousian] of each and its
10 This is perhaps the canonical formulation of a principle which forms a near constant refrain in Aristotle's works. This formulation makes absolutely clear that the end in question is the attainment of the individual organism's completed or perfected form or nature.
287
essence [to ti ên autô(i) einai]. (IA 8 708a10; see Phys. ii.2 194a28-33; Met. v.16 1021b24-9; cf. Met. vi.3 1027b8-10, PN 478b23-5)
The teleological end is normative in the sense that it provides a standard by reference to which
other ends (=final stages) may be evaluated for their goodness (or health) and badness (or
disease). This normative conception of the natures of living things expresses itself in Aristotle's
doctrine that the positions (up, down, front, back, left and right) of parts of animals are
determined by their natures rather than relative to the universe.11 12
As further evidence yet for the commitment to this type of end (i.e., the attainment of
individual form), note that in discussing Empedocles' misguided conception of fire as a
proximate cause of life, Aristotle claims that the possession of a limit to growth is a sign of life.
[W]hile the growth of fire goes on without limit so long as there is a supply of fuel, in the case of all complex wholes formed in the course of nature there is a limit [peras] or ratio [logos] which determines their size and increase, and limit and ratio are marks of soul but not of fire, and belong to the side of account rather than that of matter. (DA ii.4 416a15-8; see GA ii.6 745a5-10, iv.4 771b33-772a29)13
11 See IA 4 705a27-b2, 705b5, 705b19, 5 706b10-16, OH ii.2 284b30-3, 285a14-9, ii.13 293b-9, PA iii.4 665b18-21, DA ii.4 416a5. 12 The connection between the normativity of the final stage and Aristotle's conception of the places of organs relative to function is this. In neither case does Aristotle accept an absolute or statistical account of the norm or standard in question. In both cases, the measure for a thing's good state (or place) is determined not with reference to some nonrelative/absolute standard but with reference to the thing's nature and intrinsic teleological ends: hearts are in the center of the body no matter where they are found in an absolute sense because their function is central to life (see PN 468b28-9a9, PA ii.1 647a22-30, iii.7 670a23-27, iv.5 678b1-6). 13 Interestingly, a modern biologist � Richard Dawkins � employs Aristotle's metaphor to a similar end. Dawkins distinguishes between reproduction and heredity, noting that a fire may 'reproduce' in the sense that one fire may be the cause of distinct fires' starting. Despite this fact, however, the 'daughter' fires will not inherit characteristics from the 'parent' fire � the color of the flame, etc., will all depend on the local conditions of the offspring fire rather than having been inherited from the parent. Dawkins says, "But in fact fires don't resemble their parents any more than they resemble the general run of fires dotted around the prairie. An individual flare-up gets its characteristic qualities, its flame size, smoke colour, crackle volume and so on, from its surroundings; from the kind of grass that happens to be growing where the spark lands; from the dryness of the grass; from the speed and direction of the wind. These are all qualities of the local area where the spark lands. They are not qualities of the parent fire from which the spark came." (Dawkins 1996, pp. 88-9) In Aristotelian terms, the fire lacks limit (peras) in the senses that it does not (a) tend to grow to a particular mature form; (b) tend to preserve that form when achieved; (c) aim at maintaining that form in a flourishing state; or (d) have a goal of reproducing offspring of the same form. These are all features of biological organisms that fires lack, and the lacks are indicative of the fire's lack of teleological directedness. Indeed, at Physics ii.1 193b7-12 these are the characteristics of living things which inspire Aristotle to argue against reductive materialistic accounts of living things.
288
The teleological nature of the limit that soul puts on growth is clearly indicated by a passage
already quoted above: "all living things both move and are moved for the sake of something, so
that this is the limit of all their movement � that for the sake of which" (MA 6 700b15-6; see
also GA ii.6 742a19-21, v.1 778b4-6). We ought now to acknowledge an enriched conception of
the normativity of the limit of living things. The forms toward which a juvenile's growth is
directed is a standard for evaluative assessment of the its growth; the last point of its growth to
maturity may be excellent or may fall short by reference to the norm imposed by the teleological
ends of growing for creatures of that kind.
Finally, the normativity of the end state of living beings expresses itself in Aristotle's
commitment to the thesis that the goal is not the mere attainment and preservation of form, but
the attainment of the form in its best or flourishing state. We will approach this point through
Aristotle's discussion of sleeping and waking.
[L]et the work of the soul be to produce living [ergon to zên poiein], this consisting in employment [chrêsis] and being awake � for slumber is a sort of inactivity and rest. (EE ii.1 1219a24-5)
Since the employment of different types of soul will consist in different activities corresponding
to the capacities of each type of soul, the actualization of different types of soul will be different
(see IA 2 704b15-6, IA 8 708a10, Met. v.16 1021b20, Phys. vii.3 246a10-15, quoted above).
Aristotle then specifies the goal we seek (relative to beings with thought and perception):
[L]ife is defined in the case of animals by the power of perception, in that of man by the power of perception or thought; and a power is referred to the corresponding activity, which is the essential [kurion] thing; therefore life seems to be essentially [kuriôs] perceiving or thinking. (NE ix.9 1170a15-9)14
[T]he waking state is the goal [to telos], since the exercise of sense-perception or of thought is the goal [to telos] for all beings to which either of these appertains; inasmuch as these are best, and the goal is what is best [to de telos beltiston]. (PN 455b22-4)
14 Note that ends in this sense will be taken by Aristotle to be definitive of particular types of life; this point is emphasized below. In this section I am merely pulling out the various strands of Aristotle's teleological conception of living things � that some ends are definitional whereas others are not does not lessen the interest in noting the richness of Aristotle's teleological conception of life.
289
For to be <for animals and man> was to perceive or to think. [to gar einai ên aisthanesthai ê noein]. (NE ix.9 1170b1; cf. DA ii.4 415b13)
Similar conclusions will obtain for plants; their goal will be the flourishing peculiar to their
natures, and such a flourishing state will not consist in the actualization of perceptual or
cognitive capacities but in the actualization of nutritive and reproductive capacities with
reference to the healthy state of each kind of plant.15
[E]xcellence [aretê] is a completion [teleiôsis]; for each thing is complete [teleion] and every substance [ousia pasa] is complete, when in respect of its proper kind [eidos] of excellence it lacks no part of its natural magnitude. (Met. v.16 1021b20)
[E]xcellence [arete] is a completion [teleiôsis] (for when anything acquires its proper excellence [tên autou aretên] we call it complete [teleion], since it is then really in its natural state.) (Phys. vii.3 246a12-15)
The end of living creatures is thus not merely the attainment of an excellent form, but the proper
exercise of that kind of form's peculiar capacities; the best state is the active state, the fulfillment
of natural capacities.
Again, this thesis that the end state is the flourishing state of the living thing is reflected
in Aristotle's general conception of the final cause as the cause that acts qua good.
That for the sake of which actions and changes and movements take place, they assert to be a cause in a way, but not in this way, i.e. not in the way in which it is its nature to be a cause.... In the same way those who say the One or the existent is the good, say that it is the cause of substance, but not that substance either is or comes to be for the sake of this. Therefore it turns out that in a sense they both say and do not say the good is a cause; for they do not call it a cause qua good but only incidentally. (Met. i.7 988b8-15; cf. Met. xii.10 1075a37-b1)16
Aristotle frequently claims that the end aimed at is the best state with reference to the type of
being in question (Phys. ii.2 194a28-33, ii.3 195a23-5, ii.7 198b8-9; Pol. i.2 1252b34-5; EE i.8
15 See HA viii.1 588b24; GA i.4 717a21, i.8 718b9-10, i.23 731a25-731b7, ii.1 735a15-7. 16 Aristotle is criticizing his predecessors' views in this passage and others like it, but such passages often give us the best insight into Aristotle's own position: Aristotle's criticism of an opponent for failing to account for some feature X of the phenomena in terms Y is strong prima facie evidence that Aristotle himself considered it a condition of adequacy on satisfactory accounts that they account for X in terms of Y. I defend this passage as indicative of Aristotle's own conception of the nature of the final cause in chapter four, above.
290
1218b9-11, ii.1 1219a9-11; Met. i.3 983a31-2). We have, thus, a very robust conception of the
normativity of the end in Aristotle's biology and general teleology.
The goal of the attainment of form is further specifiable, however. The goal of attaining
form manifests itself not only in the (normative) goal of attaining the individual form, but of
preserving the form once achieved. In discussing the functions of the nutritive soul in
Nichomachean Ethics i.13 Aristotle claims that this same soul is possessed alike by embryos,
nurslings and mature members of species and that its excellence is common to all living things
(NE i.13 1102a32-b4). Mature beings will not be growing, but will take on nutrition for the
maintenance of their mature state; doing this will be a function of the nutritive soul, and doing it
well will be an excellence of the nutritive soul (see NE i.7 1097b26-8). Thus, Aristotle says,
Generation is the initial participation, mediated by warm substance, in the nutritive soul, and life is the maintenance of this participation. (PN 479a28-9)17
[T]he psychic power which we are now studying may be described as that which tends to maintain whatever has this power in it of continuing such as it was. (DA ii.4 416b17-18)
This goal of maintaining the mature form of an organism is clearly endorsed by Aristotle.
Aristotle, then, specifies a number of distinct yet related ends for living things. In the
most general sense, the end for (mortal) living things is the attainment of soul. This general end
can be seen to divide into nonproximate ends for eternally preserving species form, proximate
reproductive ends for producing immediate descendants, and nutritive ends of growth into a
mature state, preservation of individual life, and the attainment of flourishing states in the
individual organism where those states of flourishing are relative to the psychic capacities of the
species in question.
There is no incompatibility in suggesting that all of these things constitute ends for
living things; indeed, Aristotle's use of teleological notions in this systematic web of
17 Recall Aristotle's numerous identifications of living with the possession of the nutritive soul. See DA ii.2 413b1-2, ii.1 412a14, ii.4 415a24-5; PN 479a28-9; GA ii.4 741a1.
291
interconnected ways constitutes strong evidence for his acceptance of a teleological
characterization of being alive.
7.1.2 Evidence that teleology defines life.
In the previous sub-section I argued that Aristotle closely connects a nested series of
ends with the possession of life. The arguments of that section do not, however, establish that
Aristotle defined life teleologically rather than simply finding life and teleology to be closely
related. The purpose of this sub-section is to establish the strong claim that Aristotle in fact
endorses the definition of living in teleological terms. This project proceeds in two steps. First I
make explicit the argument for the claim that life and teleology are coextensive in Aristotle, then
I move to establish the stronger claim that Aristotle actually defines life in terms of the
possession of teleological directedness.
In chapter three I argued that we lack good reason to suppose that Aristotle held any
nonliving beings to possess intrinsic teleological directedness. In this section, I argue that
Aristotle held that the being alive is necessary and sufficient for being teleologically directed.
That being alive is sufficient for teleological directedness is made clear in Aristotle's treatment of
the movements of the heavens. In Aristotle's view, once we recognize the heavens to be living
beings we must recognize that despite their uniform shape they have functional distinctions
which set off their 'right' and 'left' sides (OH ii.2 285a28-31). Further, being alive is, in the same
context, treated as a necessary condition for being teleologically directed. We can understand the
movements of the heavens to be goal directed only if we recognize them as being alive (OH ii.12
292a19-292b24). Further, as Aristotle says elsewhere, we cannot suppose something to have
nonteleological desire (orechis) (DA iii.10 433a15),18 but we cannot suppose anything with such
18 Again, it is important to recall that Aristotle frequently uses mentalistic sounding terms such as 'desire' to describe processes which he finds to occur in beings which have, on his view, no minds. The teleological use of 'desire' is broader than its use as a psychological notion used in explaining behavior mentalistically.
292
orechis not to be alive (EE i.8 1218a27): orechis is teleological, and sufficient for life.19 Further
evidence comes from the fact that health and disease are coextensive with living (PN 436a18,
Rhet. i.6 1362a30), and these are teleological concepts (HA x.1 633b16-20, and Rhet. i.5 1361b3-6
with NE ii.6 1106a15-7 and EE ii.1 1218a38). These considerations show that not only do we
lack grounds to extend teleology beyond the living in Aristotle, we have grounds to suppose
being alive is necessary and sufficient for being teleologically directed.20
Evidence that Aristotle endorses the stronger position that life is defined teleologically
comes from application of a principle of kind individuation to the case of life. Aristotle commits
himself to a metaphysical principle of kind individuation which we may call21 the functional
determination thesis. This view is paradigmatically stated in the following passage:
What a thing is is always defined by its function ((h)apanta d' estin (h)ôrismena tô(i) ergô(i)): a thing really is itself when it can perform its function. (Meteor. iv.12 390a10-11)
The commitment is widely affirmed throughout the corpus, however.
[E]ach thing is defined by its end. (EE iii.7 1115b23)
Everything which has a function exists for its function. (OH ii.3. 286a8)
[T]hings are defined by their function and power. (Pol. i.2 1253a23)
[S]ince it is right to call things after the ends they realize... (DA ii.4 416b23)
[I]t is right to call anything that which nature intends [bouletai] it to be, and which belongs to it, rather than that which it is by constraint and contrary to nature. (OH ii.14 297b21-2)
Aristotle is committed to a thesis concerning kind individuation which would, if applicable,
resolve the issue of whether life was defined by or merely associated with teleological
19 The references, quoted above, from DA ii.4 415a27-b2, b15-20 and MA 6 700b15-6 all constitute additional support for the claim that life is sufficient for teleological directedness. 20 The necessity claim may be doubted given that artifacts possess ends but are not alive; I deal with this objection below when I argue that Aristotle distinguishes between intrinsic and derived teleology. 21 Following Shields (1999).
293
commitment. We must, of course, ask whether Aristotle applied this metaphysical principle to
the case of living beings.
It is clear that he does. We may here cite again Aristotle's broad commitment to the
claim that living beings' last state (in the normative sense) is their nature, the attainment of their
mature form, and their end. Thus,
[T]he end is developed last ((h)ustaton gar ginetai to telos), and the peculiar character of the species is the end of the generation in each individual. (GA ii.3 736b3-4. See also Phys. ii.7 198b5-9; IA 2 704b15-6, 8 708a10; Met. v.16 1021b20; Phys. vii.3 246a10-15, quoted above.)
Further, Aristotle applies this thesis to the lives of plants, animals, and men in the famous
function argument in Nichomachean Ethics i.7.
[I]n general, for all things that have a function (ergon ti) or activity (praxis), the good and the 'well' is thought to reside in the function, so would it seem to be for man, if he has a function. . . . What then can this be? Life (to zên) seems to be common even to plants, but we are seeking what is peculiar to man. Let us exclude, therefore, the life of nutrition and growth. Next there would be a life of perception, but it also seems to be common even to the horse, the ox, and every animal. There remains, then, an active life of the element that has a rational principle. (NE i.7 1097b26-7, 1097b33-1098a4)
Life is here identified as a function common to all living things. Specific functions for particular
types of living things are specified by reference to the most complete activities of beings of the
same type.
As a final support for this definitional thesis we may note again that Aristotle identifies
the life of a living thing with its essence.
That it is substance (ousia) is clear; for in everything the substance ((h)ê ousia) is identical with the cause of its being (to aition tou einai), and here, in the case of living things, their being is to live (to de zên... to einai estin), and of their being and their living the soul in them is the cause or source. (DA ii.4 415b12-4)
294
The essence of a living creature is its life, and it is the essence that we specify in the definition of
a thing.22 This is identified with the final cause of living beings at de Anima ii.4 415b15-20 and
Generation of Animals i.1 715a4-8 (quoted above).
The end of creatures in the relevant sense is the attainment of their mature form (with
all the capacities and activities that accompany this). But the attainment of form for distinct
creatures is actualized in the attainment of different forms. The attainment of form is the
actuality of living creatures,23 and actualities are identified with their objects.24 But the objects
(= ends) of the various souls will be different: the object of the nutritive soul is to cause
nourishment and growth to the attainment of mature form; the object of the reproductive soul is
to reproduce the species form, etc.
If these arguments are correct, then Aristotle should recognize distinctions in the kinds
of ends possessed by what he considers to be different classes of living things, and this is exactly
what we find. The ends and definitions of life for different living things ought to be relative to
their specific natures as the kinds of living things they are. Applying this to living things on a
case-by-case basis we have the following results.
Those things which cannot reproduce, but live (i.e., spontaneously generated things)
possess the nutritive soul only, and are defined by the peculiar ends of that soul (DA ii.1 413a31-
b1, iii.12 434a23; PN 454a13). Plants which possess both the nutritive and reproductive souls
are defined by their possession of the ends of the reproductive soul (HA viii.1 588b24-5, GA
i.23 731a25-8). Animals possess the nutritive and sensitive souls (and usually the reproductive)
and are defined by the ends of their sensitive soul (PN 454b24, 469b4, PA ii.5 651b4, ii.8
653b23, iii.4 666a35, GA i.23 731b3, NE ix.9 1170a15). Humans possess all these souls plus the
22 See APo. ii.10 93b29-30; Top. i.4 101b20-2, i.5 101b31, i.8 103b9-10, vii.3 153a15-22, vii.5 154a31; Met. viii.1 1042a16-7. 23 Form is actuality. See Phys. ii.1 193b7; DA i.1 412a9, 412a21, ii.2 414a15; Met. xii.5 1071a8-9, ix.8 1050b2. 24 See DA ii.4 415a16-21; NE vi.1 1139a9, ix.9 1170a15-19; Met. ix.8 1049b4-16.
295
rational (NE ix.9 1170a15-8, 1170b1, Pol. vii.15 1334b15). Finally, god possesses only the
contemplative soul (Met. xii.7 1073a4-13 and NE x.8 1178b20-3).
We have ample evidence, then, that Aristotle defined life in teleological terms. Life and
teleology are coextensive in Aristotle's view and Aristotle consciously applied a metaphysical
principle of kind individuation that invoked teleological ends to the case of living things. This
application, moreover, was no passing thought but finds itself deployed systematically in his
remarks about the natures of different types of living things.
7.1.3 Puzzles.
Despite the amount and variety of evidence that may be produced in favor of the thesis
that Aristotle defines living in terms of being teleologically directed, a number of puzzles present
themselves. The puzzles I address here take a familiar form given our earlier discussion of
philosophical problems raised by life's definition (see chapter one) and surrounds the question of
how a definition of life can succeed in bringing unity to the diversity of things which are or could
be alive.
In an Aristotelian context, this question may be raised for our understanding of a class
of living things which we no longer recognize: spontaneously generated living things which
cannot reproduce. How can such creatures be alive given the teleological account of life as I
have explained it thus far? Further, and of greater contemporary interest, the problem of the
unity of definition arises in Aristotle with reference to god's life. How are we to understand
god's living on the proposed teleological account of life: given the widely shared intuition that
nonmaterial entities such as gods or angels would be alive if they did exist, how can Aristotle �
who shares a commitment to this intuition � explain such lives?
7.1.3.1 Spontaneously generated living things.
Aristotle's commitment to the existence of a class of living things which are
spontaneously generated is well known. Aristotle says, for example:
296
Other insects are not derived from living parentage, but are generated spontaneously [automata]: some out of dew falling on leaves, by nature in spring-time, but not seldom in winter too when there has been a stretch of fair weather and southerly winds; others grow in decaying mud or dung; others in timber, green or dry; some in the hair of animals; some in the flesh of animals; some in excrements: and some from excrement after it has been voided, and some from excrement yet within the living animal, like the intestinal worms. (HA v.19 550b34-551a10. See also GA iii.11, Met. vii.7 1032a30-2, etc.)
Spontaneously generated organisms have goals (MA 6 700b4). For example, testaceans are
spontaneously generated (GA iii.11 763a24 ff.) and have sensation (HA iv 535a8), which is an
indication of the presence of desire (DA ii 413a24), and all desire is teleological (DA iii.10
433a15). Further, spontaneously generated organisms have parts with functions (see HA iv.7
532a5, a11; PA ii.8 653b38-654a2, iv.5 678b11-14, 679b35-6, iv.7 683b10-11). Nevertheless,
such creatures do not propagate after their own kind: "other bloodless animals generate indeed,
but not offspring of the same kind; such are all that come into being not from a union of the
sexes, but from decaying earth and excrements" (GA i.1 715a23-5).25 Given then that such
creatures lack reproductive ends, how can it be the case that they are living things if being alive is
analyzed teleologically?26
The solution to the puzzle is contained in our earlier discussion of the structure of ends
Aristotle recognizes for living things. Aristotle recognizes individual ends for growth to mature
form, maintenance of that form, and the flourishing possession of that form for individual living
25 Aristotle's explanation of the fact that the spontaneously generated things do not propagate after their own kind is given at Generation of Animals i.1 715b4-15. There he explains that if they had reproductive ends, then they would generate after their own kind � which they do not. If, contrary to fact, they generated after their own kinds then we should have expected the same to have been true of the spontaneously generated original living thing, but according to the hypothesis that these are spontaneously generated things that supposition is false. Aristotle therefore appears willing to give up the claim that such creatures have reproductive ends more readily than he will give up the claim that things which generate after their own kind have such an eventuality as an end or goal. 26 This problem has contemporary relevance despite the fact that science has shown there not to be a class of spontaneously generated living things. It is relevant because whether or not such beings exist in the actual world, they appear to be possible, and insofar as an account of life ought to account for not only all actual but all possible living beings, the case continues to have relevance to our account of life. My thanks to Chris Shields for emphasizing this point.
297
things, and spontaneously generated things may have such ends; as such they will qualify as living
things.
The fact that nutritive ends can in this way be distinguished from reproductive ends
strongly suggests that Aristotle's claim that the nutritive soul is the reproductive soul (DA ii.4
416a20) should be understood to commit him only to the weaker claim that the two souls are (in
general) coextensive. This hypothesis is borne out by Aristotle's explanation of the unity of the
nutritive and sensitive souls at Generation of Animals ii.4 740b29 ff. In that passage Aristotle
argues that because the reproductive soul is the same entity that gives to the new embryo the
powers of the nutritive soul, the two must be one and the same from the start. Clearly, however,
this argument commits Aristotle only to the coextensiveness of nutritive and reproductive souls
in the normal case of non-spontaneously generated living things.
The hypothesis that the ends of the nutritive soul alone are sufficient for being alive also
explains Aristotle's repeated claim that it is the possession of the nutritive soul (not the
reproductive soul) which explains the life of mortal things (DA ii.1 412a14, 413a31-b1, iii.12
434a23; PN 454a13). The common factor in the life of (mortal) living things with and without
reproductive souls is their teleological directedness towards growth, maturity, maintenance of
form, and flourishing biological functioning.
While the existence of creatures without reproductive ends presents a prima facie puzzling
case for an Aristotelian account of life in terms of teleology, I conclude that further investigation
reveals the case to add support to the hypothesis of a teleological definition.27
27 It adds support because an account's ability to handle apparent problems from within its existing theoretical apparatus properly understood (i.e., not through the theoretical apparatus supplemented by ad hoc auxiliary assumptions) is a strong indication of a flourishing theory. In Aristotle's words, "Of some such kind are the difficulties that arise; some of these points must be refuted and the others left in possession of the field; for the solution of the difficulty is the discovery of the truth" (NE vii.3 1146b7-8; see also NE vii.14 1154a22-4, EE i.3 1215a6, Rhet. iii.17 1418b5, Phys. iv.4 211a6-11). To illustrate with a common contemporary example, the ability of Newton's theory of celestial mechanics to solve from within its own resources the problem of Uranus' orbit (by postulating, then finding, Neptune) constituted a powerful proof of the theory. Of course, such proofs are defeasiable in light of further insuperable problems; they nevertheless constitute strong evidence for particular theories.
298
7.1.3.2 God's life.
This same conclusion about the centrality of teleological directedness to living can be
seen at the opposite end of the spectrum of living things in the case of god's life. Aristotle
frequently remarks that the capacity to think seems to be an independent substance, an activity
possibly separate from material embodiment (see DA i.4 408b18, ii.2 413b25, iii.4 429b4; GA ii.3
736b27; NE x.8 1178a21, iii.10 1117b29). When Aristotle lists the powers associated with living,
the power to think is included in the lists (see DA ii.2 413a21-5, ii.3 414a29-32; see also Top. vi.10
148a27 ff.). The clear implication is that a living thing which had no body but that could
nonetheless think would count as living, and Aristotle's theoretical commitment to god's life bear
this out.
[I]f you take away from a living being action, and still more production, what is left but contemplation? Therefore the activity of God, which surpasses all others in blessedness, must be contemplative; and of human activities, therefore, that which is most akin to this must be most of the nature of happiness. (NE x.8 1178b20-3)
[L]ife also belongs to God; for the actuality of thought is life, and God is that actuality; and God's essential actuality is life most good and eternal. We say therefore that God is a living being, eternal, most good, so that life and duration continuous and eternal belong to God; for this is God. (Met. xii.7 1072b26-30; see also Met. xii.7 1073a4-13, OH i.9 279a19-22, and ii.3 286a8-9)
As I mentioned in chapter one in reference to this commitment, the intuition that if there are
such things as god(s) or angels or other such immaterial beings which do think then such beings
are alive, is both widely shared and has a long history in Western thought. If we take intuitions
as evidence in forming our theories28 then there is a strong presumption in favor of an account
of life being able to explain how it is that beings such as Aristotle's god are alive.
Can the teleological account do this? We would suppose not only if we believed that
thinking has no end, that it is not directed toward any object as its goal. On Aristotle's view,
however, this is false.
28 George Bealer argues a number of papers that this is the appropriate stance to take towards intuition in. See Bealer (1992; 1996a; 1996b; 1998; 1999). See also Bonjour (1995; 1998) and Katz (1995; 1998).
299
[T]he end of theoretical knowledge is truth, while that of practical knowledge is action. (Met. ii.1 993b21)
The function of both the intellectual parts, then, is truth. (NE vi.2 1139b12)
[O]f the intellect which is contemplative, not practical nor productive, the good and the bad state are truth and falsity (for this is the function of everything intellectual). (NE vi.2 1139a26)
[S]ince there are two parts of the soul, the excellences are divided correspondingly, those of the rational part being the intellectual, whose function is truth, whether about a thing's nature or genesis, while the others belong to the part irrational but appetitive � for not any and every part of the soul, supposing it to be divisible, is appetitive. (EE ii.4 1221b28-30)
There exist plausible ends for contemplative activity, and Aristotle attributes such ends to
immortal beings who would have none of the mortal psychic functions (i.e., the nutritive,
reproductive and the sensitive).
Once again, the common feature uniting the various examples of living things
mentioned by Aristotle is the possession of teleological directedness. This, coupled with our
earlier result (in chapter three) that teleology is not present in nonliving things in Aristotle's view,
presents more confirmation that Aristotle's account of being alive is at base teleological. The
hypothesis of a teleological account of life explains the unique cases at both ends of Aristotle's
hierarchical conception of living things passing up from spontaneously generated things (which
have the nutritive power and ends, but not the reproductive power), through normal living
things such as plants (which have nutritive and reproductive powers and ends) and animals
(which have these plus sensitive capacities and ends), humans (which have all of the previous and
can also think) and finally god (which again possesses only one life capacity � that of thinking).
7.1.4 Section summary.
In this section I have argued on textual grounds that Aristotle not only systematically
associates life and teleological directedness, but that he actually defines living in teleological
terms. Aristotle's metaphysical thesis that at least some kinds are determined by their ends or
functions is applied by him in the case of living things such that the ends of the normal mature
living being determine the essence to be specified in the account of a life of that sort.
300
7.2 The unity and diversity of living things.
Thus far in this chapter I have argued for two theses. First, ample evidence suggests
that Aristotle correlates teleology and life in systematic ways. Second, Aristotle actually defines
living in terms of the possession of teleological directedness. The substance or content of
Aristotle's account of life is therefore in place; all that remains is to become clear about the form
or structure Aristotle's definition of life takes.
My discussion thus far suggests that Aristotle finds there to be significant unity among
the cases of living things, and that this unity is provided by life's teleological directedness. On
the provisional definition I have been supporting:
LAPD: a is alive =df a is a teleological system.
It may be objected however that, whether or not this can suffice as a contemporary account of
life, such an analysis is too neat to be Aristotle's. The problem for my account is raised by
Aristotle's apparent commitment to the homonymy of life. Aristotle says,
Now life (zên) is said in many ways, and provided any one alone of these is found in a thing we say that thing is living � viz. thinking or perception or local movement and rest, or movement in the sense of nutrition, decay and growth. (DA ii.2 413a22-5; translation slightly altered from ROT)
Life seems not to be said according to one form ((h)ê de zôê ou kath (h)en eidos dokei legesthai), but to belong differently to animals and to plants. (Top. vi.10 148a29-30, translation mine; see also DA ii.3 414a29).
As noted earlier in chapter two, Christopher Shields takes Aristotle's commitment to the
homonymy of life in these passages as essential to a proper understanding of Aristotle on the
nature of life. If life is 'spoken of in many ways' or is 'not of one kind', then it would appear that
the sort of interpretation I have given thus far cannot be correct, for Aristotle recognizes more
diversity among types of life than my account allows.
In this section I argue that my account of Aristotle on the nature of life has the
resources to account for Aristotle's theses concerning both the unity and the multiplicity of types
of life. My defense rests upon finding in Aristotle a form of definitional relation suitable for
301
acknowledging a univocal general account of life (in terms of teleology) and at the same time
distinct peculiar natures for types of life specified with reference to the peculiar ends of living
organisms of that type. This section is therefore devoted to the proper understanding of the
form of Aristotle's definition of life.
7.2.1 Homonymies and synonymies.
When Aristotle accepts the homonymy of life, he is at least rejecting the claim that living
things are alive synonymously.29 As we have elucidated the notions of homonymy and
synonymy,30 the question at hand has to do with whether the accounts of what it is to be alive
for different types of living things are the same (and the things are therefore synonymously
alive), or different (in which case the things are homonymously alive), and if they are different
whether the accounts have nothing (discrete homonymy) or something important (associated
homonymy) in common.
Recall that on Shields' (1999) account of associated homonymy (which I accept � with
two caveats31), there being 'something important in common' in the definitions is equated with
satisfying the conditions for being a focal homonym. These conditions are analyzed in terms of
there being core and derived cases of Fs, where derived cases of being F asymmetrically depend
for their being F on the core case of being F. The form of dependence in question is causal
dependence of one of the four sorts Aristotle recognizes.32
29 It is important to remember that for Aristotle it is things or properties rather than terms which are homonymous or univocal. See Ackrill (1963, 71) and Shields (1999, Part I). 30 On the notions of synonymy and homonymy, see the discussion of Shields' interpretation of Aristotle on the nature of life in chapter two. 31 The first caveat to my acceptance of Shields' (1999) analysis of focal homonymy concerns my rejection of his thesis that derived cases of being F can depend on core cases of being F through relations of formal causation. I explain my reasons for rejecting this particular case in chapter two of this dissertation. I explain the second caveat in this section. 32 The issue was discussed above in chapter two, but here is an illustration. Thus, both my salad and I are healthy, but we are healthy in a focally homonymous way: I am healthy in a core sense (my body is functioning well and being health =core df having a body which is functioning well) whereas my salad is healthy in a derived sense (in the sense that it is causally effective in maintaining my body in a condition of functioning well). The derived sense of being healthy has 'something' in common with the core sense: its
302
I argued above in chapter two that living things do not present us with a case of focal
homonymy either in fact or in Aristotle. There is no division among core and derived cases of
life such that the derived cases depend on the core cases for their being alive in a coherent and
textually supported four-causal fashion. Shields' account could not identify an Aristotelian core
for life and his suggestion that the appropriate causal tie between alleged core and noncore cases
of living was a formal tie failed to receive adequate textual support as a commitment of
Aristotle's, relied for its motivation on examples which were both textually and philosophically
problematic, and its suggested mode of causal dependence was insufficiently explicated for a
dialectically charged context. I therefore concluded that life is not a focal homonym.33
If this is the case, are we forced to accept the conclusion that Aristotle held life to be a
discrete homonym? Such a result would appear to be unacceptable both exegetically and
philosophically. It is exegetically unacceptable because our investigation of Aristotle's
teleological characterization of living things combined with the fact that Aristotle's teleology is
restricted in its scope to living things strongly suggests that the accounts of all living things do
share something important in common � their teleological directedness. Life in Aristotle
appears not to be a discrete homonym. It is unacceptable philosophically because even if we are
willing to grant Aristotle the claim that plants and animals are alive in different ways, it is
extremely implausible to suggest that the accounts of living for plants and animals have nothing
in common in the same way that the accounts of being cranes for birds and construction cranes
have nothing in common.34
account makes essential reference to being healthy in the core sense. The account of being F for the core case, in contrast, makes no reference to the accounts of being F in derived cases. 33 See again the discussion of Shields' account of life in chapter two for the arguments in support of these conclusions. 34 So implausible that we may wish to echo Aristotle's remark concerning the homonymy of velocities: "It would be absurd (atopon) to suppose that the motion (kinêsis) of one thing in a circle and of another in a straight line cannot be the same ((h)omoiôs), but that the one must inevitably move more quickly or more slowly than the other, just as if the course of one were downhill and of the other uphill." (Phys. vii.4 248a19-22; translation revised from ROT)
303
So long as we restrict our attention to the types of definitional relations attributed to
Aristotle thus far, we seem unable to make sense of the form of Aristotle's definition of life. In
the next section I motivate my solution to this problem. We have strong grounds to deny that
life is synonymous, discretely homonymous, and focally homonymous. Could a definition of life
take another form?
7.2.2 A starting point: the parallel between figure and soul.
Julie Ward (1996, 127) discusses a similar problem, the problem of the unity of
Aristotle's definition of soul. Since according to Aristotle the soul is the cause of life, we may
follow her discussion of the nature of the soul as a clue to answering our puzzles about the
nature of life.
Ward focuses her discussion of the problem of the definition of soul on Aristotle's
extended comparison of figure and soul in de Anima ii.3.
It is clear, then, that there would be one account of both soul and figure in the same way. For in the one case there is no figure beyond triangle and those that follow in order <quadrilateral, etc.>; in the other case, there is no soul beyond those mentioned. There might be a common account (logos koinos) for figures, which fits them all; but it will not be peculiar to any one of them. Similarly, <there might be a common account> for the souls mentioned. For this reason, it is ludicrous (geloion) to seek a common account for these and other cases, which will not be an account peculiar to any of the things which exist, nor in accordance with the proper and indivisible species, while omitting this sort of thing. (Things falling under soul are similar to what occurs concerning figures; for what is earlier in the series for both figures and ensouled belongs in potentiality to what follows, e.g. the triangle in the square and the nutritive capacity in the perceptual.) Consequently, one must inquire into each individually, what the soul of that is, for example, what the soul of a plant or a man or a beast is. One must look at the cause in virtue of which <things> stand in this serial way. (DA ii.3 414b20-34; trans. Shields)
Ward notes that in this passage Aristotle claims that figures and souls constitute series and she
seeks to uncover in this parallel a significant form of unity within multiplicity for both figures
and souls.35 Her attempt comes up short, however, and she suggests that we accept that
35 Julie Ward argues that "soul is either one by analogy or one by focal reference", and concludes that "the feasibility of explaining soul's unity adequately by means of analogical resemblance seems remote" (1996, 125-6). I agree with the reasons Ward offers for rejecting the analogical solution, and would add that it is
304
Aristotle has raised a problem here (whether that problem is real or imagined) for which he has
no solution. Ward concludes that "while Aristotle aims to demonstrate that soul is a unifying
principle of organization of the living thing, the goal eludes his grasp" (1996, 127).36
I will argue that Ward's sceptical conclusion concerning Aristotle's analysis of soul is
unwarranted. The text (quoted above) that constitutes the focus of Ward's article contains the
key to the form of Aristotle's definition of life, but Ward fails to find unity in the soul in her
discussion of the passage because she neglects to consider whether the order we find in the
figure/soul case may itself provide a primitive sort of unity to a class in Aristotle or in fact.
When we consider this possibility, however, Aristotle's theoretical resources for defining
different but importantly related forms of life, figure and soul are revealed to be philosophically
rich, defensible and textually sound.
Taking the case of figure, Aristotle notes in his philosophical lexicon in Metaphysics v.6
that isosceles and equilateral triangles are one figure in the sense that their kind (genos) � figure
� "is one though distinguished by opposite differentiae" (Met. v.6 1016a24-32; see also Met. v.28
1024b1-9).37 This meshes well with Aristotle's claim in de Anima ii.3 that figure (and soul) might
be given a general definition which applies to all the cases. It is apparent that Aristotle
recognizes enough unity to the class of figures to suppose that their all being called 'figures' is
more than a mere chance happening; they do constitute a kind with sufficient unity to seek a
general account.
unclear in principle that analogical unity is sufficient for philosophical purposes. Aristotle raises the main problem himself in discussing analogical unity: "you cannot get one identical thing which pounce and spine and bone should be called; but there will be things that follow them too, as though there were some single nature of this sort" (APo. ii.14 98a21-3). According to Metaphysics v.6 two things are one by analogy which are "related as a third thing is to a fourth" (1016b35), but Aristotle places no constraints on the ways in which such things may be related, and without such constraints analogical unity appears distinctly ill-suited to providing the unity of definition. I have argued against the hypothesis that life is a focal homonym above. 36 While Ward's article is focused on the unity of Aristotle's definition of soul, the same problem arises for life given that soul is the cause of life. 37 Similarly, at Posterior Analytics ii.13 97b31-36 shape is claimed to have a genus.
305
But if there is such unity, what does Aristotle indicate in arguing that life � and by
implication soul and figure � is homonymous? This question is most profitably approached
through focus on the fault Aristotle finds with the supposed general definitions of figure and
soul. If we can understand why Aristotle might claim that such a general account fails in these
cases we will have an understanding of the diversity and the unity Aristotle finds in life.
Common explanations for why it is that a general definition of figure or soul will be
uninformative are incorrect. Hamlyn claims that the core of the problem for the unity of these
general items has to do with the relations of priority and posteriority that obtain between the
instances of these kinds of things: "The difficulty in fact arises whenever a general term covers
things that form a progression or hierarchy" (1993, note to 414b20). The problem with the
general account, Hamlyn claims, is that "it will omit the crucial point that figures and souls form
a progression" (ibid.). Hicks concurs. He says:
A. holds the triangle to be the simplest rectilinear figure (seeing that two straight lines cannot enclose a space). Given the triangle, the quadrilateral, pentagon etc. can be derived from it by combining two, three, or more triangles; while, if there were no such thing as a triangle, the other rectilinear figures, thus dependent upon it, could not exist. The triangle has thus a logical priority among rectilinear figures: it is presupposed by them. (1907, commentary on 414b29)
On this analysis as well, the problem is that the general definition fails to indicate the logical
relations or priority between the distinct kinds of figures.38
38 The cases of figure and soul have another feature which might be thought to be responsible for the failure of definition Aristotle mentions. Aristotle conceives of types of figures and souls as forming a series such that earlier members of the series are present potentially in later members of the series. Ordinary genus-species relationships are such that species under one genus exclude one another, but in the cases of figure and soul the distinct types are not mutually exclusive, at least in a weak sense. While no thing may be simultaneously actually a triangle and actually a square, it may be at the same time actually a square and potentially two triangles. While the cases of figure and soul do have this feature and it is true that standard genus-species relationships do not have the feature, it is not this feature that creates the problem for the general definition indicated by Aristotle in de Anima ii.3. There are at least two reasons for this. First, the 'lowest order' and most widely shared type of each series cannot be definitive of the series, because it is contained only potentially in the later members whereas kinds are specified by what they are actually. Second, Aristotle's general definitions of soul in de Anima ii.1 do not give the definition of the most widely shared soul, they seem rather to give a high level general definition which will apply to all souls, albeit in different ways depending on which type of naturally organized body is actualized in a particular case.
306
These explanations are incorrect, however. An indication of this fact is the
philosophical oddness of suspecting that there is a failure in a general definition for not revealing
the relations of priority and posteriority within the instances which fall under the concept. What
Aristotle actually says about things among which the relation of priority and posteriority is
present is that in such cases the genus cannot exist apart (Met. 999a6-7). While Aristotle makes
this claim in de Anima ii.3 passage on figures and souls (414b21), there is no suggestion that it is
this perfectly general fact about the genus that explains the problem with which we are
concerned. The reason the above analysis cannot provide the failure of the general definition is
that on Aristotle's view no genus can exist apart (see Met. vii.8 1033b20-34a8, vii.12 1037b29-
38a9, vii.14 1039a24-32, vii.15 1040a8-21), but some admit of adequate general definitions.
There is some other failure present in the cases of figure and soul which explains why the general
definition cannot be fully informative on its own.
The proof that these explanations are incorrect comes from a parallel case where there is
no relation or priority suggested or plausibly present. Aristotle says,
Now, sailors have different functions, for one of them is a rower, another a pilot, and a third a look-out man, a fourth is described by some similar term; and while the precise account of the excellence of each <type> will be peculiar <to that type> ((h)o men akribestatos (h)ekastou logos idious estai tês aretês), there is, at the same time, something common (koinos tis) applicable to them all. For they have all of them a common task (ergon), which is safety in navigation. (Pol. iii.4 1276b20-7, revised from the ROT; see also SE 31 181b33-4)
In this passage Aristotle indicates that the existence of common definitions applicable to all
things in a kind is compatible with sub-kinds' possessing precise definitions peculiar to particular
sub-kinds. A rower's function is to row, and this fully specifies the rower's definition (qua
rower).39 Nevertheless, rowing is a primitive way of serving the end of safety in navigation; by
being a rower one ipso facto satisfies the end of safety in navigation. Rowers provide safety in
navigation simply by doing their work, rowing, well.
307
There is no suggestion that this is a case of a progressive or hierarchical ordering, but as
with the case of figure and soul we possess a general definition which will fail to reveal any
portion of the precise nature of any particular kind of thing falling under it.40 Hierarchical order
is therefore unnecessary for the problem Aristotle indicates when he suggests that while a general
account is possible it will fail because it will fail to specify the nature of any particular kind of
thing.
Take another example. In Politics i.13 Aristotle discusses the excellences proper to
different sorts of human beings, and objects to Socrates' view41 that there is a good general
definition of excellence (aretê) which will apply to, e.g., men, women, and slaves. Aristotle says,
[T]hose who say generally that excellence consists in a good disposition of the soul, or in doing rightly, or the like, only deceive themselves. Far better than such definitions is the mode of speaking of those who, like Gorgias, enumerate the excellences. (Pol. i.13 1260a25-7)
Aristotle is committed to there being a difference in kind between ruler and subject, master and
slave (Pol. i.13 1259b37), and states in the passage just quoted that this results in a multiplicity of
excellences sufficient to justify enumerating particular accounts (just as in the figure/soul
passage). The difference in kind does not result in discrete homonymy between the excellences
of different types of humans, however, for:
[A]lthough the parts of the soul are present in all of them, they are present in different degrees. For the slave has no deliberative faculty at all; the woman has, but it is without authority, and the child has, but it is immature. So it must necessarily be supposed to be with the excellences of character also; all should partake of them, but only in such manner and degree as is required by each for the fulfillment of his function. (Pol. i.13 1260a11-16)
39 See Meteor. iv.12 390a10-11, EE iii.7 1115b23, OH ii.3. 286a8, Pol. i.2 1253a23, DA ii.4 416b23, OH ii.14 297b21-3, quoted below. 40 Likewise, the example displays the same failure of the general definition which appears in the cases of figure and soul but without it being the case that one type of sailor is potentially another type, without them constituting series such as figures and souls do. Again, this shows that it is not their constituting a series where earlier members are contained potentially in later members that causes the failure in the general definition which Aristotle finds in de Anima ii.3. 41 See Meno 71e-73c.
308
What appears to happen is this. There is a general account or excellence (i.e., good disposition
of the soul), but this general account is such that there are different primitive ways of satisfying
the conditions of the account as displayed by the different types of souls human beings may
have.
The excellences of each type of human being will therefore be different; the general
account of excellence offered by Socrates fails in failing to specify any portion of the nature of
the particular kinds of good disposition of the soul which are themselves definitive of the
particular kinds of human excellence. One needs to know not merely that a person's excellence
is the good disposition of the person's soul � one needs to know what type of soul is in
question, for the proper dispositions will be dramatically different in different cases (i.e., man,
woman, slave, child, etc.).42 It is absurd, then, to give only the general account, and Gorgias'
procedure is better (1260a27).
On the basis of these examples, then, we should conclude that the problem with general
definitions in the cases of figure, soul and living beings is not anything having to do with failing
to indicate logical relations between instances but is instead exactly what Aristotle says it is: the
general definition fails to specify any part of the specific nature of any particular instance (DA
ii.3 414b25-7). On this (i.e., Aristotle's) diagnosis of the problem, Aristotle recommends that in
such situations we look to the individual accounts if we want a satisfactory understanding of our
subject matter.
The connection between the diagnosis and the recommendation is presumably this: if
the general accounts were synonymous, then the accounts would apply in exactly the same sense
to each distinct species of figure, soul, or living being � it would provide a part of the
42 As will be common in explaining Aristotle's theoretical apparatus in this section, I should warn the reader that the importance or legitimacy of the type of definitional relation Aristotle believes obtains in the case of excellence does not hang on each of his alleged examples of that definitional relation actually being cases of the definitional relation. Presumably no one reading this dissertation will distinguish between types of human soul as Aristotle does here; nevertheless, my concern is not here evaluating his choice of examples but understanding the distinction he wishes to draw using those examples.
309
specification of the essence of such things which would appear without alteration in the
definitions of each particular species and would for that reason be informative, albeit not
completely informative about the particular natures.43 In the case of sailor, however, each
individual kind of sailor has a precise definition which is wholly distinct from the general
definition: while the end of rowing serves the end of safety in navigation, that it so serves safety
in navigation is no part of the definitional account of being a rower.44 Thus, the general account,
while applicable to rowers, will fail to specify any portion of the definatory account of being a
rower. Even if we are in possession of the general account of being a sailor (or figure or soul),
then, it makes eminent sense if we wish to understand sailors to go on to investigate the natures
peculiar to each particular kind of sailor.
This account of the problem diagnosed in de Anima ii.3 creates its own problem, for it
seems to reintroduce the specter of discrete homonymy. If the accounts of the particular kinds
of figures, souls, or sailors shares nothing important, but each has an individual nature all its
own, then these are paradigm cases of discrete homonyms. But if we rest satisfied with this
result, we will have captured the multiplicity Aristotle finds in these kinds of things only by
abandoning the search for the unity he claims exists as well.
The key to our problem appears through the contrast between the cases at hand (figure,
soul, and living being) and paradigm cases of the genus-species relationship.45 The genus-species
relationship is a paradigm case of a synonymous relation in Aristotle's conception: the account of
the genus applies univocally to each of the species.46 In the univocal genus-species case Aristotle
43 See Cat. 5 3b2-8, Top. ii.2 109b4-6, iv.1 121a11-5, iv.3 123a29. 44 Thus, note that Aristotle attributes to the heart a supreme function in the body parallel to the captain's role on a ship: other organs with functions or sailors with functions serve tasks subordinate to the ruling function (PN 468b28-469a9). Nevertheless, when Aristotle engages in defining particular organs according to their function, no reference is made to their subserving the ends of the heart; they have primitive functions all their own the exercise of which constitutes their serving the higher end of the heart (see, for example, DA ii.8 420b23-5, GA v.2 781b23-28, Phys. ii.8 198b17-33). 45 As Aristotle says, "contrast makes observation easy" (OH ii.6 289a10). 46 Top. ii.2 109b4-6, iv.1 121a11-5, iv.3 123a29; Cat. 5 3b2-8. cf. Cat. 1 1a6-12.
310
maintains that while defining the genus will not reveal the complete nature of the thing,47
nevertheless such a specification does not fail in the way the generic definition of figure and soul
fail � our task is to understand why.
I suggest that the reason is as follows. Where two things are synonymously F (animals,
say), while it is true that the specification of the genus of the animal will not reveal the complete
nature of the animal (noting that a thing is a bird does not explain what type of bird it is), it will
reveal a property whose specification will appear as an unmodified part of the specification of
the essence of the lowest level species. With cases such as figure, soul, and sailor, however, this
is not the case: specifying the general or generic nature does not reveal a part of the specification
of the essence of any sub-kind that will survive unmodified in the account of that thing's being
and which needs only supplementation by an independently specifiable property in order to have
a complete specification of the essence. The types of figures, souls, and sailors are primitive (not
independently distinguishable) ways of instancing the general property. This account is
developed and defended in what follows.
7.2.3 The account of unity.
We must not stop at the level of understanding of Aristotle's use at which we have
already arrived. What we need is a way of understanding this form of unity and multiplicity
within a kind in Aristotelian terms. I will argue that Aristotle supplies this lack in theoretical
terms using his apparatus of homonymy and synonymy. Aristotle recognizes a kind of
homonymy which displays the features illustrated above and which I will call 'account48
homonymy'. The account of unity within multiplicity which we find here will, I argue, provide
us with an understanding of Aristotle's attitude towards the unity and diversity of types of life.
47 It will not include the differentia. See Top. i.8 103b14, vi.4 141b26, vii.3 153b2, 153b14, Met. vii.12 1037b30 ff. 48 Called 'account' or 'logos' homonymy on the basis of Top. i.15 107b6-7.
311
7.2.3.1 Two modern models.
It will prove fruitful, before we turn to the complexities of Aristotle's account, if we
have models before us of the types of definitional relation I claim underlie account homonyms.
In proposing these models I do not argue that Aristotle clearly identified or distinguished either
of the forms of definitional relationship specified here; I do claim, however, that the cases of
account homonyms he does recognize fall neatly into these definitional patterns.
The first model is the relationship between determinables and determinates. David
Sanford describes a determinable as
a general characteristic or property analogous to a genus except that while a property independent of a genus differentiates a species that falls under the genus, no such independent property differentiates a determinate that falls under the determinable. (1995; see also Searle 1967).
Color is a determinable, and specific colors such as red and blue are determinates under that
determinable. Colors such as red and blue are primitives in the sense that one cannot specify
independently of characterizing them as, for example, red, which particular color they are; color
is not a genus to which one can add any differentia (which does not simply indicate the color
itself) to give the account of red. Nevertheless, it is true and informative to group blue and red
together as colors; there is good reason to hope for a general definition of color which will capture
something essential to the nature of any color (i.e., it is not an accidental but an essential feature
of red that it is a color) despite the fact that the accounts of individual colors will not express this
essential nature as an unmodified element of the specification of their definition. Being red is
just a primitive way of being a color.
The second model I will employ comes from second order definition schemas. Such
schemas present an account of a thing which contains a variable within it which may be variously
filled. Thus, consider the following set of (rough) definitions.
Figure: x is a figure =df x is a y line enclosed area.
a is a circle =df a is a 1 line enclosed area.
312
a is a triangle =df a is a 3 line enclosed area.
a is a rectangle =df a is a 4 line enclosed area.
etc.
Figure fails to appear unmodified in the particular definition of any kind of figure. Such a
definition captures what it is to be a figure in a generic way � without specifying the nature of
any particular kind of feature.
My hope is that these models will help us to understand Aristotle's terminology in
describing certain cases, below. Aristotle clearly does not employ the terminology I have used to
explain these cases; nevertheless, it is imperative to realize that the relations specified above find
no home in a theoretical apparatus restricted to discrete and focal homonymy and synonymy, for
these relations are distinct from any of those definitional relations. What we must ask ourselves
first is whether Aristotle recognizes definitional relations such as the ones I have indicated here.
Supposing that he does, we must ask whether he conceives of the definition of life along these
lines.
7.2.3.2 Aristotle's apparatus of account homonymy.
The question which concerns me is whether Aristotle, in his theoretical discussions of
the form of definitions ever distinguishes a type of definitional relation or relations which
follows the models above and fits the cases I have described from the Politics. I argue that he
does in Topics i.15, a lengthy theoretical discussion of homonymy. Our starting point is
Aristotle's observation in the Topics that homonymies often trail into accounts unnoticed49 (en
autois tois logois lanthanei parakolouthoun to homônumon, Top. i.15 107b7; see also Phys. vii.4 248b17-
20). Even this statement, however is unclear. What does Aristotle mean when he says that
homonymy may trail into an account?
49 The phrase is from Shields (1999, 19).
313
On my interpretation, when homonymy trails into an account we will have a higher
order univocal account which specifies a genuine natural kind � the definition of this general
feature picks out real uniformity in the way things are in the world. Nevertheless, when
homonymy trails into the account, there will be discretely distinct ways of satisfying the analysans
of that property. Thus, to use Aristotle's example (Top. i.15 107b9-12), we may propose (as
Aristotle himself does not) to define health as follows:
x is healthy =df x has proper balance.
If the definition were correct, it would count as a univocal specification of the property of being
healthy. Nevertheless, there would be distinct ways of being healthy along the lines of distinct
ways of having proper balance. Thus, I would be healthy if my functional parts were working
together in proper balance, but my salad would be healthy if its constituents were balanced such
that they contained the right proportions of things to maintain health in a thing like me.50 If the
above general definition of health here were adequate, then health would be an account
homonym along the lines indicated.
Aristotle explains this point further in a difficult passage51 in Physics vii.4 devoted to the
problem of incommensurable univocals. Aristotle has already given a general definition of
change (kinêsis):52
50 While I have no general account of when we ought to give an account homonymous definition of a particular term, this account fails in Aristotelian terms because its specification here reveals a more perspicuous definition as a focal homonym, and it is this definition that Aristotle himself seems to accept (see Top. i.15 106a1-8, 106b29-107a2; Met. iv.2 1003a35-7, xi.3 1061a1-7). What kind of account is to be given in particular cases is a matter to be settled on a case-by-case basis. 51 Ross begins his commentary on the passage with this rather ominous warning. "This is a particularly difficult chapter. The text is somewhat corrupt; the expression is even terser than usual. The discussion is highly aporematic; suggestions and objections follow each other with great rapidity, and the turns of thought are unusually difficult to follow." (1936, 677) While I agree with Ross about the difficulty of many of the particulars of the text, there is a clear general flow to the argument of the chapter. My interpretation, above, seeks to capture the theoretical commitments without sweating Aristotle's often obscure details. 52 I translate kinêsis as 'change' rather than 'motion' when context makes apparent that it is kinêsis as a general notion that is in question. When kinêsis is used more narrowly to mean what we would naturally call 'motion' rather than qualitative change I will translate it as such.
314
We have distinguished in respect of each class between what is in fulfillment and what is potentially; thus the fulfillment of what is potentially, as such, is change [kinêsis] � e.g. the fulfillment of what is alterable, as alterable, is alteration. (Phys. iii.1 201a10-3; slightly revised from ROT)53
In Physics vii.4 Aristotle asks whether every change is commensurable with every other change or
not, and finds himself in the embarrassing situation of holding a set of beliefs which appear to be
inconsistent.54
First, there is a sense in which are changes are � and Aristotle holds them to be � the
same. Aristotle has a general definition of change applicable to all types of change (quoted
above), and besides,
It would be absurd (atopon) to suppose that the motion (kinêsis) of one thing in a circle and of another in a straight line cannot be the same ((h)omoiôs),55 but that the one must inevitably move more quickly or more slowly than the other, just as if the course of one were downhill and of the other uphill. (Phys. vii.4 248a19-22; translation revised from ROT)
Second, Aristotle believes that some changes are incommensurate: we cannot call rates of
qualitative and quantitative change equal, for instance (248a13-5). Finally, Aristotle's account of
incommensurability suggests that things are incommensurate qua Fs in virtue of being
homonymous qua Fs ((h)oti (h)omônuma, ou sumblêta, 248b8-9; see also Cat. 8 11a5-13, Top. i.15
107b17, GC ii.6 333a22). Thus, if changes are incommensurate qua changes, then change is
homonymous; but Aristotle holds there to be a single definition that applies to all changes in the
same way � the general definition is adequate.
53 See also Physics 201a28, 201b6, 202a7, 251a9. 54 Theses peculiar to Aristotle generate a number of the cases in what follows and I anticipate � from my own reaction � that the reader may feel no sympathy for at least some of the cases that generate Aristotle's potential embarrassment. Nevertheless, our concern is not with evaluating the theses that push Aristotle to what are, to us, strange conclusions. Our concern is with Aristotle's way of extracting himself from these troubles; as we shall see, his solution is far more philosophically plausible than some of the cases that generate his problem. 55 The ROT translates (h)omoiôs here as 'similar' rather than as 'same', as I have done. The example of two things moving up- and down-hill at the end of the quoted section shows that it is more than mere similarity Aristotle has in mind, however; the two velocities can be equal, literally the same. Ross (1936) agrees in his analysis: "It would be absurd if the motion of one thing in a straight line and that of another in a circle could not be of the same speed" (425).
315
In sum, Aristotle appears to hold an inconsistent triad of theses.
1. Change has an account which applies univocally to all types of change.
2. Some changes are incommensurable with other changes.
3. Things are incommensurable qua Fs in virtue of being homonymously F.
It would appear that something has to go: either Aristotle's commitment to there being a single
account of change or to the incommensurability of the cases at hand, or his account of
incommensurability.
But Aristotle will not take the easy way out. Instead, he preserves all of these theses, and
explains how he can do this consistently in Physics vii.4. The solution involves a deeper
understanding of both Aristotle's doctrines concerning incommensurability and, more to the
point of this chapter, his understanding of homonymy and synonymy and their relations.
Aristotle's solution appeals to the fact, alluded to previously, that homonymy may
appear in a definition. To employ one of Aristotle's examples, suppose that A's being double B
consists in A's standing in relation to B in the proportion of two-to-one (Phys. vii,4 248b14-5).
Then, Aristotle thinks, the account of double can apply to things in the same way even though
particular doubles will be incommensurable � though my kiwis are double the apples (in price)
and my phone is double my modem (in weight), my phone and my kiwis are not commensurable
qua doubles in the way that my blue shirt and blue pants are commensurate qua blue. Aristotle
explains that this incommensurability is the result not of their being homonymous qua double,
but of an homonymy trailing into double's account. The account of double employs the notion
of being one, but what it is to be one in the relevant sense is to be measured by a particular unit,
and one will therefore signify different things as we pick different units (pounds, dollars, inches,
etc.). To be one is to be measured once by a particular unit (in the example, a unit of price and a
unit of weight are used),56 and if so, different doubles will be incommensurable because of the
56 I develop this account from that of Ross (1936, 679, note to 248a17-20).
316
primitive differences in ways of satisfying the analysans. Although the account of being double
is the same in each case of being double (being in the proportion of two-to-one), what it is to be
double is different in the two cases since the units of measurement that underlie the proportion
are different (i.e., what units are we considering when we way that a particular group of things
stands in the proportion two-to-one to one another?).57
Aristotle concludes the passage by noting that non-homonymy (mê (h)omônuma) is not a
sufficient condition for commensurability; some other necessary condition must also be fulfilled
(249a3-5; see also the instance of 'mê (h)omônuma' at 248b13). Given that Aristotle holds � and
he clearly indicates that he does � that being double applies univocally to all cases of doubles
(248b14),58 why does Aristotle refrain from calling this a case of synonymy and choose instead to
call it a case of non-homonymy?
I claim that it is because Aristotle here distinguishes between synonymy and account
homonymy. Cases like double present a form of definition Aristotle recognizes which falls
between the cracks of his standard distinctions between homonymy and synonymy. Double
applies univocally to all doubles � the general account of the proportion applies to them all.
Moreover, the general account of double reveals � albeit in an abstract fashion � something
essential about what it is to be double. But double is nonetheless non-synonymous in Aristotle's
technical sense, for the account of the proportion in different cases of being double is itself
different; homonymy trails into the account when we go to specify the units of measurement
relevant to As' standing in relation to Bs in the proportion of two-to-one. The account of a
57 Apostle notes in his commentary that the incommensurability of Aristotle's examples � water and air � as regards to being double results from there being no common unit which measures water and air (1969, 306n17). The idea of comparing water and air qua double arises, I suppose, as follows. Collect all blue things together in class A, and divide class A into sub-classes B and C. Now it should be the case that the members of B and C are commensurate qua blue. However, group all doubles together in class A' and divide A' into sub-classes B' and C'. In this case commensurability between sub-classes fails whenever there fails to be a common unit between the sub-classes. 58 Velocity applies univocally to velocities despite the fact that the same problem arises in that case; see Phys. vii.4 248a20.
317
particular case of doubles will inevitably refer to a particular way of standing in the relationship
of two-to-one (i.e., qua number, or qua price, or...) and these ways constitute primitive ways of
being double another thing.
It has been relatively easy to describe how account homonymy relates to Aristotle's
conception of synonymy. Things are more complicated when we turn to understand this
definitional relation in the context of Aristotle's doctrines on homonymy. Shields (1999)
distinguishes between three forms of homonymy in Aristotle: two forms of discrete homonymy
and one form of associated homonymy � focal homonymy. The most common examples of
discrete homonyms in Aristotle are what we may call 'mere' homonyms such as 'bank'. Calling
them 'mere' homonyms marks the fact (supposing it is a fact) that it is a mere accident of
language that the word 'bank' is used to indicate both river banks and savings banks.59 Other
important cases of discrete homonyms, however, are called by the same name not accidentally,
but to mark important but non-definitional continuity within a thing. The homonymy of axe and
eye in de Anima ii.1 serve as examples here. A dead eye is discretely homonymously an eye, for it
no longer has its function and eyes are defined by their functions. But it is no mere accident of
language that we call a dead eye an eye, for there is a salient material continuity between the
functioning eye and the dead eye.60 The third and final class of homonyms on Shields' account,
associated homonyms, are such that different ways of being F are definitionally distinct but
related in some important way. On Shields' analysis, the only 'important' way to be associated is
as focal homonyms on his account � through asymmetrical relations of (four-) causal
dependence.
59 See Aristotle's example: "if they have not the same Form, they will have only the name in common, and it is as if one were to call both Callias and a piece of wood 'man', without observing any community between them" (Met. xiii.4 1079b2-3). 60 See Meteor. iv.12 390a10-15, DA i.1 412b17-24, Pol. i.2 1253a20-5, PA i.1 640b30-641a6, GA ii.1 734b25-7, Int. 11 21a21-2.
318
I believe, however, that in Topics i.15 107b7, Physics vii.4 and elsewhere,61 Aristotle
recognizes a fourth kind of homonym, the kind I have been calling account homonyms.
Account homonyms are homonyms in virtue of the fact that they satisfy the definition of
homonymy � the accounts of the particular instances of account homonyms are different.
Account homonyms do not form a class of discrete homonyms, however, for there is an
important definitional connection between the discrete ways in which distinct types of F are F.
The general account does specify, albeit in an abstract way, what it is to be an F for F things. In
cases of account homonymy we use the same name, 'F', to apply to different kinds of F things to
mark the fact that there are essential definitional ties between particular types of Fs despite the
fact that neither synonymy nor focal connection obtains. Instead, in such cases there will be a
common definition which will apply univocally (but not synonymously in Aristotle's technical
sense) to all particular types of Fs; there will be primitive distinct ways of satisfying this common
account; and those primitive distinct ways will be reflected in the fact that the peculiar types of F
are not all of one form (Top. vi.10 148a29-30).
Building on the accounts of different types of homonymy discussed in chapter two, I
propose to distinguish between two forms of associated homonymy as follows.
x and y are associated homonyms F iff (i) x is F and y is F; (ii) the accounts of being F in 'x is F' and in 'y is F' are different; and (iii) the accounts of F in 'x is F' and in 'y is F' have something important62 in common.
61 Thus far I have mentioned cases in the Politics, de Anima and Metaphysics; see also Eudemian Ethics iii.5 1232a20-7. 62 We can think of the two ways of being associated homonyms set out below as associated homonyms of the account sort: there exist distinct primitive ways of satisfying the condition of containing something important in common instanced by focal and account homonymy.
I take it that this is a genuine advantage of my account, for Shields has trouble answering the question of whether account (logos in the relevant sense) is homonymous or synonymous on his interpretation. There exist distinct types of definitional accounts, synonymous and focally homonymous, in his view. But definitional account cannot be synonymous, since the account of being a definitional account will be different for the different types of accounts. Nor (apparently) can it be focally related, for there exists neither a plausible core account of definitional account nor a plausible relation of four-causal definitional dependence to obtain between alleged non-core on core. It would be an advantage if our account was both textually supported and contained room to answer questions such as how the types of definitional account are related � my interpretation claims to be both.
319
x and y are F as focal homonyms iff (i) x is F and y is F; (ii) the accounts of being F in 'x is F' and in 'y is F' are different; and (iii) the accounts of F in 'x is F' and in 'y is F' display asymmetrical four-causal dependency relations between core and non-core ways of being F.
x and y are F as account homonyms iff (i) x is F and y is F; (ii) the accounts of being F in 'x is F' and in 'y is F' are different; nevertheless; (iii) there exists a general account, A, of being F which applies univocally but non-synonymously63 to x and y; and (iv) A specifies an essential feature of Fs qua Fs.
This definition requires a certain amount of explication. First, note that I leave what it is to
apply "univocally but non-synonymously" without an independent definition. To apply
univocally in the intended sense is to fail to apply synonymously, discretely homonymously, or
focally homonymously. Examples of things with accounts which apply in such a way are
available in modern terms with determinables and in Aristotelian terms with his examples of
equality in velocity, double, much and one in Physics ii.4 and sailor and excellence from the
Politics. Further, I have given above a theoretical exposition of the nature of the relations
intended.
Nevertheless, my leaving this notion primitive may appear to cause problems for my
account as follows. Shields introduced focal homonymy by noting that a dialectical opponent
could attempt to unify obvious cases of discrete homonyms by specifying a spurious form of
connection between alleged core and noncore cases: 'bank' in 'savings bank' may be defined as
'financial institution within five hundred miles of a river bank'. Shields sought (and found) a way
of ruling out such spurious connections in his account of legitimate (i.e., four-causal) definitional
connections. If I have no positive account of the type of univocity intended, however, will my
proposal not be open to the same sorts of spurious attacks?
Well, yes, it will. But this is not a problem for the following reason. As an ad hominem
point, even Shields has not avoided this problem altogether with his four-causal account: the
objections I make in chapter two to Shields' claim that the formal cause constitutes a legitimate
320
form of focal definitional connection show that even his current account relies at some point
upon the judgment of those evaluating the account. To take things one step further, what
Shields would need is a principled way of distinguishing real from merely apparent causal
connections. Suppose such an account is provided. Then the problem will arise again in
determining whether that account is satisfied in particular cases, and so on.
As a perfectly general point, no amount of theoretical work can prevent this and it is no
failing of Shields' account nor my own that it succumbs to this problem.64 In short, all accounts
must at some point rely upon the judgement of intelligent, honest, and critically informed
inquirers.65 I have no guiding theory that will indicate in a perfectly general way when univocity
in the sense intended is achieved, but at some point any theory inevitably exhausts its resources.
I rely at this point on the judgement of honest inquirers critically informed by the theoretical
discussion and examples of account homonymy provided above. Such inquirers are satisfactorily
placed by the account already given to determine whether the univocity condition as explained
and illustrated above is satisfied on a case-by-case basis.
Second, similar remarks apply with reference to what it means for a particular account to
satisfy the general account 'in a primitive way'.66 Once again I lack a general theory to explain
63 The condition 'but non-synonymously' is redundant here given condition (ii) in the definition. I include it, however, to emphasize again the fact that synonymy in this context is a technical term which is not equivalent to univocity. 64 The point in the text applies in a perfectly general way to any possible account of any concept or property and does not constitute any particular problem for this or any other account. There exist general sceptical worries about the possibility of giving informative and correct accounts, but this dissertation does not address those sceptical worries, and they constitute no special problem for the account offered here. 65 As Aristotle warns us: "[I]n every inquiry there is a difference between philosophic and unphilosophic argument; therefore we should not think even in political philosophy that the sort of consideration which not only makes the nature of the thing evident but also its cause is superfluous; for such consideration is in every inquiry the truly philosophic method. But this needs much caution" (EE i.6 1216b35-1217a1). Aristotle recognizes that inquiry depends on the intelligence and good will of honest inquirers when he distinguishes between those with good judgement or acumen (Top. viii.14 163b13; APo. i.34 89b10 ff.; NE iii.4 1113a32, vi.9 1142b2-5, vi.11 1143a20 ff., vi.12 1144a23-9) and those inquirers who 'want culture' (EE i.6 1217a1-10; Met. iv.4 1006a5). 66 Other primitive types of satisfaction relation include the way in which a species satisfies the requirements for belonging to its genus, the way a determinate satisfies the requirements for falling under
321
what it is in perfectly general terms for an account to satisfy the conditions laid out in a higher
level account. Lacking theory, however, I will point again to the contrast between the way a
genus applies to a species (synonymously with independently specifiable differentia) and the way
a determinable applies to the determinates that fall under it (primitively). The second order
account schema for figures, above, also illustrates a way in which satisfying accounts may be
basic: the particular kinds of figures are ordered in a serial fashion following the number series.
Disputes may arise concerning particular cases and these will be resolved using standard
philosophical methods of critical argumentation.
Finally, a word ought to be said in defense of the general account. It is true that
Aristotle tells us in the de Anima that it would be 'ludicrous' or 'absurd' (geloion) to cease our
inquiry into figure or soul having obtained only the general account without going forward to
give accounts of the particular natures that fall under that general account. The brunt of this
section has been taken up with explaining what Aristotle means to indicate by this absurdity: the
general account would leave us without even a partial specification of the peculiar nature of the
types of F falling under the general account.
As Aristotle says elsewhere (and frequently), the account would be too 'abstract' or
'general' (logikên, katholou, GA ii.8 747b29, 748a8-15). An account may be 'empty' (kenos) if it is
too far removed from the first principles of the proper inquiry (GA ii.8 748a8-9), and an account
which is too general can obscure this failure. But the fact that general accounts are open to this
failing should not obscure from us the fact that good general accounts have a legitimate role to
play in serious inquiry, since "a grasp of any universal facilitates the understanding of its specific
forms" (GC ii.9 335a26). For this reason, "the investigation of special attributes comes after that
of the common attributes" (Phys iii.1 200b24).67 There is nothing wrong in principle with giving
its determinable, the way a derived case of a focal homonym satisfies the requirements for being a focal homonym, etc. 67 See Phys. i.7 189b30-2 and PA i.4 644a33-b7 (with PA i.1 639a23-4 and i.5 645b11-2) for accounts of how and why investigation should proceed starting from the universal. Aristotle starts a number of his
322
general or abstract accounts; they are in fact essential to competent and complete inquiry. While
it would be absurd to stop after giving only the general account there is no absurdity in wanting
to give a general account in the first place.
7.2.4 Applications.
I have argued above that Aristotle recognizes a sort of homonymy not commonly
recognized, account homonymy, and have given both Aristotelian and modern exegesis of
examples satisfying this notion. It would do no good in the context of this dissertation,
however, if this apparatus were inapplicable to the case we are interested in, the case of life. I
shall argue in this section that the apparatus of account homonymy developed and explained in
the preceding sections finds natural application in Aristotle's theoretical remarks concerning the
nature of life. Life is an account homonym.
7.2.4.1 Figure.
As we began our investigation of Aristotle's parallel between figure and soul in de Anima
ii.3, we shall begin here with figure. At On the Heavens ii.4 Aristotle suggests that the general
definition of a figure is a line-enclosed-area (see 286b15-6).68 Such a definition is satisfactory as
far as it goes, but it fails to specify the peculiar nature of any specific figure. Thus, and only
roughly, a circle is a one-line-enclosed-area, a triangle is a three-line-enclosed-area, etc.69 Figure
has a general account which will apply to all particular kinds of figures despite the fact that the
central theoretical projects in this way. Thus, the Metaphysics vii account of ousia begins with 'logical remarks' before moving forward (vii.4 1029b10-16); the account of soul in de Anima ii begins with a general account in the first chapter, moving in the second to state the cause of this account from the point of view of what is more familiar (ii.2 413a11-20; see also ii.3 415a1); finally, the account of excellence in the Eudemian Ethics starts from a general account before moving in iii.1 to an investigation of the particular cases. 68 In his commentary on 415b23 R.D. Hicks writes: "The corresponding definition would be for figure generally, 'that which is enclosed by one or more lines,' and for rectilinear figure (which A. has in mind: cf. Simpl. 106, 36sq.) 'that which is enclosed by three or more straight lines.'" (1907). 69 Hicks continues: "Such definitions convey no information about any of the particular figures. Nor does the general definition of II.,c.1 give any information about any particular kind of soul. The only vital functions which it recognizes were those common to all animals and all plants (412a14)" (1907, comment on 415b23).
323
specifications of the essences of the various instances of figures will have nothing but (at most) a
general pattern in common. Instead, the distinct kinds of figures there are instantiate primitive
ways of satisfying the general account. This fact is reflected in modern accounts of figures in the
fact that the relationship between figure and figures is taken to be a paradigm case of that
between determinables and determinates (see Searle 1967).
This interpretation satisfies the requirements of the definition of account homonymy.
(i) One may correctly say of triangles, rectangles, etc., that they are figures. (ii) The accounts of a
triangle's being a figure and a square's being a figure are different (a triangle's being a figure
consists in its being a 3-line enclosed area, a rectangle's in its being a 4-line enclosed area). (iii)
There exists a general account of figure which applies univocally but not synonymously (in the
technical sense) to both triangles and rectangles and indicates (in a general way) the what it is to
be a figure: being a line enclosed area. And finally, (iv), the general account specifies an essential
feature of triangles and squares; we mark real distinctions in the world when we distinguish
between figures and nonfigures.
If we were to attempt to undermine this account, we might try to define triangle as
follows:
a triangle isdf a line enclosed area enclosed by three lines.
The problem with this attempt, and others like it, is that we repeat ourselves in specifying an
enclosed area enclosed in a particular way.70 Putting the point in more modern terms, there is no
way of specifying a differentia for types of figures which will not itself already contain the
information relevant to the definition of figure in general.71
70 Aristotle would say that in specifying this definition the definer has started to babble; see Top. v.2 130a29-130b10. 71 John Searle gives the following test for satisfying the determinable-determinate relationship. "[W]hereas we can say, 'All humans are animals which are rational,' no analogous statement can be made beginning, 'All red things are colored things which are . . . .' Any term which could fill the gap would have to be synonymous with 'red.' Red things do not possess some trait other than their redness which, when conjoined with their coloredness, makes them by definition red." (1967, 358)
324
7.2.4.2 Soul.
Similarly for soul. Aristotle has indicated a general account of soul applicable to all
things with souls in the de Anima.
If, then, we have to give a general formula applicable to all kinds of soul, we must describe it as an actuality of the first kind of a natural organized body (entelecheia (h)ê prôtê sômatos phusikou organikou). (DA ii.1 412b4-6)72
Actualities of the first sort may, roughly, be conceived as capacities fully developed but not
currently actively employed (e.g., knowing English but remaining silent) as distinguished from
capacities possessed fully and currently employed (e.g., speaking English).73 Since the actuality
of a living being is the attainment of its end,74 and its end is identified with its peculiar nature as
a living thing,75 there will exist distinct primitive ways of satisfying the definition of a soul
corresponding to the distinctive ends of each type of living nature:76 plant, animal, and human.77
Aristotle's account of soul satisfies the definition of account homonymy. (i) We may say
that plants and animals, etc., have souls. (ii) What it is for plants and animals to have souls is
different: in the one case it is for a natural body with the capacity to nourish and reproduce to be
72 See also DA ii.1 412a19-22, 412a27-9, 412b10-1; see also DA ii.2 414a28; PA i.1 641a25-9; GA ii.4 738b27; Met. vii.10 1035b14-31, vii.11 1037a5, viii.3 1043a35. 73 See DA ii.5 417a22-b1, 417b31-418a2, ii.1 412a10, 412a22-6, 412b27-413a2, iii.4 429b6-9; APr. ii.21 67b4-5. For further distinctions cf. NE vii.3 1147a11, 1147b11; Phys. vii.3 247b14-6, 248a3-5 and APo. i.12 77b24-5; Phys. viii.4 255a34-b4. 74 See EE ii.1 1219a24-5; NE ix.9 1170a15-9, etc, quoted above. 75 See GA ii.3 736b3-4. See also Phys. ii.7 198b5-9; IA 2 704b15-6, 8 708a10; Met. v.16 1021b20; Phys. vii.3 246a10-15, quoted above. 76 We need to recall, of course, that the truth of this statement does not imply the falsity of the thesis of the material continuity of life and nonlife, which Aristotle also endorses (see HA viii.1 588b4-16; PA iv.5 681a12-5). Living beings are instanced in a continuum such that at their borders it is difficult to distinguish whether we have a case of a plant or an animal (for instance), or a living or non-living thing, but this is compatible with each type of living thing possessing a specifiable essence. See the argument of chapter one for more detail on this point. 77 Thus, the spontaneously generated � DA ii.1 413a31-b1, iii.12 434a23; PN 454a13; plants � HA viii.1 588b24-5; GA i.23 731a25-8; animals � PN 454b24, 469b4; PA ii.5 651b4, ii.8 653b23, iii.4 666a35; GA i.23 731b3, NE ix.9 1170a15; and humans � NE ix.9 1170a15-8, 1170b1; Pol. vii.15 1334b15. Not having a natural organic body to actualize god falls into another category: see Met. xii.7 1073a4-13 and NE x.8 1178b20-3.
325
actualized, in the other case it will be for a natural body with the capacities to nourish, reproduce,
and sense to be actualized. (iii) There exists a general definition of the soul which applies
univocally but not synonymously (in the technical sense) to all the types of souls (see DA ii.1
412b4-6). And finally, (iv) this general account picks out a feature essential to souls.
Nevertheless, no part of the specification of the general definition of soul survives unmodified in
the peculiar definitions of each type of soul.78 To exercise the sensory capacity just is to have
soul in the same way that to be red just is to be colored.
7.2.4.3 Life.
Finally, we turn to the case of life. This account will follow the account of soul very
closely. I argued above that Aristotle's commitment to the functional determination thesis79 is
explicitly invoked in the case of the definition of living things.80 Living things are defined by the
activities they can engage in when they have reached a mature state and may flourish as members
of their kind. This end is identified with nature as a living thing (See GA ii.3 736b3-4. See also
Phys. ii.7 198b5-9, IA 2 704b15-6; IA 8 708a10; Met. v.16 1021b20 and Phys. vii.3 246a10-15,
quoted above). Thus given that a thing's nature is its actuality81 as I argued above, there ought to
(and do) exist distinct primitive ways of satisfying the definition of living corresponding to the
distinctive ends of each type of living nature: plant, animal, and human.82
Given its crucial place in the argument of this dissertation, it is important to emphasize
once again that this is not an obscure result of Aristotle's theoretical metaphysics that he may
78 The specification of the genus, for instance, appears unmodified in the account of the species. See Cat. 1 1a6-12, 5 3b2-8; Top. ii.2 109b4-6, iv.1 121a11-5, iv.3 123a29. 79 See Meteor. iv.12 390a10-11; EE iii.7 1115b23; OH ii.3. 286a8, Pol. i.2 1253a23, DA ii.4 416b23, OH ii.14 297b21-2, quoted above. 80 See NE i.7 1097b26-7, b33-a4. See also GA ii.3 736b3-4. See also Phys. ii.7 198b5-9; IA 2 704b15-6; 8 708a10; Met. v.16 1021b20; Phys. vii.3 246a10-15, quoted above. I employed further arguments in favor of the definitional thesis above. 81 Since nature in the relevant sense is form, and form is actuality. See Phys. ii.1 193b7; DA i.1 412a9, 412a21, ii.2 414a15; Met. xii.5 1071a8-9, ix.8 1050b2. 82 See the references above in n. 77.
326
plausibly be thought to have missed. In fact, the train of thought explained above finds itself
instanced over and over again in Aristotle's thoughts on living things and it is this that allows us
to be extremely specific and textually detailed as we specify the ends of various kinds of lives.
The lives of spontaneously generated things are defined by their actualization of the nutritive
soul only (DA ii.1 413a31-b1, iii.12 434a23, PN 454a13). The life of plants which possess both
the nutritive and reproductive souls is defined by their possession of the activities of that soul
(HA viii.1 588b24-5, GA i.23 731a25-8). Animals possess the nutritive and sensitive souls (and
usually the reproductive) and their life is defined by the activities of their sensitive soul (PN
454b24, 469b4; PA ii.5 651b4, ii.8 653b23, iii.4 666a35; GA i.23 731b3 and NE ix.9 1170a15).
Humans possess all these souls plus the rational (NE ix.9 1170a15-8, 1170b1 and Pol.
vii.15 1334b15). As is well known, Aristotle's thesis that the rational soul is in some sense a
separable substance83 leads Aristotle into a number of difficulties, not least with the
identification of the ends of human life.84 It is beyond the scope of my project to adjudicate this
issue here. Suffice it to say, first, that the ends of humans are determined by the ends of their
rational capacities, and that human lives are therefore distinct in kind from animal and plant
lives. But, second, rational capacities occur in humans as the activity of a functionally specified
natural body (DA ii.1 412b4-6) and this fact places constraints on human ends which will
distinguish human life from god's life. For, just as Aristotle acknowledges different ends for
(alleged) different types of humans (Pol. i.13 1260a11-16, quoted above) based on their differing
capacities and natures, Aristotle ought by parity of reasoning not to identify human ratiocinative
goals with god's contemplative goals. God possesses only contemplative activity, and god's life is
83 See DA i.1 413a4-6, i.4 408b18, ii.2 413b25-8, iii.4 429b4, GA ii.3 736b27; NE iii.10 1117b29-30, x.8 1178a21-4. 84 This is amply illustrated by the contrast, often noted and discussed, between Nichomachean Ethics books i-ix and book x.
327
identified with the unimpeded active contemplation of the best objects (Met. xii.7 1073a4-13 and
NE x.8 1178b20-3).85
These results may be applied in the context of our definition of account homonymy. (i)
Aristotle clearly endorses claims to the effect that spontaneously generated beings, plants,
animals, humans and god are alive. (ii) The accounts of being alive in the case of each particular
kind of living being will be distinct in virtue of the fact that the kinds are identified with
reference to the activities of the mature forms of beings of that kind. Thus, we may specify:
1. The life of spontaneously generated organisms =df the activity of the nutritive soul.
2. The life of plants86 =df the activity of the reproductive soul.
3. The life of animals =df the activity of the sensitive soul.
4. The life of humans =df the activity of the contemplative soul.
5. The life of god =df the activity of uninterrupted contemplation.
Each of these lives and activities are identified with the actualization of distinct ends and
therefore constitute distinct ways of being alive.
But (iii), there is a general account of being alive which applies univocally but
nonsynonymously (in the technical sense) to all kinds of life. I have thus far specified this
general account in terms of life's provisional definition which I will repeat here.
LAPD: a is alive =df a is a teleological system.
Finally, (iv), this general account specifies an essential feature of being alive, the possession of
teleological directedness. Despite the fact that among mortal living beings the souls fall into a
hierarchical ordering such that 'lower level' souls are always possessed potentially within 'higher
85 I recognize that saying even as much as I have in this brief paragraph raises as many questions as it answers. I do not claim to have solved the problem of Aristotle's position concerning the dual natures of humans. For, on the account of the distinctions between human and godly life given here, it would appear that each is alive synonymously but (roughly) to a different degree on the parallel with Aristotle's claim that men and women do not form a different species since their forms are the same and their differences are explained with reference to their matter (see Met. x.9 and GA i.23 730b32, iii.8 748a1; HA iv.11 538a13). Problems remain in this contested area.
328
level' souls, each way of being alive constitutes a distinct primitive way of being alive in the sense
that no part of the general definition of life will survive unmodified in the specific accounts
(above) of the distinct types of life.
7.2.5 Section summary.
I have argued that my interpretation of Aristotle's account of life has the resources to
account both for the unity and the multiplicity Aristotle evidently finds in the orders of living
things. The unity of life is given by the general teleological account of what it is to be alive, an
account which applies univocally but not synonymously (in the technical sense) to each kind of
living thing. Further, given the analysis of account homonymy offered above, this account of the
nature of life explains both in Aristotle's own terminology and with coherent contemporary
parallels the nature of the failure of unity Aristotle identifies in the cases of soul and figure at de
Anima ii.3. Life can have a general definition that specifies what it is to be alive � albeit in
nonspecific terms � but which nevertheless fails to give any independently statable part of the
essence of particular kinds of living things because there exist distinct primitive ways of satisfying
the condition stated in the general definition. I conclude that an interpretation of Aristotle on
the nature of life in terms of teleology gains substantive support both for its content (teleology)
and for its form (account homonymy) from Aristotle's systematically interconnected theses on
living things.
7.2.6 The contemporary adequacy of the account's form.
However, since the project of this dissertation is not simply to give an account of
Aristotle's definition of life, but to assess that account's contemporary relevance, we ought to
end by asking whether Aristotle is correct in thinking that life is an account homonym.87 It
86 Those plants which can reproduce, that is. 87 Support for my claim that Aristotle recognizes life as an account homonym is provided by my exegesis of the failure Aristotle diagnoses in the general definitions of figure and soul in de Anima ii.3. That failure is, I have argued, to be understood in terms of account homonymy as here explained.
329
would be difficult to attempt a general answer to this question, for such an answer would require
substantive inquiry into the adequacy of the philosophical principles concerning kind
individuation Aristotle employs in his substantive metaphysics. Are living things best identified
by the activities characteristic of their mature, functionally specified goal states? Ought we
accept the functional determination thesis as applied to living beings?
I will not propose general answers to such questions here, but will instead note simply
that Aristotle's framework for a solution contains the resources to satisfy both those who seek an
account of life which makes life 'all of one kind' (i.e., those who think that all living things have
life synonymously) and those who side with Aristotle and hold that living reveals significant
multivocity that must be reflected in our account. Deciding between these two approaches does
not affect the content of Aristotle's general analysis, but it will affect both the form of the final
account we accept and the specific handling of individual cases of living things. It is enough for
this thesis to reveal the contours for further debate on this subject; I leave it as an open question
how one ought to evaluate the metaphysical principles Aristotle employs in forwarding his
account homonymous definition.
7.3 A first attempt at a definition.
I have now textually supported the substantive claim that Aristotle defines life
teleologically and have given an interpretation of the form of definition Aristotle believes
structures that substantive account. It remains to offer a formal Aristotelian definition of life.
As a first approximation to a satisfactory account of life, then, we may turn back to what I earlier
called the 'provisional account' of life.
L1: a is alive =df a is a teleological system.
L1 captures the unrefined core content of the account of life offered in this dissertation.
Nevertheless, the account cannot stand as it is since it is open to counterexamples.
330
The counterexamples I have in mind come from the fact that living things are not the
only teleological systems in existence. Artifacts have functions and ends � they satisfy the
conditions for being alive laid down in the definition � but are not alive.
Paul Taylor encounters a similar problem in his teleological account of what it is to be
an organism. Taylor says,
We conceive of the organism as a teleological center of life, striving to preserve itself and realize its good in its own unique way. To say it is a teleological center of life is to say that its internal functioning as well as its external activities are all goal-oriented, having the constant tendency to maintain the organism's existence through time and to enable it successfully to perform those biological operations whereby it reproduces its kind and continually adapts to changing environmental events and conditions.... All organisms, whether conscious or not, are teleological centers of life in the sense that each is a unified, coherently ordered system of goal-oriented activities that has a constant tendency to protect and maintain the organism's existence. (1986, 121-2)
Taylor here defines being a biological organism in terms of being a teleological system.88 While
his account is narrower than the account of life we seek � given that his account applies only to
biological organisms and not to living things generally � the account's obvious similarity89 to
the account of this chapter makes it instructive.
Against this conception of being an organism, Taylor notes that we might think complex
machines such as satellites and assembly-line robots which are constructed "to function in a
quasi-autonomous, self-regulating manner in the process of accomplishing certain purposes"
(1986, 123), and that they therefore count as organisms on the account offered. Such complex
artifacts present particularly plausible counterexamples to a teleological definition of life or
organisms, it should be noted, because their internal structures and behaviors display the same
sorts of material and structural complexity that organisms and living things display. We saw in
88 Jonathan Jacobs develops a similar account. He says, "the basic conception of what it is to be a certain kind of organism is that the thing is a quantity of certain types of matter the integrity of which depends upon its maintaining certain organization and functional properties which it has by virtue of its own internally-regulated activity" (1984, 26). His more formal account is developed in stages on pp.17-24. 89 The similarity is not complete, for Taylor does not offer an analysis of the nature and possibility of teleological directedness of the type he requires. I have attempted such an analysis and defense in chapters five and six.
331
chapter five that attempts to analyze teleology in terms of structure or behavior failed because
they were open to counterexamples in the form of spontaneously formed material objects that
possess structures and/or behaviors similar to genuine teleological systems, but which lack
ends.90 Nevertheless, it does seem true to say in some sense that robots � i.e., a kind of entity
with complex structures and behaviors like those of organisms but which do not, in standard cases,
occur spontaneously � have the function or goal of creating their products. The crucial distinction
appears to be that spontaneously formed complex structures lack teleology, but designed structures
(and paradigm biological organisms) do not. But an account which allows designed satellites and
robots as organisms or living is an obvious failure � these examples constitute counterexamples
to L1.
Taylor's solution to the problem is to distinguish between 'original' and 'derived' ends.
He says,
The goal-oriented operations of machines are not inherent to them as the goal-directed behavior of organisms is inherent to them. To put it another way, the goals of machines are derivative, whereas the goals of a living things are original. (1986, 124)91
Given this distinction between what I will call 'intrinsic' and 'derived' teleology we have a
response to the counterexamples offered by artifacts. For something's end-directedness to be
derived is for that thing's end-directedness not to be causally determined by its physical
instantiation base; instead, its being teleological depends causally at least in part on its relation to
contexts of use � the ends to which it is put.92 Only the possession of intrinsic teleological
directedness is sufficient for being alive (or being an organism); the possession of teleological
90 The paradigm system (refined as the structural/behavioral analyses get more complex) is a marble sitting in the bottom of a bowl � the marble's goal state is not being in the bottom of the bowl despite the fact that the bowl-marble system maintains that state throughout significant environmental perturbations. See chapter five and Mark Bedau's essays (1992a; 1992d) for the arguments. 91 Christopher Shields develops his account of being an intentional system along identical lines after consideration of the same types of counterexamples. See Shields (1999, pp.190-2). 92 This account is indebted to Lynne Rudder Baker's (1995, 195) account of the way in which artifacts have intentional properties.
332
directedness derivative upon the intrinsic ends of living beings � such as robots and guided
missiles possess � is not sufficient for being alive.
This distinction between intrinsic and derived teleological directedness is common in the
contemporary literature on teleology. Considering the purposes of a man running to catch a
train and a knife, respectively, Andrew Woodfield notes:
'[P]urpose' seems to be used in different ways in these two examples. The man's purpose is his intention or goal, but a knife does not have intentions. Knives have a purpose in virtue of the fact that men create and use knives for a purpose. . . . In general, the purposes that artifacts have are borrowed from the purposes of their users and designers, and they 'have' them in a different sense. (1976, 27)
Despite the fact that the distinction is common currency in contemporary accounts,93 however,
we may well wonder whether Aristotle drew the distinction we need in order to be able to
ground a satisfactory contemporary account of life in his work.
I will conclude this section by arguing that Aristotle explicitly draws this distinction, and
further that Aristotle's remarks on the functions and living of organic parts of mortal living
beings confirms his acceptance of the relevant distinction. Such a distinction reveals itself in the
following extended passage.
But the prior (to proteron) is already [said] in many ways. [E.g.] [i] that for the sake of which a thing is and [ii] the thing which is for its sake are different and the latter [ii] is prior in coming to be, but the former [i] prior in being (ousia(i)). Further, that which is for the sake of an end has two divisions, first, [ii'] that whence motion is, and second [ii''] that which is used by that for the sake of which. I mean, first, [ii'] something that generates, and second, [ii''] the organ or tool used for the thing generated. . . . So there are three things: first, [i] the end or that for the sake of which; second, [ii'] the things for the end � the source of motion and generation (for what is productive or generative is, as itself, relative to (pros) what it makes or generates); third, [ii''] the useful or those which the end uses. Thus, first some part which contains the source of motion [ii'] must come to be..., next after this the whole and end [i] [comes to be], and third and last the
93 This is not to say, of course, that there is no controversy over the status of the distinction. Eliminitivists about teleology will claim that there is no such thing as intrinsic teleological directedness, and will therefore deny the distinction, claiming that neither 'derivative' nor 'intrinsic' teleology reflects anything more than our spreading of our concepts onto the structure of the world (where they find no home). Dennett and Searle have for a time now engaged in a debate about the viability of a distinction between intrinsic and derived intentionality. See Searle (1992, pp.78-82) for Searle's positive views, and the articles cited in Dennett (1987, 288) for the debate between them over intrinsic and derived intentionality.
333
organic parts serving this [i.e., the end] for certain uses [ii'']. (GA ii.6 742a21-36, with omission. My translation, following Peck)
This passage, note, claims that both [ii'] the generative principle and [ii''] the organs of living
beings are posterior in being to the being or form which is the end of generation and whose ends
regulate the activity of the organs. Now, the being of an organ is determined by its function, by
what it is for (GA i.2 716a23-5; see Meteor. iv.12 390a10-11, EE iii.7 1115b23, OH ii.3. 286a8, Pol.
i.2 1253a23, DA ii.4 416b23, OH ii.14 297b21-2). Generation of Animals ii.6, then, states that
organs derive their being ( = their ends) from the form; the teleology of organs is derived
teleology. The teleological function which defines the organs and generative principle is posterior
in being, that is, that it derives its end from the prior ends of the mature form.
We know from elsewhere that the mature form has ends intrinsically; as Aristotle says,
"We also are in a sense an end" (Phys. ii.3 194b33-4). The sense in which we are ends is as beings
in whose interest or for whose benefit things may be done (DA ii.4 415b2-3, ii.4 415b20-1; Met.
xii.7 1072b1-2; cf. EE viii.3 1249b15-6). Organs, however, do not have ends in this sense. As
Aristotle says,
[S]ince there is the same relation between soul and body, artisan and tool, and master and slave, between each of these pairs there is no partnership; for they are not two, but the first term in each is one, and the second a part of this one. Nor is the good to be divided between the two, but that of both belongs to the one for the sake of which the pair exists. For the body is the soul's natural tool, while the slave is as it were a part and detachable tool of the master, the tool being a sort of inanimate slave. (EE vii.9 1241b18-24)
The concurrence of the saw and the art that uses it is of another sort; for it is not for some end common to both � it is like instrument and soul � but for the sake of the user. It is true that the tool itself receives attention, and it is just that it should receive it, for its function, that is; for it exists for the sake of its function. (EE vii.10 1242a13-7)
[W]here there are two things of which one exists for the sake of the other, they have nothing in common except that the one receives what the other produces. Such, for example, is the relation in which workmen and tools stand to their work. (Pol. vii.8 1328a29-32; see also Pol. i.10 1258a22-7, quoted above.)
The body and its parts exist for the sake of the soul (DA ii.1 415b17-18, PA i.5 645b15-20, PA
i.1 642a12-4; see GA ii.22 730b21-2), but the soul does not likewise exist for the sake of the parts
of the body (see PA iv.12 694b14, quoted below). The body's parts, defined functionally (GA i.2
334
716a23-5), are defined in terms of their contribution to the intrinsic well-being of the whole
whose ends they serve. But, as we saw above in Generation of Animals ii.6 742a21-36, the whole is
prior in being to the parts and does not derive its ends or goods from any prior ends.
This is reflected in Aristotle's commitment to the seemingly odd claim that:
[N]ature makes the organs (ta organa) for the function (to ergon), and not the function for the organs. (PA iv.12 694b14; see also PA iv.8 684a28-9, PA iv.10 687a11-15, Pol. iii.12 1282b35-7. cf. Top. vi.6 145a26-7)
The meaning of this passage is simply that a being's intrinsic ends (maturity, flourishing,
reproduction) determine the functions of the various organs the organism will have, and its form
determines their ends. Organs and their functions exist for the sake of the wholes of which they
are parts; wholes do not exist for the sake of the organs and their ends.94 This is yet another
reflection of Aristotle's recognition that organisms have and are ends in a sense (as beneficiaries
of actions) that the tools and instruments of organisms do not have them; the ends of the tools
are derivative in their being upon the prior ends of the wholes they serve.
Further, the ends of the organs depend for their existence on the ends of the organism
to which they belong. As is well known, Aristotle is committed to the claim that the 'eye' of a
dead person is only an eye homonymously.
What a thing is is always determined by its function: a thing really is itself when it can perform its function; an eye, for instance, when it can see. When a thing cannot do so it is that thing only in name, like a dead eye or one made of stone, just as a wooden saw is no more a saw than one in a picture. (Meteor. iv.12 390a10-15)95
Without the form, the organ caeases to have a function � its function is derived from the prior
ends of the nature it serves. The instruments of the soul � the body and all its functionally
defined parts � derive their ends from the ends of the living creature which they serve.
94 Pace Richard Dawkins (Dawkins 1982; Dawkins 1989). Dawkins is an eliminitivist about teleology, but is willing to speak of Cummins functions (see chapter five). On his view, then, genes 'use' organisms or (more generally) phenotypic expressions of genetic instructions as 'vehicles' for promoting their own reproduction. On the units of selection problem, see the above texts and the articles gathered in part III of Hull and Ruse (1998). 95 See also DA i.1 412b17-24; Pol. i.2 1253a20-5; PA i.1 640b30-641a6; GA ii.1 734b25-7; Int. 11 21a21-2.
335
I have argued that Aristotle distinguishes between intrinsic and derived ends in exactly
the way needed to refine our earlier definition of life, L1, in such a way that it can avoid
counterexamples from things such as artifacts that possess non-intrinsic or derived ends.
Aristotle clearly recognizes a class of such derived ends in the generative principles and the
organic parts of living things: these parts and principles derive their ends from the ends of the
organism as a whole.
This interpretation receives further confirmation from the fact that Aristotle attributes a
derivative kind of life to the organs of living things.
There is, then, no need of soul in each part: it is in some governing origin of the body, and other parts live because they are naturally attached, and do their tasks because of nature (poiein de to ergon to autôn dia tên phusin). (MA 10 703a36-7b; trans. Nussbaum)
We can say, then, that despite the fact that organs have only derivative teleology, there is a sense
in which it is correct to say that they are alive (in a derivative sense): they are alive because they
are naturally connected to a being that is alive in the primary sense, that possesses intrinsic
teleological ends.
7.4 An Aristotelian account of life.
We are, at last, prepared to offer a final Aristotelian definition of life.
L2: a is alive =df a possesses intrinsic ends.
We may also specify a secondary sense in which organs may be said to be alive (i.e., derivatively):
L2': a is alive =df a is naturally connected to a thing which possesses intrinsic ends.
In this definition, L2 picks out the only intrinsic sense of being alive in Aristotle and in fact.
Organisms (plants, animals, humans) and god are alive in this sense. The parts of organisms
(hearts, lungs, kidneys, etc.) are alive, in a derivative sense, L2': they are naturally connected to
beings with intrinsic ends.
Further, the definition admits of distinct primitive ways of being satisfied, indicated
earlier.
336
1. The life of spontaneously generated organisms =df the activity of the nutritive soul.
2. The life of plants96 =df the activity of the reproductive soul.
3. The life of animals =df the activity of the sensitive soul.
4. The life of humans =df the activity of the contemplative soul.
5. The life of god =df the activity of uninterrupted contemplation.
Activities of each of the sorts indicated are intrinsically teleologically directed, and satisfy L2,
ensuring that paradigm living things live in the core sense.
It is important to recall at this stage that the kind of teleology invoked in this definition
is teleology considered as a sui generis sort of real causal factor in the structure of the world.
Teleology or end-directedness is a basic or primitive part of the fabric of the world in the sense
that it is irreducible to any other constituents or factors in the world that underlie or explain it.
The fact that teleology as employed in this definition constitutes such a primitive factor,
however, does not entail that we cannot give an account of it � it only entails that we cannot
give a reductive account. If an account of the teleology employed here is desired, I offer the
following.
x is a final cause =df x is teleologically directed.
x is teleologically directed =df x is end-directed.
x is end-directed =df x is directed upon an object.
None of these definitions claims to be informative in the sense of offering an analysis of
teleology or end-directedness in terms which are themselves nonteleological. Nevertheless, it is
possible to give non-reductive yet informative accounts of entities and relations by specifying the
relations between closed families of terms.97
96 Those plants which can reproduce, that is. 97 Thus, possibility and necessity are interdefined and cannot be given a definition outside of the circle of modal terms. Alvin Plantinga offers this sort of nonreductive account of teleology in his (1993).
337
The definition of being end-directed offered above deserves some attention. Note that
as stated, and given traditional understandings of intentionality, propositions qualify as end-
directed and hence as final causes on this account; this is unacceptable in the end, but revealing.
The argument is as follows. Propositions possess intentionality � they are about certain states
of affairs. Possessing intentionality is, roughly being directed upon an object. But, being end-
directed is being directed upon an object. Therefore, propositions are end-directed on the
definitions specified.
This result is not accidental. There is a close connection between mentality,
intentionality, and teleology, as can be well seen given traditional philosophical debates
concerning whether all (and only) mental states are intentional and whether all teleology is
mentalistic. While I believe neither that all mental states are intentional (there are, after all,
affective states such as ennui which appear not to be about anything) nor that all teleological
directedness is mentalistic (living beings without minds have intrinsic ends, after all), I do believe
that at bottom that there is an affinity between teleology and intentionality that reveals itself
when one attempts to give a nonreductive account of either. I propose that we understand there
to be two fundamentally distinct ways of being directed upon an object: being directed upon an
object as information content and being directed upon an object as goal. It is the latter sort of
direction upon an object to which I refer in my definition of being end-directed, above.
7.5 Emergence in Aristotle.
I have now completed the tasks of specifying the form and content of Aristotle's
definition of life, and our investigation of Aristotle's account of life is therefore nearly complete.
Aristotle's definition of life as revealed in the account above is teleological through and through.
In chapters three and four I argued that teleology was in Aristotle's view a sui generis causal factor
in the structure of the world. In chapters five and six I argued that this notion had surprising
contemporary relevance given the independent support available for the emergentist ontology
necessary to make sui generis Aristotelian teleology intelligible in a modern context.
338
If Aristotle's account is to have contemporary relevance, then according to the
theoretical results of our investigation into the nature of teleology in a contemporary context
Aristotle should not only accept sui generis final causes but should also be committed to the claim
that such teleology is an emergent property,98 at least insofar as it is found among material
beings.99 On the analysis of the last chapter, then,100 we will need to establish that teleology is a
sui generis property in Aristotle; that this property is thought by Aristotle (at least in the case of
mortal beings) to emerge from the interactions of definite kinds of physical parts and processes;
and that teleology has downward causal efficacy in Aristotle.
7.5.1 The emergence of teleology.
The first claim, that teleology is considered by Aristotle to be a sui generis real causal
factor in the world was argued in chapters three and four. The second claim, that this sui generis
property is an emergent property in Aristotle will be argued in this sub-section. I will first argue
that Aristotle is an emergentist for at least some properties before moving to establish the
narrower thesis that Aristotle accepts teleology as an emergent property of certain complex
systems with material realization bases.
98 Presenting the issue in this manner may make it appear that I am searching to read emergentism (independently arrived at and defended) back into Aristotle's texts no matter what Aristotle himself may have held. While I acknowledge the appearance of interpretive misconduct here, I do not believe the charge can stick. The defense below presents the barest sketch of what I believe to be an enormously strong case to be made for Aristotle's acceptance of the core tenets of emergentism as explicated in the appendix to chapter six. Supporting this claim fully would require at least a full article length interpretive essay in itself. Such a thorough defense is beyond the scope of the project of this dissertation. 99 The ontology of strong emergence provides a way of understanding the existence and possible causal efficacy of sui generis features of the physical world. That such an understanding is possible does not entail that all instances of properties that are emergent in this world emerge from complex material structures. The emergence of mental properties from the interactions of physical kinds would be compatible with the existence of an immaterial god who also possessed mental properties. The god's mental properties would not be emergent properties of physical stuffs in the way that our mental properties are, but would belong to an immaterial being instead. The dependency relations between that being's parts need not be settled for the purposes of this chapter. 100 According to that analysis: A property P of a structure X with components a1...an is strongly ontologically emergent if, and only if (i) P is novel and (ii) P has causal powers which are absent from a1...an and their interactions independently of entities of the same type as X.
339
There is strong reason to believe that Aristotle accepted a class of emergent properties
or forms in his work. Aristotle says,
[T]he elements [stoicheiôn, 409b24] are not the only things; there are many others, or, more exactly, an infinite number of others, formed out of the elements. Let us admit that the soul knows and perceives the elements out of which each of these composites is made up; but by what means will it know or perceive the composite whole, e.g. what god, man, flesh, bone (or any other compound) is? For each is, not merely the elements of which it is composed, but those elements combined in a determinate mode or ratio, as Empedocles himself says of bone, "The kindly Earth in its broad-bosomed moulds/ Won of clear Water two parts out of eight/ And four of Fire; and so white bones were formed." Nothing, therefore, will be gained by the presence of the elements in the soul, unless there be also present there the ratios and the composition. (DA i.5 409b28-8)
Since, however, some things are potentially while others are actually, the constituents can be in a sense and yet not-be. The compound may be actually other than the constituents from which it has resulted; nevertheless each of them may still be potentially what it was before they were combined, and both of them may survive undestroyed. (GC i.10 327b23-7)
The passage from Generation and Corruption is particularly relevant to establishing Aristotle's
acceptance of emergent properties. This passage contains a precursor of C.D. Broad's
contention that we may speak of emergents as latent in their emergent bases if we so choose
(1918-19, 114). What an emergentist must emphasize if she chooses to speak this way, and what
Aristotle does emphasize in this passage, is the novelty of the emergent result. As Aristotle says,
"The compound may be actually other than the constituents from which it has resulted." Each
passage contains a prima facie commitment to the emergence of novel properties as an
Aristotelian thesis.
This commitment is confirmed in Aristotle's theoretical remarks on mixture. Aristotle
believes that mixtures exist, but also believes that fundamental obstacles arise which make it
difficult to understand how this is possible. On Aristotle's account, the problems are as follows:
[A]ccording to some thinkers, it is impossible for one thing to be combined with another. They argue that if the combined constituents continue to exist and are unaltered, they are no more combined now than they were before, but are in the same condition; while if one has been destroyed, the constituents have not been combined � on the contrary, one constituent is and the other is not, whereas combination demands uniformity of condition in them both; and on the same principle even if both the combining constituents have been destroyed as the result of their coalescence, they cannot be combined since they have no being at all. (GC i.10 327a35-327b6)
340
Aristotle explains a dilemma: if the parts which are, by hypothesis, mixed survive in the mixture,
then there is no genuine mixture but they are only juxtaposed. If, however, the parts no longer
exist in the mixture, then the parts have not mixed at all but have given way to an entirely new
thing.
The details of Aristotle's solution to the problem of mixture are beyond the scope of
this paper.101 It suffices for my point to note that Aristotle conceives the above dilemma to be a
false one. A mixture is the coming into being of a new thing without its being the case that what
comes to be is a simple juxtaposition of parts or a destruction of parts and the creation of a
wholly new and independent being. Instead, the resultant mixture is emergent in a way
dependant upon the interactions of the elements which remain (potentially � GC i.10 327b23-5)
in the final product. Unlike vitalists, Aristotle does not posit an end result of mixture which is
independent of the elements which composed it; unlike reductivists he does not believe that
mixed things are nothing but102 recombinations of independently existing microparts.103 The
middle ground between these two positions which Aristotle occupies is emergentist through and
through.
Having established Aristotle's general commitment to emergentism, we must continue to
ask whether teleology is emergent in Aristotle. There is strong evidence for this commitment in
Aristotle's doctrines concerning the necessity for a form to come to be in an appropriate kind of
matter.
101 On the problem of mixture and Aristotle's solution, see Sharvy (1983), Bogen (1996), Code (1996), and Fine (1996; 1999). 102 What one means when one says that one phenomena is nothing but another kind of phenomena varies widely with one's conception of reduction. In weak senses of this phrase Aristotle may accept that a mixed thing is nothing but its component parts, but in stronger senses this will not be true. The result of a genuine mixture constitutes a true addition to the ontological inventory of the world (compared to the inventory taken only with reference to the unmixed parts that compose the mixture). See the appendix to chapter six for more details on conceptions of reduction. 103 Aristotle holds the reasonable view, of course, that not all properties of mixtures and other emergent entities are emergent � some are, quite reasonably, held to be fully reducible. Thus, Aristotle maintains that the magnitude of a complex material object is reducible to the magnitudes of its parts (OH i.5 271b18-271b25; see also Phys. viii.7 260b8-11; GC ii.2 329b17 ff. and PA ii.1 646a18-20).
341
[E]vidently one chance thing will not receive another. It comes about as reason requires: the actuality of any given thing can only be realized in what is already potentially that thing, i.e. in a matter of its own appropriate to it. (DA ii.2 414a25-8)
It is as absurd as to say that the art of carpentry could embody itself in flutes; each art must use its tools, each soul its body. (DA i.3 407b24-5)
[T]he product cannot come to be without things which have a necessary nature, but it is not due to these (except as its material); it comes to be for an end. For instance, why is a saw such as it is? To effect so-and-so and for the sake of so-and-so. This end, however, cannot be realized unless the saw is made of iron. (Phys. ii.9 200a8-11; see also GA ii.1 734b20-2, Met. viii.4 1044a27)
These passages show that Aristotle accepts a general thesis concerning the necessity for a thing
embodied in a complex material structure to come into being from material with an appropriate
nature.
The following passage shows that this general thesis is accepted in the case of living
beings.
[I]t is not anything whatever that is made into flesh or bone by the heat, but only something naturally fitted for the purpose; nor is it made in any place or time whatever, but only in a place and time naturally so fitted. For neither will that which exists potentially be made except by that moving agent which possesses the actuality, nor will that which possesses the actuality make anything out of anything whatever; the carpenter would not make a box except out of wood, nor will a box be made out of the wood without the carpenter. (GA ii.6 743a21-6; see also PA ii.1 646a28-36)
[A] given seed does not give rise to any chance living being, nor spring from any chance one; but each springs from a definite parent. (PA i.1 641b27-8; see Phys. ii.4 196a31)104
Again, there is wide support for the thesis that in Aristotle certain properties are dependent in a
strong and sensible way on facts concerning the nature of the materials going to make up its
physical realization base.
Life, indeed appears to be dependant upon the existence of particular structurings of
matter in exactly the way indicated for emergence.
Suppose that a tool, e.g. an axe, were a natural body, then being an axe would have been its essence, and so its soul; if this disappeared from it, it would have ceased to be an axe, except in name. As it is, it is an axe; for it is not of a body of that sort that what it is to
104 See also PA i.1 640a25-6, 640b1-4, ii.1 646a35; GA iv.3 767b35; Phys. ii.1 193b8, ii.2 194b13, ii.7 198a25; Met. vii.7 1032a22, xii.3 1070a29, xii.5 1071a20, 1071a25, xiii.10 1087a21
342
be, i.e. its account, is a soul, but of a natural body of a particular kind, viz. one having in itself the power of setting itself in movement and arresting itself. (DA ii.1 412b12-17)
Since then the complex here is the living thing, the body cannot be the actuality of the soul; it is the soul which is the actuality of a certain kind of body. Hence the rightness of the view that the soul cannot be without a body, while it cannot be a body; it is not a body but something relative to a body. That is why it is in a body, and a body of a definite kind. It was a mistake, therefore, to do as former thinkers did, merely to fit it into a body without adding a definite specification of the kind or character of that body, although evidently one chance thing will not receive another. It comes about as reason requires: the actuality of any given thing can only be realized in what is already potentially that thing, i.e. in a matter of its own appropriate to it. From all this it is plain that soul is an actuality or account of something that possesses a potentiality of being such. (DA ii.2 414a16-28; see also GA ii.6 743a21-6 and PA ii.1 646a15-21, 28-36)
Being alive depends upon possessing material composition of a very definite kind.
Further, small changes in the material constitution of things account for large changes in
the end result of processes of growth and development according to Aristotle. Aristotle makes a
general point about the nature of the dependence of living things for their natures on fine details
of material composition in the following passage.
[E]ven among men themselves if we compare children with adults, or such adults as are of dwarf-like shape with such as are not, we find that, whatever other superiority the former may possess, they are at any rate deficient as compared with the latter in intelligence. The explanation, as already stated, is that in many their psychical principle is corporeal and impeded in its motions. Let now a further decrease occur in the elevating heat, and a further increase in the earthly matter, and the animals become smaller in bulk, and their feet more numerous, until at a later stage they become footless and extended full length on the ground. Then, by further small successions of change, they come to have their principal organ below; and at last the part which answers to a head becomes motionless and destitute of sensation. Thus the animal becomes a plant, that has its upper parts downwards and its lower parts above. For in plants the roots are the equivalents of mouth and head, while the seed has an opposite significance, for it is produced above at the extremities of the twigs. (PA iv.10 486b23-487a2)
In this passage Aristotle notes that in both the general and the specific cases higher functions in
animals and plants depend for their existence upon the existence of suitable material
underpinnings for those functions. Changes in the material nature of things brings about
dramatic effects in the formal nature which depends on that material nature, even to the point of
explaining (through many iterations of such small changes) the difference between when a thing
is a plant and when it is an animal. We can reasonably extrapolate from these cases that life itself
343
emerges through gradual stages from material structures only poorly suited for giving rise to life
activities all the way to highly articulated living things which depend for their existence on
incredibly complex and sophisticated material structures.
The changes in the material nature need be neither great nor obvious to have dramatic
effects.
[A] small change in a first principle is usually attended by changes in many of the things depending on it. (GA i.2 716b4-5)
[A]s one part of first-rate importance changes, the whole system of the animal differs greatly in form along with it. This may be seen in the case of eunuchs, who, though mutilated in one part alone, depart so much from their original appearance and approximate closely to the female form. The reason of this is that some of the parts are principles, and when a principle is moved many of the parts that go along with it must change with it. (GA iv.1 766a23-29; see also MA 7 701b25-33, HA viii.2 589b29-590a11, PN 463a18, GA iv.8 776b15-7, NE ii.1 1103b24-6)105
In each of these passages Aristotle clearly indicates the nature of the dependence of end states in
living things on the fine details of the constitutions of the material parts of the earlier stages in
the development of such living things. As we saw above in PA iv.10 486b23-487a2, such
intricate levels of material dependence find themselves reflected in the material continuity of
living things (on which see HA viii.1 588b4-16 and PA iv.5 681a12-5).
Given the dependence of end states on the material processes involved in living things,
however, we have established the desired result that the property of being teleologically directed
is an emergent property in Aristotle. For, we have already established (in chapters three four)
that teleology is for Aristotle a novel property, and all that remained to be done was to show its
causal dependence upon the material structures which make it up. The existence of things
directed toward particular ends depends in complex ways upon the existence of very particular
material structures as realization bases.
105 This physical principle has a logical or methodological analogue in Aristotle. See OH i.5 271b8-12; Pol. v.3 1303a21-3, v.4 1303b29-32, v.7 1307a40-2; NE i.7 1098b4-9; Phys. i.2 185a10-12 and GC i.2 315b26-8.
344
I have argued in this sub-section for the truth of Aristotle's commitment to teleology
being an emergent novelty. Further, I have situated the argument for this thesis in the context of
a general argument for Aristotle's commitment to a wide range of novel emergent properties; the
commitment to this wider set of theses makes even more plausible the claim that being alive, i.e.,
being teleologically directed, is an emergent property in Aristotle.
7.5.2 The downward causal efficacy of teleology.
In order to establish the claim that teleology in Aristotle is emergent in the sense
necessary to make his account of life relevant on the contemporary scene I need to establish one
further thesis: that Aristotle's account of the emergent novelty of teleology has a place for the
downward causal efficacy of final causes. This is not a particularly controversial thesis, however,
given Aristotle's clear and well known commitment to the priority of formal to material
causation in the processes of living things (see PA i.1 640b22-9), and I will spend comparatively
little time establishing my case � much much more could be said.
Aristotle says,
[I]t is the presence of the soul that enables matter to constitute the animal nature, much more than it is the presence of matter which so enables the soul. (PA i.1 641a30-1; see also DA ii.2 414a16-28)
It is well known that Aristotle holds form to be responsible for the being or identity of the
material it informs (see Met. vii.17; see also Met. vii.10 1035a7 and PA i.1 640b22-9). It may be
doubted, however, that form is causally active on material processes in the way requisite for a
proper understanding of teleological causation.106
Such doubts may be laid to rest once and for all by Aristotle's discussion of the roles of
the male (who provides the form) and the female (who provides the material) in biological
reproduction as described in Generation of Animals books i and ii. As Aristotle says,
106 See the introduction to chapter six for the distinction between two explanatory roles for teleology, one 'static' and the other 'dynamic'. I argued there that no defense of sui generis teleology could be fully
345
[T]he female always provides the material, the male that which fashions it, for this is the power that we say they each possess, and this is what it is for them to be male and female. Thus while it is necessary for the female to provide a body and a material mass, it is not necessary for the male, because it is not within what is produced that the tools or the maker must exist. While the body is from the female, it is the soul that is from the male, for the soul is the substance of a particular body. (GA ii.4 738b19-26; see also GA i.20 729a9-14, i.21 729b6-21, iv.1 765b10-15)
Aristotle is emphatic that it is the formal nature which shapes the material nature in the process
of biological reproduction and growth, and this is precisely the sort of downward causal
influence necessary for the account offered by this dissertation, since the process of generation
has as its end the goal of the mature, flourishing individual form and the continuance of the cycle
of reproduction. There can be no doubt that Aristotle recognizes downward causal influence of
the sort required by the hypothesis that his account of life is in terms of the emergence of
teleological directedness.
7.5.3 Section summary.
I have argued that Aristotle recognized the central notion in his account of life,
teleology, to be a sui generis irreducible causal factor in the structure of the world (chapter four). I
argued further in chapter six that such a conception of teleology, far from being in conflict with
the results of modern science as is nearly invariably supposed, actually finds a natural home in
the ontology of the world that science appears to have disclosed to us. I have argued in this
section that Aristotle's conception of sui generis teleology is not a vitalistic conception on which
teleology 'floats free' of material constituents; it is, rather, a fully natural product of the causal
interactions of physical properties as they occur in complex structures. This conception of
teleology is important for substantiating the claim, made earlier in chapter one, that Aristotle's
account of life holds promise as a 'middle way' between vitalistic and mechanistic accounts of
life. Aristotle's sophisticated emergentism about teleology and life means that his account of life
is not refuted by modern science, as vitalist positions appear to be, and that it nevertheless
adequate unless it explained the possibility of a 'dynamic' role for teleology in the causation of events the physical world.
346
accepts the appearance of the sui generis nature of life's teleological features more adequately than
any sophisticated mechanist account devised thus far. Aristotle's account, then, is both a
scientifically respectable and a philosophically rich and interesting alternative to such
contemporary accounts as exist.
7.6 Conclusion.
In chapter one I argued that contemporary accounts of life took the task of defining life
insufficiently seriously, and noted that even a brief survey of contemporary accounts raised a
host of interesting philosophical problems that deserve greater attention than they have received
to date. I suggested further that a return to Aristotle's account promised to be useful in
contemporary terms given the unique set of methodological and substantive theses he held
concerning living things and the rigorousness with which he investigated the problem. This
dissertation has been devoted to establishing these general theses by drawing out and defending
Aristotle's account of life.
I have argued in this chapter that there is ample evidence that Aristotle defined life in
terms of teleology, and that this supposition can explain a number of puzzles concerning
Aristotle's account of life which revolve around the general problem of the unity of the
definition of life. I have argued, further, that Aristotle has the resources to draw the distinctions
necessary to make a teleological definition of life plausible in his distinction between intrinsic and
derived ends. In previous chapters I have established the nature and contemporary philosophical
and scientific acceptability of the sort of teleology Aristotle's definition of life requires. I
conclude, therefore, that Aristotle's account provides a sophisticated handling of the tough set of
interconnected philosophical and scientific problems surrounding the nature of life surveyed in
chapter one.
This account has a number of virtues. The account respects our modal intuitions about
what it is to be alive; it does not assume either that science tells us everything or that science tells
us nothing about the nature of life, but makes a serious � and surprisingly successful � attempt
347
at respecting both science and modal intuition as sources of evidence relevant to our account.
Further, the account does not place unwarranted epistemological constraints upon our account
of life; our account is guided by metaphysical concerns centering on life's nature rather than
epistemological concerns for drawing lines in the world between particular kinds of things.
Further still, the account resulting from this focus respects our intuitions about what things
count as clear cases of living things (cats, dogs, the grass outside) and unclear or borderline cases
(viruses, etc.).
The account offered here is falsifiable by the progress of science in just the way
hypotheses concerning the nature of chemical kinds are. If we should some day discover that
the nature of water is not H2O but is, instead, XYZ (say), then we will have falsified our current
account of water as H2O. Likewise, if we should someday come up with either (a) a viable and
plausible reductive account of teleology or (b) convincing empirical reasons to suppose that our
world � contrary to its appearance to us now utilizing our best scientific theories � contains no
ontologically emergent properties in the sense required for this definition, then we would have
strong grounds for reconsidering the account as a satisfactory account of life as it is. Even in
such a situation, however, we may retain the account as capturing what would have
(hypothetically) proven to be an outmoded conception of life that we would then have reason to
revise in favor of a better definition or no definition at all.
348
Bibliography
Ackrill, J.L. (trans.). 1963. Aristotle: Categories and De Interpretatione. Oxford: Clarendon Press.
Allen, C., and M. Beckoff. 1995. Biological Function, Adaptation, and Natural Design. Philosophy of Science 62:609-22. Reprinted in Allen, Colin, Marc Bekoff, and George Lauder, eds. 1998. Nature's Purposes: Analyses of Function and Design in Biology: The MIT Press. pp. 571-87. Also reprinted in Buller, David J., ed. 1999. Function, Selection, and Design. Albany: State University of New York Press. pp. 243-56.
Allen, C., and M. Bekoff. 1995. Function, Natural Design, and Animal Behavior. In Perspectives in Ethology 11: Behavioral Design, edited by N. S. Thompson. New York: Plenum.
Allen, Colin, Marc Bekoff, and George Lauder. 1998a. Introduction. In Nature's Purposes: Analyses of Function and Design in Biology, edited by C. Allen, M. Bekoff and G. Lauder. Cambridge, MA: The MIT Press.
Allen, Colin, Marc Bekoff, and George Lauder, eds. 1998b. Nature's Purposes: Analyses of Function and Design in Biology: The MIT Press.
Alston, William P. 1985. Thomas Reid on Epistemic Principles. History of Philosophy Quarterly 2 (4):435-452.
Apostle, Hippocrates G. 1969. Arisotle's Physics. Translated with Commentaries and Glossary. Bloomington, IN: Indiana University Press.
Balme, David. 1987. Teleology and Necessity. In Philosophical Issues in Aristotle's Biology, edited by A. Gotthelf and J. B. Lennox. New York: Cambridge University Press.
Balme, David. 1992. Aristotle's De Partibus Animalium I and De Gerneratione Animalium I (with passages from II.1-3). Trnaslated with Notes by D.M. Balme, with a report on Recent Work and an Additional Bibliography by Allan Gotthelf. Edited by J. L. Ackrill and L. Judson, Clarendon Aristotle Series. Oxford: Clarendon Press.
Balme, David (trans.). 1991. History of Animals Books vii-x: Edited and Translated by D.M. Balme; Prepared for Publication by Allan Gotthelf. Edited by G. P. Goold, The Loeb Classical Library. Cambridge: Harvard University Press.
Barnes, Jonathan, ed. 1984. The Complete Works of Aristotle: The Revised Oxford Translation. 2 vols, Bollingen Serioes. Princeton, NJ: Princeton University Press.
Bealer, George. 1987. The Philosophical Limits of Scientific Essentialism. In Philosophical Perspectives, edited by J. E. Tomberlin.
Bealer, George. 1992. The Incoherence of Empiricism-I. Aristotelian Society, Supp 66:99-138. reprinted in Naturalism: A Critical Appraisal. Steven J. Wagner and Richard Warner eds. 1993. University of Notre Dame Press, Notre Dame, IN: pp. 163-196.
Bealer, George. 1993a. Materialism and the Logical Structure of Intentionality. In Objections to Physicalism, edited by H. Robinson. New York: Oxford University Press.
Bealer, George. 1993b. Universals. The Journal of Philosophy XC (1):5-32.
Bealer, George. 1994. Mental Properties. The Journal of Philosophy:185-208.
349
Bealer, George. 1996a. On the Possibility of Philosophical Knowledge. Philosophical Perspectives 10.
Bealer, George. 1996b. A Priori Knowledge and the Scope of Philosophy. Philosophical Studies 81 (2-3):121-42.
Bealer, George. 1998. Intuition and the Autonomy of Philosophy. In Rethinking Intuition: the Psychology of Intuition and Its Role in Philosophical Inquiry, edited by M. R. DePaul and W. Ramsey. New York: Rowman & Littlefield Publishers, Inc.
Bealer, George. 1999. The A Priori. In The Blackwell Guide to Epistemology, edited by j. Greco and E. Sosa. Oxford: Blackwell Publishers.
Bechtel, William, and Robert C. Richardson. 1998. Vitalism. In Routledge Encyclopedia of Philosophy, edited by E. Craig. London/New York: Rougledge.
Beckner, Morton O. 1967. Vitalism. In The Encyclopedia of Philosophy, edited by P. Edwards. New York: Macmillan Publishing Co., Inc. & The Free Press.
Bedau, M. 1986. Cartesian Interaction. Midwest Studies in Philosophy 10:483-502.
Bedau, M. 1990. Against Mentalism in Teleology. Americal Philosophial Quarterly 27:61-70.
Bedau, M. 1991. Can Biological Teleology be Naturalized? The Journal of Philosophy 85:647-55.
Bedau, M. 1992a. Goal Directed Systems and the Good. The Monist:34-49.
Bedau, M. 1992b. Naturalism and Teleology. In Naturalism: A Critical Appraisal, edited by S. Warner and R. Wagner. Notre Dame, IN: University of Notre Dame Press.
Bedau, M. 1992c. Philosohical Aspects of Artifical Life. In Toward a Practice of Autonomous Systems, Proceedings of the First European Conference on Artificial Life, edited by F. J. V. P. Bourgine. Cambridge, MA: The MIT Press/ Bradford Books.
Bedau, M. 1992d. Where's the Good in Teleology? Philosophy and Phenomenological Research 52:781-805. Reprinted in Allen, Colin, Marc Bekoff, and George Lauder, eds. 1998. Nature's Purposes: Analyses of Function and Design in Biology: The MIT Press. pp. 261-91.
Bedau, M. 1996. The Nature of Life. In The Philosophy of Artificial Life, edited by M. A. Boden. Oxford: Oxford University Press.
Bedau, M. 1997. Weak Emergence. Philosophical Perspectives 11.
Bedau, M. 1998. Philosophical Content and Method of Artificial Life. In The Digital Phoenix, edited by T. W. Bynum. Cmabridge: Blackwell.
Bedau, M., and Norman Packard. 1991. Measurement of Evolutionary Activity, Teleology, and Life. In Artificial Life II, edited by C. Langton, C. Taylor, D. Farmer and S. Rasmussen. Redwood City, CA: Addison Wesley.
Bigelow, J., and R. Pargetter. 1987. Functions. The Journal of Philosophy 84:181-96. Reprinted in Allen, Colin, Marc Bekoff, and George Lauder, eds. 1998. Nature's Purposes: Analyses of Function and Design in Biology: The MIT Press. pp. 241-60.
Boden, Margaret. 1996. Introduction. In The Philosophy of Artificial Life, edited by M. A. Boden. Oxford: Oxford University Press.
Bogen, James. 1996. Fire in the Belly: Aristotelian Elements, Organisms, and Chemical Compounds. In Form, Matter and Mixture in Aristotle, edited by F. A. Lewis and R. Bolton. Oxford: Blackwell Publishers.
350
Bonabeau, Eric, and Guy Theraulaz. 1995. Why Do We Need Artificial Life? In Artificial Life: an overview, edited by C. Langton. Cambridge, MA: The MIT Press/ A Bradford Book.
BonJour, Laurence. 1995. Toward a Moderate Rationalism. Philosophical Topics 23 (1):47-78.
BonJour, Laurence. 1998. In Defence of Pure Reason. New York: Cambridge University Press.
Boorse, C. 1976. Wright on Functions. Philosophical Review 85:70-86.
Bradie, Michael, and Fred D. Jr. Miller. 1984. Teleology and Natural Necessity in Aristotle. History of Philosophy Quarterly 1:133-46.
Brandon, R.N. 1981. Biological Teleology: Questions and Explanations. Studies in the History and Philosophy of Science 12:91-105. Reprinted in Allen, Colin, Marc Bekoff, and George Lauder, eds. 1998. Nature's Purposes: Analyses of Function and Design in Biology: The MIT Press. pp. 79-97.
Broad, C.D. 1918-19. Mechanical Explanation and its Alternatives. Proceedings of the Aristotelian Society 19:86-124.
Broad, C. D. 1925. The Mind and its Place in Nature, International Library of Psychology, Philosophy and Scientific Method. London: Routledge & Kegan Paul.
Buller, David J., ed. 1999a. Function, Selection, and Design. Albany: State University of New York Press.
Buller, David J. 1999b. Introduction: Natural Teleology. In Function, Selection, and Design, edited by D. J. Buller. Albany: State University of New York Press.
Burnyeat, M.F. 1981. Arisotle on Understanding Knowledge. In Aristotle on Science: 'The Posterior Analytics', edited by E. Berti. Padua.
Cameron, Rich. 1998. Review of Irwin's Aristotle's First Principles. unpublished manuscript.
Campbell, D. 1974. Evolutionary Epistemology. In The Philosophy of Karl Popper, edited by P. A. Schilpp. LaSalle, Ill.: Open Court Publishing.
Campbell, Keith. 1970. Body and Mind. Garden City, N.Y.: Doubleday & Company, Inc.
Chalmers, David J. 1996. The Conscious Mind: in Search of a Fundamental Theory. Oxford: Oxford University Press.
Charles, David. 1988. Aristotle on Hypothetical Necessity and Irreducibility. Pacific Philosophical Quarterly 69:1-53.
Charles, David. 1991. Teleological Causation in the Physics. In Aristotle's Physics: a Collection of Essays, edited by L. Judson. Oxford: Oxford University Press.
Charlton, William (trans.). 1992. Aristotle Physics Books I and II. Translated with Introduction, Commentary. Note on Recent Work, and Revised Bibliography by William Charlton. Edited by J. L. Ackrill and L. Judson, Clarendon Aristotle Series. Oxford: Clarendon Press.
Churchland, Paul M. 1988. Matter and Consciousness. Revised Edition ed. Cambridge: A Bradford Book: The MIT Press.
Code, Alan. 1996. Potentiality in Aristotle's Science and Metaphysics. In Form, Matter and Mixture in Aristotle, edited by F. A. Lewis and R. Bolton. Oxford: Blackwell Publishers.
351
Cohen, S.M., and D. Keyt. 1992. Analysing Plato's Arguments: Plato and Platonism. In Oxford Studies in Ancient Philosophy, edited by J. C. S. Klugge, N.O.
Cooper, John. 1982. Aristotle on Natural Teleology. In Language and Logos: Studies in Ancient Greek Philosophy Presented to G.E.L. Owen, edited by M. Schofield and M. Nussbaum. Cambridge: Cambridge University Press.
Cooper, John. 1987. Hypothetical Necessity and Natural Teleology. In Philosohical Issues in Aristotle's Biology, edited by A. Gotthelf and J. B. Lennox. New York: Cambridge University Press.
Cornman, James W., Keith Lehrer, and George S. Pappas. 1992. Philosophical Problems and Arguments: an introduction. Fourth Edition ed. Indianapolis: Hackett Publishing Company.
Crane, Tim. 1994. Physicalism (2): Against Physicalism. In A Companion to the Philosophy of Mind, edited by S. Guttenplan. Oxford: Basil Blackwell.
Crane, Tim. 1995. The Mechanical Mind: A Philosophical Introduction to Minds, Machines and Mental Representation. New York: Penguin Books.
Crane, T., and D.H. Mellor. 1990. There is No Question of Physicalism. Mind 99:185-206.
Crick, F. 1981. Life Itself: Its Origin and Nature. New York: Simon and Schuster.
Cummins, Robert. 1975/1984/1998. Functional Analysis. Journal of Philosophy 72:741-65. Reprinted, with changes, in Allen, Colin, Marc Bekoff, and George Lauder, eds. 1998. Nature's Purposes: Analyses of Function and Design in Biology: The MIT Press.
Curzer, Howard J. 1998. An Argument in Physics II.8. Philosophia: Philosophical Quarterly of Israel 26 (3-4):359-382.
Dawkins, Richard. 1982. The Extended Phenotype. San Fransicisco: Freeman, Cooper and Co.
Dawkins, Richard. 1986/1996. The Blind Watchmaker: Why the Evidence of Evolution Reveals a Universe without Design. New York: W.W. Norton & Company.
Dawkins, Richard. 1989. The Selfish Gene. 2nd ed. Oxford: Oxford University Press.
Dawkins, Richard. 1996. Climbing Mount Improbable. New York: W.W. Norton & Company.
Dennett, Daniel. 1987. Evolution, Error, and Intentionality. In The Intentional Stance. Cambridge: The MIT Press.
Dennett, Daniel. 1989. Setting Off on the Right Foot. In The Intentional Stance. Cambridge: The MIT Press.
Dennett, Daniel. 1995a. Artificial Life as Philosophy. In Artificial Life: An Overview, edited by C. Langton. Cambridge, MA: The MIT Press/ A Bradford Book.
Dennett, Daniel. 1995b. Darwin's Dangerous Idea: Evolution and the Meanings of Life. New York: Simon & Schuster.
Descartes. 1967. The Philosophical Works of Descartes. Edited by E. S. Haldane and G. R. T. Ross. 2 vols. London: Cambridge University press.
Dobzhansky, Theodosius. 1973. Nothing in Biology Makes Sense Except in the Light of Evolution. Americal Biology Teacher 35:125-9.
352
Dobzhansky, Theodosius, Francisco J. Ayala, G. Ledyard Stebbins, and James W. Valentine. 1977. Evolution. San Francisco: W.H. Freeman and Company.
Doepke, Frederick C. 1982. Spatially Coinciding Objects. Ratio XXIV (1):45-60.
Dretske, Fred. 1988. Explaining Behavior: Reasons in a World of Causes. Cambridge, MA: The MIT Press.
Ducasse, C.J. 1961. In Defense of Dualism. In Dimensions of Mind, edited by S. Hook. New York: Collier Books.
Emmeche, C. 1992. Life as an Abstract Phenomenon: Is AL Possible? In Toward a Practice of Autonomous Systems, Proceedings of the First European Conference on Artificial Life, edited by F. J. V. P. Bourgine. Cambridge, MA: The MIT Press/ Bradford Books.
Engels, Friedrich. 1880/1978. Socialism: Utopian and Scientific. In The Marx-Engles Reader, edited by R. C. Tucker. New York: W.W. Norton & Company.
Engerg-Pedersen, T. 1979. More on Aristotelian Epagoge. Phronesis 24:301-19.
Falk, Arthur E. 1981. Purpose, Feedback, and Evolution. Philosophy of Science 48:198-217.
Falk, Arthur E. 1995. Essay on Nature's Semeiosis. Journal of Philosophical Research 20:298-348.
Farmer, D., and A. Belin. 1992. Artificial Life: the Coming Evolution. In Artificial Life II, edited by C. G. Langton, C. Taylor, J. D. Farmer and S. Rasmussen. Redwood City, CA: Addison-Wesley.
Fine, Kit. 1996. The Problem of Mixture. In Form, Matter and Mixture in Aristotle, edited by F. A. Lewis and R. Bolton. Oxford: Blackwell Publishers.
Fine, Kit. 1999. Mixing Matters. In Form and Matter: Themes in Contemporary Metaphysics, edited by D. S. Oderberg. Oxford: Blackwell Publishers.
Foster, John. 1991. The Immaterial Self: a Defense of the Cartesian Dualist Conception of the Mind. New York: Routledge.
Furley, David. 1985. The Rainfall Example in Physics II.8. In Aristotle on Nature and Living Things: Philosophical and Historical Studies Presented to David M. Balme on his Seventieth Birthday, edited by D. M. Balme and A. Gotthelf. Pittsburgh, Pa.: Mathesis Publications Inc.
Furley, David. 1996. What Kind of Cause is Aristotle's Final Cause. In Rationality in Greek Thought, edited by M. Frede and G. Striker. Oxford: Clarendon Press.
Futuyma, Douglas. 1983. Science on Trial: The Case for Evolution. New York: Pantheon.
Godfrey-Smith, Peter. 1993. Functions: Consensus without Unity. Pacific Philosophical Quarterly 74:196-208. Reprinted in Hull, David L., and Michael Ruse, eds. 1998. The Philosophy of Biology. Oxford: Oxford University Press. pp. 280-92. Also reprinted in Buller, David J., ed. 1999. Function, Selection, and Design. Albany: State University of New York Press. pp. 185-97.
Gotthelf, Allan. 1976/ 1987. Aristotle's Conception of Final Causality: with Postscript 1986. In Philosohical Issues in Aristotle's Biology, edited by A. Gotthelf and J. B. Lennox. New York: Cambridge University Press.
Gotthelf, Allan. 1980. review of M.C. Nussbaum, Aristotle's De motu animalium. The Journal of Philosophy 77:365-78.
353
Gotthelf, Allan. 1988. The Place of the Good in Aristotle's Natural Teleology. In Proceedings of the Boston Colloquium in Ancient Philoosphy, edited by J. J. Cleary and D. C. Shartin. Lanham, Md.: University Press of America.
Gotthelf, Allan. 1989. Teleology and Spontaneous Generation: A Discussion. In Apeiron: Special Issue. Nature, Knowledge and Virtue: Essays in Memory of Joan Kung., edited by R. Kraut and T. Penner.
Gotthelf, Allan. 1997. Understanding Aristotle's Teleology. In Final Causality in Nature and Human Affairs, edited by R. F. Hassing. Washington, D.C.: Catholic University of America Press.
Goudge, T.A. 1967. Emergent Evolutionism. In The Encyclopedia of Philosophy, edited by P. Edwards. New York: Macmillan Publishing Co., Inc. & The Free Press.
Gould, Stephen Jay. 1983. Darwin's Untimely Burial--Again! In Scientists Confront Creationism, edited by L. Godfrey. New York, London: W.W. Norton & Company.
Griffiths, Paul E. 1993. Functional Analysis and Proper Functions. British Journal for the Philosophy of Science 44:409-22. Reprinted in Buller, David J., ed. 1999. Function, Selection, and Design. Albany: State University of New York Press. pp. 143-58.
Hamlyn, D.W. 1976. Aristotelian Epagoge. Phronesis xxi:167-84.
Hamlyn, D.W. (trans. with introduction and notes). 1993. Aristotle: De Anima. (With passages from Book I) With a report on recent work and a revised bibliography by Christopher Shields. Oxford: Clarendon Press.
Harnad, Stevan. 1995. Levels of Functional Equivalence in Reverse Bioengineering. In Artificial Life: An Overview, edited by C. Langton. Cambridge, MA: The MIT Press/ A Bradford Book.
Hasker, William. 1982. Emergentism. Religious Studiens 18:473-88.
Hasker, William. 1999. The Emergent Self. Ithaca, NY: Cornell University Press.
Healey, Richard A. 1991. Holism and Nonseparability. Journal of Philosophy 88:393-421.
Hempel, Carl. 1965. The Logic of Functional Analysis. In Aspects of Scientific Explanation. New York: Free Press.
Hempel, Carl. 1966. Philosophy of Natural Science. Upper Saddle River, NJ: Prentice Hall.
Hicks, R.D. 1907. Aristotle's De Anima, with Translation, Introduction, and Notes. Cambridge: Cambridge University Press.
Hobbes, Thomas. 1994. Leviathan. Edited by E. Curley. Indianapolis: Hackett Publishing Company, Inc.
Horgan, Terence E. 1994. Physicalism (1). In A Companion to the Philosophy of Mind, edited by S. Guttenplan. Oxford: Basil Blackwell.
Hull, David. 1965. The Effect of Essentialism on Taxonomy: 2000 Years of Stasis. British Journal for the Philosophy of Science 16:1-18.
Hull, David. 1974. Philosophy of Biological Science. Englewood Cliffs, New Jersey: Prentice-Hall, Inc.
Hull, David L., and Michael Ruse, eds. 1998. The Philosophy of Biology. Oxford: Oxford University Press.
354
Hume, David. 1777. An Enquiry Concerning Human Understanding. In Classics of Western Philosophy, edited by E. Steinberg. Indianapolis: Hackett Publishing Company.
Hume, David. 1975. Enquiries Concerning Human Understanding and Concerning the Principles of Morals. Edited by L. A. Selby-Bigge and P. H. Nidditch. Oxford: Clarendon Press.
Hume, David. 1990. A Treatise of Human Nature. Edited by L. A. S.-B. a. P. H. Nidditch. second edition ed. Oxford: Oxford University Press.
Humphreys, Paul W. 1995. Understanding in the Not-So-Special Sciences. Southern Journal of Philosophy XXXIV, Supp.:99-114.
Humphreys, Paul W. 1996. Aspects of Emergence. Philosophical Topics 24:53-70.
Humphreys, Paul W. 1997a. Emergence, Not Supervenience. Philosophy of Science 64 (4):s337-s345.
Humphreys, Paul W. 1997b. How Properties Emerge. Philosophy of Science 64:1-17.
Irwin, Terence. 1988. Aristotle's First Principles. New York: Oxford University Press.
Jackson, Frank. 1986. What Mary Didn't Know. The Journal of Philosophy LXXXIII (5):291-5.
Jacobs, Jonathan. 1984. Teleology and Essence: An Account of the Nature of Organisms. Nature and System 6:15-32.
Jacobs, Jonathan. 1986. Teleology and Reduction in Biology. Biology and Philosophy 1:389-99.
Joachim, Harold H. 1926. Aristotle on Coming-to-Be and Passing-Away. A Revised Text with Introduction and Commentary. Oxford: The Clarendon Press.
Johnson, Phillip E. 1993. Darwin on Trial. Downers Grove, Il.: Intervarsity Press.
Judson, Lindsay. 1991. Chance and 'Always or For the Most Part' in Aristotle. In Aristotle's Physics: a Collection of Essays, edited by L. Judson. Oxford: Oxford University Press.
Kahn, C. 1985. The Place of the Prime Mover in Aristotle's Teleology. In Aristotle on Nature and Living Things, edited by A. Gotthelf. Pittsburgh: Mathesis Publications, Inc.
Katz, Jerrold J. 1982. Common Sense in Semantics. Notre Dame Journal of Formal Logic 23:174-218.
Katz, Jerrold J. 1990. The Metaphysics of Meaning. Cambridge, Mass: The MIT Press: A Bradford Book.
Katz, Jerrold J. 1992. The New Intensionalism. Mind 101 (404).
Katz, Jerrold J. 1995. What Mathematical Knowledge Could Be. Mind 104 (415):491-532.
Katz, Jerrold J. 1998. Realistic Rationalism. Cambridge, MA: The MIT Press.
Kemeny, J., and P. Oppenheim. 1956. On Reduction. Philosophical Studies 7:6-19.
Kim, Jaegwon. 1992. 'Downward Causation' in Emergentism and Nonreductive Physicalism. In Emergence or Reduction, edited by A. Beckermann, H. Flohr and J. Kim. Berlin: De Gruyter.
Kim, Jaegwon. 1993. The Non-Reductivist's Troubles with Mental Causation. In Mental Causation, edited by J. Heil and A. Mele. Oxford: Clarendon Press.
355
Kim, Jaegwon. 1998. Mind in a Physical World: An Essay on the Mind-Body Problem and Mental Causation. Cambridge, MA: The MIT Press.
Kim, Jaegwon. 1999. Making Sense of Emergence. Philosophical Studies 95:3-36.
Kitcher, Philip. 1982. Abusing Science. Cambridge: The MIT Press.
Kitcher, Philip. 1993. Function and Design. Midwest Studies in Philosophy xvii:379-397. Reprinted in Hull, David L., and Michael Ruse, eds. 1998. The Philosophy of Biology. Oxford: Oxford University Press. pp. 258-79. Also reprinted in Buller, David J., ed. 1999. Function, Selection, and Design. Albany: State University of New York Press. pp. 159-83.
Klee, Robert. 1984. Micro-Determinism and Concepts of Emergence. Philosophy of Science 51:44-63.
Klee, Robert. 1997. Introduction to the Philosophy of Science: Cutting Nature at its Seams. New York: Oxford University Press.
Kornblith, Hilary. 1994. Naturalism: Both Metaphysical and Epistemological. In Midwest Studies in Philosophy, edited by P. A. French, T. E. J. Uehling and H. K. Wettstein. Notre Dame, IN: University of Notre Dame Press.
Kornblith, Hilary. 1998. The Role of Intuition in Philosophical Inquiry: An Account with No Unnatural Ingredients. In Rethinking Intuition: the Psychology of Intuition and Its Role in Philosophical Inquiry, edited by M. R. DePaul and W. Ramsey. New York: Rowman & Littlefield Publishers, Inc.
Kraemer, Eric Russert. 1984. Teleology and the Organism-Body Problem. Metaphilosophy 15:45-54.
Küppers, B.-O. 1985. Molecular Theory of Evolution: Outline of a Physico-Chemical Theory of the Origin of Life. Berlin: Springer.
Lange, Marc. 1996. Life, 'Artificial Life,' and Scientific Explanation. Philosophy of Science 63:225-44.
Langton, Christopher. 1989/ 1992/ 1996. Artificial Life. In Lectures in Complex Systems: Santa Fe Institute Studies in the Sciences of Complexity, Lectures, 4, edited by L. Nadel and D. Stein. Reading, Mass.: Addison-Wesley.
Langton, Christopher G. 1995. Editor's Introducion. In Artificial Life: an overview, edited by C. Langton. Cambridge, MA: The MIT Press/ A Bradford Book.
Lederberg, J. 1960. Exobiology: Approaches to Life Beyond Earth. Nature 132:393-400.
Lesher, J. 1973. The Role of Nous in Aristotle's Posterior Analytics. Phronesis 18:44-68.
Lewis, David. 1983. Extrinsic Properties. Philosophical Studies 44:197-200.
Lewis, David. 1986a. On the Plurality of Worlds. Oxford: Blackwell.
Lewis, David. 1986b. Philosophical Papers. Vol. 2. Oxford: Oxford University Press.
Lewis, F. 1988. Teleological and Material/Efficient Causes in Aristotle. Pacific Philosophical Quarterly 69:54-98.
Lovejoy, Arthur. 1927. The Meaning of 'Emergence' and its Modes. In Proceedings of the Sixth International Congress of Philosophy. London: Longmans, Green, and Co.
Lowe, E.J. 1989. Kinds of Being: A Study of Individuation, Identity and the Logic of Sortal Terms. Oxford: Basil Blackwell.
356
Lowe, Ernest J. 1996. Subjects of Experience. New York: Cambridge University Press.
MacDonald, Scott. 1989. Aristotle and the Homonymy of the Good. Archiv für Geschichte der Philosophie 71:150-74.
Matthen, Mohan. 1989. The Four Causes in Aristotle's Embryology. In Apeiron: special issue. Nature, Knowledge and Virtue: Essays in Memory of Joan Kung., edited by R. Kraut and T. Penner.
Matthen, Mohan. 1991. Naturalism and Teleology. The Journal of Philosophy 85:656-7.
Matthews, Gareth B. 1977. Consciousness and Life. Philosophy 52:13-26. Reprinted in Rosenthal, David M., ed. 1991. The Nature of Mind. New York: Oxford University Press. pp. 63-70.
Matthews, Gareth B. 1992. De Anima 2. 2-4 and the Meaning of Life. In Essays on Aristotle's De Anima, edited by M. C. N. A. O. Rorty. New York: Oxford University Press.
Maynard Smith, J. 1986. The Problems of Biology. New York: Oxford University Press.
Mayr, Ernst. 1982. The Growth of Biological Thought: Diversity, Evolution, and Inheritance. Cambridge, Mass: Harvard University Press.
Mayr, Ernst. 1988. The Multiple Meanings of Teleological. In Toward a New Philosophy of Biology : Observations of an Evolutionist. Cambridge, Mass.: Belknap Press of Harvard University Press.
McGinn, Colin. 1991. The Problem of Consciousness: Essays Towards a Resolution. Oxford: Blackwell.
McLaughlin, Brian P. 1992. The Rise and Fall of British Emergentism. In Emergence or Reduction, edited by A. Beckermann, H. Flohr and J. Kim. Berlin: De Gruyter.
Medawar, P.B., and J.S. Medawar. 1977. The Life Science. New York: Harper and Row.
Miller, Fred. 1995. Aristotle's Politics Reconsidered. In Nature, Justice and Rights in Aristotle's Politics. Oxford: Oxford University Press.
Miller, Kenneth R. 1999. Finding Darwin's God: A Scientist's Search for Common Ground Between God and Evolution. New York: Cliff Street Books: An Imprint of HarperCollins Publishers.
Millikan, Ruth G. 1984. Language, Thought and Other Biological Categories. Cambridge, Massachusetts: MIT Press.
Millikan, Ruth G. 1989. In Defense of Proper Functions. Philosophy of Science 56:288-303. Reprinted in Allen, Colin, Marc Bekoff, and George Lauder, eds. 1998. Nature's Purposes: Analyses of Function and Design in Biology: The MIT Press. pp. 295-312.
Misak, C.J. 1995. Verificationism: Its History and Prospects. New York: Routledge.
Mitchell, M., and S. Forrest. 1995. Genetic Algorithms and Artificial Life. In Artificial Life: an overview, edited by C. Langton. Cambridge, MA: The MIT Press/ A Bradford Book.
Monod, J. 1972. Chance and Necessity. New York: Knopf.
Mooney, P.R. 1990. On Folkseed and Life Patents. Advances in Biotechnology: Proceedings of an International Conference Organized by the Swedish Council for Forestry and Agricultural Research and the Swedish Recombinant DNA Advisory Committee, 11-14 March 1990.
357
Mullen, John D. 1995. Hard Thinking: The Reintroduction of Logic into Everyday Life. Lanham: Rowman & Littlefield Publishers, Inc.
Nagel, Ernest. 1961. The Structure of Science. New York: Harcourt.
Nagel, Ernest. 1979. Teleology Revisited. In Teleology Revisited and Other Essays in the Philosophy and HIstory of Science, edited by E. Nagel. New York: Colubia University Press.
Nagel, Thomas. 1986. The View from Nowhere. New York: Oxford University Press.
Neander, Karen. 1991a. Functions as Selected Effects: The Conceptual Analyst's Defence. Philosophy of Science 58:168-84. Reprinted in Allen, Colin, Marc Bekoff, and George Lauder, eds. 1998. Nature's Purposes: Analyses of Function and Design in Biology: The MIT Press.
Neander, Karen. 1991b. The Teleological Notion of 'Function'. Australasian Journal of Philosophy 69:454-68. Reprinted in Buller, David J., ed. 1999. Function, Selection, and Design. Albany: State University of New York Press. pp. 123-41.
Neander, Karen. 1995. Misrepresenting and Malfunctioning. Philosophical Studies 79:109-41. Reprinted (in part) in Buller, David J., ed. 1999. Function, Selection, and Design. Albany: State University of New York Press. pp. 221-31.
Newman, David V. 1996. Emergence and Strange Attractors. Philosophy of Science 63:245-61.
Nissen, Lowell. 1997. Teleological Language in the Life Sciences. New York: Rowman & Littlefield Publishers, Inc.
Nussbaum, Martha C. 1978. Aristotle's de Motu Animalium: Text with Translation, Commentary and Interpretive Essays. Princeton: Princeton University Press.
Nussbaum, Martha C. 1980. Book Review: Substance, Body, and Soul:Aristotelian Investigations. Edwin Hartman. The Journal of Philosophy 77:355-65.
O'Connor, T. 1994. Emergent Properties. American Philosophical Quarterly 31:91-104.
Owens, Joseph. 1968. Teleology of Nature in Aristotle. Monist 52:159-73.
Penrose, R. 1995. Shadows of the Mind. Oxford: Oxford University Press.
Pettit, Philip. 1993. A Definition of Physicalism. Analysis 53:213-23.
Plantinga, Alvin. 1987. Two Concpets of Modality: Modal Realism and Modal Reductionism. Philosophical Perspectives 1:189-231.
Plantinga, Alvin. 1993. Warrant and Proper Function. New York: Oxford University Press.
Polanyi, M. 1968. Life's irreducible structure. Science 160:1308-1312.
Popper, Karl R., and John C. Eccles. 1977. The Self and its Brain. Berlin, New York: Springer International.
Post, John. 1995. Naturalism. In The Cambridge Dictionary of Philosophy, edited by R. Audi. Cambridge: Cambridge University Press.
Preus, Anthony. 1975. Science and Philosophy in Aristotle's Biological Works, Studien und Materialien zur Geschichte der Philosophie : Kleine Reihe ; Bd. 1. Hildesheim ; New York: G. Olms.
Prior, Elizabeth. 1985. What is Wrong with Etiological Accounts of Biological Function? Pacific Philosophical Quarterly 66:310-28.
358
Randall, John Herman Jr. 1960. Aristotle. New York: Columbia University Press.
Ray, Thomas S. 1979. Slow-motion world of plant 'behavior' visible in rainforest. Smithsonian 9 (12):121-30.
Ray, Thomas S. 1992a. An Approach to the Synthesis of Life. In Artificial Life II, edited by C. G. Langton, C. Taylor, J. D. Farmer and S. Rasmussen. Redwood City, CA: Addison-Wesley.
Ray, Thomas S. 1992b. Foraging behaviour in tropical herbaceous climbers (Araceae). Journal of Ecology 80:189-203.
Ray, Thomas S. 1995. An Evolutionary Approach to Synthetic Biology: Zen and the Art of Creating Life. In Artificial Life: An Overview, edited by C. Langton. Cambridge, MA: The MIT Press/ A Bradford Book.
Reeve, C.D.C. 1992. Practices of Reason: Aristotle's Nicomachean Ethics. Oxford: Clarendon Press.
Reichenbach, Hans. 1957. The Rise of Scientific Philosophy. Berkeley and Los Angeles: University of California Press.
Reid, Thomas, ed. 1983. Thomas Reid's Inquiry and Essays. Edited by R. E. Beanblossom and K. Lehrer. Indianapolis: Hacktt Publishing Company, Inc.
Rensberger, Boyce. 1996. Life Itself: Exploring the Realm of the Living Cell. New York: Oxford University Press.
Rorty, Richard M. 1992a. Introduction: Metaphilosophical Difficulties of Linguistic Philosophy. In The Linguistic Turn, edited by R. M. Rorty. Chicago: The University of Chicago Press.
Rorty, Richard M., ed. 1992b. The Linguistic Turn: Essays in Philosophical Method. Chicago: The University of Chicago Press.
Rosenberg, Alex. 1996. A Field Guide to Recent Species of Naturalism. The British Journal of the Philosophy of Science 47:1-29.
Rosenberg, Jay F. 1993. Comments on Peter van Inwagen's Material Beings. Philosophy and Phenomenological Research LIII (3):701-8.
Rosenblueth, A., N. Wiener, and J. Bigelow. 1943. Behavior, Purpose, and Teleology. Philosophy of Science 10:18-24.
Ross, W.D. 1924. Aristotle's Metaphysics: A Revised Text with Introduction and Commentary. II vols. Oxford: The Clarendon Press.
Ross, W.D. 1936. Aristotle's Physics: A Revised Text with Introduction and Commentary. Oxford: The Clarendon Press.
Rudder Baker, Lynne. 1995. Explaining Attitudes: A Practical Approach to the Mind. Cambridge: Cambridge University Press.
Rudder Baker, Lynne. 2000. Persons and Bodies: A Constitution View. Cambridge: Cambridge University Press.
Ruse, Michael. 1998. Taking Darwin Seriously: A Naturalistic Approach to Philosophy. 2nd ed. Amherst, NY: Prometheus Books.
Sanford, David H. 1995. Determinable. In The Cambridge Dictionary of Philosophy, edited by R. Audi. Cambridge: Cambridge University Press.
359
Sauvé Meyer, Susan. 1992. Aristotle, Teleology, and Reduction. The Philosophical Review 101 (4):791-825.
Scaltsas, Theodore. 1988. Commentary on Gotthelf. In Proceedings of the Boston Colloquium in Ancient Philoosphy, edited by J. J. Cleary and D. C. Shartin. Lanham, Md.: University Press of America.
Schröder, Jürgen. 1998. Emergence: Non-Deducibility or Downwards Causation? The Philosophical Quarterly 48 (193):433-52.
Searle, John R. 1967. Determinables and Determinates. In The Encyclopedia of Philosophy, edited by P. Edwards. New York: Macmillan Publishing Co., Inc. & The Free Press.
Searle, John R. 1980. Minds, Brains, and Programs. The Behavioral and Brain Sciences iii (3):417-24.
Searle, John R. 1983. Intentionality: An Essay in the Philosophy of Mind. Cambridge: Cambridge University Press.
Searle, John R. 1992. The Rediscovery of the Mind. Cambridge, Mass: The MIT Press/ A Bradford Book.
Searle, John R. 1995. The Construction of Social Reality. New York: The Free Press.
Searle, John R. 1997. The Mystery of Consciousness. New York: The New York Review of Books.
Sedley, David. 1991. Is Aristotle's Teleology Anthropocentric? Phronesis 36:179-96.
Sharvy, Richard. 1983. Aristotle on Mixtures. Journal of Philosophy 80:439-57.
Shields, Christopher. 1999. Order in Multiplicity: Homonymy in the Philosophy of Aristotle. Oxford: Clarendon Press.
Silberstein, Michael, and John McGeever. 1999. The Search for Ontological Emergence. The Philosophical Quarterly 49 (195):182-200.
Singer, Peter. 1993. How are We to Live? Amherst, NY: Prometheus Books.
Sober, Elliot. 1980. Evolution, Population Thinking, and Essentialism. Philosophy of Science 47:350-383. Reprinted in Sober, Elliott, ed. 1993. Conceptual Issues in Evolutionary Biology. 2nd ed. Cambridge: MIT Press. pp. 161-89.
Sober, Elliot. 1985. Panglossian Functionalism and the Philosophy of Mind. Synthese 64:165-93.
Sober, Elliott. 1992. Learning from Functionalism--Prospects for Strong Artificial Life. In Artificial Life II:, edited by C. G. Langton, C. Taylor, J. D. Farmer and S. Rasmussen. Redwoond City, CA: Addison-Wesley.
Sober, Elliott. 2000. Philosophy of Biology. 2nd ed. Boulder, CO: Westview Press.
Sorabji, Richard. 1980. Necessity, Cause, and Blame: Perspectives on Aristotle's Theory. Ithaca: Cornell University Press.
Spafford, Eugene H. 1995. Computer Viruses as Artificial life. In Artificial Life: An Overview, edited by C. Langton. Cambridge, MA: The MIT Press/ A Bradford Book.
Spencer-Smith, Richard. 1994. Reductionism and Emergent Properties. Proceedings of the Aristotelian Society 95:113-29.
Stairs, A. 1990. Quantum Mechanics, Mind and Self. In Essays on Personal Identity, edited by C. Taylor. Harvard: Harvard University Press.
360
Stapp, H.P. 1993. Mind, Matter and Quantum Mechanics. Munich: Springer.
Sterelney, Kim, and Paul E. Griffiths. 1999. Sex and Death: An Introduction to the Philosophy of Biology. Chicago and London: The University of Chicago Press.
Strong, D.R., and T.S. Ray. 1975. Host tree location behavior of a tropical vine (Monstera gigantea) by skototropism. Science 190:804-6.
Taylor, Paul W. 1986. Respect for Nature: A Theory of Environmental Ethics. Princeton: Princeton University Press.
Teller, Paul. 1986. Relational Holism and Quantum Mechanics. British Journal for the Philosophy of Science 37:71-81.
Teller, Paul. 1995. Reduction. In The Cambridge Dictionary of Philosophy, edited by R. Audi. Cambridge: Cambridge University Press.
van Inwagen, Peter. 1990. Material Beings. Ithaca and London: Cornell University Press.
Walsh, Denis M., and André Ariew. 1996. A Taxonomy of Functions. Canadian Journal of Philosophy 26:493-514. Reprinted in Buller, David J., ed. 1999. Function, Selection, and Design. Albany: State University of New York Press. pp. 257-79.
Ward, Julie K. 1996. Souls and Figures: Defining the Soul in De Anima ii 3. Ancient Philosophy 15:113-28.
Wardy, Robert. 1993. Aristotelian Rainfall or the Lore of Averages. Phronesis xxxviii (1):18-30.
Warren, Mary Anne. 1997. Moral Status: Obligations to Persons and Other Living Things. Oxford: Clarendon Press.
Wieland, Wolfgang. 1962. The Problem of Teleology. In Articles on Aristotle, I, edited by J. Barnes, M. Schofield and R. Sorabji. London.
Wiggins, David. 1980. Sameness and Substance. Cambridge, MA: Harvard University Press.
Wolterstorff, Nicholas. 1983. Thomas Reid on Rationality. In Rationality in the Calvinian Tradition, edited by J. V. H. Hart, and N. Wolterstorff. Lanham, Md.: University Press of America.
Woodfield, Andrew. 1976. Teleology. Cambridge ; New York: Cambridge University Press.
Wright, Larry. 1973. Functions. Philosophical Review 82 (2):139-168. Reprinted in Allen, Colin, Marc Bekoff, and George Lauder, eds. 1998. Nature's Purposes: Analyses of Function and Design in Biology: The MIT Press. Also reprinted in Buller, David J., ed. 1999. Function, Selection, and Design. Albany: State University of New York Press. pp. 29-55.
Wright, Larry. 1976. Teleological Explanations: An Etiological Analysis of Goals and Functions. Berkeley: University of California Press.
Yablo, Stephen. 1993. Is Conceivability a Guide to Possibility? Philosophy and Phenomenological Research 53:1-42.
361
Index Locorum
Cat.
1a1 ff. ...................................................................... 65 1a1-7........................................................................ 65 1a6-12............................................................309, 325 3b2-8 .............................................................309, 325 11a5-13 .........................................................120, 314 12b38..................................................................... 126
Int.
21a21-2 .........................................................317, 334 22b36-23a6........................................................... 126
AnPr.
46a17-26 ............................................................... 114 67b4-5 ................................................................... 324
AnPo.
67b12...............................................................71, 160 76a31-6 ................................................................. 117 77b24-5 ................................................................. 324 81a38-b9 ............................................................... 114 89b10 ff. ............................................................... 320 93b29-30............................................................... 294 94b28....................................................................... 41 94b32-3 ................................................................... 92 94b37-95a3............................................................. 92 96a12 ff. ................................................................ 101 97b31-36............................................................... 304 98a21-3 ................................................................. 304
Top.
100a13-14 ............................................................. 113 101a25-b4............................................................... 45 101b20-2............................................................... 294 101b31 .................................................................. 294 102a18..................................................................... 95 103b9-10............................................................... 294 103b14 .................................................................. 310 105a3 ....................................................................... 45 106a1-8 ................................................................. 313 106b29-107a2 ...................................................... 313 107b6-7 ................................................................. 310 107b7.............................................................312, 318 107b9-12............................................................... 313 107b17 ..........................................................120, 314 109b4-6 .........................................................309, 325 121a11-5 .......................................................309, 325 123a29...........................................................309, 325 130a29-130b10 .................................................... 323 133b31-6.........................................................71, 160 133b33 ............................................................71, 160 141b26 .................................................................. 310 145a26-7 ............................................................... 334
148a23 ff..................................................................45 148a26-31................................................................70 148a27 ff............................................................... 298 148a29.........................................................54, 70, 75 148a29-30.....................................................300, 318 148b16-22 ...............................................................26 153a15-22............................................................. 294 153b2 .................................................................... 310 153b14 .................................................................. 310 154a31................................................................... 294 157a34-5 ............................................................... 111 157b31-3............................................................... 111 160a6........................................................................45 160a38-b13........................................................... 111 160b7-8................................................................. 116 161a37-b1............................................................. 111 163b13 ............................................................45, 320
SE
181b33-4............................................................... 306
Phys.
185a10......................................................................45 185a10-12............................................................. 343 185a14-16............................................................. 115 186a30-1 .........................................................71, 160 188a32-4 ............................................................... 109 188b28-30 ............................................................ 116 188b28-9............................................................... 226 189a28-33................................................................41 189b30-2............................................................... 321 192a25-33............................................................. 156 192b9-11.............................................. 118, 124, 128 192b21-3.......................................................119, 157 193a4 ff....................................................................45 193b7 ..................................156, 157, 287, 294, 325 193b7-12.............................................. 108, 112, 287 193b7-8................................................................. 156 193b8 ................................................... 106, 109, 341 194a28-33............................................ 149, 287, 289 194a29-33............................................................. 286 194a30-3 ........................................................... 94, 95 194a33-6 ............................................................... 130 194a35................................................................... 133 194b9-10............................................................... 156 194b13 ..........................................................106, 341 194b33-4............................................................... 333 194b36-195a2 ...................................................... 129 195a23-5 .......................................................149, 289 195b22 ..........................................................283, 284 196a30...........................................................108, 109 196a31................................................................... 341 196a31-2 ............................................................... 108 196b18-20 ............................................................ 283 196b33-6............................................................... 102
362
197a3-5 ................................................................. 102 197b7-9 ................................................................. 121 198a25.................................106, 109, 144, 157, 341 198b5-9 ................................................286, 293, 325 198b8-9 .........................................................149, 289 198b10-11............................................................... 90 198b16-199a8 ...................................................... 129 198b16-21............................................100, 101, 102 198b17-198b33...................................................... 90 198b17-33............................................................. 309 198b24-9............................................................... 104 198b26-32.....................................................105, 106 198b33-199a8 ...................................................... 110 198b34 ................................90, 96, 98, 99, 102, 105 198b34-199a8 ...................................90, 96, 98, 102 198b34-6................................................................. 99 198b35 .................................................................... 97 199a1-2 ...........................................................92, 101 199a2-3 ................................................................. 102 199a9-14 .......................................................129, 130 199a9-19 ............................................................... 111 199a15-8 .......................................................129, 130 199a20-33 ............................................................. 111 199a30-3 ............................................................... 286 199a30-33 .....................................................144, 157 199a33-b6............................................................. 107 199a34-b6............................................................. 111 199b14 ff. ............................................................. 106 199b14-18............................................................. 107 199b14-26............................................................. 110 199b14-8............................................................... 109 199b17-8................................................................. 99 199b19 ............................................................93, 107 199b25-6.......................................................107, 109 199b27-30............................................................. 111 199b27-31.....................................................123, 283 200a1-5 ................................................................... 92 200a8-11 ............................................................... 341 200b11 .................................................................. 157 200b24 .................................................................. 321 201a10-3 ............................................................... 314 201a28................................................................... 314 201b6..................................................................... 314 202a7 ..................................................................... 314 202a10................................................................... 127 202b9-10.........................................................71, 160 202b20-1.........................................................71, 160 211a6-11 ............................................................... 297 213a2 ..................................................................... 157 219a20-1 .........................................................71, 160 219b18-9.........................................................71, 160 229a19.............................................................71, 160 229a28.............................................................71, 160 241a34................................................................... 157 246a10-15 ................................... 288, 293, 324, 325 246a12-15 ............................................................. 289 247b14-6............................................................... 324
248a3-5 ................................................................. 324 248a17-20............................................................. 315 248a19-22.....................................................302, 314 248a20................................................................... 316 248b6-10............................................................... 120 248b8-9................................................................. 314 248b13 .................................................................. 316 248b14 ..........................................................315, 316 248b14-5............................................................... 315 248b17-20 ............................................................ 312 249a3-5 ................................................................. 316 251a9..................................................................... 314 252a10.................................................. 107, 109, 157 252a32-3 ........................................................... 94, 95 253b9 .................................................................... 157 255a6-10 ..................................................................71 255a34-b4............................................................. 324 257b8-10............................................................... 109 257b8-9................................................................. 127 260b8-11............................................................... 340 260b22 .................................................................. 117 260b23 .................................................................. 114 262a20.............................................................71, 160 263b7-8...........................................................71, 160 263b12-4.........................................................71, 160
OH
268b15 .................................................................. 157 271a33................................................................... 124 271b8-12............................................................... 343 271b18-271b25.................................................... 340 279a19-22............................................................. 298 279a22-30............................................................. 127 279a29......................................................................43 279b4-5....................................................................94 283b33-284a1 ..............................................111, 117 284b30-3............................................................... 287 285a14-9 ............................................................... 287 285a28-31............................................................. 291 286a8............. 52, 95, 284, 292, 298, 307, 325, 333 286b15-6............................................................... 322 288b15-18 ...............................................................95 289a10................................................................... 309 292a19-292b24 .................................................... 291 292a22-b25........................................................... 127 292b17-18 ............................................................ 285 293b-9 ................................................................... 287 296b6 ff. ............................................................... 125 296b8 ff. .........................................................71, 160 297b21 ............................................................95, 284 297b21-2.............................................. 292, 325, 333 297b21-3............................................................... 307 299b20 .................................................................. 127 301b18 .................................................................. 157 306a16-18............................................................. 114 312a16-21............................................................. 157 312a19-21.......................................................71, 160
363
312a30-2 .........................................................71, 160
GC
314b1-3 ................................................................... 99 315b26-8............................................................... 343 316a5-13 ............................................................... 114 320b13-4.........................................................71, 160 320b23-5.........................................................71, 160 327a35-327b6 ...................................................... 339 327b23-5............................................................... 340 327b23-7.........................................................41, 339 329b17 ff. ............................................................. 340 333a22...........................................................120, 314 333b4-16............................................................... 109 333b4-17............................................................... 109 333b16-7............................................................... 109 335a15-6 ................................................................. 52 335a26................................................................... 321 335b6.............................................................144, 157 336a13-22 ............................................................. 122 336a13-24 ............................................................. 120 336b25-34.....................................................119, 127 336b25-35.....................................................282, 284 336b27-337a7 ...................................................... 125 337a21................................................................... 125
Meteor.
338b12-19............................................................. 285 339a19-32 ............................................................. 125 346b20-4.......................................................102, 120 347b12 ff. ............................................................. 101 359b34-360a5 .....................................101, 102, 120 369a24-33 ............................................................... 92 378b31-4................................................................. 92 380b13-4................................................................. 92 384b15-18............................................................... 92 389b27-9............................................................... 156 390a10........................................... 95, 284, 307, 333 390a10-11 .....................................................292, 325 390a10-15 .....................................................317, 334 390b2-14................................................................. 92
DA
402b21-5............................................................... 114 403b23-8................................................................. 48 405b11 .................................................................... 48 407b24-5............................................................... 341 408b18 ..........................................................298, 326 409b24 .................................................................. 339 409b28-8.........................................................41, 339 411a24-b4............................................................... 49 412a9 .....................................69, 156, 157, 294, 325 412a10................................................................... 324 412a14.....................................51, 77, 290, 297, 322 412a19-22 .......................................................49, 324 412a21..................................................156, 294, 325 412a22..................................................................... 69
412a22-6 ............................................................... 324 412a27-9 .........................................................49, 324 412b4-6..........................................49, 324, 325, 326 412b10-1.........................................................49, 324 412b12-17 ............................................................ 342 412b17-24 ....................................................317, 334 412b27 .....................................................................69 412b27-413a2 ...................................................... 324 413a4-6 ................................................................. 326 413a11-20.......................................................49, 322 413a20-6 ........................................................... 49, 70 413a21-2 .....................................................54, 70, 75 413a21-5 ............................................................... 298 413a22-5 ............................................................... 300 413a24................................................................... 296 413a31-b1................................... 294, 297, 324, 326 413a32-4, .................................................................51 413b1-2...........................................................51, 290 413b4-13..................................................................51 413b11-13 ...............................................................49 413b25 .................................................................. 298 413b25-8............................................................... 326 414a15......................................... 156, 157, 294, 325 414a16-28.....................................................342, 344 414a25-8 ............................................................... 341 414a28.............................................................49, 324 414a29...........................................................298, 300 414a29-32............................................................. 298 414a29-415a12........................................................51 414b20 ..........................................................303, 305 414b20-34 ............................................................ 303 414b20-415a1 .........................................................55 414b21 .................................................................. 306 414b25-7............................................................... 308 414b29 .................................................................. 305 415a1..................................................................... 322 415a16-21.....................................................145, 294 415a18................................................................... 158 415a24-5 .........................................................51, 290 415a24-b2............................................................. 124 415a25-b7.......................................................79, 103 415a27-30............................................................. 285 415a27-b2.............................. 55, 57, 59, 62, 64, 83,
84, 281, 282, 284, 292 415b2-3................................................................. 333 415b8-14..................................................................71 415b9-11.............................................. 285, 324, 325 415b10-12 ....................................................144, 157 415b12-14 ............................................................ 285 415b12-4............................................................... 293 415b13 ..................................................... 45, 71, 289 415b15-20 ..........................282, 284, 285, 292, 294 415b15-21 ............................................................ 285 415b17-18 ....................................................134, 333 415b20-1............................................................... 333 415b22-7............................................................... 285 416a5..................................................................... 287
364
416a15-8 .......................................................157, 287 416a20................................................................... 297 416a20-1 ................................................................. 52 416b17-18.....................................................149, 290 416b23 ................ 95, 149, 284, 292, 307, 325, 333 417a18................................................................... 127 417a22-31 ............................................................... 69 417a22-b1............................................................. 324 417b31 ff. ............................................................... 69 417b31-418a2 ...................................................... 324 420b23-5............................................................... 309 424a18-24 ............................................................... 68 424a32-b3............................................................... 68 425b23 .................................................................... 68 429a10-1 .........................................................71, 160 429b4.............................................................298, 326 429b6....................................................................... 69 429b6-9 ................................................................. 324 431a3 ..................................................................... 127 431a9 ...............................................................71, 160 431a13-4 .........................................................71, 160 431a28.............................................................71, 160 431b28-32a2........................................................... 68 432a-b4 ................................................................... 59 432a20.............................................................71, 160 433a15..................................................123, 291, 296 433b24 ............................................................71, 160 434a23......................................... 294, 297, 324, 326
PN
436a18...........................................................120, 292 449a14-9 .........................................................71, 160 454a11-454a18....................................................... 51 454a13......................................... 294, 297, 324, 326 454b24 .................................................294, 324, 326 455b20-1............................................................... 116 455b22-4............................................................... 288 457b31 ff. ............................................................. 101 462b29-30............................................................... 91 463a2-3 ................................................................. 107 463a18................................................................... 343 463b11 ..........................................................107, 109 465a14-5 ................................................................. 95 467b12-25............................................................... 71 467b26-7.........................................................71, 160 468a18-21 ............................................................... 93 468b28-9a9........................................................... 287 469b4....................................................294, 324, 326 474b10-12............................................................... 51 478b23-5............................................................... 287 479a28-9 ..........................................51, 77, 149, 290
HA
486b23-487a2 ..............................................342, 343 532a5 ............................................. 93, 107, 150, 296 532a11........................................... 93, 107, 150, 296 535a8 ..................................................................... 296
538a13................................................................... 327 550b34-551a10 ............................................107, 296 588b4-16.......................................................324, 343 588b4-23........................................................... 43, 44 588b24 ........................................ 289, 294, 324, 326 588b24-5..................................... 103, 294, 324, 326 588b25-6.......................................................149, 284 589b29-590a11 .................................................... 343 596b21-3............................................................... 103 633b16-20 ............................................................ 292
PA
639a23-4 ............................................................... 321 640a12-16............................................................. 115 640a19-25.....................................................108, 109 640a22-4 ..................................................................99 640a25-6 .............................................. 106, 109, 341 640b1-4.........................................................106, 341 640b22-9............................................................... 344 640b30-41a17 .........................................................67 640b30-641a6 ..............................................317, 334 641a1-5 ....................................................................99 641a10-15................................................................99 641a15-8 ..................................................................99 641a20-24................................................................99 641a25-9 .........................................................49, 324 641a26-7 ..................................................................99 641a30-1 ............................................................... 344 641b5-8....................................................................99 641b10-642a1 ..............................................113, 137 641b24-8.......................................................108, 109 641b27-30 ...............................................................99 641b27-8............................................................... 341 641b34-6..................................................................99 642a1-2 ....................................................................93 642a12-4 .............................................. 129, 134, 333 642a19-20.....................................................116, 226 642a27-8 .......................................................116, 226 644a33-b7............................................................. 321 645b11-2............................................................... 321 645b15-20 ........................................... 129, 134, 333 646a12-b10........................................................... 156 646a15-21............................................................. 342 646a18-20............................................................. 340 646a28-36.....................................................341, 342 646a31-5 ............................................................... 109 646a35.................................................. 106, 109, 341 646b11 .................................................................. 156 647a22-30............................................................. 287 651b4 ................................................... 294, 324, 326 653a2 ff................................................................. 101 653b23 ................................................. 294, 324, 326 653b38-654a2 ...............................93, 107, 150, 296 654a32......................................................................93 655b21 .................................................................. 284 658a18-24............................................................. 122 663a9-12 ..................................................................95
365
663b12-4................................................................. 95 664a3-8 ................................................................... 95 665a18-26 ............................................................. 122 665b18-21.....................................................122, 287 666a35..................................................294, 324, 326 670a23-27 ............................................................. 287 670a30..................................................................... 95 677a12-31 .................................................. 60, 93, 95 678b1-6 ................................................................. 287 678b11-14..................................... 93, 107, 150, 296 679b35-6....................................... 93, 107, 151, 296 681a11-15 .........................................................42, 44 681a12-5 .......................................................324, 343 683b10-11..................................... 93, 107, 151, 296 684a28-9 ............................................................... 334 687a11-15 ............................................................. 334 687a16................................................................... 122 694b14 ..........................................................333, 334
MA
698a13..................................................................... 45 700b4..................................................................... 296 700b15-6...............................83, 282, 284, 288, 292 701b25-33............................................................. 343 703a36-7b............................................................. 335 703b10-16............................................................... 93 704b11-16............................................................. 114 704b11-17............................................................. 117 704b15-6..................................... 286, 288, 293, 325
IA
705a27-b2............................................................. 287 705b5..................................................................... 287 705b19 .................................................................. 287 706b10-16............................................................. 287 708a10......................................... 287, 288, 293, 325 710a3 ff. .................................................................. 95
GA
715a4-6 .........................................................144, 157 715a4-8 .........................................................285, 294 715a8-9 .........................................................144, 157 715a9-11 ............................................................... 156 715a23-5 .......................................................150, 296 715a23-b15........................................................... 107 715b4-15.......................................................150, 296 716a23-5 ..............................................134, 333, 334 716b4-5 ................................................................. 343 717a21...........................................................103, 289 718b9-10.......................................................103, 289 722b23-4............................................................... 131 725a4-6 ................................................................... 95 729a9-14 .........................................................41, 345 729b6-21.........................................................41, 345 730b21-2............................................................... 333 730b32 .................................................................. 327 731a25-8 ..............................................294, 324, 326
731a25-b7................................... 103, 149, 284, 289 731b3 ................................................... 294, 324, 326 731b20-732a9 ...................................................... 127 731b24-32 ....................................................282, 284 731b24-732a11 .................................................... 285 734a30-1 ............................................................... 127 734a30-2 ............................................................... 109 734b20 .................................................................. 127 734b20-1............................................................... 109 734b20-2............................................................... 341 734b25-7.......................................................317, 334 734b27-30 ............................................................ 134 734b31-735a2 .........................................................92 735a15-7 ............................................................... 289 735b16-21 ...............................................................92 735b26-37 ...............................................................92 736b3-4................................................ 293, 324, 325 736b4-5.........................................................149, 286 736b27 ..........................................................298, 326 738a9-b3..................................................................95 738b19-26 ......................................................41, 345 738b25 .....................................................................71 738b27 ............................................................49, 324 739a8-13 ..................................................................95 740b29 ff. ............................................................. 297 741a1........................................................ 51, 77, 290 741a19-20............................................................. 121 742a19-21............................................................. 288 742a21-36.............................................. 86, 333, 334 743a21-4 ..................................................................99 743a21-6 .......................................................341, 342 745a5-10 ............................................................... 287 747b29 ............................................................45, 321 748a1..................................................................... 327 748a8...............................................................45, 321 748a8-9 ................................................................. 321 760a32................................................................... 157 760b29-32 ............................................................ 114 760b29-33 ............................................................ 115 762a9 ff................................................................. 107 762b29 ff. ............................................................. 107 763a24 ff............................................................... 296 764a3-7 ....................................................................99 764a12-19................................................................99 765a35-b6............................................................. 283 765b1-7....................................................................99 765b10-15 ......................................................41, 345 766a23-29............................................................. 343 767a17................................................................... 157 767b35 ..........................................................106, 341 771b33-772a29 .................................................... 287 776a15 ff..................................................................95 776b15-7............................................................... 343 777b24-778a2 ..............................................102, 120 778a29-b1................................................................92 778a32-b6................................................................60 778b4-6................................................................. 288
366
778b7-10............................................................... 165 778b10-19............................................................... 92 778b11-19............................................................... 60 781b23-28............................................................. 309 784a12-16 ............................................................. 103 788b20 ..........................................................114, 117 788b33-789a2 ...................................................... 105 789a8-b8 ................................................................. 41 789a11-b2............................................................. 105 789b1-15................................................ 41, 165, 166 789b2-15................................................................. 92
Met.
983a31-2 .......................................................149, 290 984a17-9 ............................................................... 116 984a18................................................................... 226 984b9-10............................................................... 116 984b11-4................................................................. 99 986b30-1............................................................... 116 988a18-24 ............................................................. 115 988b6-15..............................................110, 165, 166 988b8-15............................................................... 289 988b16-8............................................................... 115 993b21 .................................................................. 299 999a6-7 ................................................................. 306 1003a33-b6............................................................. 65 1003a35-7 ............................................................. 313 1003b2-3...........................................................69, 80 1003b6 .................................................................... 78 1006a5.................................................... 45, 115, 320 1012a18-20........................................................... 116 1013a36-b3........................................................... 129 1015a13-9 ............................................................. 157 1016a24-32........................................................... 304 1016b35 ................................................................ 304 1021b20 ...................................... 288, 289, 293, 325 1021b24-5............................................................... 94 1021b24-9............................................................. 287 1021b28-30 ............................................................ 94 1024b1-9............................................................... 304 1027a31-2 ............................................................. 109 1027b8-10.......................................................95, 287 1028a35-6 ............................................................. 158 1029b10-16 .......................................................... 322 1032a22................................................106, 109, 341 1032a30-2 .....................................................107, 296 1033b20-34a8 ...................................................... 306 1033b31 ..........................................................71, 160 1034a22.........................................................109, 127 1035a7................................................................... 344 1035b14-31 ....................................................49, 324 1037a5.............................................................49, 324 1037b29-38a9 ...................................................... 306 1037b30 ff. ........................................................... 310 1039a24-32........................................................... 306 1040a8-21 ............................................................. 306 1042a16-7 ............................................................. 294
1043a35...........................................................49, 324 1043b2 .....................................................................71 1044a27................................................................. 341 1044a33-1044b2..........................................283, 284 1044b8-12 ........................................................ 92, 93 1044b15-21 .......................................................... 283 1044b34-1045a2.................................................. 126 1045b18-9 ......................................................71, 160 1049b4-16 ........................................... 145, 158, 294 1046b5-24 ............................................................ 126 1049b9 .................................................................. 157 1049b17-29 .......................................................... 127 1049b24-6 ............................................................ 127 1050a15.........................................................156, 157 1050a15-16........................................................... 156 1050b1-2............................................................... 121 1050b2 ........................................ 156, 157, 294, 325 1050b3 .................................................................. 158 1050b4 .................................................................. 127 1050b5-6............................................................... 127 1050b22-30 .......................................................... 126 1051a3................................................................... 158 1061a1-7 ............................................................... 313 1065a29...........................................................93, 107 1069b19 ................................................................ 157 1070a4...........................................................109, 127 1070a4-5 ............................................................... 109 1070a29................................................ 106, 109, 341 1071a8...........................................................156, 157 1071a8-9 .............................................. 156, 294, 325 1071a20.........................................................106, 341 1071a24-9 ............................................................. 109 1071a25.........................................................106, 341 1071b26-30 .............................................................52 1071b29-31 .......................................................... 127 1072a9................................................................... 158 1072a11-18...................................................102, 120 1072b1-2............................................................... 333 1072b3-4............................................................... 122 1072b24-30 ................................................58, 59, 70 1072b26-30 .......................................................... 298 1072b26-7 ..................................................72, 76, 78 1073a4-13................................... 295, 298, 324, 327 1073a24-8 ............................................................. 124 1073b10-6 ............................................................ 115 1074a15-7 ............................................................. 115 1075a11-25............................................ 92, 131, 132 1075a16-7 ............................................................. 132 1075a18................................................................. 124 1075a37-b1........................................................... 289 1075a37-b8...................................................165, 166 1075b4-5.........................................................71, 160 1076a4................................................................... 124 1077b1-4.........................................................71, 160 1079b2-3............................................................... 317 1085b36 ................................................................ 115 1087a21.........................................................106, 341
367
1092a3...........................................................156, 157 1092a3-4 ............................................................... 156
NE
1094b27 ff. ............................................................. 45 1095a28................................................................... 45 1096a9..................................................................... 45 1097a25-b21......................................................... 283 1097b26-7.....................................................293, 325 1097b26-8............................................................. 290 1097b33-1098a4 .................................................. 293 1097b33-a4........................................................... 325 1098b4-9............................................................... 343 1100a3-4 ................................................................. 69 1102a32-b4........................................................... 290 1103b24-6............................................................. 343 1106a15-7 ............................................................. 292 1112a35-b2........................................................... 283 1113a32................................................................. 320 1115b23 ....................... 95, 284, 292, 307, 325, 333 1117b29 ........................................................298, 326 1139a9................................................................... 294 1139a26................................................................. 299 1139b12 ................................................................ 299 1142b2-5............................................................... 320 1143a20 ff............................................................. 320 1144a23-9 ............................................................. 320 1145b1-7............................................................... 117 1146b6-7............................................................... 111 1146b7-8............................................................... 297 1147a11................................................................. 324 1147b11 ................................................................ 324 1154a22-4 ............................................................. 297 1154a23-5 ............................................................. 117 1155b30-1............................................................. 120 1161b28 ................................................................ 285 1170a15.................................77, 288, 294, 324, 326 1170a15-19..........................................145, 158, 294 1170a15-8 ............................................295, 324, 326 1170a15-9 .....................................................288, 324 1170b1 ........................................ 289, 295, 324, 326 1178a21................................................................. 298 1178a21-4 ............................................................. 326 1178b20 .................................................................. 52 1178b20-3................................... 295, 298, 324, 327 1179b25 .................................................................. 45
MM
1185a17................................................................... 48
EE
1214b29 ff. ............................................................. 45 1215a6...........................................................117, 297 1216b26-8............................................................. 114
1216b35-1217a1.................................................. 320 1217a1-10............................................................. 320 1217a9......................................................................45 1218a27.........................................................123, 292 1218a38................................................................. 292 1218b22 ...................................................................45 1218b9-11 ....................................................149, 290 1219a9-11.....................................................149, 290 1219a24-5 .....................................................288, 324 1221b28-30 .......................................................... 299 1232a20-7 ............................................................. 318 1235a2......................................................................45 1241b18-24 ..................................................134, 333 1242a13-7 .....................................................134, 333 1247a32.........................................................107, 109 1247a33-4 ............................................................. 109 1247a33-6 ............................................................. 107 1247a35-6 .....................................................107, 109 1249b15-6 ............................................................ 333
Pol.
1252a27-33........................................................... 123 1252a27-9 ............................................................. 103 1252a28-30.......................................... 149, 283, 284 1252b34-5 ....................................................149, 289 1253a19 ff. ..............................................................92 1253a20-5 .....................................................317, 334 1253a23........................ 95, 284, 292, 307, 325, 333 1256b10-22 ....................................................92, 129 1256b11-21 .......................................................... 132 1258a22-7 .....................................................133, 333 1259b37 ................................................................ 307 1260a11-16...................................................307, 326 1260a25-7 ............................................................. 307 1260a27................................................................. 308 1276b20-7 ............................................................ 306 1282b35-7 ............................................................ 334 1303a21-3 ............................................................. 343 1303b29-32 .......................................................... 343 1307a40-2 ............................................................. 343 1328a24-6 ......................................................... 60, 93 1328a29-32........................................................... 333 1334b15 ............................................... 295, 324, 326
Rhet.
1356b35 ...................................................................45 1361b3-6............................................................... 292 1362a30................................................................. 292 1364b12 ff. ..............................................................45 1369b1-2.......................................................107, 109 1394b7 .....................................................................45 1395b30 ...................................................................45 1418b5 ..........................................................117, 297
368
General Index
A
analysis, conceptual ......................... 19�22, 24, 27, 30, 33, 44, 46, 182, 184, 188, 189, 192, 195, 196, 200
B
Bedau, Mark................................ 1, 2, 6, 11, 16, 17, 23�30, 36, 46, 50, 121, 154, 168, 170�75, 177, 204�6, 208, 216, 223, 226, 227, 230, 231, 240, 241, 243, 254, 258, 259, 331
Boden, Margaret................................................ 2, 28 Broad, C.D. .........................35, 163, 226, 227, 231,
233, 234, 252, 254, 257, 259, 276, 277, 339
C
Chalmers, David .................... 1, 16, 162, 191, 222, 239, 271
Churchland, Paul....................... 13, 17, 19�23, 270 Cummins, Robert..............................171, 172, 174,
180, 186�88, 194, 204, 334
D
Dawkins, Richard..........................9, 10, 14, 15, 64, 185, 187, 199, 287, 334
definition, theoretical ....................... 10, 19�24, 27, 29, 32, 33, 44, 46, 178, 182�85, 188�92, 195�97
Dennett, Daniel................................11, 15, 28, 332
E
emergentism affirmation of novelty ................................... 229 basal conditions.............................................. 262 core sense of emergence............................... 257 emergence, strong ontological ..................... 279 emergence, weak ontological ....................... 279 emergent, nomic strong ontological ........... 279 epistemological emergence........................... 259 merely resultant properties........................... 272 novelty ............................................................. 262 novelty* ........................................................... 271 ontology........................................................... 229 preformationist assumption .................228, 255 ultimate physical ontology............................ 229
F
function cybernetic approach ...................................... 174 efficient causality condition ......................... 174 etiological account ......................................... 176 Furley/Irwin account .................................... 150 normativity condition.................................... 206 the realist stance ............................................. 172 the reductivist's project ................................. 173
vs. adaptation.................................................. 211 Wright's definition ......................................... 175
Furley, David.......................... 85, 96, 97, 137, 138, 147�49, 151, 157, 164
G
Gotthelf, Allan ................................84, 97, 137�39, 141, 145, 147, 153, 155�60
H
homonymy ..............................................................65 account ............................................................ 319 associated ..................................................66, 318 discrete................................................................65 focal.................................................................. 319 four causal connection..............................66�67 general introduction ..................................64�66
Hume, David ............................ 217, 228, 230, 242, 243, 249, 255
I
Irwin, Terence .......................... 45, 84, 96, 97, 106, 137, 147�49, 151, 157
K
Kahn, Charles................. 85, 88, 124�28, 130, 142 Kim, Jaegwon ...................................................... 271
L
Langton, Christopher...................................7, 8, 29 Lewis, David......................137, 193, 239, 256, 262 life
Churchland, final definition............................20 Churchland, first attempt ................................20 cluster definitions ......................................11�14 eliminitivism about .........................9�11, 14�16 L1...................................................................... 329 L2...................................................................... 335 L2'..................................................................... 335 Matthews' account of.......................................55 of animals ........................................................ 327 of god............................................................... 327 of humans ....................................................... 327 of the spontaneously generated................... 327 provisional definition ..................... 84, 300, 327 Shields' definition..............................................73
M
Matthews, Gareth .......................12, 13, 29, 48, 50, 54�64, 75, 82�85, 281
Mayr, E. .......................... 12, 35�39, 137, 216, 217, 257, 266, 268
Millikan, Ruth G. ............................19, 27, 28, 182, 185, 187, 189, 198, 200, 201, 206, 207
369
N
naturalism............... 238�39, 244�45, 246, 345�46 Neander, Karen.............................19, 29, 114, 170,
171, 174, 176, 177, 182, 187, 192, 196�99, 201�3, 206, 207, 211, 212, 230, 238
P
Pettit, Philip ............................... 227, 236, 237, 248 Plato
Meno 71e-73e ................................................. 307 Symposium 206e-7a......................................... 62
R
reduction by coextension..........................................266�67 by property identity .................................269�71 compositional ...........................................268�69 explanatory theory ......................................... 269 generic notion of...................... 141�42, 264�65 Nagelian.....................................................267�68 weak notions of.............................................. 266
S
Scaltsas, D. ..........................................158, 159, 160 Searle, John .......................1, 6, 177, 207, 226, 227,
254, 263, 323, 332 Sedley, David .................... 85, 88, 92, 128�33, 135 Shields, Christopher ............12, 14, 19, 46, 48�50,
54, 55, 57�59, 61, 64�76, 78�85, 222, 281, 284, 292, 296, 300�303, 312, 317�20, 331
Sober, Elliot ..........................1, 5, 7, 15�17, 30, 37, 46, 173, 187, 211, 230
Socrates.........................................................307, 308 synonymy................................................................ 65
T
teleology nonreductive analysis .................................... 336 traditional objections to..........................216�17
Theses
categorial constraint on definitions ......26, 191 causal closure.................................................. 244 circularity problem............................................51 coextensiveness thesis......................................86 continuity entails eliminitivism.......................10 continuity entails failure of conceptual analysis ................................................................21 continuity plus conceptual analysis................22 Continuity plus theoretical definition ...........22 continuity thesis, the ........................................10 continuity, conceptual analysis and real definition ............................................................22 decision procedure requirement.....................31 life, emergentist accounts of ...........................40 material continuity ............................................44 mechanism, sophisticated................................40 mechanism, strong............................................36 mechanism, weak..............................................36 multivocity, problem of ...................................54 naked mind ........................................................28 natures thesis .................................................. 118 natures thesis, Kahn ...................................... 124 natures thesis, Sedley..................................... 128 no epistemeological baggage...........................31 particularism ................................................... 256 regularities thesis...............................................91 science, ontic authority of ............................ 233 small sample problem, the ................................7 unity, radical problem of .................................53 unity, the problem of .......................................50 vitalism, strong ..................................................38 vitalism, weak ....................................................38
V
van Inwagen..................................17, 18, 19, 27, 53 vitalism.......................17, 34�43, 46, 216, 219, 226
W
Ward, Julie....................................................303, 304 Wright, Larry ......................106, 148, 174�77, 191,
195, 204, 210
