Aristotle PDF

Papermacs are sold subject
to the condition that they

shall not, by way of trade or
otherwise, be lent, re-sold,
hired out, or otherwise
circulated without the
publisher's prior consent, in
any form of binding or cover
other than that in which they
are published and without a
similar condition including
this condition being imposed
on the subsequent
purchaser.
MODERN STUDIES IN PHILOSOPHY
ARISTOTLE
Modern Studies in Philosophy is a series of anthologies
presenting contemporary interpretations and evalua-
tions of the works of major philosophers. The editors
have selected articles designed to show the systematic
structure of the thought of these philosophers, and
to reveal the relevance of their views to the problems
of current interest. These volumes are intended to be
contributions to contemporary debates as well as to
the history of philosophy; they not only trace the
origins of many problems important to modern phi-
losophy, but also introduce major philosophers as in-
terlocutors in current discussions.
MODERN STUDIES IN PHILOSOPHY
ARISTOTLE
A Collection of Critical Essays
EDITED BY J. M. E. MORAVCSIK
PALGRAVE
MACMILLAN
© J. M. E. Moravcsik 1967
First published in the United States of America 1967

First published in Great Britain 1968
Published by
MACMILLAN AND CO LTD
Little Essex Street London w C 2
and also at Bombay Calcutta and };Iadras
Macmillan South Africa (Publishers) Pry Ltd Johannesburg
The Macmillan Company d Australia Pry Ltd Melbourne
ISBN 978-0-333-00436-4 ISBN 978-1-349-15267-4 (eBook)

DOI 10.1007/978-1-349-15267-4
CONTENTS
Introduction 1
I. LOGIC
Aristotle and the Sea Battle, G. E. M. ANSCOMBE 15
Aristotle's Different Possibilities,
K. J AAKKO J. HINTIKKA 34
On Aristotle's Square of Opposition, M. THOMPSON 51
II. CATEGORIES
Categories in Aristotle and in Kant,
JOHN COOK WILSON 75
Aristotle's Categories, Chapters I-V: Translation
and Notes, J. L. ACKRILL 90
Aristotle's Theory of Categories, J. M. E. MORAVCSIK 125
III. METAPHYSICS
Essence and Accident, IRVING M. COPI 149
Tithenai ta Phainomena, G. E. L. OWEN 167
Matter and Predication in Aristotle, JOSEPH OWENS 191
Problems in Metaphysics Z, Chapter 13,
M. J. WOODS 215
IV. ETHICS
The Meaning of Agathon in the Ethics of Aristotle,
H. A. PRICHARD 241
Agathon and Eudaimonia in the Ethics of Aristotle,
J. L. AUSTIN 261
The Final Good in Aristotle's Ethics,
w. F. R. HARDIE 297
Aristotle on Pleasure, J. o. URMSON 323
Notes on Contributors 334
Selected Bibliography 335
ARISTOTLE
INTRODUCTION
What is it for us to understand Aristotle?l In order

to grasp the difficulties involved in answering such a
seemingly simple question, it is necessary to consider,
briefly, some of the features of the history of philosophy
and some of the characteristics of Aristotle's philoso-
phy. Even if this necessarily sketchy introductory dis-
cussion fails to yield answers satisfactory to everyone,
it should at least help to make clear why this volume
is the way it is, and how it is viewed by its editor.
The'mere existence of recent and contemporary com-
mentaries on Aristotle might seem to call for an ex-
planation if not an apology. At first glance at least, it
might seem that after more than two thousand years
Plato and Aristotle should be as well explained as they
are ever likely to be, and that historians of philosophy
should be turning their attention to the interpretation
of less well known figures. And yet the facts' are that
there is exciting work being done on Aristotle today,
that in the view of numerous philosophers and his-
torians of philosophy much of the interpretation done
in the past, even when of excellent quality, is not
wholly satisfactory for our understanding, and that
lThe views expressed in this introduction about the nature of

the history of philosophy and about Aristotle are those of the
editor and are not to be imputed to the contributors to this
volume. The editor wishes to express his gratitude to Professor
\Villiam Dray for a helpful discussion of some of these matters.
2 J. M. E. Moravcsik
future generations are likely to view with dissatisfac-
tion much of the work done today, while doing good
and interesting historical work themselves. Under these
circumstances the question arises: Why do men feel
the need to rewrite the history of philosophy over and
over again? One should attempt to answer this ques-
tion, even if it turns out to be merely a special case of
the tantalizing general question: Why must history be
rewritten over and over again? In the case of philoso-
phy, this question is sharpened by the fact that the
repeated attempts at reconstructing the past are rarely
occasioned by the appearance of drastically new em-
pirical evidence. Once an ancient temple or palace is
unearthed, it is there for all to see; subsequent excava-
tions seek new targets. The interpreters of ancient
philosophy, however, do not behave this way.
Faced with a bewildering variety of materials, one
might be tempted to conclude that each school of phi-
losophy feels the need to rewrite the history of the
profession in its own image. Perhaps with each so-
called "revolution" in philosophy-of which we have
recently had such an uncomfortably large number that
one is strongly inclined to doubt the genuineness of
some-new emphases are given to different aspects of
the thinking of the past. Such a line of thought leads
to two popular conclusions, which this introduction is
intended to combat. One of these is the claim that
there is little if any objectivity in histories of philos-
ophy written by philosophers. The other is the view
that in order to gain unprejudiced, real insight into
the philosophies of classical writers the best thing to
do is to turn to the writings of those historians who
have little or no systematic interest in current philos-
ophy or at least hold no controversial positions upon
this subject.
One way of showing these conclusions to be non
sequiturs is to demonstrate that the peculiarly peren-
Introduction 3
nial nature of the history of philosophy can be ac-
counted for on the basis of four considerations. First,
an illuminating historical account involves comparisons
(not necessarily evaluative ones) between the thought
of the past and that of the present. (It is, after all,
mostly by contrast and comparison that we learn about
the thoughts of contemporaries as well.) Such com-
parisons tend to cast light both ways, and function
as a conceptual bridge. Second, much of the explain-
ing of a classical philosopher's thought involves ex-
planations in our own terms, or at least in terms fa-
miliar to us. No historical interpretation of an author's
works can be complete and adequate if it never breaks
out of the circle, no matter how wide, of that author's
terminology. These two ingredients of the historian's
work, ingredients that link the past to a perpetually
changing "present," and a perpetually changing set of
terms, "our terms," help to explain the need for re-
writing history without abandoning the claim of ob-
jectivity.
A third ingredient of the interpreter's task is the
tracing of implications and consequences of some of
a classical author's key claims. Such logical analysis is
aided by the cumulative experience of the centuries,
and it is an unending task, since such implications and
consequences have no natural boundary. Finally, if we
give a wide enough interpretation to the notion of
empirical evidence, one can say that the appearance of
new evidence fairly frequently plays an important role
in new efforts at historical reconstruction. Such evi-
dence may be a large number of quotes embedded in
later texts that escaped notice for a long time, or new
findings concerning the predominant mode of commu-
nication, oral or written, of a certain period. New dis-
coveries in such related fields as the history of science
or the history of literature also may lead to revisions
4 ,. M. E. Moravcsik
in the history of philosophy. In view of these four fac-
tors, one may hold that the history of philosophy is
an empirical enterprise with a corresponding claim of
objectivity,2 partly cumulative in its progress and part-
ly repetitive for reasons not necessarily linked to lack
of objectivity or the influence on the historian of con-
temporary philosophic positions.
At the same time these considerations begin to show
the complexity of the task of the historian of philos-
ophy, and they hint at the variety of skills required for
success. Both the complexity and the variety can be
better appreciated if one considers the differences
among the related but distinct questions that a his-
torian can ask about a given text. Examples of these
questions are: (i) What did Aristotle have in mind
when he said that ... ? (ii) What did Aristotle's state-
ment that ... mean to his contemporaries? (iii) To
what did Aristotle commit himself when he said
that ... ? (iv) How should one take Aristotle's state-
ment that ... in relation to the problems of our own
times?
What skills are required in order to answer these
questions? In the context of this brief introduction
only one thing needs special emphasis, namely that one
of the requisite skills is that of being a good philoso-
pher. How else, except by being a good philosopher
in his own right, can a historian construct a sufficient
variety of interesting possible interpretations adequate
to the original thought of the subject? Robert Lowell
once said that no translation of a classic in poetry is
adequate unless the translation of a fine poem yields
what is a fine poem in its own right. This point could
easily be extended to the interpretation of classic phi-
2Some of what Aristotle said is likely to be true or false a priori,

but when the historian claims that Aristotle expressed such and
such an a priori claim, the historian's claim is empirical.
Introduction 5
losophers. Such an extension, however, must meet the
possible objection that while poetry is, in a sense, time-
less, in an interpretation of a classic philosopher we
must distinguish between (a) the treatment of, e.g.
Aristotle, as an attempt to rediscover certain thoughts,
and (b) the attempt to use this treatment as part of
a general inquiry into expressions of timeless verities.
With regard to (a) the "Lowell criterion" need not
hold, for what is good philosophy in one period might,
when viewed from the vantage point of a later period,
appear no more than an "adolescent version" of a
promising view. Against this objection it can be said
that the sharp dichotomy implied by (a) and (b) has
already been undermined by our discussion of the in-
terrelatedness of questions (i) to (iv). With regard
to the claim that some of the earlier philosophic views
may seem "adolescent" to us, the following must be
said. Most of the philosophy of our civilization has
been organized around a set of loosely connected prob-
lems, most of which are not empirical in nature, at
least in their original formulation. At times the ac-
cumulation of new empirical evidence leads to a re-
formulation of some of these problems so that they
are seen either as not genuil}e, or as admitting of em-
pirical solutions. In other cases the advances in logic
or the availability of other tools of conceptual clarity
allow us to arrive at reformulations which have greater
clarity in the sense that these reformulations suggest
answers that would not have been conceivable under
the earlier formulations. There are still other problems,
however, which remain with us through the centuries
in spite' of continuous reformulations. Thus philoso-
phy is cumulative in some ways while in others it is
not; some of its problems are perennial while others
are not. Much of what Aristotle wrote is centered
around issues which are alive today, and thus Robert
6 1. M. E. Moravcsik
Lowell's criterion of adequacy is applicable to most
interpretations of Aristotle.
This part of the introduction sheds, one may hope,
some light on the nature of the history of philoso-
phy. Our last point also sheds some light on some of
the considerations that have guided the construction
of this volume. For one of the key considerations in
the selection of the material has been that the inter-
pretations should be not only fine pieces of scholar-
ship, but also fine pieces of philosophizing in their
own right.
Aristotle's life and its historical context will not be
conSIdered here; these matters are discussed adequately
in scores of other books. Nor will this introduction
summarize Aristotle's views; it may be gathered from
what has been said so far, and it will become even
clearer from what follows, that according to the edi-
tor's view Aristotle's philosophizing does not admit of
a summary. Instead, an attempt will be made to eluci-
date the following questions: (a) Why should one re-
gard Aristotle as a classic philosopher? (b) How does
the nature of Aristotle's philosophizing compare with
the ways in which more modern philosophers have
conducted their inquiries? (c) To what extent and in
what ways does Aristotle's philosophy have unity?
With regard to the first question, one should note
the variety of ways in which Aristotle's influence has
been significant. One of these is manifested in the
frequency with which his successors react, positively
or negatively, to his writings. Regardless whether the
topic is the rules for deductive inferences, or the prop-
er analysis of what it is to explain natural phenomena,
or the alleged primacy of material substances, in al-
most every age philosophers dealing with these topics
have conducted a dialogue with Aristotle or with an
Aristotelian view. Another manifestation of his influ-
Introduction 7
ence is that he has been associated, rightly or wrongly,
with a coherent view of reality which is supposed to
give philosophic foundations to Christian doctrines
about God, man, and nature. One may rightly ques-
tion the extent to which this view can be attributed
to Aristotle, but that he inspired it is undeniable. A
third, and perhaps most important, way in which Aris-
totle's influence has been crucial can be seen when we
realize the extent to which he set the problems of sub-
sequent philosophy. These problems set for philosophy
tasks that have yet to be completed. Such tasks are:
collecting in a systematic fashion all the rules of valid
inference, giving a proper analysis of the notion of re-
sponsibility, clarifying concepts basic to our under-
standing, such as the concepts of change, action, thing
(or- object), mass (or heap), and delineative proper
ways of classification. This manifold nature of Aris-
totle's influence explains why he is a figure to be dealt
with by the logician and semantic analyst as well as
by the Thomistic theologian, by the student of meta-
physics as well as the philosopher of science.
This partial answer to our first question also gives
us some conception of the breadth of Aristotle's phi-
losophy, and this-in turn-tells us something about
the nature of his philosophizing, the subject of our
second question. Aristotle is one of those rare phi-
losophers who wrote about an enormous range of
topics within the fields of logic, ethics, philoso-
phy of education, philosophy of science, philosophy
of art, philosophy of language, epistemology, and met-
aphysics. But it is a mistake to think that his philos-
ophizing adds up either to a system or to a world view
within which all the answers within all these fields are
organized and interrelated. Some of the topics with
which Aristotle deals, and his corresponding solutions,
are only loosely connected with one another, and he
repeatedly emphasizes that different types of topic re-
quire different types of approach. Still, neither are we
to think of his philosophizing as fragmentary, or as de-
signed only to catch glimpses here and there of the
nature of things. In his philosophizing Aristotle seems
to have kept in mind the ideal of the self-sufficient
man, an ideal that both Plato and he espoused in their
ethical writings; a self-sufficient man ought to be able
to deal with any problem that he encounters even
though he may not organize his answers into an axio-
matic pattern, and even though he may come to the
conclusion that different types of question require differ-
ent types of answer. Among his problems epistemolog-
ical questions did not occupy a central positiotJ., and
when he came to set for himself questions about knowl-
edge these often cut across the dichotomy of justifica-
tion and explanation which has dominated most of phi-
losophy since Descartes. But this is only a partial expla-
nation of why he cannot be classified among either
the empiricists or the rationalists. Another part of that
explanation would be that from his vantage point-as
from the vantage point of many philosophers today-
these positions seem neither mutually exclusive nor
exhaustive of the possibilities. We fare no better if
we try another dichotomy, made familiar by Professor
I. Berlin,S and applicable mainly to the contrast be-
tween the dreamers and ideologists of the nineteenth
century, and the careful piecemeal thinkers dominat-
ing professional philosophy in the twentieth century in
England and the United States. Aristotle could be de-
scribed neither as a hedgehog of the type familiar from
the post-Kantian philosophy of the nineteenth century,
nor as the intellectual ancestor of that variety of philo-
sophic thought which flourishes in the twentieth and is
SThe Hedgehog and the Fox. An Essay on Tolstoy's View of

History (London, 1953).
Introduction 9
commonly referred to as "analytic philosophy." Per-
haps this elusiveness of Aristotle's philosophy with re-
spect to subsequent classifications helps to account for
the fact that both hedgehogs and foxes are often in-
debted to him. If this brief discussion does not give
much of a positive picture of what Aristotle's philos-
ophy was like, at least it should suffice to make the im-
portant negative point, that in its nature it was very
different from the dominant ways of philosophizing of
the last four hundred years.
The foregoing also gives, in outline, the answer to
the question about the unity of Alistotle's thought.
Today most people tend to think of the great philos-
ophers of the past as system-builders who laid down
definitions and basic axioms, and then tried to account
for all the salient features of nature, man, and the rules
of conduct by means of these basic "truths" and the
theorems that can be derived from them. Or, perhaps
more loosely, people think that classical philosophers
tried to account for all the salient features of the world
and man's role in it in terms of an explanation that
blends these features into an interrelated organic whole.
It is very important to note that neither Aristotle, nor
his teacher Plato, was a system-builder of this type.
Though some of their concepts ranged over more than
one problem, there is no simple truth which either
Plato or Aristotle would have regarded as a key to
all their answers and suggestions. The only sense of
unity that applies to a discussion of Aristotle's views
is that of consistency, and even this should be re-
garded as a desideratum-both Aristotle's and the in-
terpreter's-rather than a fact to be taken for granted.
It could be said that Aristotle's metaphysical sugges-
tions were intended to underlie some of what he says
about the proper way to explain natural phenomena,
but it would be wrong to say that they underlie all
his philosophy. (It was reported to me that the late
Professor John L. Austin said on one occasion that
one of the refreshing features of Aristotle's ethics is
its almost total lack of connection with his meta-
physics.)
These reflexions suggest that the proper approach to
Aristotle's philosophy is a piecemeal, or pluralistic one.
What Aristotle says as an answer to one question may
not relate to his answers to other questions similar in
appearallce. Thus acceptance of some of Aristotle's
suggestions does not commit one to a wholesale ac-
ceptance of his philosophic claims. This remark ap-
plies both to content and to methodology. Further-
more, even if it did not have other support, this ap-
proach suggests itself as a heuristic principle, for it fa-
cilitates the emergence of dialogue between the phi-
losophers of the past and those of the present. The
conception of a classic philosopher as a system-builder
carries with it the implication that if you reject a part,
you must reject all; or conversely that partial accept-
ance must lead to wholesale acceptance. The approach
favored by the editor makes it possible for us to claim
with plausibility that one may understand some parts
of Aristotle's thought while finding others obscure, that
some of his suggestions are acceptable while others do
not seem to be, that some of his results seem naive
in the light of subsequent research, while in some oth-
er ways (e.g. in his discussion of the variety of types
of predication, his analysis of responsibility, and his
views on the nature of pleasure and enjoyment) his
suggestions are far more profound and far more fruit-
ful than those made by the vast majority of philoso-
phers in the twentieth century.
Right or wrong, these meditations on the nature of
the history of philosophy and on the nature of Aris-
totle's philosophy have influenced the construction of
Introduction 11
this volume. As we have already seen, one criterion of
inclusion was that each contribution should be a good
interpretation of Aristotle in the sense that it should
also be a good piece of philosophizing. Furthermore.
the authors were selected in such a way that there is
no common philosophical view or outlook shared by
all. An attempt was made to include papers on a va-
riety of topics, but only in this sense does this volume
try to convey much of Aristotle's thought; the essays
printed here cannot be used as a basis for reconstruct-
ing "the philosophy" of Aristotle. Finally. the volume
was put together in this manner in the hope that the
reader will come to discover the perennial nature of
some of the problems with which Aristotle wrestled.
But this hope need not be shared by the reader. Per-
haps it is only a sign of the editor's Platonistic tend-
encies.
Thanks and gratitude is due to all those who con-
sented to have their papers reprinted. Special thanks
is due to Miss Anscombe who made some corrections
in her essay for this volume, and to Mr. Michael
Woods and Mr. Urmson for being so kind as to un-
dertake the task of writing new papers. The editor also
wishes to thank Mrs. J. L. Austin for having con-
sented to the inclusion of a previously unpub-
lished paper by the late Professor Austin. By far
the greatest debt is owed to Mr. Urmson; for in ad-
dition to writing his fine paper, he helped to secure
permission to publish Professor Austin's manuscript.
and without his kind editorial help that manuscript
could not have been prepared for publication. Final-
ly, the cditor wishes to thank Mrs. Rorty for having
asked him to do this volume in her series.
I. LOGIC
ARISTOTLE AND THE SEA BATTLE
De Interpretatione, Chapter IX
G. E. M. ANSCOMBE
1 For what is and for what has come about, then,

it is necessary that affirmation, or negation, should
be true or false; and for universals universally
quantified it is always necessary that one should
be true, the other false; and for singulars too, as
has been said; while for universals not universally
quantified it is not necessary. These have bcen
discussed.
For what is and for wllat has come about: he has in
fact not mentioned these, except to say that a verb or
a tense-sc. other than the present, which he regards
as the verb par excellence-must be part of any prop-
osition.
it is necessary: given an antiphasis about the present
or past, the affirmative proposition must be true or
false; and similarly for the negative. An antiphasis is
a pair of propositions in which the same predicate is
in one affirmed, in the other denied, of the same sub-
ject. Note that Aristotle has not the idea of the nega-
tion of a proposition, with the negation sign outside
the whole proposition; that was (I believe) invented by
the Stoics.-What Aristotle says in this sentence is am-
biguous; that this is 'deliberate can be seen by the C011-
From Mind, LXV (1956), 1-15, Newly revised by the author.

Reprinted by permission of the author and the editor of Mind.
16 G. E. M. Anscombe
trast with the next sentence. The ambiguity between
necessarily having a truth-value, and having a neces-
sary truth-value-is first sustained, and then resolved
at the end of the chapter.
for universals universally quantified: he does not
mean, as this place by itself would suggest, that of "All
men are white" and "No men are white" one must be
true and the other false. But that if you take "All
men are white" and "No men are white" and con-
struct the antiphasis of which each is a side, namely,
"All men are white-Not all men are white" and "No
men are white-Some man is white," then one side of
each antiphasis must be true, and the other side must
be false.
for singulars too, as has been said: sc. of "Socrates is
white-Socrates is not white" one side is necessarily
true, the other necessarily false. (This is what a mod-
ern reader cannot take in; but see the "Elucidation.")
for universals not universally quantified: his example
rendered literally is "man is white-man is not white."
From his remarks I infer that these would be correct-
ly rendered "men are ...". For, he says, men are beau-
tiful, and they are also not beautiful, for they are ugly
too, and if they are ugly they are not beautiful. I be-
lieve that we (nowadays) are not interested in these
unquantified propositions.
These have been discussed: i.e. in the immediately
preceding chapters, by which my explanations can be
verified.
2 But for what is singular and future it isn't like
this. For if every affirmation and negation is true
or false, then it is also necessary for everything to
be the case or not be the case. So if one man says
something will be, and another says not, clearly
it is necessary for one of them to be speaking
truly, if every affirmation and negation is true or
3 false. For both will not hold at once on such con-
Aristotle and the Sea Battle 17
ditions. For if it is true to say that something is
white or is not white, its being white or not
white is necessary, and if it is white or not white,
it is true to sayar deny it. And if it is not the
case, then it is false, and if it is false, it is not the
case; so that it is necessary as regards either the
affirmation or the negation that it is true or false.
singular and future: sc. There will be a relevant dis-
cussion tonight; this experiment will result in the mix-
ture's turning green; you will be sent down before the
end of term.
it isn't like this: namely, that these propositions (or
their negations) must be true or false. Throughout this
paragraph the ambiguity is carefully preserved and con-
cealed.
it is also necessary for everything to be the case
or not be the case: the Greek "or" is, like the Eng-
lish, ambiguous between being exclusive and being non-
exclusive. Here it is exclusive, as will appear; hence
the "or" in the conditional "if every affirmation and
negation is true or false" is also exclusive, and to point
this he says "every affirmation and negation", not, as
in (I) "every affirmation or negation"; that "or" was
non-exclusive.
For both will not hold on such conditions: namely,
on the conditions that every affirmation is true or false.
This condition is not a universal one; it does not apply
to the unquantified propositions, though if the "or" is
non-exclusive it does. But if the conditions hold, then
just one of the two speakers must be speaking the
truth.
It is true to say or deny it: ~v is the common philo-
sophical imperfect.
4 So nothing is or comes about by chance or
'whichever happens'. Nor will it be or not be, but
everything of necessity and not 'whichever hap-
18 C. E. M. Anscombe
pens'. For either someone saying something or
someone denying it will be right. For it would
either be happening or not happening accord-
ingly. For whichever happens is not more thus or
not thus than it is going to be.
'whichever happens': the Greek phrase suggests both
"as it may be" and "as it turns out". "As the case
may be" would have been a good translation if it
could have stood as a subject of a sentence. The 'scare-
quotes' are mine; Aristotle is not overtly discussing the
expression "whichever happens".
is not more thus or not thus than it is going to be:
as the Greek for "or" and for "than" are the same, it
is so far as I know a matter of understanding the ar-
gument whether you translate as here, or (as is more
usual) e.g.: "isn't or (sc. and) isn't going to be rather
thus than not thus". But this does not make good
sense. Aristotle is arguing: "We say 'whichever hap-
pens' or 'as the case may be' about the present as well
as about the future; but you don't think the present
indeterminate, so why say the future is?" Or rather (as
he is not talking about the expression): "Whatever
happens will be just as determinately thus or not thus
as it is."
5 Further, if something is white now, it was true
earlier to say it was going to be white, so that it
was always true to say of any of the things that
have come about: "it is, or will be." But if it was
always true to say: "it is, or will be", then: im-
possible for that not to be or be going to be. But if
it is impossible for something not to come about,
then it is unable not to come about. But if some-
thing is unable not to come about it is necessary
6 for it to come about. Therefore it is necessary
that everything that is going to be should come
about. So nothing will be 'whichever happens' or
by chance. For if by chance, then not by necessity.
But it it is impossible tor something not to come
about, then it is unable not to come about: the reader
who works through to the end and understands the
solution will examine the dialectic to see where it
should be challenged. It will turn out that the point
is here, in spite of the equivalence of the two Greek
expressions. It is impossible for the thing not to come
about, i.e. necessary that it should come about by
necessitas consequentiae, which does not confer the
character of necessity necessitas consequentis, on what
does come about. A necessary consequence of what is
true need not be necessary.
Still, it is not open to us, either, to say that
neither is true, as: that it neither will be nor will
7 not be. For firstly, the affirmation being false the
negation will not be true, and this being false the
affirmation won't be true. -And besides, if it is
true to say that something is big and white, both
must hold. And if they are going to hold to-
morrow, they must hold tomorrow. And if some-
thing is neither going to be nor not going to be
tomorrow, 'whichever happens' won't be. Take a
sea-battle, for example: it would have to be the
case that a sea-battle neither came about nor
didn't come about tomorrow.
Still, it is not open to us, either, to say tllat neither
is true: And yet Aristotle is often supposed to have
adopted this as the solution.
For firstly: this goes against what he has shown at
the end of (3): "if it is false, it does not hold." So
much, however, is obvious, and so this is not a very
strong objection if we are willing to try whether
neither is true. What follows is stronger.
And if they are going to hold tomorrow: from here
to the end of the paragraph the argument is: if it is
the case that something will be, then it will be the
20 c. E. M. Anscombe
case that it is. In more detail: you say, or deny, two
things about the future. If what you say is true, then
when the time comes you must be able to say those
two things in the present or past tenses.
'whichever happens' won't be: i.e. 'whichever hap-
pens' won't happen.
8 These are the queer things about it. And there
more of the sort, if it is necessary that for every
affirmation and negation, whether for universals
universally quantified or for singulars, one of the
opposites should be true and one false, that there
is no 'whichever happens' about what comes
about, but that everything is and comes about of
necessity. So that there would be no need to
deliberate or take trouble, e.g.: "if we do this,
9 this wiII happen, if not, not." For there is noth-
ing to prevent its being said by one man and
denied by another ten thousand ye~rs ahead that
this will happen, so that whichever of the two was
then true to say will of necessity happen. And in-
deed it makes no difference either if people have
said the opposite things or not; for clearly this is
how things are, even if there isn't one man saying
something and another denying it; nor is it its
having been asserted or denied that makes it
going to be or not, nor its having been ten thou-
sand years ahead or at any time you like. So if in
10 the whole of time it held that the one was the
truth, then it was necessary that this came about,
and for everything that has been it always held,
so that it came about by necessity. For if anyone
has truly said that something will be, then it can't
not happen. And it was always true to say of
what comes about: it will be.
These are the queer things about it. And: I have di-
verged from the usual punctuation, which leads to the
rendering: "These and similar strange things result,
if...." This seems illogical.
E.g.: often rendered "since": "since if we do this,
this will happen, if not, not." This does not appear
to me to make good sense. The Oxford translator sits
on the fence here.
So it in the whole of time it held: one must beware
of supposing that Aristotle thinks the conclusion stated
in the apodosis of this sentence follows from the con-
dition. It only follows if the previous arguments are
sound. He is going to reject the conclusion, but there
is no reason to think that he rejects the condition: on
the contrary.
11 Now if this is impossible! For we see that things
that are going to be take their start from deliberat-
ing and from acting, and equally that there is in
general a possibility of being and not being in
things that are not always actual. In them, both
are open, both being and not being, and so also
both becoming and not becoming. And plenty of
12 things are obviously like this; for example, this
coat is capable of getting cut up, and it won't get
cut up but will wear out first. And equally it is
capable of not getting cut up, for its getting worn
out first would not have occurred if it had not
been capable of not getting cut up. So this
13 applies too to all other processes that are spoken
of in terms of this kind of possibility. So it is clear
that not everything is or comes about of necessity,
but with some things 'whichever happens', and
the affirmation is not true rather than the nega-
tion; and with other things one is true rather and
for the most part, but still it is open for either to
happen, and the other not.
take their start: literally: "there is a starting point
of things that are going to be." The word also means
"principle". A human being is a prime mover (in the
engineer's sense), but one that works by deliberating.
As if a calculating machine not merely worked, but
was, in part, precisely qua calculating, a prime mover.
But Aristotle's approach is not that of someone en-
quiring into human nature, but into causes of events
and observing that among them is this one.
acting: he means human action, which is defined in
terms of deliberation; see Nicomachean Ethics VI
1139: there he repeats the word "6:PX~": "~ ,[OlOlJlTl
6:PX~ CivSpulTIoc;": the cause of this sort is man. An
animal too or a plant, is a prime mover. Hence his
thought is not that there are new starting points con-
stantly coming into existence; that would not matter.
It is first of all the nature of deliberation that makes
him think that the fact of human action proves the
dialectic must be wrong. I cannot pursue this here;
though I should like to enter a warning against the
idea (which may present itself): "the nature of delibera-
tion presupposes freedom of the will as a condition."
That is not an Aristotelian idea.
things that are not always actual: things that are al-
ways actual are the sun, moon, planets and stars. Aris-
totle thought that what these do is necessary. The
general possibility that he speaks of is of course a con-
dition required if deliberation and 'action' are to be
possible. If what the typewriter is going to do is ncc-
essary, I cannot do anything else with the typewriter.
Not that this is Aristotle's ground for speaking of the
general possibility. That is shown in his consideration
about the coat: the assumption that the coat will be
worn out does not conflict with our knowledge that it
can be cut up. We know a vast number of possibilities
of this sort.
in terms of this kind of possibility: I take it that we
have here the starting point for the development of
Aristotle's notion of potentiality. The sentence con-
firms my view of the point where he would say the
dialectic went wrong.
witlI otlIer tlIings onc is true rather and for the most
part: as we should say: more probable.
14 The existence of what is when it is, and the non-
existence of what isn't when it isn't, is necessary.
But still, for everything that is to be is not neces-
sary, nor for everything that isn't not to be. For
it isn't the same: for everything that is to be of
necessity when it is, and: for it simply to be of
neccssity. And the same for what isn't. And the
same reasoning applies to the antiphasis. For it
is nccessary that everything should be or not, and
should be going to be or not. But it is not the
15 case, separately speaking, that either of the sides
is necessary. I mean, e.g. that it is necessary that
there will be a sea-battle tomorrow or not, but
that it is not necessary that there should be a sca-
battle tomorrow, nor that it should not happen.
But for it to come about or not is necessary. So
that since propositions are true as the facts go, it
is clear that whcre things are such as to allow of
'whichever happens' and of opposites, this must
hold for the antiphasis too.
TlIe existence of wlIat is wlIcn it is ... is necessary:
i.e. it cannot be otherwisc. A modern gloss, which Aris-
totle could not object to, and without which it is not
possible for a modern person to understand his argu-
ment, is: and cannot be shown to be otherwise. It will
by now have become very clear to a reader that the
implications of 'necessary' in this passage are not what
he is used to. But see the "Elucidation".
simply to be of necessity: there is a temptation to
recognise what we are used to under the title "logical
necessity" in this phrase. But I believe that Aristotle
thought the heavenly bodies and their movements
were necessary in this sense. On the othcr hand, he
seems to have ascribcd something like logical necessity
to them.
B
24 c. E. M. Anscombe
But it is not the case, separately speaking, that ei-
ther of the sides is necessary: the ambiguity of the
opening "it is necessary that an affirmation (or nega-
tion) should be true or false" is here resolved. And
we learn that when Aristotle said that, he meant that
if p is a statement about the present or the past, then
either p is necessary or not-p is necessary. But this
means that in order to ascribe necessity to certain prop-
ositions (the ones, namely, that are not 'simply' neces-
sary) we have to be informed about particular facts.
So, one may ask, what has this necessity got to do
with logic? -Aristotle, however, states no facts, past,
present, or future. (I do in what follows; 1 hope this
will not prove misleading: the purpose is only didactic.)
His results could perhaps be summarised as follows:
we use indices p and I to the propositional sign to in-
dicate present and past time references on the one
hand, and future time reference on the other. Then for
all p, p vel not-p is necessary (this covers the unquan-
tified propositions too) and pp is necessary vel not-pp
is necessary; but it is not the case that for all p, PI is
necessary vel not-P! is necessary.
This is how it is for what is not always existent
16 or not always non-existent. For such things it is
necessary that a side of the antiphasis should be
true or false, but not this one or that one, but
whichever happens; and that one should be true
rather than the other; but that does not mean
that it is true, or false. So it is clear that it is not
necessary for every affirmation and negation that
this one of the opposites should be true and that
one false; for it does not hold for what does not
exist but is capable of being or'not being; but it is
as we have said.
whichever happens: sc.: it is a matter of whichever
happens.
that one should be true rather than the other: d.
"rather and for the most part" above; note that this
is governed by "it is necessary"; I infer that Aristotle
thought that correct statements of probability were true
propositions.
but that does not mean: ~f>TJ, logical, not temporal;1
fJf>TJ works rather like the German "schon" (only here
of course it would be "noch nicht"). fJf>TJ in a non-
temporal sense is, like oUKETl, frequent in Greek litera·
ture. English translators of philosophical texts usually
either neglect to translate it or mistranslate it. For ex-
amples, see Theaetetus 201e4, Physics 187a36, De In-
terpretatione 16a8, Metaphysics 1006a16. Bonitz gives
some more examples.
AN ELUCIDATION OF THE FOREGOING

FROM A MODERN POINT OF VIEW
A. The Vice Chancellor will either be run over next

week or not. And therefore either he will be run
over next week or he will not. Please understand
that I was not repeating myself I
B. I think I understand what you were trying to do;
but I am afraid you were repeating yourself and,
what is more, you cannot fail to do so.
A. Can't fail to do so? Well, listen to this: The Vice
Chancellor is going to be run over next week ...
B. Then I am going to the police as soon as I can.
A. You will only be making a fool of yourself. It's
not true.
B. Then why did you say it?
A. I was merely trying to make a point: namely, that
I have succeeded in saying something true about
the future.
11 am indebted to Miss M. Hartley of Somerville College for
pointing this out to me.
B. What have you said about the future that is true?
A. I don't know: but this I do know, that I have said
something true; and I know that it was either
when I told you the Vice Chancellor would be
run over, or on the other hand when I said he
wouldn't.
B. I am sorry, but that is no more than to say that
Either he will or he won't be run over. Have you
given me any information about the future? Don't
tell me you have, with one of these two remarks,
for that is to tell me nothing, just because the
two remarks together cover all the possibilities.
If what you tell me is an Either/Or and it em-
braces all possibilities, you tell me nothing.
A. Can an Either/Or be true except by the truth of
one of its components? I seem to remember Quine
speaking of Aristotle's "fantasy", that "It is true
that either p or q" is not a sufficient condition
for "Either it is true that p or it is true that q."
Now I will put it like this: Aristotle seems to think
that the truth of a truth-functional expression is
independent of the truth values of the component
propositions.
B. But that is a howler! The "truth" of Either p or
not p is determined, as you know very well, by
its truth value's being T for all possible combi-
nations of the truth possibilities of its compo-
nents; that is why its "truth" gives no informa-
tion. Having set out the full truth-table and dis-
covered that for all possibilities you get T in the
final column, you need make no enquiry to af-
firm the truth of pv"" p-any enquiry would be
comic. If on the other hand you tell me pv"" q
(q being different from p) you do give me some
information, for certain truth-combinations are ex-
cluded. There is therefore the possibility of en-
quiring whether your information is correct. And
that I do by discovering which of the truth-pos-
sibilities is fulfilled; and if one of the combina-
tions of truth-possibilities which is a truth-condi-
tion for pv"" q is fulfilled, then I discover that
your information is correct. But to tell me "It
will rain, or it won't", is not to tell me of any
truth-possibility that it is-or, if you like, will be,
satisfied. Now will you actually tell me something
about the future?
A. VelY well. Either you are sitting in that chair or
it will not rain tomorrow.
B. I agree, that is true, because I am sitting in this
chair. But still I have been told nothing about the
future, because since I know I am sitting in this
chair I know what I have been told is true wheth-
er it rains tomorrow or not-i.e. for all truth pos-
sibilities of "It will rain tomorrow." But do you
mind repeating your information?
A. Either you are sitting in that chair or it will not
rain tomorrow.
B. (Having stood up.) I am glad to be told it will be
fine-but is it certain? Do you get it from the
meteorologists? I have heard that they are some-
times wrong.
A. But surely we are talking about truth, not certain-
ty or knowledge.
B. Yes, and I am asking whether your information
-which I agree is information this time-is true.
A. I can't tell you till some time tomorrow; perhaps
not till midnight. But whatever I tell you then
will have been so now-I mean if I tell you then
'True', that means not just that it will be true
then but that it was true now.
B. But I thought it was the great point against Aris-
totle that 'is true' was timeless.
A. Yes-well, what I mean is that if I tell you-as I
shall be able to-'True' tomorrow-I mean if I am
able to, of course-why, then it will have been, I
mean is now correct to say it is true.
B. I understand you. If it is going to rain tomorrow
it is true that it is going to rain tomorrow. I
should be enormously surprised if Aristotle were
to deny this.
A. But Aristotle says it isn't true that it is going to
rain tomorrowl
B. I did not read a single prediction in what Aristotle
said. He only implied that it didn't have to be
true that it will rain tomorrow, i.e. it doesn't have
to rain tomorrow.
A. What? Even if it is going to rain tomorrow?
B. Oh, of course, if it is going to rain tomorrow, then
it necessarily will rain tomorrow: (I? ::;) p) is neces-
sary. But is it going to?
A. I told you, I can't say, not for certain. But why
does that matter?
B. Can't you say anything for certain about tomor-
row?
A. I am going to Blackwell's tomorrow.
B. And that is certain?
A. Yes, I am absolutely determined to go. (Partly be-
cause of this argument: it is a point of honour
with me to go, now.)
B. Good. I fully believe you. At least, I believe you
as fully as I can. But do I-or you-know you
will go? Can nothing stop you?
A. Of course lots of things can stop me-anything
from a change of mind to death or some other
catastrophe.
B. Then you aren't necessarily going to Blackwell's?
A. Of course not.
B. Are you necessarily here now?
A. I don't understand you.
B. Could it turn out that this proposition that you,
NN., are in All Souls today, May 7, 1954, is un-
true? Or is this certain.
A. No, it is quite certain-My reason for saying so
is that if you cared to suggest any test, which
could turn out one way or the other, I can't see
any reason to trust the test if, situated as I am,
I have any doubt that I am here. I don't mean
I can't imagine doubting it; but I can't imagine
anything that would make it doubtful.
B. Then what is true about the present and the past
is necessarily true?
A. Haven't you passed from certainty to truth?
B. Do you mean to tell me that something can be
certain without being true?-And isn't what is true
about the present and the past quite necessary?
A. What does 'necessary' mean here, since it obvious-
ly doesn't mean that these are what we call nec-
essary propositions?
B. I mean that nothing whatever could make what
is certain untrue. Not: if it is true, it is necessary,
but: since it is certainly true it is necessary. Now
if you can show me that anything about the fu-
ture is so certain that nothing could falsify it,
then (perhaps) I shall agree that it is necessarily
true that that thing will happen.
A. Well: the sun will rise tomorrow.
B. That is so certain that nothing could falsify it?
A. Yes.
B. Not even: the sun's not rising tomorrow?
A. But this is absurd! When I say it is certain I am
here, am I saying it wouldn't falsify it for me not
to be here? But I am here, and the sun will rise
tomorrow.
B. Well, let me try again: Could anything that can
happen make it untrue that you are here? If not,
I go on to ask: Could anything that can happen
make it untrue that the sun rises tomorrow?
A. No.
B. If we continued in darkness, the appearance of
the night being continued for the rest of our lives,
all the same the sun will have risen; and so on?
A. But that can't happen.
B. Is that as certain as that you are here now?
A. I won't say. -But what does Aristotle mean when
he says that one part of the antiphasis is neces-
sarily true (or false) when it is the present or the
past that was in question? Right at the beginning,
when I said "The Vice Chancellor will either be
run over or not, therefore either he will be run
over or he will not" you said that I was repeating
myself and could not fail to be repeating myself.
And then you referred to the Truth-table-tautolog-
ical account of that proposition. But does not pre-
cisely the same point apply to what Aristotle says
about "Either p or not p" when p is a proposi-
tion about the present or the past?
B. You could have avoided repeating yourself if you
had said "The Vice Chancellor will either be run
over or not, therefore either it is necessary that he
should be run over or it is necessary that he should
not be run over." But as you would have been dis-
inclined to say that-seeing no possible meaning
for an ascription of necessity except what we are
used to call 'logical necessity'-you could not
avoid repeating yourself.
Thus Aristotle's point (as we should put it) is that
'Either p or not p' is always necessary, and this neces-
sity is what we are familiar with. But-and this is from
our point of view the right way to put it, for this
is a novelty to us-that when p describes a present or
past situation, then either p is necessarily true, or .-
p is necessarily true; and here 'necessarily true' has a
sense which is unfamiliar to us. In this sense I say it
is necessarily true that there was not-or necessarily
false that there was-a big civil war raging in England
from 1850 to 1870; necessarily true that there is a Uni-
versity in Oxford; and so 011. But 'necessarily true' is
not simply the same as 'true'; for while it may be true
that there will be rain tomorrow, it is not necessarily
true. As everyone would say: there may be or may not.
We also say this about things which we don't know
about the past and the present. The question presents
itself to us then in this form: does "may" express mere
ignorance on our part in both cases?
Suppose I say to someone: "In ten years' time you
will have a son; and when he is ten ycars old he will
be killed by a tyrant." Clearly this is something that
may be true and may not. But equally clearly there is
no way of finding out (unless indeed you say that wait-
ing and seeing is finding out; but it is not finding out
that it will happen, only that it does happen.)
Now if I really said this to someone, she would ei-
ther be awe-struck or think me dotty; and she would be
quitc right. For such a prediction is a prophecy. Now
suppose that what I say comes true. The whole set of
circumstances-the prophecy together with its fulfil-
ment-is a miracle; and one's theoretical attitude (if
one has one at all) to the supposition of such an oc-
currence ought to be exactly Lhe same as one's theo-
retical attitude to the supposition that one knew of
someone's rising from the dead and so on.
As Newman remarks, a miracle ought not to be a
silly trivial kind of thing-e.g. if my spoon gets up one
day and dances a jig on my plate, divides into several
pieces and then joins up again, it qualifies ill as a mira-
cle, though it qualifies perfectly well for philosophical
B'"
discussion of naturally impossible but imaginable oc-
currences. Similarly if one were discussing impossible
predictions one would take such an example as the
following: Every day 1 receive a letter from someone
giving an accurate account of my actions and experi-
ences from the time of posting to the time 1 received
the letter. And whatever 1 do (I do random, absurd
actions for example, to see if he will still have written
a true account) the letter records it. Now, since we are
dealing in what can be imagined and therefore can be
supposed to happen, we must settle whether this would
be knowledge of the future: its certainly would sure-
ly be a proof that what I did I did necessarily.
It is interesting to note thae Wittgenstein agrees
with Aristotle about this problem, in the Tractatus.
"The freedom of the will consists in the fact that fu-
ture actions cannot be known. The connexion of know-
ing and the known is that of logical necessity. 'A
knows that p' is senseless, if p is a tautology." We are
therefore presented with the logical necessariness of
what is known's being true, together with the logical
non-necessity of the kind of things that are known.
The "logical necessity" of which he speaks in the re-
mark on knowledge is thus not just truth-table neces-
sariness. It is the unfamiliar necessariness of which
Aristotle also speaks. "A knows that p" makes sense
only if p describes a fact about the past or present;
so it comes out in Wittgenstein, and in Aristotle: past
and present facts are necessary. (In more detail, by the
Tractatus account: if A knows p, for some q (q ~ p)
is a tautology, and q expresses a fact that A is 'ac-
quainted' with.)
Then this letter about my actions would not have
been knowledge even if what it said was always right.
However often and invariably it was verified, it would
still not be certain, because the facts could go against
Aristotle and the Sea Battlc 33
it. However Aristotle's considerations are not about
knowledge and certainty but about necessity and its
correlative possibility. They have probably been made
more difficult to understand because they are an in-
stance of his willingness to deny what it does not make
sense to assert-for the affirmation of the necessity of
p is equivalent to a denial of the possibility of not-po
The possibility in question relates only to the future,
hence by some current conceptions the negation of
such possibility also relates only to the future.
E.g. 'This plaster can be painted for the next eight
hours and not after that since by then it will have set
too hard.' Neither the affirmation nor the negation of
this sort of possibility can be constructed relative to the
past. 'It can be painted yesterday' demands emenda-
tion perhaps to 'It may have been painted yesterday,'
perhaps to 'It could have been painted yesterday.' (The
contingency of the past is that something was possi-
ble. not that it is possible. The openness of the future
is that something is possible. This is not the same as
saying that it will be possible, since the possibility may
be extinguished. It is a present state of affairs that the
plaster can be painted for the next few hours.) Now
Aristotle who readily uses 'No men are numbers' as
a premise would pass from the denial that possibility
of this sort holds in relation to the past (or present)
to asserting that the past and present are necessary.
ARISTOTLE'S DIFFERENT POSSIBILITIES
K. JAAKKO J. HINTIKKA
1. The interrelations of modal notions in Aristotle.

The results of our examination of the varieties of am-
biguity in Aristotle (see Inquiry, [1959], 137-51) can
be used to analyze his notion of possibility. This no-
tion is closely connected with the other modal notions,
notably with those of necessity and impossibility. Since
these notions are somewhat more perspicuous than that
of possibility, it is advisable to start from them.
Aristotle knew that the contradictory (negation) of
'it is necessary that p' is not 'it is necessary that not
p' but rather 'it is not necessary that p' (De Int. 12,
22a4 fl.). The last two phrases are not contradictories,
either, for they can very well be true together (De Int.
13, 22bl if.), the latter being wider in application than
the former. By parity of form, 'it is not necessary that
not p' is wider in application than 'it is necessary
that p'.
These relations are conveniently summed up in the
following diagram (which is not used by the Stagirite):
necessary that p not nccessary that p
(i)
,
_----~----~"~--------~A~--------~
....----------"I.....I.----~"r---I
'----------~y
not nccessary that not p nccc~~ary that not p
From Inquiry, III (1960), 18-28. Reprinted by permission of

the author and the editors of Inquiry.
Aristotle's Different Possibilities 3,
According to Aristotle, 'impossible' behaves like 'nec-
essary' (De Int. 12, 22a7 if.). \Ve can therefore illus-
trate it by means of a diagram similar to (i). In fact.
the diagram will be virtually the same as the one for
'necessary', for "the proposition 'it is impossible' is
equivalent, when used with a contrary subject. to the
proposition 'it is necessary'." (See De Int. 13, 22M.)
In other words, 'impossible that p' is equivalent to 'nec-
essary that not p', 'impossible that not p' equivalent
to 'necessary that P'. etc. We can therefore complete
the diagram (i) as follows:
necessary that p = not necessary that p =

impossible that not p not impossible that not r
r-----~----~,(r----------~-----------
(ii)
~--------~y~-----------)'~----~y~------
not necessary that not p = ncccssary that not r =
not impossible that p impossible that p
2. The two notions of possibility. The problem is to

fit the notion of possibility into the schema (ii). In
this respect, Aristotle was led by two incompatible im-
pulses. On one hand. he was naturally tempted to say
that 'possible' and 'impossible' are contra.dictories;
something is possible if and only if it is not impossi-
ble (See e.g. De Int. 13, 22aI6-18, 32-38.) Under
this view, we get the following diagram:
not possible that not p = possible that not p =

nccc;s,uy tha t p not lin possible that not p
\( ,
(iii)
~--------~y~-----------"~------v-----~
possible that p = not possible that p =
110timpossible that p impossible that p
36 K. Jaakko J. Hintikka
But this temptation is not the only one. In ordinary
discourse, saying that something is possible often serves
to indicate that it is not necessary. Aristotle catches
this implication. For him, "if a thing may be, it may
also not be" (De lnt. 13, 22b20; see also 22a14 ff.).
Essentially the same point is elaborated in the Topica
II 7, 112b1 ff. There Aristotle says that "if a neces-
sary event has been asserted to occur usually, clearly
the speaker has denied an attribute to be universal
which is universal and so has made a mistake."
Under this view, our diagram will have this look:
necessmy that p = possible that p = impossible that p =

impossible that not p possible that not p necessary that not p
___---"A~----,\r__--.A----.,\~
(iv)
v
not impossible that p
It is seen that (iii) and iv) differ in that in (iii)

the range of possibility comprises everything that is
necessary, while in (iv) possibility and necessity are
incompatible.
It appears from De lnterpretatione that Aristotle did
not immediately see that the assumptions underlying
(iii) and (iv) are incompatible. Not surprisingly, he
ran into difficulties which he discusses in a not entire-
ly clear way in De lnt. 13, 22b11-23a7 (although I
suspect that the confusion of the usual translations of
this passage is not altogether Aristotle's fault). He per-
ceives clearly enough that the gist of the difficulty lies
in the relation of possibility to necessity (De lnt. 13,
22b29 ff.). And at the end he is led to distinguish two
senses of 'possible' one of which satisfies (iii) and the
other (iv). (De lnt. 13, 23a7-27.) However, Aristotle
does not make any terminological distinction between
Aristotle's Different Possibilities 37
the two. Insofar as the distinction is vital, I shall call
the notion of possibility which satisfies (iii) 'possibility
proper' and the notion which satisfies (iv) 'contingency'.
3. Homonymy v. multiplicity of applications. Now
we can see why the distinction between a diversity of
applications and homonymy is vital for this essay. \Ve
have found a clear-cut case of homonymy: the notions
of contingency and of possibility proper have different
logical properties. They cannot be covered by a single
term 'possibility' except by keeping in mind that this
word has different meanings on different occasions.
Their relation is therefore one of homonymy (cf. sec-
tion 10 of the first paper).
But in addition to this duality of 'contingency' and
'possibility proper', there is a different kind of distinc-
tion. One of these two logically different notions, viz.
possibility proper, covers two kinds of cases. 'Vhen one
says that p is possible (in the sense of possibility prop-
er), one sometimes could also say that p is contingent
and sometimes that p is necessary. This does not mean,
of course, that the term 'possible' is ambiguous; it
merely means that its field of application falls into two
parts. It was for Aristotle therefore a typical case of
multiplicity of applications as distinguished from ho-
monymy (cf. section 10 of the first paper). The fol-
lowing diagram makes the situation clear:
necessary contingent impossible

~r,------~A'---__~,~
(v)
~------~y----------~
possible
The distinction between the different applications

of 'possibility proper' loomed large for Aristotle be-
cause he tended to emphasize the distinction between
necessity and contingency. Thus Aristotle argues in
Met. X 10 that the perishable and the imperishable
are different in kind (£looc; and yEVOC;). In an earlier
paper ("Necessity, Universality, and Time in Aristot-
le," Ajatus, XX [1957], 65-90), I have argued that the
distinction between contingency and necessity is for
Aristotle equivalent to that between what is perish-
able and imperishable. The field of application of 'pos-
sibility proper' therefore falls into two parts which are
different in kind. We have already seen in sections 10
and 12 of the first paper that Aristotle viewed situa-
tions of this kind with suspicion, although he grudg-
ingly admitted that no logical harm need result. This
is probably one of the reasons why Aristotle in Ana-
lytica Priora preferred the notion of contingency to
that of possibility proper.
4. Aristotle's definition ot contingency. According to
the results of our examination of the ambiguities of
Aristotelian ambiguity, we may expect that the Stagi-
rite usually refers to the distinction between possibil-
ity proper and contingency by means of oflwvuflla
and that he always refers to the distinction between
the different cases of possibility proper (i.e. between
necessity and contingency) by means of some other lo-
cution, e.g. OlX(;)C; "-EY£·Wl, KaTO: 560 Tp6trouc; AEY£lal
or 'ITOAAax(;)c; AEY£Tal. An examination of the text
will bear out this expectation as far as the first, the
second, and the fourth expressions are concerned. Sim-
ilarly, the third locution is, we shan find, used by Aris-
totle in distinguishing two kinds of cases of contin-
gency (see section 6 below).
In Aristotle's discussion of the notion of possibility,
the key passage is in An. PI. I 13, 32al8-21. It is re-
ferred to by Aristotle repeatedly as the definition of
possibility (e.g. An. PI. I 14, 33b24; 15, 33b28; 15, 34b
28; 17, 37a27). The 'definition' is clear enough (I shan
not discuss here why Aristotle thinks of it as a defi-
nition):
I use the tenus 'possibly' and 'the possible' of that
which is not necessary but, being assumed, results in
nothing impossible.
This is clearly the notion I have called contingency.
However, it is not the only variant of possibility, for
Aristotle continues:
'to yap aVUYKUlOV OflCilVUflCil<; EvoExw6m AEyoflEV
That is, to say of the necessary that it is possible is to
use the term 'possible' homonymously. This explana-
tion obviously serves to motivate the qualification
"which is not necessary" in Aristotle's definition. The
use 'Of the word OflCilvuflCil<; shows that he knows that
he is making a choice between two incompatible mean-
ings of EvoExw6m (to be possible). The second mean-
ing, under which even necessary things are called pos-
sible, is the notion of possibility which satisfies (iii)
and which I have called possibility proper.
\Ve have thus reached two important conclusions:
(a) the main notion of possibility employed by Aris-
totle in An. PI. is what I have caned contingency; (b)
Aristotle is aware of the existence of the other notion
(possibility proper) which is different from contingency
to the degree that the same term can be applied to
them only homonymously.
These results are confirmed by other passages. A
glance at (iv) shows that contingency is symmetrical
with respect to negation: p is contingent if and only
if not-p is also contingent. They are therefore convert-
ible to each other. Aristotle makes the same observa-
tion and applies it to syllogistic premises in An. Pro
I l3, 32a30 ff. This shows that his 'possibility as de-
fined' agrees with my 'contingency'. Essentially the
same point is made in An. Pro I 17, 37a22 ff.
Aristotle's awareness of the ambiguity of possibility
is also demonstrated by the development of his syl-
logistic. He frequently points out that the conclusion
of a certain syllogism is valid only if one does not un-
derstand possibility in the sense defined (i.e. in the
sense of contingency) but in a sense in which it is the
contradictory of impossibility (e.g. An. PI. I 15, 33b30-
33; 15, 34b27-32; 16, 35b33; 17, 36b33; 20, 39aI2).
5. An analysis of An. Pro I 13, 32a21-28. The fact
that Aristotle was aware of the different logical prop-
erties of contingency as distinguished from possibility
proper seems to me to be in agreement with what Aris-
totle writes immediately after. the passages I have
quoted (An. Pro I 13, 32a21-28). This passage has been
censured by the recent commentators in spite of the
fact that it occurs in all the MSS and is recognized
both by Alexander and by Philoponus (W. D. Ross,
Aristotle's Prior and Posterior Analytics· [Oxford, 1949],
p. 327). However, it seems to me that the passage can
be understood as it stands by making due allowance
for Aristotle's conspicuous conciseness. I shall offer a
paraphrase of the passage, enclosing explanatory addi-
tions as well as my Own comments in brackets. The
superscripts refer to further comments.
Aristotle has just explained his sense of EV!)EX6flEVOV
(possible) and distinguished it from the homonymous
notion of possibility proper. He goes on:
That this [= Aristotle's definition] is the meaning of
'possible' is obvious from the opposing affirmations
and denials. 1 For [in the other sense of 'possible'] 'it is
not possible to apply', 'it is impossible to apply' and 'it
is necessary not to apply' are either the same or imply
each other. 2 Consequently their contradictories3 'it is
possible to apply', 'it is not impossible to apply' and 'it
is not necessary not to apply' are the same or imply
each other. For either the affirmation or the negation
always applies. [This is not correct, however, for we
mean by possibility something more than the absence
of impossibility.4] That which is necessary will there-
fore not be possible, and that which is not necessary
[nor impossible5 ] will be possible.
Further comments: (1) This elliptic sentence poses
two questions:
(a) What are these affirmations and denials affirma-
tions and denials of?
(b) What kind of opposition is Aristotle here referring
to?
As regards (a), the sequel shows that Aristotle is not
dealing with affirmations and denials of possibility in
the sense (of contingency) just defined. It turns out
(cf. (2) below) that the affirmations and denials per-
tain to the other sense of possibility (possibility prop-
er). Since Aristotle is obviously trying to justify his
own definition, it may be concluded that he is here
starting a reductio ad absurdum argument.
As regards (b), a comparison with the occurrences of
aV"n KElllEvo<; later in the passage (cf. (3)) suggests that
this word-which is Aristotle's vaguest and most gen-
eral term for opposition of any kind-here refers to
contradictory opposition. The alternative would be to
understand the sentence as referring to the opposition
between the two kinds of possibility; this would suit my
interpretation quite as well as the other reading.
(2) This is exactly what we get by accepting the
other sense of Ev5EX6IlEVOV, i.e. by not excluding neces-
sity from the range of possibility: 'not possible that p'
will be equivalent to 'impossible that p' which is (cf.
diagram (ii)) tantamount to 'necessarily not p'.
(3) The following sentence shows that these contra-
dictories are the aV1:lKElflEva referred to here.
(4) It was already pointed out above (in section 2)
that Aristotle took this view. See especially the refer-
ence to the Topica, loco cit. It appears from thecpavEpov
42 K. Taakko f. Hintikka
in 32a20 that Aristotle thought of this point as being
perfectly obvious; so obvious, indeed, that he neglected
to make it explicit here.
(5) The second part of the last sentence seems
strange. The addition I have indicated is a most tempt-
ing way of making the passage correct. It is very likely,
however, that the passage is Aristotle's as it stands. He
knew that his notion of possibility (i.e. contingency)
is symmetrical with respect to negation in the sense
which best appears from diagram (iv). He may have
thought that this symmetry justifies the transition from
'what is necessary is not possible' to 'what is not neces-
sary is possible'. This leads to a reading of 'not neces-
sary' as an elliptic form of 'not necessary either way',
i.e. 'neither necessary nor impossible'. In the sequel,
we shall find more indications that this was Aristotle's
reading; see section 7 intra.
Here I shall only point out that my interpretation is
supported by what we find in De Intcrpretatione. If
it is true that 'not necessary' sometimes does duty for
'neither necessary nor impossible', it may be expected
that 'not necessarily not', i.e. 'not impossible', will
sometimes mean 'neither impossible nor necessary'.
When this is so, 'not impossible' will entail (in fact,
it will be equivalent to) 'not necessary'. And this is
exactly what we find in De Int. 13, 22bl4-16, where
Aristotle infers 'not necessary' from 'not impossible'.
This inference is very difficult to explain otherwise.
The inference is based on the sequence of implications
(equivalences?) set up by Aristotle in De Int. 13,
22a16 ff., where again !l~ 6:MVO:TOV EtV<Xl entails !l~
6:Va:yKO:lOV ElVO:l.
6. A subdivision ot contingency. Having made these
distinctions, Aristotle goes on to say (in An. Pro I 13,
32b5 ff.) that possibility has two applications (the ex-
pression he uses is K<XTO: Mo Tp61touc; MYET<Xl). On
One hand, it is used to describe what generally happens
but falls short of being necessary, on the other hand it
is used to describe the indeterminate, that which can
be 'thus or not thus' without the prevalence of either
alternative. Now the distinction plainly has nothing to
do with the difference between possibility proper and
contingency. Neither of the two uses distinguished by
Aristotle covers what happens necessarily. \Vhat we
have here is therefore a subdivision of contingency.
Aristotle's use of the expression KenO: Mo -rp61touC;
AEYE-ral suggests that he is not distinguishing two
meanings of EVf>EX6!lEVOV but rather two kinds of cases
to which it can be applied. This is verified by his re-
marks on the conversion of statements of contingency.
He says that in both cases the possible premise can be
converted into its opposite premise, i.e. 'p is contin-
gent' into 'not-p is contingent'. This is trivial in the
case of a p which is contingent because it is 'indeter-
minate'. But Aristotle also holds that the conversion
applies to contingency in the sense of that which 'gen-
erally happens'. This may seem mistaken; if p happens
generally but not necessarily, we certainly cannot infer
that not-p happens generally. \Vhat Aristotle means
is that even in this case not-p is neither necessary nor
impossible and hence contingent in the sense of his
definition. If 'what happens generally but not neces-
sarily' were one of several meanings of 'contingent',
Aristotle would not be able to say that 'contingent'
always converts with its opposite. What he means is
that in each of the different cases that fall under the
term 'contingent' we have a conversion to the opposite
of some case-not necessarily of the same case-
covered by the term.1 Hence, he is not dealing with
1Aristotle's awareness of the fact that a case of a concept may

be converted into another case of the same concept is also shown
by his remarks on TO EiKO<; in Rhet. II 23, 1402a9 If.
44 K. 'aakko f. Hintikka
different meanings of EvBEX6f.lEVOV, but only with dif-
ferent applications of the term. 'Contingent' is not
homonymous although it covers different kinds of
cases.
7. An analysis ot An. Pro I 3, 25a37-25b19. Some
of the passages I have just discussed are referred to by
Aristotle earlier in Analytica PriOla in connection with
the conversion of problematic (possible) premises
(An. Pr. I 3, 25a37-25b19). We are now in a position
to understand the context of these references.
In An. Pr. I 3, 25bl8 Aristotle refers to his later dis-
cussions of the conversion of problematic premises. All
the remarks on this subject later in Analytica PriOla
pertain to contingent premises. This suggests that the
notion Aristotle has in mind in 25bl8 is his 'possibility
as defined' or contingency. This is confirmed by the
way Aristotle explains the notion of possibility which
he is here dealing with: "But if anything is said to be
possible because it is the general rule and natural ..."
(An. Pr. I 3, 25b14 ff.). This recalls one of the different
cases of contingency discussed in An. Pr. I 13, 32b5 if.
(vide supra). And when Aristotle says that this "is the
strict sense we assign to possible" (Ross's translation),
he is obviously anticipating his definition of contingen-
cy, in 32b18-20.
I conclude, therefore, that in An. Pro I 3, 25bl4-18
Aristotle is thinking of contingency rather than pos-
sibility proper. Now the passage which was just quoted
shows that this variant of possibility is contrasted to
the one employed in the immediately preceding passage
(An. Pro I 3, 25b8--13). We may therefore expect that
this latter notion of possibility is what I have called
possibility proper. Aristotle's examples show that this
is in fact the case. In one of his examples the term
'man' necessarily does not apply to any horse, while in
another the term 'white' does not necessarily apply to
any coat. This shows that the meaning of possibility
which is used here covers cases of necessity as well as
cases of contingency. (The examples are both negative
in form because he is discussing the conversion of neg-
ative premises.)
Although the testimony of Aristotle's examples thus
unambiguously shows that in 25b8-13 he is discussing
possibility proper, one may still be puzzled by his own
explanation of the variant of possibility he is using:
"Whatever is said to be possible because it is necessary
or because it is not necessary admits of conversion like
other negative statements ...." For what one would
expect here is 'neither necessary nor impossible' instead
of 'not necessary'. Some commentators have tried to
emend the passage by in'serting the negative particle
!IT] so as to make it read 'not necessarily not', although
there is no real support for such an insertion in the
MSS (see Ross, op. cit.). Moreover, this insertion has
the disadvantage of making the clause 'because it is
necessary' superfluous. In any case, the emendation is
quite unnecessary, for we have already found independ-
ent reasons for suspecting that Aristotle sometimes uses
'not necessary' (1:0!I~ o:vayKOlOV) and, by analogy,
'not necessarily' (!I~ tf, av6:YKY}C;) as elliptic expressions
for 'neither necessary nor impossible' and 'neither neces-
sarily nor impossibly', respectively (see section 5, com-
ment (5) supra). This suspicion is now confirmed by
the fact that the same explanation works here: on my
reading the quoted passage says just what one is en-
titled to expect.
Here one may ask whether my reading is contradicted
by the fact that in his second example Aristotle says
that it is not necessary that 'white' applies to any coat
(1:0 ot OOK avayKY} lm6:pXElv). If Aristotle were con-
sistently using the elliptic mode of expression, should
he not use double negative OOK d:v6:YKY} !l~ lm6:pXElv.
since he is here dealing with negative premises? To
this it may be answered that 'neither necessary nor im-
possible' is symmetrical with respect to negation, so
that no extra f.L~ is needed even if the elliptic mode of
expression is used. Besides, one of Bekker's MSS as
well as Philoponus do have the missing f.L~ (see Ross,
op. cit.), so that Aristotle may very well have been
even pedantically consistent in his usage.
We may conclude that in his treatment of the con-
version of negative problematic premises Aristotle first
discusses premises in which the notion of 'possibility
proper' is used and then those in which the notion of
contingency is used. In contrast, both these notions are
lumped together in Aristotle's discussion of the con-
version of positive problematic premises (25a37 - 25b2).
He indicates this as follows: ." ETIElO~ TIOAAaxWC;
MYE'Wl '[0 EVOEXEo8m (Kat yap '[0 avaYKalOV Kat '[0
f.Ll'J &VCXYKalOV Kat '[0 OUVa1:0V EVOEXE08al MYOf.LEV) .
. . . Here the words TIOAAaxWC; MYE'[<Xl suggest that he
is not exclusively concerned with the different mean-
ings of EVOEX0f.LEVOV. In fact, it has been pointed out by
Ross that the three cases listed in the parenthetical
clause cannot possibly be as many different meanings of
EVOEX0f.LEVOV. However, it seems to me that it cannot
be said, either, that they are just three different cases
to which the notion of possibility can be applied, The
first two are clear; we have encountered '[0 &vaYKai:ov
and '[0 f.Ll'J &vaYKalov before as the two cases covered
by the notion of possibility proper. The recurrence for
the fourth time of the elliptic expression '[0 f.Ll'J &vay-
KalOV (or of one of its variants) where one expects
'neither necessary nor impossible' gives further support
to my interpretation of this phrase. But '[0 ouva'[ov
cannot very well be a third case to which the notion
of possibility is applied, for there is no third case com-
parable with the two already listed. Rather, we must
understand .0 avayKa'lov Kat .0 ,"Ul avayKalov as re-
ferring to the notion of possibility proper, and under-
stand .6 ouya.6v as referring to the other notion of
possibility, viz. contingency. This, in fact, seems to be
the way Ross understands the passage. Its meaning may
hence be expressed somewhat as follows: " ... seeing
that possibility has many applications (for we caU pos-
sible both that which is necessary or is not necessary
either way and that which is capable of being) ...."
This interpretation is supported by the fact that the
context shows that Aristotle is here treating both the
variants of possibility at the same time. If they are here
mentioned in the same order in which they are sub-
sequently treated (in the connection of the conversion
of negative problematic premises) we can scarcely sep-
arate Aristotle's references to the two variants in any
way different from the one just suggested.
Our interpretation also agrees with the way ouvm6v
is used elsewhere in Analytica Priora. The most im-
portant passage in which this term occurs is An. Pro
I 15, 34a6 if. And it is indicated by Aristotle (in 34a14)
that the arguments he there gives pertain to possibility
with respect to generation. Now this variant of possi-
bility is very likely just our contingency. For some-
thing which is generated will sometimes be (viz. after
having been generated) and sometimes not be (viz.
before it is generated). It is therefore possible in the
very sense (in that of contingency) which we wanted
to give ouva.6v in An. Pro I 3, 25a39.
8. Remarks on An. Pro I 13, 32b25-32. What we
have found in this paper and its predecessor confirms
my earlier analysis of An. Pro I 13, 32b25-32 (in
"Necessity, Universality, and Time in Aristotle," pp.
86-88). Here I shall only briefly outline the argu-
ment, adding such new evidence as was not mentioned
in the earlier paper.
In the passage under discussion Aristotle seems to
be saying that
(P) it is possible for A to apply to all B
is ambiguous in that it may mean either
(PI) it is possible for A to apply to everything
to which B in fact applies
or
(P2 ) it is possible for A to apply to everything
to which B possibly applies.
This cannot be his meaning, however. For one thing,
he never seems to use (PI) but only (P2 ) in his sub-
sequent discussion of syllogisms from possible prem-
ises. He seems even to say that (P2 ) is what (P) was
defined to mean (An. Pro I 14, 33a24-25). For another,
the term Aristotle uses is EllXWC;, which strongly sug-
gests that he is not at all distinguishing two different
meanings of (P). Rather, he is saying that (P) covers
two kinds of cases, i.e. that (P) is tantamount to the
conjunction of (PI) and (P2). This suffices to explain
everything that Aristotle says and does. It may be ex-
pected that the variant of possibility Aristotle is using
in (P2 ) is the one he usually employs, viz. contingency.
Expressed as explicitly as possible, Aristotle's point
therefore is that (P) is equivalent to the conjunction
of (PI) and (P21), where the latter is
(P21) it is possible for A to apply to everything
to which B applies contingently.
Now this conjunction is clearly equivalent to what we
get by assuming that the variant of possibility used in
(P2) is my 'possibility proper':
(P22) it is possible for A to apply to everything
to which B possibly applies (in the sense of
'possibility proper'), i.e. to everything to which
B applies necessarily or contingently.
This explains why Aristotle seems to deal exclusively
with (P z) in his syllogistic theory; for what he is really
dealing with is (P 22) which is equivalent to the con-
junction of (PI) and (P 21 ) and therefore also to (P).
Further evidence is perhaps found in An. Pro I 29,
45b31-32. Having just explained how the different
kinds of assertoric syllogisms are established, Aristotle
goes on to say that apodeictic (necessary) and problem-
atic (possible) syllogisms are established in the same
way. But he adds a warning:
In the case of problematic propositions, however, we
must include those terms which, although they do not
apply, might possibly do so; for it has been shown that
the problematic syllogism is effected by means of
these .... (H. Tredennick's translation in the Loeb
Library edition)
Prima facie this is completely tautologous. For prob-
lematic syllogisms contain by definition terms which
do not apply but may apply. What can Aristotle mean
here? It is clear that the predicate term A of a prem-
ise like (P) may apply possibly but not actually. But
it is not equally obvious whether the subject term B
is to be taken to apply possibly or actually; whether,
in other words, (P) is to be understood as being equiv-
alent to (PI) or to (P 22 ). Unless we assume that Aris-
totle's statement is pointless, we can scarcely interpret
it except as a repetition of the point which we found
him making in An. Pro I 13, 32b25-32, viz. as identify-
ing (P) and (P 22 ). Notice in particular that there is
no semblance here of a distinction between two mean-
ings.
9. Concluding remarks. \Ve have discussed the most
important passages of An. Pro I which turn on the dis-
tinction between the various notions of possibility
used by Aristotle. Insofar as we have been successful
in applying the results of my earlier analysis of the
ambiguities of ambiguity in Aristotle, our success con-
versely serves as a further confirmation of the earlier
analysis. In particular, it supports what was said in
section 10 of the first paper. Aristotle's own defini-
tion of contingency (see supra, section 4) establishes
a connection between contingency and possibility prop-
er: contingent is that which is (properly) possible but
not necessary. If homonymy were tantamount to the
absence of any common element in definition, contin-
gency and possibility proper would not be homonyms.
The fact that Aristotle calls them homonyms shows
that there is more to his notion of homonymy than
that.
ON ARISTOTLE'S SQUARE
OF OPPOSITION
M.THOMPSON
Arguments have recently been offered that purport-

edly save Aristotle's square of opposition from the
charges which modern logicians have brought against
iU The main argument in this defense, as I under-
stood it, is the contention that the principle of ex-
cluded middle, or the principle that "every meaningful
statement is either true or false," is subject to certain
qualifications. These qualifications have nothing to
do with multivalued logics or the adoption of some
new formal principle, but conccrn an alleged "feature
of ordinary speech." In brief, this feature is that in or-
der for an empirical descriptive sentence to be mean-
ingful it is not necessary for it to yield a true or false
statement on every occasion of its utterance. Thus, the
sentence "All Smith's children are girls" is meaning-
ful because it would yield a true and not also a false
statement if uttered on an occasion when Smith had
From The Philosophical Review, LXII (1953), 251-65. Re-
printed by permission of the author and the editor of The Philo-
sophical Review.
lThe most lengthy statement of this defense of Aristotle is
given by H. L. A. Hart, "A Logician's Fairy Tale," The Philo-
sophical Review, LX (1951), 198-212. Hart says he owes the
"substance" of his criticism of "modern logical doctrine" to P. F.
Strawson, and refers to the latter's article, "On Referring," Mind,
LlX (1950), especially pp. 343-44. For a shorter statement of
a similar defense of Aristotle, see P. T. Geach, "Subject and
Predicate," Mind, LIX (1950), esp. 480.
52 M. Thompson
children all of whom were girls. It would likewise yield
a false and not also a true statement if uttered on an
occasion when Smith had at least one child who was
a boy. But if uttered on an occasion when Smith had
no children at all it would yield neither a true nor a
false statement-"the question of its truth or falsity
simply would not arise."
Bearing in mind this feature of ordinary speech, we
are to interpret Aristotle's square of opposition as ap-
plicable to empirical descriptive sentences only on
those occasions when their utterance would yield a
statement that it is either true or false but not both.
Thus, when an Aristotelian logician affirms that the
logical relation of "All ogres are wicked'" to "Some
ogres are wicked" is one of super-implication, he is not
also affirming that some ogres exist. He is merely af-
firming that the two sentences are so related that on
any occasion when an utterance of the first yields a
true statement, an utterance of the second must also
yield a true statement. The fact that we believe there
could be no such occasion is beside the point. The
sentences are meaningful because we can imagine what
would have to obtain on such an occasion. An Aris-
totelian logician is then no more guilty of affirming
ogres into existence than is the teller of a fairy tale
who makes statements about ogres. For the logician
like the story teller speaks only of what would be the
case if there were ogres.
The modern interpretation which permits us to say
that "All ogres are wicked" is true though "Some ogres
are wicked" is false is thus not only unnecessary in
order to avoid the paradoxes of existential import, but
it also leads to a misunderstanding of ordinary speech.
For it assumes that every meaningful empirical descrip-
tive sentence yields a true or false statement on every
occasion of its utterance, and overlooks those occasions
On Aristotle's Square of Opposition 53
when common sense would say that the question of
its truth or falsity simply does not arise. If we inter-
rupted the story teller with the question, "But is it
true that all ogres are wicked?" this would be a clear
indication that we had misunderstood his use of lan-
guage. Similarly, if we argued about the truth or fal-
sity of "All Smith's children are girls" when we knew
that Smith had no children, we could hardly be said
to have understood the conditions under which this
statement would ordinarily be taken as true or false.
Now I want first of all to argue that these proposed
qualifications of the principle of excluded middle are
of dubious value since they make any application of
the supposedly simple and elementary logical relations
in the square of opposition extremely complex.
I
Classical logicians from Aristotle on have taken the
relations in the square of opposition as connecting en·
tities which must be either true or false, even though
these entities have been variously identified as linguis-
tic expressions, as propositions, and as judgments.
While the relations themselves may on occasion have
been defined without reference to truth or falsity, such
a mode of definition does not mean that sometimes
the relations may hold when neither truth nor falsity
is applicable. The prerogative of the logician extends
only to the point of not having to specify in a given
case which one obtains, truth or falsity; it does not
extend to the point of casting aside the question of
truth and falsity altogether. Yet the latter appears to
be what the new defenders of Aristotle are proposing.
For according to the new defense, we are to say that
"All ogres are wicked" and "Some ogres are not
wicked" are related as contradictories, even though
neither truth nor falsity is applicable to either of them.
54 M. Thompson
Any assertion here of the commonplace that one of a
pair of contradictories must be true and the other false
will take the form of a contrary-to-fact conditional. If
there were ogres, i.e. if the question of the truth or fal-
sity of these statements were to arise, then one would
have to be true and the other false.
Thus, what has traditionally been a simple and
straight-forward specification of relations of truth con-
ditions has now been made to involve the complicated
logical distinctions required to cope with contrary-to-
fact conditionals. Even if we succeed in defining the
relations in the square without reference to truth or
falsity, we must resort to contrary-to-fact conditionals
in order to specify how any of the relations are appli-
cable to all ordinary statements in a way that logical
relations have traditionally been supposed applicable.
The decision of modern logicians to regard the truth
conditions of a universal affirmative statement as sat-
isfied when nothing of the kind named exists enables
us to preserve the traditional simplicity in the appli-
cation of the relation of contradictoriness. Even though
this decision does away with the applicability of con-
trariety, sub-contrariety, and super- and sub-implica-
tion, it seems preferable to a decision which makes
the application of any of the relations exceedingly
complex. The traditional square can be retained as a
valid schema relating by contradictoriness four of the
most frequent types of quantified statements. The re-
maining relations indicated by the square can be de-
fined by other schemata in accordance with which
they are always applicable, and the traditional simplic-
ity of application for any of the relations can be re-
tained.
Even if we agree with the new defenders of Aristotle
that the decision which leads to the modern analysis
is repugnant to ordinary speech, we can still argue that
this is more desirable than a decision repugnant to log-
ical analysis itself. For surely a decision which makes
any specification of the conditions under which a sim-
ple logical relation is universally applicable take the
form of a contrary-to-fact conditional sins against our
conception of what logical analysis should achieve. It
seems to me fairly easy to show that the sin is against
Aristotle's as well as against our modern conception.
I propose in the remainder of this paper to argue
(a) that the account of the square of opposition given
in Aristotle's On Interpretation is fundamentally dif-
ferent from what is usually taken as the traditional
analysis of the square,2 and (b) that Aristotle's account
is free from the contradictions and absurdities which
arise from the latter, this freedom being achieved, not
by qualifying the principle of excluded middle, but by
Aristotle's own peculiar restrictions on the existential
import of statements. For reasons that should soon be-
come apparent, I shall begin with some remarks about
singular statements.
II
The applicability of truth or falsity to statements
when the subject named does not exist is explicitly de-
clared by Aristotle to hold for singular statements.3
2By "traditional analysis" I mean simply the sort that is pre-
sented today in most logic texts as the Aristotelian one. As will
be made clear in the sequel, the essential features of this analysis
are (1) the usual doctrine of immediate inferences, including
obversion, contraposition, and inversion; (2) all the logical rela·
tions purportedly expressed in Aristotle's square of opposition.
Vvhcn I speak of contradictions and absurdities in this analysis,
I do not mean to deny that some classical logicians may have
intended some special assumptions not in Aristotle (such as the
proposed qualifications of the excluded middle) which keep the
analysis consistent. I am concerned only with the analysis as it
stands without such assumptions.
SCat. 13b26-35. All quotations from Aristotle, unless otherwise
specified, are from the Oxford translation.
c
56 M. Thompson
But in the case of affirmation and negation, whether
the subject exists or not, one is always false and the
other true. For manifestly, if Socrates exists, one of the
two propositions 'Socrates is iII: 'Socrates is not ill: is
true, and the other false. This is likewise the case if he
does not exist; for if he does not exist, to say that he is
ill is false, to say that he is not ill is true.
This clearly suggests that any affirmation about a sin-
gular subject implies that the subject exists, while a
negative statement about such a subject cannot have
this implication. Yet in another passage Aristotle seems
to contradict what he says here by arguing that "Homer
is a poet" does not imply "Homer is." "Take the prop-
osition 'Homer is so-and-so,' say 'a poet'; does it fol-
low that Homer is, or does it not? The verb 'is' is
here used of Homer only incidentally, the proposition
being that Homer is a poet, not that he is, in the in-
dependent sense of the word."4
The crucial point in understanding this second pas-
sage is the meaning of "is, in the independent sense
of the word," i.e. of "is per se." Clearly this is "is" in
the substantive sense of the word and means the same
as "is a substance." What Aristotle is denying in this
passage, then, is that "Homer is a substance" follows
from "Homer is a poet." Now in the first passage
quoted above the word "exists" is used to translate
ontos, which of course literally means just being and
not necessarily being a substance. Hence we should
take "exists" as meaning the same as "being" without
the specification of the sense of being, as being a sub-
stance or being an accident, and then there is no con-
tradiction. "Homer exists" does follow from "Homer
is a poet," since if Homer did not exist, i.e. if he were
simply non being, it would not be true to say that he
is anything. This point is suggested by Aristotle's re-
40 n Interpretation 21a26-28.
marks almost immediately following the second pas-
sage quoted above. "But in the case of that which is
not, it is not true to say that because it is the object
of opinion, it is; for the opinion held about it is that
it is not, not that it is." In other words, the fact that a
non-existent Homer may be the object of opinion, as
he would be if we were to construct a myth about him.
does not mean that it is true to say Homer is some-
thing. 5 The assertions "Homer is merely the object of
opinion" and "Homer is a mythical being" are about
Homer only in the sense of denying that he is in fact
anything. While we might say that "HomeI is a poet"
is true in fiction. what is true in this case is true of
the myth and not of Homer. And the myth does exist.
even though not per se as a substance does. 6
An affirmation about a singular subject, then, is false
and its contradictory is true when either of the fol-
lowing conditions obtains.
Cl The subject exists. either per se as a substance
or as something dependent on a substance (as an
accident of a substance), but does not possess the
predicate affirmed of it (e.g. when Socrates exists
and is healthy. "Socrates is ill" is false. and when
the color of this table exists but is not dull. "The
color of this table is dull" is false).
liThe Oxford translation of the passage just quoted above
(21a32-33) obscures the fact that the question is whether non·
being is something because it is the object of opinion. The Greek
just before the semicolon in this translation reads on ti. The
Loeb translation gives "it is not true to say that it 'is' somewhat,
because it is matter of opinion."
6'fhe myth may be said to exist as an artificial substance, de-
pendent on its maker. With this interpretation of the text, we
must of course deny that Aristotle would allow a realm of ficti-
tious things as a realm of being. This point admittedly calls for
further consideration, but it will be passed over here as simply
one of the assumptions required in order to make sense out of
Aristotle's remarks that are relevant to his account of the square
of opposition.
58 M. Thompson
C2 This subject does not exist, i.e. is neither a sub-
stance nor something dependent on a substance.
There is thus no need to qualify the principle that
every meaningful statement is either true or false, as
this principle applies to statements about singulars.
There is one further point to be considered before
we turn to the square of opposition. \Vhat about the
relation of "Socrates is not-ill" to "Socrates is not ill"?
As Aristotle analyzed this relation it is an implication
with the first statement as antecedent; it is not an
equivalence. "Socrates is not-ill" counts as an affirma-
tion, although one of a very peculiar sort. For the pred-
icate "is not-ill" is not strictly what Aristotle calls a
verb, and after complaining that there is no specified
name for such a predicate, he proposes to call it an
indefinite (or infinite) verb, since it applies "equally
well to that which exists and to that which does not."7
Thus, we might say "Socrates is not-ill" when we wish
to affirm that Socrates is a not-ill man, though we
might also use this same affirmation when we want
merely to deny that Socrates is ill. When C-2 obtains,
our assertion is false in the first case but not in the
second. 8 We, of course, could overcome this indefinite-
ness by adopting the convention that in logic a sin-
gular statement with an indefinite verb as predicate
will always be taken in one sense and not the other.
But this would be to retreat from the problem rather
than solve it. Statements of the sort in question are
716b12-16.
8Aristotle remarks in 20a25-27 that if the negative answer to
"Is Socrates wise?" is true, then an inference affirming that Soc-
rates is unwise (or not-wise) is correct. In the light of 13b26-35
(quoted above) we must interpret this remark as assuming that
the question would not be askcd if Socrates were nonexistent.
The purpose of such an assumption at this point is to show a
difference between universal and singular statements. Even though
men exist, a negative answer to "Is every man wise?" does not
allow the inference that every man is unwise.
logically indefinite and the logician must accept them
as such and make what he can of them. Aristotle
pointed out that if "Socrates is not-ill" is true, then,
whether this is equivalent to affirming that Socrates
is a not-ill man or merely to the denial that he is ill,
"Socrates is not ill" must be true. However, because
of the indefiniteness of the antecedent in this impli-
cation the relation does not hold the other way around.
When C-2 obtains, the antecedent is false in its strict-
ly affirmative sense while the consequent remains true,
and the relation thus cannot be an equivalence.9
The recognition of statements with indefinite terms
as peculiar entities for the logician to cope with rather
than to obliterate by convention is essential to Aris-
totle's account of the square of opposition.
Aristotle does not say explicitly that a universal af-
firmative or A statement is false when nothing of the
kind named exists, but what he does say about the log-
ical relations of quantified statements seems to me to
make the best sense when we take this interpretation
of the universal affirmative. Let us see, then, how far
we can go toward making sense out of Aristotle's ac-
count when we begin with the seemingly rash assump-
tion that an A statement is false when either of the
following conditions obtains.
C'-1 At least one thing of the kind named exists,
i.e. is either a substance or something depend-
ent on a substance, but does not possess the
predicate affirmed of it.
C'-2 Nothing of the kind named exists, i.e. is either
a substance or something dependent on a sub-
stance.
We must first make sense out of saying that the
corresponding 0 statement, the contradictory of A, is
9This point is made in the An. Pr. 51b36-52a17.
60 M. Thompson
true when C'-2 obtains. Now it is clear that Aristotle
regarded 0 as denying precisely what A affirmed, and
that he took these statements as fonns of simple affir-
mation and denial, respectively. Yet the usual render-
ing of 0 as "Some S exists and is not P" certainly
does not express what Aristotle meant by a simple
denial. We have instead a compound statement in
which one thing is affirmed and another denied. A lit-
eral translation of his examples of 0 gives "Not every
S is P," which is to be taken as simply the denial that
P is truly affirmed universally of S. In this simple de-
nial we do not affirm anything, and hence if C'-2 ob-
tains, our denial is true because if there are. no S's, it
is not true to affirm P universally of them, any more
than it is true to affirm illness of Socrates when Soc-
rates does not exist. We must next explain how the
corresponding E statement, the contrary of A, is like-
wise true when C'-2 is the case. For if E is false its
contradictory I must be true and we would have the
absurdity that "Some 8 is P" is true when there are
no 8's. But E and 0 do not differ in being more or
less simple, or more or less a denial. They are both
simple denials, though E denies that P can be partic-
ularly affirmed (affirmed in at least one instance) of
S while 0 denies that P can be universally affimled.
But clearly, then, if C'-2 obtains E is true since P
cannot be particularly affirmed.
With this interpretation of the square, the existen-
tial import of a statement is determined by its qual-
ity rather than its quantity,lo While this means that
lOCo S. Peirce remarks, "It is probable that Kant also under·

stood the affirmative proposition to assert the existence of its
subject, while the negative did not do so: so that 'some phoenixes
do not rise from their ashes' would be true, and 'All phoenixes do
rise from their ashes' would be false" (Collected Papers, Vol. II,
par. 381). Peirce refers, Vol. III, par. 178, to the view that ex-
in any case at least one of the two particular statements
in the square must be true, it does not result in the
absurdity of forcing us to affirm the existence of what-
ever may be the subject of a statement. In all cases
where C'-2 obtains, both affirmative statements are
false and we do not affirm the existence of anything.
This interpretation of the square, however, is im-
possible if we accept the usual treatment of what Aris-
totle called indefinite (or infinite) nouns and verbs. As
they are normally used in logic books these indefinite
terms permit us to assert equivalences between affirma-
tive and negative statements, so that the quality of a
statement remains relative to a particular mode of ex-
pression and clearly cannot serve as the determination
of its existential import. But Aristotle seems to deny
these equivalences in his discussion of indefinite terms
as they occur in universal and particular statements.
We noted above that, according to Aristotle, the
logical relation between a singular affirmation with an
istential import is determined by the quality of the statement as
a view "usually understood" in the traditional account.
Peirce does not cite evidence in support of this interpretation
of Kant, but it is not difficult to find passages in the Critique of
Pure Reason that suggest it. For example, "As far as logical form
is concerned, we can make negative any proposition we like; but
in respect to the content of our knowledge in general, which is
either extended or limited by a judgment, the task peculiar to
negative judgments is that of rejecting error" (A 709, B 737; tr.
N. Kemp Smith; italics in original). Viewed in this way, then,
negative judgments simply reject affirmative judgments as errone-
ous and affirm nothing about objects in the world. The remark
that, as regards logical form, we can make negative any proposi-
tion we like, is relevant to the considerationj in the remainder of
the present paper. We should note in passing that in accordance
with this remark the qnality of a statement as determining its
existential import is not, for Kant, to be identified with its qnality
as determined by its logical form.
Evidence that Kant assnmed affirmative statements to have
existential import may be found in his well-known denial that
"existence" is a predicate, plus his insistence on the distinction
between categorical and hypothetical judgments.
62 M. Thompson
indefinite verb as predicate and the corresponding de-
nial with respect to the same subject was not equiv-
alence but implication with the first statement as an-
tecedent. Although Aristotle's account of universal and
particular statements with indefinite terms is difficult
in its details and the text seems corrupt at a few places,
I believe there is little room for doubt about the fol-
lowing two points. (1) Whenever a statement contain-
ing no indefinite terms and representing one of the
four categorical forms is logically related to another
statement containing indefinite terms, the relation
seems to be implication rather than equivalence.n
(2) In every case the affirmative statement is the an-
tecedent of the implication.
The following four implications are clearly indicated
by Aristotle's remarks in Chapter X of his On Inter-
pretation (the implications run horizontally and the an-
tecedent is always first).
(1) Every man is just. XNot every man is not-just.
(2) Every man is not-just. Not every man is just.
(3) Every man is not-just. No man is just.
(4) Some men are just. Not every man is not-just.
The first two pairs in this list occur in a context where
Aristotle is concerned primarily with the contradictory
oppositions indicated by the diagonals rather than the
implications and the latter emerge from the schema of
11The verb used to express this relation is akolouthein, which

means literally "to follow," "to go after," or "with." The most
obvious translation when the verb occurs in a logical treatise is
"to follow from," "to be implied by," and this is the practice
followed by the Oxford and Loeb translators. However, many
commentators have read Aristotle as often using the verb with
the force of "to be equivalent to." This reading is ruled out by
the present interpretation of Aristotle, but the exclusion rests on
an attempt to make sense out of his logical doctrine rather than
on the claim that it is necessitated by the Greek alone. When,
on the other hand, the text gives tauton semainein the logical
relation must of course be equivalence rather than implication.
oppositions. The last two pairs are stated separately as
implications and are not presented as part of a schema
of oppositions. Since an A statement with an indefi-
nite verb as predicate implies both an a and E state-
ment, a definite A statement would also seem to have
two corresponding implications. Further, since the list
of antecedents includes both the definite and indefinite
forms of A it would seem that this list should also in-
clude both forms of the I statement. It might thus
appear that we can add the following two implications
which are not explicitly stated in the text.
(5) Every man is just. No man is not-just.
(6) Some men are not-just. Not every man is just.
However, I believe we can argue plausibly that, with
Aristotle's analysis, these two additional implications
are fundamentally different from the preceding ones,
and that his failure to include them here is probably
not an oversight. We should note first of all that in
none of the four implications listed by Aristotle is there
a statement with an indefinite term as subject. All the
indefinite terms are predicates. Secondly, the two cate-
gorical statements that convert simpliciter, viz. E and
I do not occur in the list with an indefinite term. Yet
our proposed additions to the list involve such occur-
rences. By simple conversion, we get the following
equivalent forms of our new implications.12
(5.1) Every man is just. No not-just is a man.
(6.1) Some not-just is a man. Not every man is just.
The difficulty here is in explaining how indefinite
nouns function as subjects. \Ve should note next that
12It is assumed here that for Aristotle simple conversion (as

distinct from obversion) yields an equivalent statement. Justifica-
tion for this assumption can be found in Aristotle's use of con-
version in the reduction of syllogisms. He of course did not lise
obversion in the reductions.
c*
64 M. Thompson
after (1) and (2) are presented as pairs of oppo-
sites Aristotle lists two more such pairs that occur
when "not-man" is taken as "a kind of subject." He
cautions that this new set of opposites "should re-
main distinct from those which preceded it, since it
employs as its subject the expression 'not-man'," The
two further pairs of opposites (with oppositions indi-
cated by the diagonals) are:
(7) Not-man is just.
(8) Not-man is not-just.
X Not-man is not not-just.
Not-man is not just.
Aristotle does not comment on the quantification of
these statements, nor does he give any explicit indica-
tion that implications hold here as in (I) and (2), but
a little later in the same chapter he gives the following
pair as equivalentsP
(9) Every not-man is not just. No not-man is just.
Yet if treated as an ordinary categorical statement, the
first member of this equivalence is ambiguous and
might mean "Not every not-man is just" instead of
"No not-man is just." In order for the equivalence to
hold, "every" and "no" must be taken with the force
of "everything" and "nothing." \Ve should thus under-
stand the equivalence as between
(9.1) Everything that is Nothing that is not man
not man is not just. is just.
The necessity for this interpretation of "every" and
"no" arises from the indefiniteness of the subject "not~
man." \Vith this term as subject the only things about
which we can make assertions are those which are not
some definite kind of thing. \Ve can thus never make
an assertion about every mcmber of a collection, but
l3The Greek here is tauton semainein as opposed to akolou·

thein. Cf. note J 1. The translation in (9) is literal. The Oxford
translation is that given in (9.1),
only about every member that is not such and such.
Our assertions will be like the statement a teacher
might make after collecting examination papers. "Every
paper in this group that does not receive a score of
50 or above is not passing," or in other words, "Every
not-50-or-above-paper is not passing." But when we use
an indefinite noun as subject without reference to some
particular collection, such as a group of examination pa-
pers, we make a reference to the collection that com-
prises the totality of real things. Thus, the statements
in (9) and (9.1) are equivalent to "Everything in the
totality of real things that is not a man is not just."
This reference to the totality of real things results from
the peculiar way that an indefinite term signifies things.
These considerations should make it apparent that
the statements in question do not have existential im-
port in so far as they are assertions about those mem-
bers of a collection that are not such and such. The
teacher's statement does not imply that there is at
least one not-50-or-above-paper in the group nor do (9)
and (9.1) imply that there is at least one not-man.
From this point of view, the statements are compound
denials rather than affirmations-they deny that there
is at least one member of the collection which is both
not-50-or-above and passing, both not-man and just.
This equivalence to a compound denial is again a re-
sult of the peculiar signification of indefinite terms,
and is no less the case when the statement appears to
be a universal affirmative. "Every not-man is just" is
equivalent to the denial that there is at least one thing
which is both not-man and not just. Particular state-
ments with indefinite terms as subjects, on the other
hand, are always equivalent to compound affirmations
of existence, even though they appear to be particular
negatives. "Not every not-man is just" is equivalent to
"Something in the totality of real things is both not-
66 M. Thompson
man and not just." The statement must be construed
in this way if it is to contradict the compound denial
expressed by "Every not-man is just."
The above remarks may suffice to explain why Aris-
totle regarded indefinite terms as providing only "a
kind of subject," and cautioned that the resulting state-
ments should remain distinct from the ones previously
considered. His failure to comment on the quantifica-
tion of the statements in (7) and (8) is perhaps due
to the fact that the contradictory oppositions, which
seem to have been his primary interest here, are not
altered when the statements are quantified as in (1)
and (2). A list of the quantified forms would have
been superfluous unless he intended also to add a spe-
cific account of the peculiarities which resulted from
the indefinite subject.
A similar explanation may be given for the failure
to mention the implications stated in (5) and (6). Un-
like any of the preceding, each of these implications
related a simple affirmation or denial to one whose sub-
ject is an indefinite term. Had Aristotle mentioned
these two implications, it certainly would have been
an oversight on his part if he had then omitted the
other implications of this sort. But since he did not
give the two in question, it seems fair to conclude
that he did not intend here to consider any of them.
In light of the above considerations, we now have
the following account of the immediate inferences that
arise from the introduction of indefinite terms, i.e. of
the processes usually called obversion and contrapo-
sition.
An A statement implies, but is not equivalent to,
its obverse, its partial contrapositive, and its full con-
trapositive. Thus, "If every S is P, then no S is not-P,
no not-P is S, and every not-P is not-S." Each of these
three statements in the consequent is equivalent to the
other two. "No not-P is S" is equivalent to its obverse,
because unlike the original A or a simple E, it has an
indefinite term as subject so that neither it nor its ob-
verse has existential import. The inference from "Every
not-P is not-S" to "Some not-P is not-S," the so-called
full inverse of "Every S is P," is illegitimate because
the antecedent in this case does not have existential
import while the consequent does.
An E statement is implied by, but is not equivalent
to, its obverse, just as the definite denial "Socrates is
not ill" is implied by the indefinite affirmation "Soc-
rates is not-ill." Hence the inference from an original
E to its contrapositive or its inverse (whether these
are full or partial) is never permissible. We cannot in-
fer from "No mathematicians are circle-squarers" that
some nonmathematicians are circle-squarers. In order
to obtain this conclusion we need independently the
premise, "Every circle-squarer is a nonmathematician."
This premise is the obverted converse of the original
E, and while with the present analysis it implies the
E, it is logically independent of the obverse of E, since
"Every mathematician is a noncircle-squarer" does not
affirm anything about circle-squarers.
An I statement implies its obverse but is not equiv-
alent to it (see (4), while an 0 statement is implied
by its obverse but is not equivalent to it (see (6)). An
o is also implied by its partial and full contrapositives,
which are both equivalent to its obverse. These equiv-
alences obtain because particular statements with in-
definite terms as subjects always have existential im-
port, in contrast to the original 0, which asserted a
simple denial.
The fundamental difference between this account of
immediate infercnces and the one which is usually ac-
cepted as characteristic of traditional logic is the treat-
ment given the distinction between statements with
68 M. Thompson
definite and indefinite subjects. Before we consider
how this account avoids the contradictions and absurd-
ities in the traditional analysis, there is one further
point to be noted about Aristotle's analysis.
Aristotle seems to have begun construction of his
square with an A statement taken as a simple affimla-
tion about every member of a collection rather than
about every member that is not such and such. In
other words, with a statement like "Every man is
just" as opposed to "Everything that is not a man is
just." As we have already indicated, with Aristotle's
analysis we should take the first but not the second of
these statements as having existential import. Yet
granting this point, we still have the problem of de-
ciding which of the following two statements we shall
take as the contrary of our A statement.
(10) No man is just.
(10.1 ) Every man is not-just.
Either of these statements is so related to our orig-
inal A that it cannot be true when the A is true,
though it can be true or false when the A is false. If
this reference to truth conditions is our only criterion
for determining the contrary, we should have to ad-
mit both statements, even though they are not equiv-
alent, as contraries of our original. Yet clearly, (10) is
the only one that will satisfy the present interpreta-
tion of Aristotle's square of opposition. If this inter-
pretation is correct, we would expect Aristotle to offer
arguments for selecting (10) rather than (10.1), and
this is precisely what he does in the last chapter of
On Interpretation. He devotes this chapter to the prob-
lem he poses in the opening sentence: "The question
arises whether an affirmation finds its contrary in a de-
nial or in another affirmation; whether the proposition
'every man is just' finds its contrary in the proposition
'no man is just', or in the proposition 'every man is
unjust'."H
It is unnecessary for our purposes here to analyze
separately each of the various arguments Aristotle of-
fers for selecting the denial rather than the affirmation
as the contrary. His main point may be summed up
by saying that since (10) is implied by but does not
imply (10.1), (10) is more opposed to the original A
because when the latter is false (10) may be true even
though (10.1) is also false. I do not propose to debate
the logical merits of this decision, as the relevant point
here is simply that with Aristotle's analysis, (10) and
(10.1) are not equivalent and that (10) alone represents
an E statement in the square of opposition. 111is deter-
mination of E does not prevent it from implying 0,
since the latter is likewise a simple denial without exis-
tential import. Thus, the logical relations which Aris-
totle claimed for his square of opposition hold if we
grant that the statements related in the square are of
the peculiar sort that he seems to have intended them
to be.
We may now turn to the contradictions and absurd-
ities in what is usually taken as the traditional anal-
ysis of the square. When an A statement is taken as
equivalent to its obverse, Aristotle's distinction be-
tween statements with definite and indefinite terms as
subjects is ignored. A statement which has existential
import with his analysis is made equivalent to one
that cannot have this import as he construed it. The
traditional analysis then becomes in~onsistent by ac-
1423a28-31. The logical relations in question here remain the
same if the privative "unjust" is replaced by the indefinite "not·
iust." Cf. An. Pro S2alS-I7. While One might intend by this
replacement to make the statement equivalent to (10), he does
not succeed in doing so because the result, (10.1), remains an
affirmation and thus, like the original "every man is unjust,"
implies but is not implied by (10).
70 M. Thompson
cepting in toto Aristotle's account of the square of op-
position along with this equivalence which Aristotle
rejected. It must then be affirmed that "Every S is P"
implies "Some S is P," and denied that "Every S is
P" can be false while "No S is not-P," "No not-P is
S," and "Every not-P is not-S" are all true. This proce-
dure becomes inconsistent when one admits, as Aris-
totle does, that any of these last three statements is
compatible with the falsity of "Some S is P." Incon-
sistency is avoided in Aristotle's account, not by quali-
fying the principle of excluded middle, but by restrict-
ing existential import in accordance with the quality
of the statement and the definite or indefinite charac-
ter of the subject term.
The traditional procedure also leads to the absurdity
that any object of opinion (anything thinkable) must
be said to exist. We have characterized statements
with indefinite terms as subjects as like statements
about every member that is not such and such in a
certain collection, but clearly as far as existential im-
port is concerned, they are also like statements about
every member that is such and such. "Every paper in
this group that scores 50 or above is passing" is just
as free of existential import as our former example.
There is obviously something absurd about a logic that
allows us to infer from this that some paper in the
group scores 50 or above and is passing. Yet this is
analogous to what the traditional account allows us to
do. For a statement like
(11 ) Every ogre is wicked.
is construed as equivalent to "Every not-wicked is not-
ogre," which clearly means the same as "Everything
in the totality of real things that is not wicked is not
an ogre." But then with this equivalence (11) must
mean the same as
(ILl) Everything in the totality of real things that
is an ogre is wicked.
\Ve thus end with the absurdity that we cannot as-
sert (ILl) without implying that ogres exist. And even
worse, we cannot deny (11.1) without making the same
implication. For
(12) Not every ogre is wicked.
is also assumed to have existential import.
Aristotle's account avoids such absurdity by denying
the equivalence between (11) and (ILl) and holding
that (12) has existential import only when taken as
the contradictory of the latter. In this way there is no
need to qualify the application of the principles of ex-
cluded middle. In order to prcserve Aristotle's square
in toto we must of course deny the equivalence be-
tween (lO) and (10.1), and accept only the former as
properly an E statement. The contradictory of such a
statement has existential import and is implied by the
corresponding A. In contrast, modern analysis accepts
the traditional position that (lO) is equivalent to
(10.1), and (11) to (ILl), but remains consistent and
avoids absurdity by refusing to accept Aristotle's square
without qualification.
I do not propose here to examine the adequacy of
the Aristotelian account or to subject it to criticism
in the light of modern analysis. But it should be noted
in conclusion that if the present interpretation of Aris-
totle is correct, his analysis cannot, as the traditional
one can, be taken simply as an inconsistent account
which modern analysis has rectified. In accepting the
equivalences mentioned above. the traditional doctrine
of immediate inferences has already taken an essen-
tial step toward the modern position. Consistency can
be restored (without special assumptions, such as that
restricting the applicability of the principle of excluded
72 M. Thompson
middle) only by completing the break with Aristotle
or by returning to his analysis. There is thus a funda-
mental difference between the Aristotelian and mod-
ern analyses which I believe it would be profitable to
examine. This difference is concerned in part at least
with the role of indefinite terms in procedures of quan-
tification and with the existential import of affirma-
tions as distinct from denials. When we say that
"-(x)fx" is equivalent to U(:!Ix)_fx," or that U_(:!Ix)
fx" is equivalent to U(x)_fx," we have replaced what
seems to be a denial by an affirmation with an indefi-
nite predicate, and we are not talking about the cate-
gorical statements Aristotle intended when he con-
structed his square of opposition.
II. CATEGORIES
CATEGORIES IN ARISTOTLE
AND IN KANT
JOHN COOK WILSON
§ 438. If our reasoning is correct, the universalities

of the various species of a genus are not particulars of
the universality of the genus but kinds (or parts) of it.
Suppose now an abstraction beginning from individual
things till we come to so-called summa genera, or, as
they would be more correctly termed, summae species.
The universalities of all other universals would be
comprised in the universalities of these, not as their
particulars but as kinds, forms, parts, or aspects of
them. If, then, we considered these summae spe-
cies as kinds which Being or Reality must take,
where Being is more accurately the Being of Things
(that is, the universal of which Things are the partic-
ulars), the Being of Things is in the position of genus
to these summae species and their several universalities
are all comprised in the universality of this genus. In
regard to this we must avoid the fallacy of creating a
universal of universalities, with an infinite regress. That
fallacy has been already sufficiently exposed. There is
no universal of universalityness of which the univer-
salities of universals would be particulars, for the uni-
versality of one genus universal as distinguished from
that of another lies not in its being a unity which is
From Statement and Inference (Oxford, 1926), Vol. II, pp.
696-706. Reprinted by permission of the Clarendon Press, Ox·
ford.
76 John Cook Wilson
particularized but in being particularized in these pre-
cise individuals, in being the particularization of its
own peculiar quality or characteristic; but this is pre-
cisely itself, and its universality is indistinguishable
from itself in the fullness of its being. Thus the clas-
sification of universalities can only be, if possible at
all, the classification of universals; there is no universal
of universalityness of which the universalities of uni-
versals would be particulars.
§ 439. This genus, then, of the summae species is
not mere Being but the universal of the being of
Things or Substances, and is therefore the generaliza-
tion of Substance, Substanceness, or Substance-in-gen-
eral. Now in this system every universal has things for
its particulars, and thus the universals of attribute and
relation will not appear in the system. Nevertheless,
every variety of being, attribute, and relation as well
as substance, must be comprised in the system, be-
cause all these are comprised in the existence of things.
The nature of substance involves in itself attributes
and relations; its universal therefore involves the uni-
versal of possession of attributes and the universal of
relatedness. We have now to consider the relation of
such a universal as attributeness in general or the uni-
versal of a particular attribute such as colour(edness)
to the universal of substance in general and to the uni-
versals of substances. We may note in passing that we
can understand why Aristotle said that Being was not
a genus if we remember that the genera par excellence
(in the Categories, for example) are in fact universals
whose particulars are substances. 1
Attributeness means being an attribute of a sub-
stance, hence the universal of substance involves in its
own being the being of attributeness. Nevertheless,
the latter is not an element in the universal substance-
leaf. 2all if.
Categories in Aristotle and in Kant 77
ness, the corresponding element is 'having-attributes'-
ness. Moreover, the latter is not a differentiation of
substanceness. The question, then, is whether these
universals are capable of being unified by any unity
beside the unity that one involves the existence of the
other. They cannot be differentiations of one and the
same universal because differentiations of the same uni-
versal must have identical particulars with that uni-
versal. Nor can they be particulars of the same uni-
versal, for if two universals are particulars of the same
universal, they must be differentiations of a common
universal though not of that universal. Thus, though
we can state of attributeness and substanceness that
they are, Being is not a universal of which they are
differentiations, nor is Being a universal of which they
are particulars.
§ 440. Thus the form'S is', which as opposed to'S
is P' does not occur in ordinary linguistic usage wheth-
er in ancient Greek or in modern languages, does not
represent in its subject particulars of a universal which
is Being or Beingness. If, then, Aristotle had carried
out fully the thought which appears to be implicit in
chapter 5 of the Categories, with the distinction there
made between 'said of a particular subject' and 'exist-
ing in a particular subject', the result would have been
a system of universals classified as in the section above
and based on his view that the only true independent
reality is the individual thing (or person). The sum-
mum gcnus would be Thingness, and for this the only
word in his terminology that appears to be suitable is
~eing (ouo(a), a word which both in the Categories
and in Metaphysics, Book Z, is sometimes the name
for Thing as such, or First (Primary) Being. This, how-
ever, is nowhere unmistakably stated; he appears in-
deed to have virtually stopped in his classification at
78 John Cook Wilson
certain highest genera and his own expression, 'the
highest of the genera? favours this interpretation.
§ 441. For his view that Being is not a genus, the
important passages are three in the Metaphysics with
which one in the Topics agrees, a passage in the De
Interpretatione, and one in the Posterior Analytics. 3
The first of the passages from the Metaphysics does
not turn on his view that the genera par exceIIence are
secondary essences and is a little difficult to interpret.
It runs as follows: 'But it is not possible for One or
Being to be a genus. For each difference of a genus
must have one and being said of it (sc. when we speak
of it), whereas it is impossible that either the species
of the genus or the genus without the species should
be said of the differences. Thus, if One or Being were
a genus, none of their differences could have one or
being stated of it'. By difference he here means not
the differentiated universal (e.g. rectilinearity) but the
differentia (straight, etcY This suggests that he had
not realized that, according to his own theory, Being
would be the proper name for the highest genus. The
passage from the Posterior Analytics, 'to exist is not
the being (essence) of anything, since the being existent
(that which exists) is not a genus', does seem to be
connected with his doctrine that the individual is the
only real existent and that the true genera have 'be-
ings' for their particulars. Finally, the passage in the
De Interpretatione runs as follows: 'for not even "to
be" or "not to be" is a sign of a thing, not even if
you mention "being" by itself merely. In itself it is
nothing, but it signifies in addition a kind of com-
pounding which cannot be understood without the
2Met. 998b18.
3Met. 998b22, l053b23, l059b30; Top. 121a16; De Int.
16b22; An. Po. 92b13.
4Cf. his language in the Topics 144a37 If.
things so compounded'. This is most important for
the linguistic point involved in our problem.
It is singular that Aristotle, who in the Metaphysics
is attacking those who made Bcing and Unity the es-
sence of things, does not adopt the seemingly clear
and decisive argument based upon his own theory of
true being as the individual and the true genus as that
of which the individuals are the particulars. His argu-
ment, however, and this is important, is directed to
showing that Being as a universal cannot be differ-
entiated, though he has not realized what seems to
be the direct proof required (he gives instead a single
reductio ad absurdum), viz. that if being were the ge-
nus or class universal of the universals which we say
'are', those universals, as we have shown, would ncc-
essarily have the same kind of particulars as it and con-
sequently the same kind as one another. But, obvious-
ly, the universals of a substance, an attribute, a rela-
tion, cannot have the same particulars. To reach this
positive point of view Aristotle would have required
to have had before him the point of view from which
a universal could be represented as a particularization,
not a differentiation, of another universal.
§ 442. We miss, then, two things in Aristotle's dis-
cussion. The line of thought which led him to say 'be-
ing is not a genus', at least when it took the shape it
seems to have in the Posterior Analytics, might have
made him recognize that though the universal 'being'
was not a highest genus, yet in his own terminology
such a genus was exactly 'essence' as the being of
things or, more accurately, 'the being a thing'. How
far he was from this may be seen when he says,5 'Unity
cannot be a genus for the same reasons that neither
being (existent) nor essence can be'. Secondly, he might
have been led to see that though Being is not a clas-
5Met. l053b23.
80 JolIn Cook Wilson
sifying genus which unified everything in that way, yet
there is a unifying principle in all reality. Instead he
rested content with the negative statement that Be-
ing is not such a principle, where being is that 'is'
which is universally predicable. He may have been
prejudiced by his justifiable criticism of Plato who had
sought this unifying principle in the Idea of goodness.
Had he followed up his thought he might have re-
flected that just as a genus demands its own differen-
tiation into species and individuals, so by the same in-
ward necessity the unity of reality demands the kinds
and things into which reality actually is differentiated.
§ 443. Taking his own categories, they are obvious-
ly unified in the reality of (primary) Being or Sub-
stance, and he does elsewhere recognize that the other
categories depend on the first. But he never put this
as the unity and the real unity whi~h corresponds to
what is common to the statement of 'being'. In fact
he never cleared up his mind about it, or he could
not, so to speak, have so degraded 'being', as he does
in the passage translated from the De Interpretatione,
and have merely left it at that. What he has said there
is true and important, but it is misleading, as it stands.
Moreover, the formula 'Being is not a genus', al-
though it shows from one point of view an accurate
insight into the nature of classification, is extraordi-
nary and misleading when considered in relation to his
own terminology. He must speak of the categories as
categories of being; this being cannot be merely the
common predicate of everything, if we are to take cate-
gories literally as predicates. For we cannot state of
this 'being-in-general' that it is a substance. On the
other hand, if it does mean 'being-in-general', cate-
gories would surely have to mean species or kinds, so
that being would indeed be a genus and the formula
be contradicted. It is most natural to leave category
its proper meaning and then 'being' will stand for
'that which is', not for being in general. This again,
if all the categories are asserted of it, as the formula
'categories of that which is' naturally implies, could
only be complete being, that which is in the fullest
sense. Now that with Aristotle is the individual thing.
This agrees with the fact that the categories are given
not as abstract universals in the noun form (whiteness,
as an example of quality) but in the adjectival form
(white, double, etc.). If so, what is meant by the cate-
gories of being (that which is) is that each of them is
an attribute of the true and real being of the individ-
ual thing, the primary being of this treatise. It is the
thing in fact of which we can properly state that it is
a substance, two cubits long, in the market-place, etc.
The only word form which causes any difficulty in
this interpretation is that of the adverb of time (e.g.
yesterday). This, however, is again not put as an ab-
stract universal but as it would occur in a statement,
and it is true (and the truth) that as every happening
belongs to a substance, so the temporal qualification
ultimately also belongs to a substance (e.g. this sub-
stance was in the market-place yesterday). This appears,
then, to be the meaning, for it is difficult to see how
anything else could be meant.
§ 444. But now, if this is so, in the categories of be-
ing, of 'that which is', the latter expression is used in
the general sense, and 'that which is' represents the
universal of all particular 'beings' and so is the uni-
versal of substanceness. It is just in relation to the dis-
tinction of the categories that the meaning of 'being
is not a genus' becomes of great importance. On the
above view 'that which is' is equivalent to substance
(being) when the categories are termed categories of
that which is. Thus 'that which is' is not only a genus
of which real complete things are the particulars (com-
82 John Cook Wilson
plete, that is, in the Aristotelian sense as equivalent
to things) but the highest and most comprehensive
genus, though not the genus of the categories, includ-
ing all reality whatever. This discussion seems to con-
firm the hypothesis that Aristotle did not pursue the
train of thought which his view that 'being' taken in
one sense is not a genus might have suggested to him.
§ 445- What, then, is the fallacy in Aristotle's proof
that being is not a genus? He contemplates two ways
of stating the genus X in regard to a given subject S.
Either a species AX is stated in S is an AX and there-
fore mediately the genus S is an X, or the genus is
stated 'without its species', and we say immediately S
is an X. In neither way, he says, can we state the genus
of the difference A. He seems to mean that odd being
the differentia of odd number, we can't say that odd
is an odd number and so a number, nor directly that
odd is a number.
But is the argument free from verbal fallacy? When
we state that S is a species, we mean that S is an AX,
and when we say that the genus belongs to the spe-
cies, we state X, the genus universal of a member of
the species AX, we mean an AX is an X. If, then,
we state the species of the differentia we ought to
mean that we state that an A is an AX. Now if the
differentia necessarily presupposes X the genus, the
statement must be true. We do not say that odd num-
berness is numberness, but that an odd number is a
number. So if we state odd number of odd, we mean
that an odd thing is an odd number, which is neces-
sarily true. It looks as though Aristotle meant that
linearity cannot be predicated of straightness because
straightness is not a line; but neither is rectilinearity
a line, and thus he would appear to have fallen into a
verbal fallacy.
§ 446. There is a danger that in appreciating the
insight shown by Aristotle in his dictum 'being is not
a genus', we may ourselves fall into the same one-sided-
ness. There are certain characteristics of Being which,
if not identical with those of a true genus, are paral-
lel to them in a remarkable way and must therefore
not be neglected. These characteristics are not de-
stroyed by the discovery that being is not a true genus;
only a certain way of regarding them, i.e. as particu-
larizations or differentiations of a universal, is destroyed.
An attribute or a relation may rightly be said to have
being as well as a substance, and this being is not iden-
tical with that of substance. We can study the attri-
bute in abstraction from the subject, as we do in the
mathematical sciences, and this proves that it is dis-
tinguishable in being. So of relation, it is essential to
relation that it should be 'between' things which have
some other nature than that of standing in a given
relation. Hence it is natural at first sight to say that
Being is a genus with its species 'being of attribute',
etc. Again we think that Being in its own nature ne-
cessitates these forms of itself, and this is parallel to
the self-determination of a universal in its differentia-
tions. If, then, the forms of Being are not species nor
kinds as ordinarily understood, nor of course particu-
lars of a universal, how should they be described ac-
curately? We have to revise one of our usual concep-
tions and either to refuse to call these universals or
to admit universals which have no particulars and no
differentiation, the universality of one quality (attri-
bute) being identical in kind with the universality of
every other. In this difficulty we may provisionally
term them common principles. They are live univer-
sals because they are a unity in a manifold; univer-
sality in each universal, particularity in each particular,
and so on, but different from true universals because
the manifold is not a particularization or differentia-
84 Jo11n Cook ,,"Tilson
tion of the common principle. This will meet the dif-
ficulty we found in regard to classes. We say of every
class that it is a class, yet we saw that classness is not a
true universal of which classes are particulars. Class has
no differentiation and no particulars. We have a unity
in a manifold, but the manifold is not a particulariza-
tion of the unity. Instead, then, of our ordinary view
we have to recognize that there are some universals
which admit of differentiation and particularization but
not of individualization, and that some (which we have
called common principles) do not admit of either par-
ticularization or differentiation.
§ 447. In Being, then, called provisionally a common
principle, we recognize a unity in a manifold which is
sui generis, just as much as the unity of a true uni-
versal in its particulars and its species is sui generis.
Being, like a true universal, also determines its own
manifoldness but in a different manner. If we are to
seek an illustration or analogy we may refer to the self,
which is an absolute unity in its different thoughts, or
a body, which is identical in its different positions and
aspects, but neither the self nor the body is the uni-
versal of those differences nor they its particulars.
§ 448. If we now consider the forms of Being, the
manifold which it must assume and which simulates
the differentiation of a universal, we may perhaps find
the true significance of the philosophical classifications
called systems of categories. These categories are ob-
viously of a very special kind and are philosophical
and not scientific. They are, that is, though compre-
hensive, not a classification and could not be reached
by abstracting successively from the whole field of in-
dividual substances. In this way we should attain to
the classifications of the natural sciences but not to
categories. Now Kant's criticism of Aristotle's catego-
ries has shown that while there may be agreement
about some of the main categories, there may be the
greatest disagreement about the real meaning and ob-
ject of them as a whole. Kant believes that Aristotle
was led to his grouping of the categories without real-
izing its true character, and that in consequence he did
not carry it out consistently. \Vhether, then, Kant's
own view of the categories was right or wrong (and
surely it was not right) this suggests that there may be
a certain instinctive impulse to search for categories,
without full consciousness of its nature. The impulse
will lead to an arrangement of a quite peculiar kind,
and reflection must then supervene in order to under-
stand the impulse and to correct its imperfect work.
Kant naturally makes the categories forms of the uni-
fying understanding, because of his dominant confu-
sion of the apprehension and the apprehended. Aris-
totle's h:ndency is far sounder, for necessity of appre-
hension can, after all, only mean apprehension of a
necessity in the object.
The explanation, then, of systems of categories may
well be that we come in time, by reflection on the use
of the verb 'to be', to recognize a corresponding unity
of being, that the totality of particulars in all their va-
riety is a unity. Long before we have recognized this
unity in particular sets or varieties of being. Then
comes the philosophic impulse to determine the forms
which being in general must take, suggested by the
analogous determination of such a unity as the section
of a cone, and so to cover exhaustively the differentia-
tions of being in general in the whole of existence.
This impulse need not be fully aware of itself. Aris-
totle, for example, doesn't consciously go to work in
this way or, if he did so begin, he probably gave it
up when through his clear idea of differentiation and
classification he realized that Being is not a genus. Still
it may lJUvt: bccn the fact that we state of everything
86 John Cook Wilson
that it is, which suggested the idea of an absolutely
unified system of being.
§ 449. Thus the characteristic of Aristotle's system
is that it is a sort of exhaustive attempt to covcr the
whole variety of reality. 'Everything is either substance
or an affection of substance' is its implicit meaning.
These are in fact his two main categories. But he gets
into difficulties about substance, so that his thought
in the end is that everything is either substance, or
quality of substance, or quantity of substance (or of
what belongs to substance), or relation of substance
(or of what belongs to it), and so on. Now it is clear
from our analysis above that the impulse to determine
a sort of differentiation of Being must produce an al-
together unique system. For the general forms of Be-
ing are not true differentiations of it, nor are the in-
dividuals, which 'are' particulars of Being but of other
universals. This explains the fact that the classification
(as it first seems to be) of all being could not be at-
tained by any abstraction from individuals or any clas-
sification of them; because this, the natural form which
classification first takes, proceeds to universals of which
the original particulars are those very individuals, and
the successive universals are differentiations of one an-
other in succession. This systematization is by means of
differentiation and particularization, the other (the sys-
tem of the categories) is achieved by neither. Hence the
latter acquires its uniqueness; for the ordinary scientific
method, even when carried to the utmost unity, will
not bring to light one single category except substance
itself. If continued ideally upwards, it cannot even
bring to light the universal 'mere being' of which the
categories are the unfolding. 1 'he same is of course true
of any classification proper which starts from the indi-
vidual, whether thing or individual attribute. Attribute-
ness, for instance, if regarded as the universal of in-
dividual attributes, is neither a differentiation nor a par-
ticularization of being; the most general universal which
could have individual attributes for its particulars is
just attributeness. Now substanceness, attributeness, re-
lationness, etc., cannot be treated as summa genera be-
cause they would then be members of the same clas-
sification; this they can never be, for such genera must
have the same sort of particulars, and here the partic-
ulars of the one would be substances, the particulars
of the other attributes.
§ 450. If, then, this is the true account of the phil-
osophic impulse under discussion, we see that it is a
most serious and fatal mistake to regard it as a clas-
sification, when once we understand what a classifica-
tion truly is. (That leads to the further error of vainly
attempting to adapt the system of categories to the or-
dinary classifications and to make them departments
of it.) Properly understood, it is simply the just en-
deavour to determine the manifoldness in which Be-
ing in general must unfold itself, and Aristotle pro-
ceeded correctly when he exhibited as categories such
general forms as those of Substance, Attribute, Rela-
tion, etc. Contenting himself, however, with point-
ing out that Being is not a genus and therefore could
not constitute the essence of things, he seems by his
merely negative attitude to have missed the true sig-
nificance of his own list of categories.
§§ 438-50. Translation of principal passages referred to

in the above investigation.
[Arist. Cat. 2all. 'Substance most properly and pri-
marily and especially so called is that which is neither
said of a particular subject nor is in a particular sub-
ject, for instance, a particular human being or a par-
ticular horse. Secondary substances are what the spe-
cies are called, in which the substances primarily so
D
88 John Cook Wilson
called exist, and besides them the genera of these spe-
cies; for instance, while a particular human being exists
in the species human being, the genus of the species
is animal. These, then, are called secondary substances,
for example, both human being and animal'.
De Int. 16bZZ. 'For not even "to be" or "not to be"
is a sign of a thing, and not even if you mention "be-
ing" by itself merely. In itself it is nothing but it sig-
nifies, in addition, a kind of compounding which can-
not be understood without the things compounded'.
An. Po. 9Zbl3. 'Again we say that it is necessary
that everything be proved to exist by demonstration,
unless it be essence. But to exist is not the essence of
anything, since the being existent (that which exists)
is not a genus. That a thing exists therefore will be
(subject of) demonstration. This is what the sciences
in fact do. The geometrician assumes what triangle
means, but proves that it exists'.
Met. 998bZZ. 'But it is not possible for either be-
ing one or being existent to be a genus of existents.
For while it is necessary that the differentiae of each
genus should each both exist and be one, it is impos-
sible either for the species of a genus to be said of the
appropriate differentiae or the genus <to be said of
them> without its species. Therefore if we assume be-
ing one or being existent to be a genus, no differentia
will either be or be one'.6
The same argument is used in Met. 1059b30, and.
Top. lZla16. In Met. 1053bZ3 he says, 'being one
cannot be a genus for the same reasons that neither
being existent nor essence can be'.
6J. C. W. translated: 'It is not possible for One or Being to be

a genus. For each differentia of a genus must have one and being
said of it (in statements), whereas it is impossible that either the
species of the genus should be said of its differentiae or the genus
without the species. So if One or Being were a genus, none of
their differentiae could have one, or being, stated of it:
Met. 1017a22. 'Whatever the forms of predication
signify are said to be essentially. For to be has as many
significations as there are forms. Inasmuch, then, as
the predicates signify what the subject is, others its
quality, etc., ... to be has a signification equivalent
to each of these (for there is no difference between
"the man is walking" and "the man walks)".J
ARISTOTLE'S CATEGORIES
TRANSLATED BY J. L. ACKRILL
CHAPTER 1
1a1. When things have only a name in common and
the de~nition of being which corresponds to the name
is different, they are called homonymous. Thus, for
example, both a man and a pi:ture are animals. These
have only a name in common and the definition of be-
ing 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.
1a6. When things have the name in common and
the definition of being which corresponds to the name
is the same, they are called synonymous. Thus, for ex-
ample, both a man and an ox are animals. Each of
these is called by a common name, 'animal', and the
definition of being is also the same; for if one is to
give the definition of each-what being an animal is
for each of them-one will give the same definition.
lal2. When things get their name from something,
with a difference of ending, they are called parony-
mous. Thus, for example, the grammarian gets his name
from grammar, the brave get theirs from bravery.
From Aristotle's Categories and De Interpretatione, tr. with

notes by J. L. Ackrill (Oxford, 1963), pp. 3-12 and 71-91. Re-
printed by permission of the Clarendon Press, Oxford.
Aristotle's Categories 91
CHAPTER 2
la16. Of things that are said, some involve combi-
nation while others are said without combination. Ex-
amples of those involving combination are 'man runs',
'man wins'; and of those without combination 'man',
,ox," runs,"wms.
.,
la20. Of things there are: (a) some are said of a sub-
ject but are not in any subject. For example, man is
said of a subject, the individual man, but is not in any
subject. (b) Some are in a subject but are not said of
any subject. (By 'in a subject' I mean what is in some-
thing, not as a part, and cannot exist separately from
what it is in.) For example, the individual knowledge-
of-grammar is in a subject, the soul, but is not said
of any subject; and the individual white is in a sub-
ject, the body (for all colour is in a body), but is not
said of any subject. (c) Some are both said of a sub-
ject and in a subject. For example, knowledge is in a
subject, the soul, and is also said of a subject, knowl-
edge-of-grammar. (d) Some are neither in a subject nor
said of a subject, for example, the individual man or
individual horse-for nothing of this sort is either in
a subject or said of a subject. Things that are individ-
ual and numerically one are, without exception, not
said of any subject, but there is nothing to prevent
some of them from being in a subject-the individual
knowledge-of-grammar is one of the things in a subject.
CHAPTER 3
IblO. Whenever one thing is predicated of another
as of a subject, all things said of what is predicated
will be said of the subject also. For example, man is
predicated of the individual man, and animal of man;
so animal will be predicated of the individual man
also-for the individual man is both a man and an
animal.
92 T. L. Ackri11
Ib16. The differentiae of genera which are differentl
and not subordinate one to the other are themselves
different in kind. For example, animal and knowledge:
footed, winged, aquatic, two-footed, are differentiae of
animal, but none of these is a differentia of knowl-
edge; one sort of knowledge does not differ from an-
other by being two-footed. However, there is nothing
to prevent genera subordinate one to the other from
having the same differentiae. For the higher are predi-
cated of the genera below them, so that all differentiae
of the predicated genus will be differentiae of the sub-
ject also.
CHAPTER 4
IbZ5. Of things said without any combination, each
signifies either substance or quantity or qualification
or a relative or where or when or being-in-a-position or
having or doing or being-affected. To give a rough idea,
examples of substance are man, horse; of quantity: four-
foot, five-foot; of qualification: white, grammatical; of
a relative: double, half, larger; of where: in the Lyceum,
in the market-place; of when: yesterday, last-year; of
being-in-a-position: is-lying, is-sitting; of having: has-
shoes-on, has-arm our-on; of doing: cutting, burning; of
being-affected: being-cut, being-burned.
Za4. None of the above is said just by itself in any
affirmation, but by the combination of these with one
another an affirmation is produced. For every affirma-
tion, it seems, is either true or false; but of things
said without any combination none is either true or
false (e.g. 'man', 'white', 'runs', 'wins').
CHAPTER 5
Zall. A substance-that which is called a substance
most strictly, primarily, and most of all-is that which
lRead TWY tTipc.JY YEVWV.

is neither said of a subject nor in a subject, e.g. the
individual man or the individual horse. The species in
which the things primarily called substances are, are
called secondary substances, as also are the genera of
these species. For example, the individual man belongs
in a species, man, and animal is a genus of the species;
so these-both man and animal-are called secondary
substances.
2a19. It is clear from what has been said that if
something is said of a subject both its name and
its definition are necessarily predicated of the sub-
ject. For example, man is said of a subject, the in-
dividual man, and the name is of course predi-
cated (since you will be predicating man of the indi-
vidual man), and also the definition of man will be
predicated of the individual man (since the individual
man is also a man). Thus both the name and the defi-
nition will be predicated of the subject. But as for
things which are in a subject, in most cases neither
the name nor the definition is predicated of the sub-
ject. In some cases there is nothing to prevent the name
from being predicated of the subject, but it is impos-
sible for the definition to be predicated. For example,
white, which is in a subject (the body), is predicated
of the subject; for a body is called white. But the defi-
nition of white will never be predicated of the body.
2a34. All the other things are either said of the pri-
mary substances as subjects or in them as subjects.
This is clear from an examination of cases. For exam-
ple, animal is predicated of man and therefore also of
the individual man; for were it predicated of none of
the individual men it would not be predicated of man
at all. Again, colour is in body and therefore also in
an individual body; for were it not in some individual
body it would not be in body at all. Thus all the
other things are either said of the primary substances
94 J. L. Ackrill
as subjects or in them as subjects. So if the primary
substances did not exist it would be impossible for
any of the other things to exist.
2b7. Of the secondary substances the species is more
a substance than the genus, since it is nearer to the
primary substance. For if one is to say of the primary
substance what it is, it will be more informative and
apt to give the species than the genus. For example,
it would be more informative to say of the individual
man that he is a man than that he is an animal (since
the one is more distinctive of the individual man while
the other is more general); and more informative to
say of the individual tree that it is a tree than that
it is a plant. Further, it is because the primary sub-
stances are subjects for all the other things and all the
other things are predicated of them or are in them,
that they are called substances most of all. But as the
primary substances stand to the other things, so the
species stands to the genus: the species is a subject
for the genus (for the genera are predicated of the spe-
cies but the species are not predicated reciprocally of
the genera). Hence for this reason too the species is
more a substance than the genus.
2b22. But of the species themselves-those which
are not genera-one is no more a substance than an-
other: it is no more apt to say of the individual man
that he is a man than to say of the individual horse that
it is a horse. And similarly of the primary substances
one is no more a substance than another: the individ-
ual man is no more a substance than the individual ox.
2b29. It is reasonable that, after the primary sub-
stances, their species and genera should be the only
other things called (secondary) substances. For only
they, of things predicated, reveal the primary sub-
stance. For if one is to say of the individual man what
he is, it will be in place to give the species or the
genus (though more informative to give man than ani-
mal); but to give any of the other things would be out
of place-for example, to say 'white' or 'runs' or any-
thing like that. So it is reasonable that these should
be the only other things called substances. Further, it
is because the primary substances are subjects for every-
thing else that they are called substances most strictly.
But as the primary substances stand to everything
else, so the species and genera of the primary sub-
stances stand to all the rest: all the rest are predicated
of these. For if you will call the individual man gram-
matical it follows that you will call both a man and
an animal grammatical; and similarly in other cases.
3a7. It is a characteristic common to every substance
not to be in a subject. For a primary substance is nei-
ther said of a subject nor in a subject. And as for sec-
ondary substances, it is obvious at once that they are
not in a subject. For man is said of the individual man
as subject but is not in a subject: man is not in the
individual man. Similarly, animal also is said of the
individual man as subject but animal is not in the in-
dividual man. Further, while there is nothing to pre-
vent the name of what is in a subject from being some-
times predicated of the subject, it is impossible for
the definition to be predicated. But the definition of
the secondary substances, as well as the name, is predi-
cated of the subject: you will predicate the definition
of man of the individual man, and also that of ani-
mal. No substance, therefore, i~ in a subject.
3a21. This is not, however, peculiar to substance; the
differentia also is not in a subject. For footed and two-
footed are said of man as subject but are not in a sub-
ject; neither two-footed nor footed is in man. More-
over, the definition of the ditferentia is predicated of
that of which the differentia is said. For example, if
D*
96 J. L. Ackril1
footed is said of man the definition of footed will also
be predicated of man; for man is footed.
3a29. We need not be disturbed by any fcar that
we may be forced to say that the parts of a substance,
being in a subject (the whole substance), are not sub-
stances. For when we spoke of things in a subject we
did not mean things belonging in something as parts.
3a33. It is a characteristic of substances and differ-
entiae that all things called from them are so called
synonymously. For all the predicates from them are
predicated either of the individuals or of the species.
(For from a primary substance there is no predicate,
since it is said of no subject; and as for secondary sub-
stances, the species is predicated of the individual, the
genus both of the species and of the individual. Simi-
larly, differentiae too are predicated both of the species
and of the individuals.) And the primary substances
admit the definition of the species and of the genera,
and the species admits that of the genus; for every-
thing said of what is predicated will be said of the sub-
ject also. Similarly, both the species and the individ-
uals admit the definition of the differentiae. But sy-
nonymous things were precisely those with both the
name in common and the same definition. Hence all
the things called from substances and differentiae are
so called synonymously.
3blO. Every substance seems to signify a certain
'this'. As regards the primary substances, it is indisputa-
bly true that each of them signifies a certain 'this'; for
the thing revealed is individual and numerically one.
But as regards the secondary substances, though it ap-
pears from the form of the name-when one speaks of
man or animal-that a secondary substance likewise
signifies a certain 'this', this is not really true; rather,
it signifies a certain qualification, for the subject is not,
as the primary substance is, one, but man and animal
are said of many things. However, it does not signify
simply a certain qualification, as white does. vVhite
signifies nothing but a qualification, whereas the spe-
cies and the genus mark off the qualification of sub-
stance-they signify substance of a certain qualification.
(One draws a wider boundary with the genus than with
the species, for in speaking of animal one takes in more
than in speaking of man.)
3b24. Another characteristic of substances is that
there is nothing contrary to them. For what would be
contrary to a primary substance? For example, there
is nothing contrary to an individual man, nor yet is
there anything contrary to man or to animal. This,
however, is not peculiar to substance but holds of many
other things also, for example, of quantity. For there
is nothing contrary to four-foot or to ten or to any-
thing of this kind-unless someone were to say that
many is contrary to few Or large to small; but still
there is nothing contrary to any definite quantity.
3b33. Substance, it seems, does not admit of a more
and a less. I do not mean that one substance is not
more a substance than another (we have said that it
is), but that any given substance is not called more,
or less, that which it is. For example, if this substance
is a man, it will not be more a man or less a man
either than itself or than another man. For one man
is not more a man than another, as one pale thing is
more pale than another and one beautiful thing more
beautiful than another. Again, a thing is called more,
or less, such-and-such than itself; for example, the body
that is pale is called more pale now than before, and
the one that is hot is called more, or less, hot. Sub-
stance, however, is not spoken of thus. For a man is
not called more a man now than before, nor is any-
thing else that is a substance. Thus substance does not
admit of a more and a less.
98 J. L. Ackri11
4alO. It seems most distinctive of substance that
what is numerically one and the same is able to re-
ceive contraries. In no other case could one bring for-
ward anything, numerically one, which is able to re-
ceive contraries. For example, a colour which is numer-
ically one and the same will not be black and white,
nor will numerically one and the same action be bad
and good; and similarly with everything else that is
not substance. A substance, however, numerically one
and the same, is able to receive contraries. For exam-
ple, an individual man-one and the same-becomes
pale at one time and dark at another, and hot and
cold, and bad and good. Nothing like this is to be seen
in any other case.
4a22. But perhaps someone might object and say that
statements and beliefs are like this. For the same state-
ment seems to be both true and false. Suppose, for ex-
ample, that the statement that somebody is sitting is
true; after he has got up this same statement will be false.
Similarly with beliefs. Suppose you believe truly that
somebody is sitting; after he has got up you will believe
falsely if you hold the same belief about him. However,
even if we were to grant this, there is still a difference in
the way contraries are received. For in the case of sub-
stances it is by themselves changing that they are able
to receive contraries. For what has become cold in-
stead of hot, or dark instead of pale, or good instead
of bad, has changed (has altered); similarly in other
cases too it is by itself undergoing change that each
thing is able to receive contraries. Statements and be-
liefs, on the other hand, themselves remain complete-
ly unchangeable in every way; it is because the actual
thing changes that the contrary comes to belong to
them. For the statement that somebody is sitting re-
mains the same; it is because of a change in the actual
thing that it comes to be true at one time and false
Notes on the Categories 99
at another. Similarly with beliefs. Hence at least the
way in which it is able to receive contraries-through
a change in itself-would be distinctive of substance,
even if we were to grant that beliefs and statements
are able to receive contraries. However, this is not true.
For it is not because they themselves receive anything
that statements and beliefs are said to be able to re-
ceive contraries, but because of what has happened to
something else. For it is because the actual thing exists
or does not exist that the statement is said to be true
or false, not because it is able itself to receive con-
traries. No statement, in fact, or belief is changed at
all by anything. So, since nothing happens in them,
they are not able to receive contraries. A substance, on
the other hand, is said to be able to receive contraries
because it itself receives contraries. For it receives sick-
ness and health, and paleness and' darkness; and be-
cause it itself receives the various things of this kind
it is said to be able to receive contraries. It is, there-
fore, distinctive of substance that what is numerically
one and the same is able to receive contraries. This
brings to an end our discussion of substance.
NOTES ON THE CATEGORIES
CHAPTER 1
lal. The word translated 'animal' originally meant just

that; but it had come to be used also of pictures or other
artistic represen,tations (whether representations of ani-
mals or not).
The tenns 'homonymous' and 'synonymous', as defined
by Aristotle in this chapter, apply not to words but to
things. Roughly, two things are homonymous if the same
name applies to both but not in the same sense, synony-
mous if the same name applies to both in the same sense.
Thus two things may be both homonymous and synony-
100 J. L. AckriII
mous-if there is one name that applies to both but not
in the same sense and another name that applies to both
in the same sense. From Aristotle's distinction between
'homonymous' and 'synonymous' one could evidently de-
rive a distinction between equivocal and unequivocal
names; but it is important to recognize from the start that
the Categories is not primarily or explicitly about names,
but about the things that names signify. (It will be neces-
sary in the translation and notes to use the word 'things' as
a blanket-term for items in any category. It often repre-
sents the neuter plural of a Greek article, pronoun, etc.)
Aristotle relies greatly on linguistic facts and tests, but his
aim is to discover truths about non-linguistic items. It is
incumbent on the translator not to conceal this, and, in
particular, not to give a misleadingly linguistic appearance
to Aristotle's statements by gratuitously supplying inverted
commas in all the places where we might feel that it is
linguistic expressions that are under discussion.
The contrast between synonyms and homonyms, be-
tween same definition and different definition, is obvious-
ly very crude. Elsewhere Aristotle recognizes that the dif-
ferent meanings of a word may be closely related. Thus
at the beginning of Metaphysics r 2 as he points out that
though the force of 'healthy' varies it always has a ref-
erence to health: a healthy person is one who enjoys
health, a healthy diet one which promotes health, a
healthy complexion one which indicates health. Similar-
ly, he says, with 'being': it is used in different ways when
used of things in different categories, but there is a pri-
mary sense (the sense in which substances have being) to
which all the others are related. Though the Categories
gives emphatic priority to the category of substance it
does not develop any such theory about the systematic
ambiguity of 'being' or 'exists'. Chapter 1 makes it seem
unlikely that Aristotle had yet seen the importance of
distinguishing between words that are straightforwardly
ambiguous and words whose various senses form a family
or have a common nucleus. (See Aristotle's suggestions
about 'good' at Nicomachean Ethics l096b26-28.)
lal2. 'Paronymous' is obviously not a term co-ordinate
Notes on tIle Categories 101
with 'homonymous' and 'synonymous', though like them
it is applied by Aristotle to things, not names. A thing
is paronymous if its name is in a certain way derivative.
The derivativeness in question is not etymological. Aris-
totle is not claiming that the word 'brave' was invented
after the word 'bravery'. He is claiming rather that 'brave'
means 'having bravery'; the brave is so called because of
(,from') the bravery he has. For an X to be paronymous
requires both that an X is called X because of something
(feature, property, etc.) which it has (or which some-
how belongs to it), and that 'X' is identical with the
name of that something except in ending. To say that
an X gets its name from something (or is called X from
something) does not necessarily imply that there is a
name for the something (loa32-b2), or that, if there is,
'X' has any similarity to that name (lob5-9)' But only
if these conditions are fulfilled does an X get its name
from something fJaronymously.
Paronymy is commonly involved when items in cate-
gories other than substance are ascribed to substances. If
we say that generosity is a virtue or that giving one's
time is a (kind of) generosity, we usc the name 'gener-
osity'; but if we wish to ascribe generosity to Callias we
do not say that he is generosity, but that he is generous
-using a word identical except in ending with the name
of the quality we are ascribing. Sometimes, indeed, the
name of an item in a category is itself used to indicate
the inherence of that item in a substance. In 'white is a
colour' 'white' names a quality; in 'Callias is white' 'white'
indicates the inherence of the quality in Callias. Here we
get homonymy or something like it, since the definition
of 'white' in the former sentence cannot be substituted
for 'white' in the latter: Callias is not a colour of a cer-
tain kind (2a29-34, 3a15-17). There are also the pos-
sibilities mentioned above: an adjective indicating the in-
herence of something in a substance may have no simi-
larity (or not the right kind of similarity) to the name
of the something, or there may be no name for the some-
thing. So the ascription of qualities, etc., to substances
does not always involve paronymy; but it very often does.
102 T. L. Ackri1I
The whole idea of an X's being called X from some-
thing (whether paronymously or not) is of importance in
the Categories. The categories classify things, not words.
The category of quality does not include the words 'gen-
erosity' and 'generous'; nor does it include two things cor-
responding to the two words. It includes generosity. 'Gen-
erosity' and 'generous' introduce the very same thing,
generosity, though in different ways, 'generosity' simply
naming it and 'generous' serving to predicate it. Aristotle
will frequently be found using or discussing distinctly
predicative expressions like 'generous', because though
they are not themselves names of items in categories they
serve to introduce such items (e.g. the item whose name
is 'generosity'). The person called generous is so called
from generosity.
CHAPTER 2
1316. What does Aristotle mean here by 'combination'

(literally, 'inter-weaving')? The word is used by Plato in
the Sophist 262, where he makes the point that a sen-
tence is not just a list of names or a list of verbs, but
results from the combination of a name with a verb; this
line of thought is taken up in the De Interpretatione
(16a9-18, 17aI7-2o). In the present passage Aristotle's
examples of expressions involving combination are both
indicative sentences, and his examples of expressions with-
out combination are all single words. Yet he ought not
to intend only indicative sentences (or only sentences)
to count as expressions involving combination. For in
Chapter 4 he says that every expression without combi-
nation signifies an item in some one category; this im-
plies that an expresion like 'white man' which introduces
two items from two categories is an expression involving
combination. Nor should he mean that all and only single
words are expressions lacking combination. For he treats 'in
the Lyceum' and 'in the market-place' as lacking combina-
tion (2al), while, on the other hand, a single word which
meant the same as 'white man' ought to count, in view of
Chapter 4, as an expression involving combination. There
seem to be two possible solutions. (a) The necessary and
sufficient condition-for an expression's being 'without com-
bination' is that it should signify just one item in some
category. The statement at the beginning of Chapter 4
is then analytic, but the examples in Chapter 2 are mis-
leadingly selective, since on this criterion a single word
could be an expression involving combination and a group
of words could be an expression without combination. (b)
The distinction in Chapter 2 is, as it looks, a purely lin-
guistic one between single words and groups of words (or
perhaps sentences). In Chapter 4 Aristotle neglects the
possibility of single words with compound meaning and
is indifferent to the linguistic complexity of expressions
like 'in the Lyceum'. Certainly he does neglect single
words with compound meaning in the rest of the Cate-
gories, though he has something to say about them in De
Interpretatione 5, 8, and 11.
la20. The fourfold classification of 'things there are'
relies on two phrases, 'being in something as subject' and
'being said of something as subject', which hardly occur
as technical terms except in the Categories. But the ideas
they express play a leading role in nearly all Aristotle's
writings. The first phrase serves to distinguish qualities,
quantities, and items in other dependent categories from
substances, which exist independently and in their own
right; the second phrase distinguishes species and genera
from individuals. Thus Aristotle's four classes are: (a)
species and genera in the category of substance; (b) indi-
viduals in categories other than substance; (c) species and
genera in categories other than substance; (d) individuals
in the category of substance.
Aristotle's explanation of 'in a subject' at 1a24-25 is
slight indeed. One point deserves emphasis. Aristotle does
not define 'in X' as meaning 'incapable of existing sep-
arately from X', but as a meaning 'in X, not as part of X,
and incapable of existing separately from what it is in'.
Clearly the 'in' which occurs twice in this definition can-
not be the technical 'in' of the definiendum. It must be
a non-technical 'in' which one who is not yet familiar
with the technical sense can be expected to understand.
104 J. L. AckrilI
Presumably Aristotle has in mind the occurrence in or-
dinary Greek of locutions like 'hcat in the water', 'cour-
age in Socrates'. Not all non-substances are naturally de-
scribed in ordinary language as ill substances, but we can
perhaps help Aristotle out by exploiting further ordinary
locutions: A is 'in' B (in the tcchnical sense) if and only
if (a) one could naturally say in ordinaryy language either
that A is in B or that A is of B or that A belongs to B
or that B has A (or that ... ), and (b) A is not a part of
B, and (c) A is inseparable from B.
The inseparability requirement has the consequence
that only individuals in non-substance categories can be
'in' individual substances. Aristotle could not say that gen-
erosity is in Callias as subject, since there could be gen-
erosity without any Callias. Only this individual generos-
ity-Callias's generosity-is in Callias. Equally, white is not
in chalk as subject, since there could be white even if
there were no chalk. White is in body, because every in-
dividual white is the white of some individual body. For
a property to be in a kind of substance it is not enough
that some or every substance of that kind should have
that property, nor necessary that every substance of that
kind should have it; what is requisite is that every instance
of that property should belong to some individual sub-
stance of that kind. Thus the inherence of a property in
a kind of substance is to be analysed in terms of the in-
herence of individual instances of the property in individ-
ual substances of that kind.
Aristotle does not offer an explanation of 'said of some-
thing as subject', but it is clear that he has in mind the
distinction between individuals in any category and their
species and genera. (Aristotle is willing to speak of species
and genera in any category, though, like us, he most often
uses the terms in speaking of substances.) He assumes
that each thing there is has a unique place in a fixed fam-
ily-tree. What is 'said of' an individual, X, is what could
be mentioned in answer to the question 'What is X?',
that is, the things in direct line above X in the family-
tree, the species (e.g. man or generosity), the genus (ani-
mal or virtue), and so on. Aristotle does not explicitly
argue for the view that there are natural kinds or that
a certain classificatory scheme is the one and only right
one.
It is often held that 'said of' and 'in' introduce notions
of radically different types, the former being linguistic or
grammatical, the latter metaphysical or ontological; and
that, correspondingly, the word translated 'subject' (lit-
erally, 'what underlies') means 'grammatical subject' in
the phrase 'said of a subject' and 'substrate' in 'in a sub-
ject'. In fact, however, it is perfectly clear that Aristotle's
fourfold classification is a classification of things and not
names, and that what is 'said of' something as subject is
itself a thing (a species or genus) and not a name. Some-
times, indeed, Aristotle will speak of 'saying' or 'predi-
cating' a name of a subject; but it is not linguistic items
but the things they signify which are 'said of a subject'
in the sense in which this expression is used in Chapter
2. Thus at 2a19 ff. Aristotle sharply distinguishes things
said of subjects from the names of those things: if A is
said of B it follows that the name of A, 'A', can be predi-
cated of B, though from the fact that 'A' is predicable
of something it does not follow that A is said of that
thing. At 2a31-34 Aristotle is careless. He says that white
is in a subject and is predicated of the subject; he should
have said that white is in a subject and its name is predi-
cated of the subject. But this is a mere s]j"p; the preced-
ing lines maintain a quite clear distinction between the
things that are said of or in subjects and the names of
those things. Being said of a subject is no more a lin-
guistic property than is being in a subject-though Aris-
totle's adoption of the phrase 'said of' to express the re-
lation of genus to species and of species to individual
may have been due to the fact that if A is the genus or
species of B it follows that 'A' can be predicated of B.
As regards 'subject', it is true that if virtue is said of
generosity as subject it follows that the sentence 'gener-
osity is (a) virtue' -in which the name 'generosity' is
the grammatical subject-expresses a truth. But 'virtue is
said of generosity as subject' is not about, and does not
mention, the names 'virtue' and 'generosity'. It would be
106 T. L. Ackri11
absurd to call generosity a grammatical subject: it is not
generosity but 'generosity' that can be a grammatical sub-
ject. Again, if A is in B as subject then B is a substance.
But this does not require or entitle us to take 'subject'
in the phrase 'in a subject' as meaning 'substance' or
'substrate'. It is the expressions 'said of' and 'in' (in their
admittedly technical senses) which bear the weight of
the distinctions Aristotle is drawing; 'subject' means nei-
ther 'grammatical subject' nor 'substance', but is a mere
label for whatever has anything 'said of' it or 'in' it. Thus
at 2b15 Aristotle explains his statement that primary sub-
stances are subjects for all the other things by adding
that 'all the other things are predicated of them or are
in them'.
The distinctions drawn in this chapter are made use
of mainly in Chapter 5 (on substance). In particular, it
is only in his discussion of substance that Aristotle ex-
ploits the distinction between individuals and species or
genera. He seems to refer to individuals in non-substance
categories at 4a1o ff., but they are not mentioned in his
chapters on these categories. \Vhy does Aristotle not speak
of primary and secondary qualities, etc., as he does of pri-
mary and secondary substances?
CHAPTER 3
Ib10. Aristotle affirms here the transitivity of the 'said
of' relation. He does not distinguish between the rela-
tion of an individual to its species and that of a species
to its genus. It does not occur to him that 'man' func-
tions differently in 'Socrates is (a) man' and '(a) man is
(an) animal' (there is no indefinite article in Greek).
Ib16. In the Topics (107b19 ff.) Aristotle gives this
principle about differentiae as a way of discovering am-
biguity. If sharpness is a differentia both of musical notes
and of solid bodies, 'sharp' must be ambiguous, since
notes and bodies constitute different genera neither of
which is subordinate to the other. At 144b12 ff. he argues
for the principle, saying that if the same differentia could
occur in different genera the same species could be in dif-
ferent genera, since every differentia 'brings in' its proper
genus. He goes on to water down the principle, allowing
that the same differentia may be found in two genera nei-
ther of which is subordinate to the other, provided that
both are in a common higher genus. In later works Aris-
totle preserves it as an ideal of classification and defini-
tion that the last differentia should entail all preceding
differentiae and genera, although he recognizes that in
practice we may fail to find such definitions and classifi-
cations (Metaphysics Z 12). In the Metaphysics Aristotle
is motivated by a desire to solve the problem of the 'unity
of definition' (De Interpretatione 17a13), but no such
interest is apparent in the TOPics and Categories. Here
he is probably influenced by the obvious cases of am-
biguity like 'sharp', and also by the evident economy of
a system of classification in which mention of a thing's
last differentia makes superfluous any mention of its genus.
Certainly the Categories gives no argument for the prin-
ciple here enunciated. The principle may help to explain
what Aristotle says about differentiae at 3a21-28, bl-9.
The last sentence probably requires emendation. As it
stands it is a howler, unless we take 'differentiae of the
predicated genus' to refer to differentiae that divide it into
sub-genera (differentiae divisivae) and 'differentiae of the
subject genus' to refer to differentiae that serve to de-
fine it (differentiae constitutivae). But there is nothing in
the context to justify such an interpretation. Only dif-
ferentiae divisivae are in question. A correct point, fol-
lowing naturally from what goes before, is obtained if
the words 'predicated' and 'subject' are transposed. That
Aristotle is willing to describe the differentiae of a genus
X as differentiae of the genus of X is clear; for he men-
tions two-footed as well as footed as a differentia of ani-
mal at 1b19, though the genus of which two-footed is an
immediate differentia is not animal but a sub-genus of
the genus animal.
108 J. L. Ackri1I
CHAPTER 4
First, some remarks about the translation. 'Substance':
the Greek word is the noun from the verb 'to be', and
'being' or 'entity' would be a literal equivalent. But in
connexion with categories 'substance' is the conventional
rendering and is used in the present translation every-
where (except in Chapter 1: 'definition of being'). 'Quan-
tity': the Greek is a word that serves both as an inter-
rogative and as an indefinite adjective (Latin quantum).
If Aristotle made use also of an abstract noun it would
be desirable to reserve 'quantity' for that; since he does
not do so in the Categories (and only once anywhere else)
it is convenient to allow 'quantity' to render the Greek
interrogative-adjective. 'Qualification': Aristotle does use
an abstract noun for 'quality' and carefully distinguishes
in Chapter 8 (e.g. loa27) between qualities and things
qualified (Latin qualia). So in this translation 'quality'
renders Aristotle's abstract noun, while his corresponding
interrogative-adjective is rendered by 'qualified' or 'qualifi-
cation'. 'A relative': Aristotle has no noun meaning 'rela-
tion'. 'A relative' translates a phrase consisting of a prep-
osition followed by a word which can function as the
interrogative 'what?' or the indefinite 'something'. In some
contexts the preposition will be rendered by 'in relation
to' or 'related to'. 'Where', 'when': the Greek words serve
either as interrogatives or as indefinite adverbs ('some-
where', 'at some time'). 'Place' and 'time' are best kept
to translate the appropriate Greek nouns, as at 4b24. 'Be-
ing-in-a-position', 'having', 'doing', 'being-affected': each
translates an infinitive (which can be used in Greek as
a verbal noun). The examples of the first two suggest that
Aristotle construes them narrowly (posture and apparel),
but the labels used are quite general. 'Being-affected is
preferred to alternative renderings because of the need to
use 'affected' and 'affection' later (e.g. 9a28 ff.) as trans-
lations of the same verb and of the corresponding noun.
The labels Aristotle uses for his ten categories are,
then, grammatically heterogeneous. The examples he pro-
ceeds to give are also heterogeneous. Man is a substance
and cutting is a (kind of) doing; but grammatical is not
a quality and has-shoes-on is not a kind of having. 'Gram-
matical' and 'has-shoes-on' are predicative expressions
which serve to introduce but do not name items in the
categories of quality and having.
How did Aristotle arrive at his list of categories?
Though the items in categories are not expressions but
'things', the identification and classification of these things
could, of course, be achieved only by attention to what
we say. One way of classifying things is to distinguish dif-
ferent questions which may be asked about something and
to notice that only a limited range of answers can be ap-
propriately given to any particular question. An answer
to 'where?' could not serve as an answer to 'when?'. Greek
has, as we have not, single-word interrogatives meaning
'of what quality?' and 'of what quantity?' (the abstract
nouns 'quality' and 'quantity' were, indeed, invented by
philosophers as abstractions from the familiar old inter-
rogatives); and these, too, would normally collect answers
from different ranges. Now Aristotle does not have a cate-
gory corresponding to every one-word Greek interrogative,
nor do all of his categories correspond to such interroga-
tives. Nevertheless, it seems certain that one way in which
he reached categorial classification was by observing that
different types of answer are appropriate to different ques-
tions. This explains some of his labels for categories and
the predicative form of some of his examples. The actual
examples strongly suggest that he thinks about answers
to questions about a man. Certainly he will have thought
of the questions as being asked of a substance. This is
why he often (though not in the Categories) uses the
label 'what is it' as an alternative to the noun 'substance'.
For what this question, when asked of a substance, gets
for answer is itself the name of a substance (cp. Cate-
gories 2b31). One must not, of course, suppose that in
so far as Aristotle is concerned to distinguish groups of
possible answers to different questions he is after all en-
gaged in a study of expressions and not things. That 'gen-
erous' but not 'runs' will answer the question 'of-what-
llO J. L. Ackrill
quality?' is of interest to him as showing that generosity
is a different kind of thing from running.
Alternatively, one may address oneself not to the var-
ious answers appropriate to various questions about a sub-
stance, but to the various answers to one particular ques-
tion which can be asked about any thing whatsoever-the
question 'what is it?' . We may ask 'what is Callias?',
'what is generosity?', 'what is cutting?'; that is, we may
ask in what species, genus, or higher genus an individual,
species, or genus is. Repeating the same question with
reference to the species, genus, or higher genus men-
tioned in answer to the first question, and continuing thus,
we shall reach some extremely high genera. Aristotle
thinks that substance, quality, etc., are supreme and ir-
reducibly different genera under one of which falls each
thing that thcre is. This approach may be said to classify
subject-expressions (capable of filling the gap in 'what
is , .. ?') whereas the previous one classified predicate ex-
pressions (capable of filling the gap in 'Callias is .. .'),
though, as before, the point for Aristotle is the classifica-
tion of the things signified by these expressions.
The only other place where Aristotle lists ten categories
is in another early work, the Topics (I 9). Here he starts
by using 'what is it' as a label for the category of sub-
stance. This implics the first approach, a classification de-
rived from grouping the answcrs appropriate to different
questions about some individual substance. But later in
the chapter the other approach is clearly indicated. It is
plain, Aristotle says, that 'someone who signifies what a
thing is sometimes signifies substance, sometimes quan-
tity, sometimes qualification, sometimes one of the other
predicates. For when a man is under discussion and one
says that what is being discussed is a man or is an ani-
mal, one is saying what it is and signifying substance;
whereas when the colour white is under discussion and one
says that what is being discussed is white or is a colour,
one is saying what it is and signifying qualification; sim-
ilarly, if a foot length is being discussed and one says that
what is being discussed is a foot length, one will be say-
ing what it is and signifying quantity'. In this passage,
where the question 'what is it?' is thought of as addressed
to items in any category, Aristotle can no longer use 'what
is it' as a label for the first category but employs the noun
for 'substance'. The whole chapter of the Topics deserves
study.
It is not surprising that these two ways of grouping
things should produce the same results: a thing aptly in-
troduced in answer to the question 'of-what-quality?' will
naturally be found, when classified in a generic tree, to
fall under the genus of quality. The two approaches in-
volve equivalent assumptions. The assumption that a given
question determines a range of answers that does not over-
lap with any range determined by a different question
corresponds to the assumption that no item when de-
fined per genus et differentiam will be found to fall un-
'der more than one highest genus. The assumption that
a certain list of questions contains all the radically differ-
ent questions that may be asked corresponds to the as-
sumption that a certain list of supreme genera contains
all the supreme genera. It should be noticed, however,
that only the second method gets individuals into cate-
gories. For one may ask 'what is it?' of an individual in
any category; but items introduced by answers to differ-
ent questions about Callias are not themselves individuals,
and a classification of such items will have no place for Cal-
lias himself or for CalIias's generosity. It has, indeed, been
suggested that individuals have no right to a place in
Aristotle's categories because the Greek word transliter-
ated 'category' actually means 'predication' or 'predicate'
(it is in fact so rendered in this translation, e.g. lOb.u).
However, it is substance, quality, quantity themselves
which are the 'categories', that is, the ultimate predicates;
items belonging to some category need not be items which
can themselves be predicated, they are items of which that
category can be predicated. Thus the meaning of 'cate-
gory' provides no reason why Callias should not be given
a place in a category, nor why non-substance individuals
should be left out.
Some general points: (1) Aristotle does not give argu-
ment to justify his selection of key questions or to show
112 J. L. Ackri11
that all and only the genera in his list are irreducibly dif-
ferent supreme genera. When speaking of categories in
other works he commonly mentions only three or four
or five (which nearly always include substance, quantity,
and quality), but often adds 'and the rest'. In one place
he does seek to show that 'being' cannot be a genus, that
is. in effect, that there must be irreducibly different kinds
of being (Metaphysics 998b22). (2) Aristotle does not
seem to doubt our ability to say what answers would be
possible to given questions or to determine the correct
unique definitions per genus et differentiam of any item
we consider. When he looks for features peculiar to a
given category (4alO, 6a26, 1la15) he does not do this
to suggest criteria for categorial classification; his search
presupposes tllat we already know what items fall into the
category in question. He assumes also that we can tell
which words or expressions signify single items rather
than compounds of items from different categories. He
does not explain the special role of words like 'species',
'predicate', etc., nor warn us against treating them, like
'animal' or 'generosity', as signifying items in categories.
(3) Aristotle does not adopt or try to establish any sys-
tematic ordering of categories. Substance is, of course,
prior to the rest; and he argues in the Metaphysics
(1088a22) that what is relative is farthest removed from
substance. (4) Aristotle does not in the Categories in-
dicate the value of the theory of categories either for deal-
ing with the puzzles of earlier thinkers or for investigating
new problems. Nor does he, as elsewhere, develop the
idea that 'is', 'being', etc. have different (though con-
nected) senses corresponding to the different categories
(Metaphysics l017a22-30, l028alO-2o, l03oa17-27;
Prior Analytics 49a7 ) .
CHAPTER 5
2all. The terms 'primary substance' and 'secondary sub-
stance' are not used in other works of Aristotle to mark
the distinction between individual substances and their
species and genera, though the distinction itself is, of
course, maintained. The discussion of substance in Meta-
physics Z and H goes a good deal deeper than does this
chapter of the Categories. Aristotle there exploits the
concepts of mattcr and form, potentiality and actuality,
and wrestles with a whole range of problems left un-
touched in the Categories.
Aristotle characterises primary substance by the use of
tem1S introduced in Chapter 2. But he does not, as might
have been expected, go on to say that secondary sub-
stances are things said of a subject but not in any sub-
ject. Instead he describes them as the species and genera
of primary substances and only later makes the point that
they are said of primary substances but not in any sub-
ject. The reason for this may be that he is going to say
(surprisingly) that the differentiae of substance genera,
though not themselves substances, are nevertheless said of
the individuals and species in the genera, and are not in
them.
'Called a substancc most strictly, primarily, and most
of all': does Aristotle mean to suggest that 'substance' is
used in two different senses? It would be difficult for-him
to allow that without upsetting his whole scheme of cate-
gorial classification. Aristotle is no doubt aware that the
distinction between primary and secondary substances is
not like that between two categories or that between two
genera in a category; 'Callias is a primary substance' is
unlike both 'Callias is a man' and 'Callias is a substance'.
But he fails to say clearly what type of distinction it is.
2a19. '\Vhat has been said' presumably refers to IbID-
15, which is taken to explain why, if A is said of B, not
only the name of A but also its definition will be predi-
cable of B. The first part of the paragraph is important
as showing very clearly that the relation 'said of . . . as
subject' holds bctween things and not words. The fact
that A is said of B is not thc fact that 'A' is prcdicable
of B. The fact that A is said of B is not even the fact
that both 'A' and the definition of A are predicable of B.
This is a fact about language that follows from that fact
about the relation between two things.
The second part of the paragraph is also of importance.
114 ,. L. AckriII
It shows that Aristotle recognizes that, for example, 'gen-
erosity' and 'generous' do not serve to introduce two dif-
ferent things (we should say 'concepts'), but introduce
the same thing in two different ways. In saying that usual-
ly the name of what is in a subject cannot be predicated
of the subject he obviously means more than that, for
example, one cannot say 'Callias is generosity'. He means
that there is something else which one does say-'Callias
is generous'-by way of ascribing generosity to Callias.
His point would be senseless if 'generous' itself were just
another name of the quality generosity or if it were the
name of a different thing altogether.
zaH. Someone might counter the claim in the first sen-
tence by pointing out that, for example, animal is said
of man and colour is in body, and man and body are
secondary substances. Aristotle therefore examines just
such cases. It is somewhat surprising that he says: 'were
it predicated of none of the individual men it would not
be predicated of man at all.' For in view of the meaning
of 'said of' he could have made the stronger statement:
'were if not predicated of all of the individual men .. .'.
However, what he does say is sufficient for the final con-
clusion he is driving at, that nothing else could exist if
primary substances did not. As for colour, Aristotle could
have argued to his final conclusion simply by using the
definition of 'in' together with tlle fact, just established,
that the existence of secondary substances presupposes the
existence of primary substances: if colour is in body it
cannot exist if body does not, and body cannot exist if
no individual bodies exist. What is Aristotle's own argu-
ment? It was suggested earlier that to say that colour is
in body is to say that every instance of colour is in an
individual body. If so, Aristotle's present formulation is
compressed and careless. For he does not mention indi-
vidual instances of colour; he speaks as if, because colour
is in body, colour is in an individual body. Strictly, how-
ever, it is not colour, but this individual instance of col-
our, that is in this individual body; for colour could exist
apart from this body (though this instance of colour could
not). Aristotle's use of a relaxed sense of 'in' may be COD-
nected with his almost complete neglect, after Chapter
2, of individuals in non-substance categories.
In drawing his final conclusion in the last sentence
Aristotle relies partly on the definition of 'in' (' ... can-
not exist separately . . .'); partly on the principle that if
A is said of B, A could not exist if B did not. The closest
he comes to arguing for this principle is at 3blO-23,
where he insists that secondary substances are just kinds
of primary substance.
Aristotle's conclusion is evidently intended to mark out
primary substances as somehow basic (contra Plato). But
the point is not very well expressed. For it may well be
doubted whether (Aristotle thinks that) primary sub-
stances could exist if secondary substances and items in
other categories did not do so. But if the implication of
existence holds both ways, from the rest to primary sub-
stances and from primary substances to the rest, the state-
ment in the last sentence of his paragraph fails to give a
special status to primary substances.
2b7. The two arguments given for counting the species
as 'more a substance' than the genus-for carrying into
the class of secondary substances the notion of priority
and posteriority already used in the distinction between
primary and secondary substances-come to much the
same. For the reason why it is more informative (2blO)
to say 'CaIlias is a man' than to say 'CaIIias is an ani-
mal' (though both are proper answers to the 'what is it'
question, zb31-37) is just that the former entails the
latter but not vice versa: 'the genera are predicated of
the species but the species are not predicated reciprocally
of the genera' (zb2o). The point of view is different at
15a4-7, where it is said that genera are always prior to
species since they do not reciprocate as to the implication
of existence: 'if there is a fish there is an animal, but if
there is an animal there is not necessarily a fish'. For
this sense of 'prior' see 14a29-35.
zbz9' Here the connexion between the 'what is it' ques-
tion and the establishment of categorial lines is made
very clear.
The second argument (from 'Further, it is be-
116 J. L. Ackri11
cause .. .') is compressed. Primary substances are subjects
for everything else; everything else is either said of or in
them (2a34, 2b15). Aristotle now claims that secondary
substances are similarly related to 'all the rest', that is,
to all things other than substances. This must be because
all those things are in secondary substances. All Aristotle
says, to establish this, is that 'this man is grammatical'
entails 'a man is grammatical'. He means to imply that
any non-substance that is in a primary substance is nec-
essarily in a secondary substance (the species or genus of
the primary substance). Since he has already argued that
all non-substances are in primary substances he feels en-
titled to the conclusion that all non-substances are in sec-
ondary substances. But it will be seen that a further re-
laxation in the sense of 'in' has taken place. It is now
implied, not only that generosity can be described as in
Callias (though generosity could certainly exist in the ab-
sence of Callias), but also that generosity can be de-
scribed as in man simply on the ground that some one
man is generous (and not, as it strictly should be, on the
ground that all instances of generosity are in individual
men).
3a7. Why is it 'obvious at once' that secondary sub-
stances are not in primary substances? It is not that they
can exist separately from primary substances (2a34-b6).
Nor does Aristotle appear to rely on the fact that a given
secondary substance can exist separately from any given
individual, that there could be men even if Callias did
not exist, so that the species man can exist separately
from Callias and is, therefore, not in him. Aristotle seems
rather to be appealing to the obvious impropriety in or-
dinary speech of saying such a thing as 'man is in Cal-
lias'. It was suggested in the note on Ia24-25 that Aris-
totle made it a necessary condition of A's being in B that
it should be possible to say in ordinary non-technical dis-
course such a thing as 'A is in B' ('belongs to B', etc.).
Now Aristotle is pointing out that this condition is not
satisfied in the case of man and Callias. If this is his point
he could have extended it to other categories; no genus
or species in any category can naturally be described as
Notes on tIle Categories 117
in (or belonging to or had by) any subordinate genus,
species or individual. What distinguishes secondary sub-
stances from non-substance genera and species is not that
they are not in the individuals, species, and genera sub-
ordinate to them but that they are not in any other in-
dividuals, species, or genera; virtue is not in generosity,
but it is in soul, whereas animal is not in man and not
in anything else either.
One cannot say 'hero is in Ca11ias' or 'father is in Cal-
lias'; but if Ca11ias is a hero and a father the definition
of 'hero' and 'father' can also be predicated of him. So
it might be suggested that the considerations advanced
by Aristotle in this paragraph imply that hero and fa-
ther are secondary substances. But Aristotle is not claim-
ing that any predicate-word which can be replaced by its
definition is the name of a secondary substance (or dif-
ferentia of substance, see below), but that a predicate-
word can be replaced by the definition of the item it
introduces if and only if the item is a secondary sub-
stance (or differentia of substance). 'Generous' can be re-
placed by definitions of 'hero' and 'father', but not by
definition of the item which 'generous' introduces, the
quality generosity. Similarly, 'hero' and 'father' can be
replaced by definitions of 'hero' and 'father', but not
by definitions of the items they serve to introduce, her-
oism and fatherhood. Aristotle gives no explicit rules
for deciding which common nouns stand for species
and genera of substance (natural kinds) and which
serve only to ascribe qualities, etc., to substances. He
would presumably rely on the 'what is it' question to
segregate genuine names of secondary substances from
other common nouns; but the question has to be taken
in a limited or loaded sense if it is always to collect only
the sorts of answer Aristotle would wish, and an un-
derstanding and acceptance of the idea of natural
kinds is therefore presupposed by the use of the ques-
tion to distinguish the names of such kinds from other
common nouns which serve merely to ascribe qualities,
etc. Surely it would often be appropriate to say 'a cob-
bler' in answer to the quesion 'What is Callias?'.
118 J. L. Ackrill
3a21. The statement that something that is not sub-
stance is nevertheless said of substance is a surprising
one, which can hardly be reconciled with the scheme of
ideas so far developed. If the differentia of a genus is not
a substance (secondary substances being just the species
and genera of substance), it ought to belong to some other
category and hence be in substance. That an item in one
category should be said of an item in another violates the
principle that if A is said of Band B of C then A is said of
C. Aristotle, indeed, positively claims that the definition
as well as the name of a differentia is predicable of the
substance falling under it, but this too seems very strange.
In a definition {Jer genus et differentiam the differentia is
commonly expressed by an adjective (or other non-sub-
stantive), and this should surely be taken to introduce an
item named by the corresponding substantive (as 'gen-
erous' introduccs but is not the name of generosity). If
we say that man is a rational animal 'rational' brings in
rationality, but ncither the name nor the definition of
rationality can be predicated of man. 11111s the differen-
tiating property satisfies a test for being in substance (cp.
2a19-34) .
Aristotle is no doubt influenced by the following facts.
( 1) Species and genera of individual substances are them-
selves called substances bccause 'if one is to say of the
individual man what he is, it will be in place to give the
species or the gcnus' (2b32). If we now consider the
question 'what is (a) man?' wc shall be strongly inclined
to mention not only the genus animal but also the ap-
propriate differentia. The differentia secms to be (Jart of
the 'what is it' of a secondary substance, and this pro-
vides a strong motive for assimilating it to substance even
while distinguishing it from species and genera. (2) The
principle enunciated at 1b16 implies that mention of a
differentia-words words which function naturally in Greek
true classification of things) any mention of the genus.
To ascribe the differentia 'two-footed' to man is as good
as to say that he is a two-footed land animal. Thus the
differentia is, in a way, the whole of the 'what is it' of
a secondary substance. (3) Aristotle uses as examples of
Notes 011 the Categories 119
differentia-words words which function naturally in Greek
as nouns (though they are strictly neuter adjectives). At
14b33-15a7 he uses the same words when speaking ex-
plicitly of species (and so they are translated. there by
'bird', 'beast' and 'fish'). Moreover, there arc in Aristotle's
vocabulary no abstract nouns corresponding to these neu-
ter adjectives (as 'footedness', 'two-footedness'). Such
facts arc far from establishing that the definition as well
as the name of a differentia is predicated of substances.
For not all differentiae are expressed by nouns or words
used as nouns, and abstract nouns corresponding to dif-
ferentia-words are not always lacking. In any case, there
are plenty of nouns (like 'hero') which Aristotle would
insist on treating as mere derivatives from the names of
the things they introduce ('heroism'); and the fact that
there is no name for, say, a quality does not exclude the
possibility that some predicative expression serves to as-
cribe that quality (though not, of course, paronymously:
loap-b5). Thus, that 'footed' is (used as) a noun
and no noun 'footedness' exists is not a justification for
refusing to treat 'footed' in the same kind of way as 'hero'
or 'generous', as introducing a characteristic neither the
name nor the definition of which is predicable of that
which is footed. Nevertheless, the above features of the
examples he hit upon may have made it somewhat easier
for him to say what he does about differentiae without
feeling the need for full explanation. For deeper discus-
sion of the relation of differentia to genus, and of the
connected problem of the unity of definition (referred
to at De Interpretatione 17aI3), see espccially Metaphys-
ics Z 12.
3a33. 'All things called from them are so called synon-
ymously': Aristotle is not denying that there are words
which stand ambiguously for either of two kinds of sub-
stance (like 'animal' in Chapter 1). Things to which
~uch a word applied in one sense would not be 'called
from' the same substance as things to which it applied
in the other sense; and Aristotle is claiming only that all
things called from any given substance are so callcd syn-
E
120 J. L. Ackrill
onymously, not that all things called by a given substance-
word are necessarily so called synonymously.
Aristotle is drawing attention again to the following
point (it will be convenient to assume that there is no
sheer ambiguity in the words used). There are two ways
in which something can be called from the quality virtue:
generosity is a virtue, Callias is virtuous; neither the name
nor the definition of virtue is predicable of Callias. There
arc two ways in which something can be called from the
quality white: Della Robbia white is (a) white, this paper
is white; the name but not the definition of white is
predicable of this paper. There is only one way in which
something can be called from man: Callias is a man,
Socrates is a man, and so on; both the name and the
definition of man are predicable of Callias and Socrates
and so on.
It is not quite clear that Della Robbia white and this
paper are homonymous with respect to the word 'white',
in the meaning given to 'homonymous' in Chapter 1. For
there the case was that the word (e.g. 'anima]') stood in
its two uses for two different things with two different
definitions. Now, however, we have 'white' in one use
standing for a thing (a quality) which has a certain defi-
nition, but in the other use not standing for a different
thing with a different definition but introducing differ-
ently the very same thing. However, an easy revision of
the account in Chapter 1 would enable one to say that
'synonymously' in the present passage contrasts with both
'homonymously' and 'paronymously': most non-substances
(like generosity) generate paronymy, a few (like the qual-
ity white) generate homonymy; no substance generates
either.
'From a primary substance there is no predicate': there
is no subject of which Callias is said or in which Callias
is. In the Analytics Aristotle speaks of sentences in which
the name 'Callias' is in the predicate place, and says that
this is only accidental predication (43a34; cpo 83al-23).
He does not make any thorough investigation of the dif-
ferent types of sentence in which a proper name may oc-
cur in the predicate place. Nor does he discuss such uses
Notes On the Categories 121
as 'he is a Socrates', 'his method of argument is Socratic'.
He would no doubt allow that these are cases of genuine
predication but deny that the predicates are 'from a pri-
mary substance': the connexion between the characteris-
tics ascribed by '. . . is a Socrates' and '. . . is Socratic'
and the individual Socrates is purely historical and con-
tingent; we should not have used '. . . is a Socrates' as
we do if there had been no Socrates or if Socrates had
had a different character, but we could perfectly well have
used a different locution to ascribe the very same char-
acteristics. A similar answer would be available if some-
one claimed that there are after all two ways in which
something may be called from a secondary substance since
while Tabitha is a cat Mrs. So-and-so is catty. It is be-
cause of real or assumed characteristics of cats that the
word 'cattiness' names the characteristics it does; but the
characteristics themselves could have existed and been
talked about even if there had never been any cats.
3blO. Aristotle has contrasted individual substances
with their species and genera. He has labelled the latter
'secondary' and has argued that their existence presup-
poses that of primary substances. Nevertheless, much that
he has said provides a strong temptation to thing of spe-
cies and genera of substance as somehow existing in their
own right like Platonic Forms. In the present passage
Aristotle tries to remedy this. It is careless of him to
speak as if it were substances (and not names of sub-
stances) that signify. More important, it is unfortunate
that he draws the contrast between a primary substance
and a secondary substance by saying that the latter sig-
nifies a certain qualification. For although he immediate-
ly insists that 'it does not signify simply a certain qual-
ification, as white does', yet the impression is conveyed
that secondary substances really belong in the category
of quality. This, of course, Aristotle does not mean. 'Qual-
ity of substance' means something like 'kind' or 'charac-
ter of substance'; it derives from a use of the question
'of what quality?' different from the use which serves to
classify items as belonging to the category of quality. 'Of
what quality is Callias?' (or 'what kind of person is Cal-
122 T. L. AckriI1
lias?') gets answers from the category of quality. But 'what
quality of animal is CaIlias?' (or 'what kind of animal is
CalIias?') asks not for a quality as opposed to substance,
quantity, etc., but for the quality-of-animal, the kind-of-
animal. It is a result of the limitations of Aristotle's vocab-
ulary that he uses the same word as a category-label and
to convey the idea of a kind, sort or character of so-and-
so. (Cp. Metaphysics 1020a33 - bl, 1024bS-6, where
'quality' refers to the differentia-in any category-not to
the category of quality.) It is also clear that he is at a
disadvantage in this passage through not having at his
disposal such terms as 'refer', 'describe', 'denote', 'con-
note'; and that he would have been in a better position
if he had from the start examined and distinguished var-
ious uses of expressions like '(a) man' instead of embark-
ing at once upon a classification of 'things there are'.
3b24. Aristotle raises the question of contrariety in
each of the categories he discusses. On the suggestion that
large and small are contraries see Sbll - 6a11.
3b33. The question of a more and a less is raised in
each category. 'We have said that it is': 2b7. There is
a certain ambiguity in 'more', since to say that a species
is more a substance than a genus is to assign it some
sort of priority but not to ascribe to it a higher degree
of some feature as one does in saying that this is more
hot than that.
The point Aristotle makes here about substances ap-
plies also, of course, to sorts which he would not recog-
nize as natural kinds: one cobbler or magistrate is not
more a cobbler or magistrate than another.
4a1o. What Aristotle gives here as distinctive of sub-
stance is strictly a characteristic of primary substances. For
he is not speaking of the possibility of man's being both
dark and pale (of there being both dark men and pale
men), but of the possibility of one and the same indi-
vidual man's being at one time dark and at another time
pale. (It will then be distinctive of secondary substances
that the individuals of which they are said arc capable
of admitting opposites.) Correspondingly, Aristotle must
be meaning to deny, not that species and genera in other
categories may in a sense admit contraries (colour may
be white or black), but that individual instances of qual-
ities, etc., can admit contraries while retaining their iden-
tity. His first example is not convincing. An. individual
instance of colour will necessarily be an instance of some
specific colour and will be individuated accordingly: if X
changes from black to white we first have X's blackness
and then X's whiteness, two individuals in the category
of quality. (To this there corresponds the fact that one
and the same individual substance cannot move from one
species to another.) What is required is to show-not that
X's blackness cannot retain its identity while becoming
white, but-that X's blackness cannot retain its identity
while having contrary properties at different times. The
sort of suggestion Aristotle ought to rebut is, for exam-
ple, the suggestion that one and the same individual in-
stance of colour could be at one time glossy and at an-
other matt, this variation not making it count as different
instances of colour. Aristotle's second example is of the
right kind, since the goodness or badness of an action
does not enter into the identity-criteria for an individual
action in the way in which the shade of colour does en-
ter into the identity-criteria for an individual instance of
colour. However, the example is still particularly favour-
able for him. For 'good' and 'bad' are commonly used
to appraise an action as a whole, and for this reason one
would not speak of an action as having been good at
first and then become bad. There are clearly very many
cases which it would be less easy for Aristotle to handle
(cannot an individual sound sustain change in volume
and tone?). The question demands a fuller scrutiny of
cases and a more thorough investigation of usage than
Aristotle attempts. It would seem that the power to ad-
mit contraries is not peculiar to individual substances but
is shared by certain other continuants, so that a further
criterion is required to explain why these others are not
counted as substances.
4a22. Aristotle of course trcats the truth and falsity of
statements and beliefs as their correspondence and lack
of correspondence to fact (4b8, 14b14-22; Metaphysics
124 J. L. Ackril1
l051b6-<)). Here he first points out that it is not through
a change in itself that a statement or belief at one time
true is at another time false, whereas an individual sub-
stance itself changes; so that it remains distinctive of pri-
mary substances that they can admit contraries by chang·
ing. He next argues (4b5) that strictly a thing should be
said to admit contraries only it it does itself undergo a
change from one to the other; so that, strictly speaking,
it is not necessary to qualify what was said at 4a1Q-1l:
only individual substances can admit contraries.
Aristotle might have argued that the alleged counter-
examples. individual statements or beliefs which change
their truth-value, fail, because my statement now that
Callias is sitting and my statement later that Callias is
sitting are not the same individual statement even if they
are the same statement (just as 'a' and 'a' are two indi-
vidual instances of the same letter). Thus they are not
examples of the very same individual admitting contraries.
Alternatively, Aristotle could have denied that the state-
ment made by 'CaIIias is sitting' when uttered at one
time is the same statement as that made by 'CaIIias is
sitting' when uttered at another time. The sameness of
a statement or belief is not guaranteed by the sameness
of the words in which it is expressed; the time and place
of utterance and other contextual features must be taken
into account.
ARISTOTLE'S THEORY OF CATEGORIES
J. M. E. MORAVCSIK
In several of his writings Aristotle presents what

came to be known as a "list of categories." The pres-
entation of a list, by itself, is not a philosophic theory.
This paper attempts a few modest steps toward an un-
derstanding of the theory or theories in which the list
of categories is embedded. To arrive at such under-
standing we shall have to deal with the following ques-
tions: \Vhat classes of expressions designate items each
of which falls under only one category? What is the
list a list of? and \Vhat gives it unity? To show this
to be a worthwhile enterprise, let us consider a few
passages in which the list of categories is introduced
or mentioned.
In Topics 103b20 If. the list is introduced as con-
taining certain kinds (gem!) within which one can find
the accidents, genus, properties, and definition of any-
thing. Thus apparently all (simple?) elements of the
nature of any entity are to be found in one of the cate-
gories. In Metaphysics 1028alO If. we are told that 'is'
has as many senses as there are categories. Thus we
see that Aristotle analyzes the ambiguity of 'is' with
the list of categories as an assumed background. What
he takes to be the systematic ambiguity of 'is' pro-
vides one of the cornerstones of his metaphysical spec-
ulations. In Physics 225b5 If. Aristotle analyzes the con-
cept of kinesis and concludes that instances of this can
be found in three of the categories.
Each of these passages presents problems; some of
these will be taken up later in this paper. This brief
preliminary survey is intended only to show that the
list of categories plays an important role in several of
Aristotle's theories, and thus it is reasonable to assume
that the list has constitutive principles and unity.
I
Which parts ot language designate an item in one
ot the categories? In Categories Ib25 ff. the list of cate-
gories is introduced as containing the kinds (e.g. qual-
ity, quantity) of those items (e.g. white, grammatical,
three cubits long) which are signified by "things said
without any combination" (Ackrill's translation, con-
tained in this volume). Chapters 2 and 4 of the Cate-
gories taken together make it quite clear that the
"things" in question are linguistic items. Our first task
is to determine which parts of language Aristotle is
referring to in the passage under consideration. Earlier,
in la16 ff. we read that "Of things that are said, some
involve combination while others are said without com-
bination. Examples of those involving combination are
'man runs', 'man wins'; and of those without combi-
nation 'man', 'ox', 'runs', 'wins',"l In view of the mea-
gerness of the examples, the key term in this account
is "combination." As Ackrill,2 and a century ago Tren-
delenburg,3 pointed out, the Greek term sumploke
used here by Aristotle had been previously used by
IJ. L. AckriII's translation here is far superior to the old Oxford
version, which renders the first part of this passage as: "Forms of
speech are either simple or composite." This suggests, wrongly,
that Aristotle is saying something about all forms of speech, and,
equally wrongly, that the distinction to be drawn is a surface
distinction between what is syntactically simple or analyzable.
2J. L. Ackrill, Aristotle's Categories and De InferpTetatione
(Oxford, 1963), p. 73. (See earlier in this volume p. 101.)
3A. Trendelenburg, Geschichfe deT KategoTienlehTe (Hilde-
sheim: Olms, 1846 [1963]), p. II.
Aristotle's Theory of Categories 127
Plato to refer not to mere conjunction or juxtaposi-
tion, but rather to the interweaving of words and
phrases into sentences.4 This would suggest that the
uncombined elements are parts of language from which
sentences can be formed. This is confirmed by 2a4 ft.,
where Aristotle says that the combination of these
items produces a true or false sentence. (Aristotle re-
gards sentences as the bearers of truth-value.) There is
an interesting parallel to this passage in Topics 101b26-
28, according to which the key elements in a statement
are property, genus, definition, and accident, and it is
emphasized that none of these by themselves make up
a statement. Thus as a reasonable first approximation
we can say that Aristotle is interested in potential ele-
ments of sentences that are true or false, or definitions.
One is tempted to add the qualification: "sentences
of subject-predicate form," for neither in the examples
given nor in the subsequent discussions are sentences
expressing identity or existence treated. Their inclusion
in the discussion would raise some interesting questions
about the extent to which these types of sentence
could be regarded as "interweaving," and about the
senses of 'is' involved. The addition of this qualifica-
tion would be in harmony with Trendelenburg's re-
mark that Aristotle in this context seems to have in
mind what Kant called judgments,5 (a notion which
carries similar limitations).
Having specified the relevant class of sentences, we
should investigate whether the Aristotelian theory en-
tails that every element of such a sentence designates
an item, and whether every element that designates is
supposed to designate an item falling into only one of
40n Plato's views on this see J. M. Moravcsik, "Sump/oke
Eidoon," Archiv filr die Geschichte der Philosophie, XLII
(1960) 117-29; also "Being and Meaning in the Sophist," Acta
Philosophica Fennica, XIV (1962), especially pp. 61-65.
5Trendelenburg, op. cit., p. 12.
the categories. An affirmative answer to either ques-
tion would place the theory in jeopardy. There is, how-
ever, evidence that:
(i) Aristotle does not think that every word or phrase
which could be part of a sentence of subject-predicate
form has the function of designating an entity.
As both Steinthal and Trendelenburg remarked,6
Aristotle does not ascribe the same type of significance
to every word within a sentence, and he does not think
that every word has a designative role. In Poetics
1456b38 ff. he separates from nouns and verbs (to
which he also assimilates adjectives) the so-called con-
nectors or auxiliary expressions. These are said to in-
clude particles and prepositions, but in order to com-
plement the theory of categories they ought to include
a great deal more: logical constants, articles, and the
ordinary language equivalents of quantifiers must also
be members of this class. There is evidence to support
the view that Aristotle intended the connectors to in-
clude a great deal more than he explicitly mentions.
For one thing, what he says about them applies to a
much larger group than his examples would suggest,
for he says that these expressions (a) do not designate
and (b) do not contribute to the content of a new
(larger) linguistic unit, i.e. the sentence. These criteria
are interesting, for they show that Aristotle did not
mark off the connectors on purely syntactic grounds.
The criteria can be taken as foreshadowing the char-
acterization of what were later called the syncategore-
matic expressions. In any case, these conditions would
allow all the above-mentioned types of expression to
qualify. It is impossible to say how many of these Aris-
totle had in mind when he wrote the Poetics, but the
6Trendelenburg, op. cit., pp. 24 If. and H. Stein thaI, Ge·

schichte der Sprachwissenschaft (Hildesheim: Olms, 1890 [1961 j),
Vol. I, pp. 263 If.
fluidity of the classification is witnessed by Rhetoric
III 5 where-as SteinthaI saw-Aristotle blends arti-
cles, pronouns, and conjunctives into one class. Stein-
thaI's interpretation of what Aristotle thought of as
the main parts of speech is of further relevance here,
for it is claimed that apart from nouns and verbs Aris-
totle recognized only one more class, i.e. the connec-
tors.7 None of this is direct evidence, but taken all
together it renders plausible the interpretation of the
class of connectors on the suggested broad basis.
With this we have narrowed down the candidates
for the "uncombined elements" to the members of the
class of designative expressions which can be elements
of those sentences that are definitions or are of sub-
ject-predicate form. Still, further restrictions are ob-
viously needed. Fortunately there is evidence to sup-
port additional qualifications.
(ii) Aristotle does not think that every noun or verb
designates an item in one of the categories.
In De Interpretatione 16a13 if. Aristotle lists as a
necessary condition for the production of what is true
or false the combination of verb and noun, and in this
context treats "being" as a verb. We see, however, from
his treatment of being in the Categories and the Meta-
physics (see the reference in the beginning of this
paper) that he does not construe being as falling into
only one of the categories. His claim that being is not
a genus, and the argument backing up this claim,8 are
sufficient ground on which to base the interpretation
that 'being'-and similar terms like 'same', 'one', etc.-
have either no designative role or a divided designative
role.
7Steinthal, op. cit., pp. 260 ff.

8The most profound discussion of this argument is to be found
in John Cook Wilson's paper "Categories in Aristotle and in
Kant," Statement and Inference, Vol. II, pp. 696-706 [above, pp.
75-89].
Before we consider further qualifications, let us look
briefly at the interesting status that definitions had for
Aristotle. It is not clear whether he regarded definitions
as true or false, and whether they counted as instances
of combination. The passage quoted from the Topics
above casts doubt on their counting as combinations.
Now in De Interpretatione, chapters 4 and 5, he argues
that what is true or false must contain such interweav-
ing or combination. Again in Poetics 1457a25 fl. Aris-
totle says that, though definitions are sentences, they
need not contain a verb. This passage, however, can
be construed as saying only that no verb need appear
explicitly in a definition; the same is true in Greek for
many other types of sentences. Again in De Anima III
6 it is pointed out that a definition does not involve a
mental task of synthesis as an assertion does. All of
this together supports the interpretation according to
which Aristotle did not regard definitions as true or
false or as produced by combination. Nevertheless, we
must assume that he thought of parts of definitions
as falling under one category.
(iii) There are some noun- or verb-phrases (other
than words like 'being', etc.) that designate items fall-
ing under more than one category.
An obvious example of a phrase designating a com-
plex that spans two categories is the expression 'in-
continent man'. In order to rule out such cases we have
to introduce further qualifications. The restriction to
be put forward here is not backed by direct textual
evidence, but it is supported by what we took to be
the significance of sumploke and the proposed broad
delineation of the class of connectors: it is to rule out
as not completely uncombined all those phrases which
the mere addition of connectors can transform into
a sentence. For though 'red colored' and 'incontinent
man' are not sentences, they can be expanded into the
sentences 'all red things are colored' and 'some men
Aristotle's Theory ot Categories 131
are incontinent' by the mere addition of connectors.
Here one can see the importance of our previous quali-
fication that only sentences of subject-predicate form
are under consideration. \Vithout it our current restric-
tion would tum out to be so strict as to leave no word
in the "uncombined" class, since the addition of "there
are some ... things" will render 'white' or 'heavy' into
sentences; yet 'white' and 'heavy' are known to be
expressions which Aristotle regards as "in no way com-
bined."
Throughout this discussion we have taken the no-
tion of a sentence for granted. It is time to see what
Aristotle has to say on this topic. But his account of a
sentence as significant sound, some part of which is
significant in separation as an expression (De Interpre-
tatione 16b26 ff. and Poetics 1457a23 ff.), is obviously
inadequate since it fails even to separate sentences from
clauses. The reason for this seems to be that Aristotle
finds himself in a curious predicament. He cannot ac-
cept the Platonic account of a sentence as the inter-
weaving of noun and verb since he recognizes counter-
examples to it. On the other hand, he knows that he
cannot use the theory of categories to explain a sen-
tence as an intercategorial connection, since he knows
that a sentence like 'all men are animals' does not
combine elements from different categories. Thus he
lacks a conceptual framework that would enable him
to give an adequate account of what a sentence is. In
view of this, we shall have to continue in this discus-
sion to take the notion of a sentence for granted.
We are not yet finished with the required restric-
tions, for we must admit the following:
(iv) Not all relevantly simple nouns and verbs des-
ignate items falling into only one category.
As Aristotle saw, one can simply define a word x as
'incontinent man' and thereby create a word which
our previous restrictions allow as uncombined but which
designates something that spans two categories. \Vbat
is clearly needed is a further restriction that would
rule out expressions like x as somehow ill-formed. There
is evidence that Aristotle did have such a qualification
in mind, for in De Interpretatione, chapters 8 and 11,
he discusses the problem of the extent to which a
predicate expression mayor may not designate a
"genuine unity." The discussions are sketchy and no
adequate interpretations of them have so far been of-
fered. Thus Ackrill wrote9 that the difficulty of decid-
ing what Aristotle regards as a genuine unity "cor-
responds" to the difficulty of deciding which simple
phrases are supposed to designate items falling into
only one category. Ackrill is quite correct in emphasiz-
ing the magnitude of the task facing the interpreter,
but this should not keep us from seeing that the notion
of a genuine unity could still be invokep by Aristotle
in an account of the "uncombined" elements of lan-
guage. For as 18a18 ff. shows, there are certain predi-
cates, e.g. that which is the equivalent of 'horse and
man', which designate items falling into one category
only, which nevertheless Aristotle would not regard as
having genuine unity. Thus whatever the correct ac-
count of genuine unity is, the passage under considera-
tion shows that it does not presuppose the correlation
between uncombined elements and the list of catego-
ries, and thus it could be invoked without circularity
to place further restriction on what is to count as an
uncombined element.
In view of these considerations it is not unreasonable
to suppose that Aristotle could describe x defined as
'incontinent man' as an expression not designating a
genuine unity; that he could do so without assuming
anything about its correlation with any of the catego-
ries, and that he could on this ground rule that it is
not one of the legitimate uncombined elements. This
9 Ackrill, op. cit., p. 126.
Aristot1e's Theory of Categories 133
final restriction leaves us with the following formulation
of Aristotle's view:
(v) Those elements of sentences of subject-predicate
form, or definitions, which (a) are not connectors;
(b) are not, like 'being', otherwise non-designative in
nature; (c) cannot be turned into sentences of subject-
predicate form, by the mere addition of connectors;
and (d) designate genuine unities, are "uncombined
elements" of language, and designate items falling in-
to one and only one category.
Intuitively restated, Aristotle's principle says that by
what we would call semantic and syntactic analysis 10
we can discover certain basic units among the elements
of sentences of subject-predicate form, and that these
turn out to designate those simple elements of reality
which fall into only one category. Thus the designative
link between these simple parts of language and the
simple parts of reality which fall into only one category
is, according to Aristotle, the key link between the
structure of language and the structure of reality.
This account is not without difficulties. The main
problem is the as yet unexplained doctrine of genuine
unity. Another problem is the fact that the restriction
ascribed to Aristotle that rests on the possibility of
expanding certain phrases into sentences with the help
of connectors is not supported by direct evidence. But
it is not a weakness of the interpretation that it seeks
to connect what is said in the Categories with some of
what is said in the Poetics and the Rhetorics. On the
contrary, to show that under a certain interpretation
these passages complement the views of the passages
in the Categories is to give that interpretation added
plausibility.
Given the difficulties, we might cast around for alterna-
tive interpretations. In this connection it is worth not-
10Aristotle, like Plato, did not distinguish between these.

ing that both the possible solutions listed by Ackrill,ll
and perceptively criticized by him, turn out to be inferi-
or to the account presented in this paper. One of these
solutions takes "without combination" to mean "des-
ignates an item falling into only one category." The
main difficulty with this view is that-as Ackrill ob-
serves-it construes the statement introducing the list
of categories in chapter 4 as analytic. If the statement
that the elements which are in no way combined des-
ignate items falling into only one category is analytic,
it is deceptively so; for this would imply that there is
no way of sorting out the uncombined elements of a
sentence except by observing whether their designata
fall into only one category. Thus the beginning of
chapter 4 could just as well have started: "some sen-
tential elements designate ... and we regard these as
being without any combination." Furthermore, accord-
ing to this interpretation what Aristotle says about the
combined and uncombined parts of language rests en-
tirely on metaphysical grounds and thus cannot be
connected with what he says elsewhere about the struc-
ture of language.
The other possible solution fares hardly better, since
according to it the distinction between what is and
what is not combined is identical with the distinction
between simple words and more complex sentential
elements. In order to accept this view we would have
to assume that in chapter 4 of the Categories Aristotle
forgets about the possibility of simple terms with com-
plex meanings. However, we saw above that this pos-
sibility is discussed in De Interpretatione, and to sup-
pose that Aristotle is not aware of this issue in a
context in which it is vital is to accuse him of too gross
a mistake. Moreover, this version suggests that Aris-
totle's distinction among different parts of language
rests on purely superficial features, whereas the passages
llAckrilI, op. cit., pp. 73-74 [pp. 102-103 above].
quoted from the Topics and the Poetics assure us that
this is not so.
To sum up, if we can regard the evidence presented
in support of the interpretation argued for in this sec·
tion as adequate, the interpretation recommends itself
on the following grounds: it links Aristotle's metaphys-
ical speculation with his view on the structure of
language; it relates the Categories to what is said in
other works on language; it helps to explore Aristotle's
views on language, which turn out to be far from sim-
ple-minded; and it sketches the structure of an explana-
tion that Aristotle would be likely to give as justifica-
tion of the claim that there are certain "uncombined"
elements in language with a-to him-vital designative
role.
II
The unity and completeness of the list of categories.
In the beginning of chapter 4 of the Categories we
find a list of ten categories. The labels given to them
are oddly heterogeneous. Some are philosophical con·
structs, some are ordinary questions, and some are
lifted out of simple singular sentences of subject·pred·
icate form,12 Thus at first glance it is not clear whether
the list is supposed to yield an ontological classification
or an analysis of the structure of propositions. Some
interpreters have gone as far as accusing Aristotle of
having failed to give the list sufficient unity. Perhaps
the most famous of these critics is Kant,13 who as-
sumes that Aristotle was interested in the same task
that he was-i.e. to give a set of necessary conditions
under which judgments are possible-and then con·
cludes that Aristotle failed in this task. A comparison
of Kant's categories with those of Aristotle, however,
suggests the not too surprising alternative that Kant and
12Ackrill, op. cit., p. 73 and Trendelcnburg, op. cit., p. 13.
13Critique of Pure Reason B 105-107.
Aristotle designed their categories with different pur-
poses in mind. Many of Kant's categories must be
construed as properties of judgments or ideas (e.g. uni-
versality or singularity) . Aristotle's list, on the other
hand, cannot be so construed. His items are either very
general properties of objects or not properties at all.
There is no reason why the two lists should coincide,
and once the difference in their aims has been dis-
cerned, the two need not be regarded as conflicting.14
One way of explaining the heterogeneity of the labels
is to assume that Aristotle is concerned primarily with
the types to be designated, and not with the manner of
designation. As we shall see, the adoption of this
assumption is rewarding. We must note at this point,
however, that it leads us to reject a claim like that
made by Trendelenburg, according to which the list of
categories is derived from grammatical considerations.15
Let us begin by considering one commonly accepted
characterization of the categories, the one that describes
them as the "highest predicables."16 In order for this
view to be even initially plausible, we must construe
'predicable' in a technical sense. For not only is the
first category ambiguous between primary and secondary
substance, but some of the examples given for the
other categories, e.g. 'in the Lyceum' and 'yesterday',
cannot be regarded as predicables in the ordinary sense
of the term. What this technical sense of 'predicable'
wide enough to embrace everything falling under each
of the ten categories could be is far from clear. But
14For a different view of the relation between Aristotle and

Kant see H. W. B. Joseph, An Introduction to Logic (Oxford,
1916), pp. 63-65. Also Cook Wilson, op, cit" p. 704 [po 851.
15Trendelenburg, op. cit., p. 33.
16As remarked but not endorsed by W. D. Ross in Aristotle,
pp. 27-28. According to Joseph, op. cit., pp. 48-49, the Greek
word for category (KaTTJyop(a) means 'predicate: but for
counterexamples see H. Bonitz, Vber die Kategorien des Aristot-
eles (Vienna, 1853), pp. 30-31.
Aristotle's Theory ot Categories 137
even if such a sense could be found, further difficulties
arise in connection with the tenn 'highest'. There are
two ways in which the metaphorical value of this term
could be captured by logical analysis. On one inter-
pretation predicate p' is higher than predicate p" if
and only if all members of the class which makes up
the range of application of p" are also members of
the class which makes up the range of application of
p', but not the other way around. In this sense animal
is higher than man. The other interpretation takes p'
as higher than p" if and only if p' is an attribute of p".
In this sense category would be a higher predicate than
quality or quantity, and colour, quality, category would
constitute a hierarchy. It is clear, however, that Aris-
totle does not want such a pyramid. He denies the
ontological reality of the higher strata here indicated.
Not to do so would leave him far closer to Platonism
than he would find comfortable. Thus the sense in
which the categories would have to be "higher" is the
former sense-as Cook \Vilson notedP That is to
say, as we go to wider and wider genera, the particulars
contained in anyone genera are on the same level as
the particulars contained in the wider genera. Given
this characterization of the categories, one is confronted
with the question: where must one stop? Why call
the categories the "highest" (actually the "widest")
genera? The only plausible Aristotelian reply is: "be-
cause by some principle the genera that would have
even wider comprehensions are not genuine genera; i.e.
they have no ontological status." Such a reply takes
us back to a metaphysical principle, and it is the ex-
plication of that principle, rather than the phrase
"highest predicables," that will shed light on the nature
of the categories.
Another current interpretation could be characterized
17Cook Wilson, op. cit., pp. 696 and 705 [pp. 75 and 86].
as the "linguistic view." Among its advocates are Ryle18
and Anscombe. 19 According to this view Aristotle uses
the list of categories to mark off the different kinds of
fairly simple things that can be said about a substance.
As such, the list could not be regarded as final or com-
plete, since there may be an indeterminate number of
ways in which one can raise questions about substances.
The most important achievement of the list turns out
to be-according to this interpretation-the anticipa-
tion of the concept of semantic category and the notion
of a "category-mistake," a confusion which supposedly
underlies a type of semantically deviant sentence. (E.g.
an answer in terms of food to a question about the
size of a substance allegedly does not make sense, even
though the sentence expressing the answer is-on the
surface at least-syntactically well formed.)
This is not the place to hold autopsy over the notion
of a category-mistake as used by recent analytic philoso-
phers. To show that this could not be the correct in-
terpretation of Aristotle's theory will have to suffice.
Three considerations prove fatal to this account. First,
according to this view once the significance of the list
is properly understood, questions about completeness
could not arise. But, as some of the passages quoted at
the beginning of this paper show, Aristotle committed
himself to claims which entail that the list should
contain mutually exclusive categories which are jointly
exhaustive of reality. (More on this point below.)
Thus either this interpretation is incorrect, or we must
18G. Ryle, "Categories," Logic and Language, second ser., cd.

A. Flew (Oxford, 1955), pp. 65-81. Ryle-like Kant-thinks
that Aristotle was really interested in what he, Ryle, is interested
in doing but that he did not do it so well. I find deeply depressing
this tendency of philosophers to think that the great men of the
past tried to do what they are doing but that they did not do it
so well.
lOG. E. Anscombe, in Three Pllilosopllers (Oxford, 1961), pp.
14-15.
suppose that Aristotle himself completely misunder-
stood the significance of his own theory. Such an as-
sumption should be adopted by a historian only as a
last resort. Secondly, the only basis for the individua-
tion of the categories would be the linguistic intuition
which allows us to detect category-mistakes. It is clear,
however, that Aristotle does not leave the individuation
of his categories to intuitions, linguistic or otherwise.
He states their differentiating characteristics quite ex-
plicitly: e.g. quality is that in virtue of which things
can be said to be similar; quantity is that which can
be said to be equal or unequaJ.2o These characteristics
do not depend on the notion of semantic anomaly.
Finally, it is difficult to suppose that had Aristotle been
yoncerned with semantic anomaly, he would have
missed the glaring fact that to describe a shape as red
or blue is semar1tically odd, even though both shape
and colour belong to the same Aristotelian category,
i.e. that of quality. Thus the linguistic interpretation
can be safely rejected.
Let us proceed to the consideration of the interpreta-
tion put forward by Professor AckrilI. 21 According to
this view Aristotle arrived at the list of categories in
two ways. One of these is the sorting out of the dif-
ferent types of question that can be asked about sub-
stances. The other is to start by asking what any given
thing is and to continue by repeating that question with
reference to whatever the previous answer revealed.
Both these approaches are supposed to come to an
end when irreducible genera are reached. Ackrill thinks
that the two ways are exemplified in Topics I 9. He
does not find it surprising that they should result in
the same list, with the exception that only the second
brings particulars into the classification. He leaves open
the question of completeness, since he does not think
20Categories 11a15-20 and 6a26.
21Ackrill, op. cit., pp. 78-81 [pp. 109-12].
140 .J. M. E. Moravcsik
that Aristotle had any grounds, in the form of a gen-
eral argument or principle, on which he could have
concluded that his list includes all and only irreducibly
different genera.
This summary shows that Ackrill's account has an
element in common with the "linguistic view." Both
interpretations assume that the list of categories is ar-
rived at by the consideration of questions, or classes of
properties, concerning substances. The only evidence
in favour of this assumption is the fact that Aristotle's
examples seem to be about a substance, i.e. man. We
must remember, however, that Aristotle employed the
category-structure in his attempt to show substances to
be prior to all else. Thus if we accept this hypothesis,
we must attribute to him the grand design of outlining
a set of categories by classifying questions about sub-
stances, and then using this structure to show that
substances are prior to all entities collected under the
other categories. Such an outrageously question-begging
procedure should not be attributed to any philosopher
except in the face of overwhelming evidence. After all,
the use of a pattern such as the one under considera-
tion would make proofs of ontological priority surpris-
ingly easy. For example, one could collect all the ques-
tions which can be raised about shadows, classify the
relevant predicates gathered through the answers to
these questions, and conclude that the items in the
categories thus formed are posterior to shadows since
they are specifications of shadows. As long as Aristotle's
general account of his list and his characterizations of
the several categories lend themselves to alternative
accounts, which avoid such question-begging, these al-
ternatives should be explored.
This objection leads to the observation that it would
be surprising indeed if the two approaches described by
Ackrill were to yield the same list. Why should th~
classification of the aspects of one kind of entity, e.g.
substance, coincide with an exhaustive classification of
the essences of all entities that make up reality? Would
this be true also of other types of entity, such as events,
qualities, numbers, etc.? Such a correlation will not
hold unless one views everything as a modification or
relational accident of the type of entity preferred. These
ways of viewing things are trivial.
Ackrill's second way of arriving at the list of catego-
ries also contains an inherent weakness. For there are
an indeterminate number of ways in which we can
classify things by the repeated asking of the question:
What is it? Furthermore, how could one decide whether
the highest genera have been reached? Intuitions, as
we saw, are not enough. To take one of Aristotle's
examples, why should one not arrive at change (kinesis)
as one of the categories? It certainly answers a "what
is it?" question. Yet, as we know, Aristotle does not
think that this is one of the categories; on the contrary,
he thinks that the list of categories must be presupposed
by the adequate analysis of this concept which shows
it to be cutting across three categories.
It does not, therefore, seem likely that Aristotle ar-
rived at his list in either of the ways that Ackrill
suggests. In view of this we should take a look at the
passage in the Topics (103b22 ff. )-mentioned above
-which Ackrill construes as containing the second way
of arriving at the list. It states, among other things, that
the essence of anything will be found in one of the cate-
gories. This statement does not entail Ackrill's second
approach, but only that the categories make up such
an exhaustive classification of reality that no real es-
sence will cut across categories. Thus it is consistent
with the possible claim that there are irreducible ulti-
mate genera within the categories, or that the categories
could be reduced in number. Most importantly, the
statement is only a necessary condition for the correct
list of categories; in itself it does not provide a pro-
cedure for arriving at the list. Nor does it give a prin-
ciple of unity for the list of categories.
In turning to the more constructive task of spelling
out a plausible alternative interpretation, let us begin
by noting-partly on the basis of the negative discus-
sion above-some necessary conditions for any adequate
interpretation. These conditions arise out of the ways
in which Aristotle employs the list of categories in his
philosophizing.22 The passages quoted previously from
the Topics, Physics, and Metaphysics, together with
Metaphysics 10 17a, show that Aristotle uses the list
in his analyses of key concepts such as being and
change, and also in his claiming priority for substances.
Thus the list has to be complete in the following ways:
(i) It must be exhaustive of all that Aristotle takes to
be existing. (ii) No reduction of the number of catego-
ries should be possible without violating the principle
upon which the list is constructed. (iii) No further
subdivision of the categories should be possible without
violating the constitutive principle. Without these
conditions Aristotle could not claim that kinesis can
be found in exactly three categories, or that by saying
that 'is' has as many senses as there are categories he
is giving a significant characterization of being.
These conditions help in resolving the question
whether the categories include only universals, or par-
ticulars as well. Underlying this issue is the debate
whether there are particulars and universals in each
category, or only universals in all but the first. 23 It is
difficult to conceive of each category as containing
universals; for example, what universal would corres-
pond to 'in the Lyceum'? On the other hand, it is
equally difficult to conceive of particulars falling under
22"Da endlich jede Lehre erst in ihren Folgcn ihrc Starke und
Schwachc offcnbart, so wird es wichtig scin, die Kategorien in der
Anwendung zu beobachten." TrendeIenburg, op. cit., p. II.
23See Ross, op. cit., p. 28, Il. 20.
Aristotle's Tl1eory of Categories 143
each category. vVhat particulars would we find under
the category of relation? The conditions laid down
above do not entail that each category must contain
both universals and particulars; they entail rather, that
the categories jointly must contain all the universals
and particulars that Aristotle would acknowledge as
existing. This leaves open the possibility that in some
categories there may be only universals, in some only
particulars, and in some both. In any case, events, proc-
esses, and abstract entities such as numbers must be
contained within the categories. The analysis of
kinesis, and the explanation of the category of quantity
reveals how Aristotle conceived of this inclusion.
The constitutive principle which we seek is likely to
be one that shows how the categories make up classes
of predicates (in a very wide sense of 'predicate') to each
of which some type of entity must be related. Apart
from the issue of question-begging, our conditions of
completeness guarantee that the type in question can-
not be that of substance; for in order to be a sensible
substance an entity must have both shape and weight,
and these two properties fall under the same Aristotelian
category of quality. Thus if the categories are those
classes of predicates of which substances must partake,
then the list as we have it would have to be subdivided
to put weight and shape into different categories, a
violation of the completeness conditions.
A more adequate interpretation can be given, based
partly on Bonitz's suggestion24 that Aristotle's list yields
a survey of what is given in sense experience, and that
each entity thus given must be related to some item
in each of the categories. According to this interpreta-
tion the constitutive principle of the list of categories
is that they constitute those classes of items to each
of which any sensible particular-substantial or other-
wise-must be related. Any sensible particular, sub-
24Bonitz, op. cit., pp. 18, 35, and 55.
stance, event, sound, etc. must be related to some sub-
stance; it must have some quality and quantity; it must
have relational properties, it must be related to times
and places; and it is placed within a network of causal
chains and laws, thus being related to the categories
of affecting and being affected. The only categories of
the complete list of ten that cause difficulties for this
interpretation are those of "having" and of "being-in-
a-position." In this connection the following should
be noted. First, these two categories are not always
included in the list Aristotle gives. Secondly, 'have'
is taken by Aristotle-as chapter 15 of the Categories
shows-in a variety of senses, one of which is the sense
of "having" parts. Given this construction, all sensible
particulars relate to that category. The category of
"position" is an obscure one, and it causes difficulties
for any interpretation, including the ones surveyed
critically above.
This proposed interpretation meets the completeness
conditions stated above. Neither a reduction nor a fur-
ther subdivision would leave the list as definitive of
that to which sensible particulars must be related. The
account also meets the exhaustiveness condition, at
least as far as Aristotle is concerned. For the Stagirite
believed that all properties, including the second-order
ones, are ultimately related to what is presented by our
senses. Finally, this account construes the list of catego-
ries as the proper background for Aristotle's investiga-
tions of being, change, and priority relations. If it is
question-begging, it is so only in the same way that
one might construe Aristotle's general preference for
what is presented to the senses-a preference never
defended-as question-begging. Thus this interpretation
fulfills the promise made at the beginning of this paper;
we have shown how the list of categories can be con-
strued as part of a theory and how this in turn serves
as a background for other Aristotelian theories.
Conclusion. The theory of categories is partly a
theory about language and partly a theory about reality.
With regard to language it states that certain elements
of a language have key-designating roles, the full under-
standing of which requires that we understand the
designata as falling within those classes which jointly
form the set definitive of that to which a sensible par-
ticular must be related. We can see from this that
Aristotle did not think of the structure of language as
mirroring the structure of reality. But he did believe
that there are specific items of language and reality
the correlation of which forms the crucial link between
the two.
III. METAPHYSICS
ESSENCE AND ACCIDENT
IRVING 1\1. COPI
The notions of essence and accident play important

and unobjectionable roles in pre-analytic or pre-philo-
sophical thought and discourse. These roles are familiar,
and need no elaboration here. Philosophers cannot
ignore them, but must either explain them or (some-
how) explain them away. My interest is in explaining
them.
If they are taken seriously, the notions of essence and
accident seem to me most appropriately discussed
within the framework of a metaphysic of substance,
which I shall accordingly assume. The account of es-
sence and accident that I wish to set forth and argue
for derives very largely from Aristotle, although it is
not strictly Aristotelian. Where it differs from Aris-
totle's account it does so in order to accommodate
some of the insights formulated by Locke in his dis-
cussion of "real" and "nominal" essences.' My discus-
sion is to be located, then, against the background of
a substance metaphysic and a realist epistemology. The
theory of essence and accident to be proposed seems to
me not only to fit the demands of the general philo-
sophical position mentioned, but also to be consistent
with the apparent requirements of contemporary scien-
tific development. I wish to begin my discussion with
some historical remarks.
From The Journal of Philosophy, LI (1954), 706-19. Re-
printed by permission of the author and editor of The Journal of
Philosophy.
150 Irving M. Copi
The earliest vVestern philosophers were much con-
cerned with change and permanence, taking positions
so sharply opposed that the issue appeared to be more
paradox than problem. If an object which changes really
changes, then it cannot literally be one and the same
object which undergoes the change. But if the chang-
ing thing retains its identity, then it cannot really have
changed. Small wonder that early cosmologists divided
into warring factions, each embracing a separate horn
of their common dilemma, the one denying perma-
nence of any sort, the othcr denying the very possibility
of change.
Aristotle discussed this problem in several of his
treatises, bringing to bear on it not only his superb
dialectical skill but an admirable, common-sense,
dogged insistence that some things do maintain their
identity while undergoing change. To explain the ob-
served facts he was led to distinguish different kinds
of change. A man does retain his identity though his
complexion may change from ruddy to pale, or though
he may move from one place to another. He is the
same man though he become corpulent in middle life
or his sinews shrink with age. In these types of change,
called alteration, locomotion, growth, and diminution,
the changing thing remains substantially or essentially
what it was before changing.
Another type of change, however, was admitted to be
more thoroughgoing. To take, for example, an artificial
substance, we can say that if a wooden table is not just
painted or moved, but destroyed by fire, we have neither
alteration, locomotion, growth, nor diminution alone,
but substantial change. The characteristic mark of sub-
stantial change is that the object undergoing the change
docs not survive that change or persist through it. but
is destroyed in the process. The ashes (and gas and
radiant energy) that appear in place of the burned table
are not an altered, moved, or larger or smaller table,
Essence and Accident 151
but no table at all. In substantial change its essential
property of being a table disappears.
It seems clear that distinguishing these different
kinds of change involves distinguishing different kinds
of attributes. Thc basic dichotomy between substantial
change and other kinds of change is parallel to that
between essential attributes or essences, and other
kinds of attributes, which may be lumped together as
accidental attributcs or accidents. (Here we diverge
rather sharply from at least one moment of Aristotle's
own terminology, in ignoring the intermediate category
of "property" or "propri urn.")
Of the various bases that have been proposed for
distinguishing between essence and accident, two stand
out as most reasonable. Thc first has already been im-
plied. If we can distinguish the different kinds of
change, then we can say that a given attribute is es-
sential to an object if its loss would result in the de-
struction of that object, whereas an attribute is a mere
accident if the object would remain identifiably and
substantially the same without it. This basis for dis-
tinguishing betwcen essence and accident, although
helpful heuristically, is not adequate philosophically,
for it seems to me that the distinctions among these
kinds of change presuppose those among thc different
kinds of attributcs.
The other, more satisfactory basis for distinguishing
essence from accident is an epistemological or methodo-
logical one. Knowledge of the essence of a thing is
said to be more important than knowledge of its other
attributes. In the Metaphysics Aristotle wrote: " ... we
know each thing most fully, when we know what it
is, e.g. what man is or what fire is, rather than when
we know its quality, its quantity, or its place ...."1 It
is the essence that is intended here, for a subsequent
passage explains that: " ... the essence is preciscly what
11028a37-1028b2. Quotations are from the Oxford translation.
F
152 Irving M. Copi
something is . ..."2 It is perhaps an understatement to
say that Aristotle held knowledge of essence to be
"more important" than knowledge of accidents, for he
latcr says explicitly that: " ... to know each thing ...
is just to know its essence ...."3 And if we confine our
attention to scientific knowledge, Aristotle repeatedly
assures us that there is no knowledge of accidents at
all,4 but only of essences. 5
Aristotle was led to draw an ontological conclusion
from the foregoing epistemological doctrine. If some
attributes of objects are epistemologically significant
and others are not, the implication is that the former
constitute the real natures of those objects, whereas
the latter can be relegated to some less ultimate cate-
gory: I must confess that I am in sympathy with the
realist position which underlies and justifies such an
inference, but to expound it in detail would take us
too far afield.
As a biologist Aristotle was led to classify things into
genera and species, holding that things belong to the
same species if and only if they share a common es-
sence. In remarking this fact we need not commit our-
selves to any position with respect to the systematic
or genetic priority of either logic or biology in Aristotle's
thought. He apparently believed these species to be
fixed and limited, and tended to ignore whatever could
not be conveniently classified within them, holding,
for example, that "the production of a mule by a horse"
was "contrary to nature,"6 a curious phrase. Some
modern writers have tended to regard this shortcoming
as fatal to the Aristotelian system. Thus Susan Stebbing
wrote: "Modern theories of organic evolution have
21030al.
31031b20.
41026b4; I027a20, 28; I064b30; I065a4. Cf. also Posterior
Analytics 75aI8-22.
575a28-30.
61033b33. But cf. 770b9-13.
combined with modern theories of mathematics to
destroy the basis of the Aristotelian conception of es-
sence...."7 It seems to me, however, that the fixity
of species is a casual rather than an integral part of
the Aristotelian system, which in its broad outlines as
a metaphysical framework can be retained and rendered
adequate to the most contemporary of scientific de-
velopments. A not dissimilar objection was made by
Dewey, who wrote that: "In Aristotelian cosmology,
ontology and logic ... all quantitative determinations
were relegated to the state of accidents, so that ap-
prehension of them had no scientific standing.... Ob-
serve by contrast the place occupied by measuring in
modern knowledge. Is it then credible that the logic
of Greek knowledge has relevance to the logic of
modern knowledge?"8 But the Aristotelian notion of
essence can admit of quantitative determination, as is
suggested by Aristotle himself in admitting ratio as
essence. 9 Hence I do not think that this criticism of
Dewey's can be regarded as any more decisive than
that of Miss Stebbing.
Having set forth in outline an Aristotelian philosophy
of essence and accident, I propose next to examine
what I consider to be the most serious objection that
has been raised against it. According to this criticism,
the distinction between essence and accident is not
an objective or intrinsic one between genuinely dif-
ferent types of attributes. Attributes are really all of the
same basic kind, it is said, and the alleged distinction
between essence and accident is simply a projection of
differences in human interests or a reflection of pecu-
liarities of vocabulary. Let us try to understand this
criticism in as sympathetic a fashion as we can.
The distinction between different kinds of change,
7A Modern Introduction to Logic, p. 433.

8Logic: The Theory of Inquiry, pp. 89-90.
9993a17-20.
154 Irving M. Copi
on this view, is subjective rather than objective. We
happen to be interested, usually, in some attributes of
a thing more than in others. When the thing changes,
we say that it persists through the change provided that
it does not lose those attributes by whose possession
it satisfies our interests. For example, our interest in
tables is for the most part independent of their colors.
Hence that interest remains satisfiable by a given
table regardless of any alteration it may suffer with
respect to color. Paint a brown table green, and it
remains substantially or essentially the same; the change
was only an accidental one. If our interests were dif-
ferent, the same objective fact would be classified quite
differently. Were our interest to lie in brown tables
exclusively, then the application of green paint would
destroy the object of our interest, would change it
substantially or essentially from something which satis-
fied our interest to something which did not. The im-
plication is that attributes are neither essential nor
accidental in themselves, but can be so classified only
on the basis of our subjective interests in them. Dewey
stated this point of view very succinctly, writing: "As
far as present logical texts still continue to talk about
essences, properties and accidents as something in-
herently different from one another, they are repeating
distinctions that once had an ontological meaning and
that no longer have it. Anything is 'essential' which
is indispensable in a given inquiry and anything is
'accidental' which is superfluous."lO
The present criticism lends itself easily to reformula-
tion in more language-oriented terms. That we regard
a table as essentially the same despite alteration in
color or movement from place to place is a conse-
quence of the peculiar nature and limitations of our
vocabulary, which has a single word for tables, regard-
less of color, but lacks special words for tables of dif-
lOOp. cit., p. 138.
ferent colors. Suppose that our language contained no
word for tables in general, but had instead-say-the
word "towble" for brown table and the word "teeble"
for green table. Then the application of green paint
to a towble would be said to change it essentially, it
might be argued, for no towble would remain; in its
place would appear a teeble. Or if there were a single
word which applied indiscriminately to tables and
heaps of ashes, say "tashble," with no special substan-
tive denoting either of them univocally, then perhaps
the destruction of a table by fire would not be regarded
as an essential change. That which appeared at the
end of the process would admittedly be in a different
state from what was there at the start, but it would
still be identifiably the same tashble. C. I. Lewis re-
gards the difference between essence and accident to
be strictly relative to vocabulary, writing: "Traditionally
any attribute required for application of a term is said
to be of the essence of the thing named. It is, of course,
meaningless to speak of the essence of a thing except
relative to its being named by a particular term."ll
I think that for our purpose these two criticisms can
be regarded as variants of a single basic one, for the
connection between human interests and human vocab-
ulary is a very intimate one. It is an anthropological
and linguistic commonplace that the concern of a cul-
ture with a given phenomenon is reflected in the vocab-
ulary of that culture, as in the several Eskimo words
which denote subtly different kinds of snow. In our
own culture new interests lead continually to innova-
tions in vocabulary; and surely it is the decline of in-
terest in certain things that leads to the obsolescence
of words used to refer to them.
Both variants of this criticism were formulated long
ago by Locke, and developed at considerable length in
his Essay. Locke paid comparatively little attention
11An Analysis of Knowledge and Valuation, p. 41.
156 Irving M. Copi
to the problem of change, but where he did discuss it
his treatment was very similar to Aristotle's. Thus we
are assured in the Essay that: " ... an oak growing from
a plant to a great tree, and then lopped, is still the
same oak; and a colt grown up to a horse, sometimes
fat, sometimes lean, is all the while 'the same
horse ... ."12 The oak " ... continues to be the same
plant as long as it partakes of the same life .. ."13 and
the identity of animals is explained in similar tenns.
Personal identity is explained in tenns of sameness of
consciousness.14 If we ignore the Cartesian dualism
implicit in that last case, and if we are not too critical
of the reappearance of the term "same" in the explana-
tion of .sameness, we can recognize these answers to
be the Aristotelian ones, for according to Aristotle the
soul is the principle of life,15 the life of a plant is the
nutritive soul,16 that of an animal its sensitive soul,17
and that of man his rational soul,18 these souls con-
stituting the substantial forms or essences of the re-
spective substances. 19 On the other hand, in his brief
discussion of identity as applied to non-living things,
Locke construes it very strictly to apply only to things
which " ... vary not at all ... ."20 But the following
passage has a characteristically Aristotelian flavor:
"Thus that which was grass to-day, is to-morrow the
flesh of a sheep; and within a few days after becomes
part of a man: in all which, and the like changes, it
is evident their real essence, i.e. that constitution,
12Bk. 2, ch. 27, 53.

13Ibid.
14Bk. 2, ch. 27, 58, 59, SIO, 516, 517,.523.
15De Anima 402a6, 415b8.
16432a29, 434a22-26; cf. also De Plantis 815b28-34.
17432a30.
18politics 1332b5.
19De Anima 412a20, 412b13, 415bl0.
20Bk. 2, ch. 27, S1.
Essence and Accidcnt 157
whereon the properties of these several things depended,
is destroyed, and perishes with them."21
Despite this partial similarity of their views, the
bases for distinguishing between the essential proper-
ties and other properties of a thing are very different for
Locke than for Aristotle. For Aristotle, the distinction
is twofold: first, the essential properties of an object
are those which are retained by it during any change
through which the object remains identifiably the same
object; and second, the essential properties of an object
are most important in our scientific knowledge of it.
For Locke, on the other hand, the real essence of a
thing is a set of properties which determine all the
other properties of that thing.22 Since all other proper-
ties depend on its real essence, any change in an object
entails a change in its real essence. Hence for Locke
the essential propertics of an object are not retained
by it during any change. This view is very different
from Aristotle's, on which the accidents of a thing are
not bound to its essence but can change independently
of it. The epistemological difference is equally striking.
Whereas for Aristotle all scientific knowledge is knowl-
edge of the essence, for Locke there is no knowledge
of the real essences of things. 23
Locke was more interested in what he called "nomi-
nal essences," which are more nearly analogous to the
Aristotelian notion of essence. Our idea of a particular
substance, according to Locke, is a complex idea com-
posed of a number of simple ideas which are noticed
to "go constantly together," plus the notion of a sub-
stratum "wherein they do subsist."24 A general or ab-
stract idea of a sort or species of substance is made
21Bk. 3, ch. 4, 519. But d. Bk. 3, ch. 6, S4, S5.

22Bk. 3, ch. 3, S15.
23Bk. 3, ch. 3, SI5, S17, SIS; ch. 6, S3, S6, S9, Sl2, SI8,
549; ch. 9, Sl2; ch. 10, SIS.
24Bk. 2, ch. 23, 51.
158 Irving M. Copi
out of our complex ideas of various particular substances
that resemble each other by leaving out "that which
is peculiar to each" and retaining "only what is com-
mon to a11."25 Such an abstract idea determines a sort
or species,26 and is caned a "nominal essence,"27 for
"every thing contained in that idea is essential to that
sort."28
The properties contained in the nominal essence of
a thing can be distinguished from the other properties
of that thing on the same basis as that on which the
Aristotelian essence is distinguished from accidents. In
the first place, a particular substance of a given species
can change with respect to some property whose idea
is not included in the nominal essence of that species,
and will continue to be recognizably the same thing;
whereas it must be regarded as a quite different thing if
it changes with respect to some property whose idea
is included in the nominal essence. 29 And in the second
place, the nominal essence is more imp~rtant in knowl-
edge than other properties. To have knowledge of a
thing is to know what sort of thing it is, and to know
the nominal essence is to know the sort. Locke says,
moreover, that the leading qualities of a thing, that is,
the most observable and hence, for Locke, the most
knowable, are ingredient in the nominal esserlce.30
Finally, it is argued in the Essay that knowledge of
nominal essences is required if we are ever to be certain
of the truth of any general proposition.31 Since Locke's
nominal essences play so similar a role to that of
Aristotle's essences, Locke's arguments intended to
prove their subjectivity and relativity to human in-
terests and vocabulary can be interpreted as applying to
Aristotle's notion as wen as his own.
25Bk. 3, ch. 3, p. 29Bk. 2, ch. 27, S28.
26Bk. 3, ch. 3, SI2. 30Bk. 3, ch. 11, SZO.
27Bk. 3, ch. 3, S15. slBk. 4, ch. 6, S4.
28Bk. 3, ch. 6, S2.
One fairly minor difference should be noted before
going on. Since Locke's nominal essences are abstract
ideas, thcy are immediately subjective in a way that
Aristotle's essences are not. But that difference is not
decisive, for substances may well have objective proper-
ties that nominal essences are ideas of, or objective
powers that correspond to them exactly.32
Locke urges that essences are subjective in a less
trivial sense. Since they are "inventions"33 or the
"workmanship"34 of the understanding, different per-
sons in fashioning abstract ideas which they signify by
the same term can and do incorporate different simple
ideas into them. Acts of choice or selection are involved
here, and people do make different choices, as proved
by the disputes that so frequently arise over whether
particular bodies are of certain species or not. 3a
That essences are relative to vocabulary is argued by
Locke in terms of an example: "A silent and a striking
watch are but one species to those who have but one
name for them: but he that has the name watch for
one, and clock for the other, and distinct complex
ideas, to which those names belong, to him they are
different species."36
That the " ... boundaries of species are as men, and
not as nature, makes them ... ,"37 proved by the verbal
disputes already referred to, is explained by the fact
that since we have " ... need of general names for pres-
ent use., ,"38 we " ... stay not for a perfect discovery
of all those qualities which would best show us their
most material differences and agreements; but we our-
32Bk. 2, eh. 23, S7.

33Bk. 3, eh. 3, S11.
34Bk. 3, eh. 3, ~12, S13, S14.
35Bk. 3, eh. 3, S14; eh. 6, S26, S27; eh. 9, S16; eh. 10, S22;
eh. 11, S6, S7.
36Bk. 3, eh. 6, S39.
37Bk. 3, eh. 6, S30.
38Ibid.
160 Irving M. Copi
selves divide them, by certain obvious appearances,
into species ...."39 Nominal essences are made for use,
and different intended uses or interests will determine
different essences. Even the noticing of similarities be-
tween distinct particulars is relative to our interest in
them, so our selection of simple ideas for inclusion in
a nominal essence is relative to such interests. These
determining interests are not scientific, for as Locke
observed, " ... languages, in all countries, have been
established long before sciences."4o The situation is
rather that the terms of ordinary discourse " ... have
for the most part, in all languages, received their birth
and signification from ignorant and illiterate peo-
ple...."41 And for the purposes or interests of those
practical people, the properties selected by them as
essential to the objects they deal with are adequate
enough. For "Vulgar notions suit vulgar discourses;
and both, though confused enough, yet serve pretty
well the market and the wake."42
Now do these arguments succeed in establishing that
the distinction between essence and accident is sub-
jective rather than objective, that is, relative to human
interests and vocabulary?
I think that the objections are not utterly destructive
of the Aristotelian doctrine, although they do call at-
tention to needed modifications of it. Locke's case, it
seems to me, depends upon his distinction between
real and nominal essences, and his belief that real es-
sences are unknowable. But his doctrine that real es-
sences cannot be known flows from two peculiarities
of his philosophy, which I see no reason to accept. One
of the bases for his belief that real essences are un-
knowable is his view that the only objects of our knowl-
edge are the ideas that we have in our.minds.43 Locke's
39Ibid. 42Bk. 3, ch. 11, SlO.

4oBk. 3, ch. 6, S25. 43Bk. 2, ch. 1, S1.
41Ibid.
other basis for his belief that real essences are unknow-
able is his doctrine that experiment and observation
yield only " ___ judgment and opinion, not knowl-
edge_ ..."44 Here the term "knowledge" is reserved for
what is certain.
I would reject these two doctrines on the following
grounds. The first of them, that knowledge is only
of ideas, is the germ of scepticism. Locke's premisses
lead necessarily to Hume's conclusions, and the p:lItial
scepticism we find explicitly set forth in Locke is but
a fragment of the complete scepticism that Hume
later showed to be implicitly contained there. It seems
to me that if a philosophy denies the very possibility
of scientific knowledge, then so much the worse for
that philosophy. As for reserving the term "knowledge"
for what is certain, that usage has but little to com-
mend it. It seems more reasonable to accept the results
of experiment and observation, although probable
rather than demonstrative, as knowledge nonetheless.
It must be admitted that the doctrine of the un-
knowability of real essences was not an unreasonable
conclusion to draw from the relatively undeveloped
state of science in Locke's day. For chemistry, at least,
if we can believe what is said of it in the Essay, was
in a very bad way in the seventeenth century. Locke
tells us of the "sad experience" of chemists " •.. when
they, sometimes in vain, seek for the same qualities
in one parcel of sulphur, antimony or vitriol, which
they have found in others. For though they are bodies
of the same species, having the same nominal essence,
under the same name; yet do they often, upon severe
ways of examination, betray qualities so different one
from another, as to frustrate the expectations of very
wary chemists."45
Contemporary science, however, presents a quite dif·
HBk. 4, ch. 12, S10; d. also Bk. 4, ch. 3, S28.
45Bk. 3, ch. 6, S8.
162 Irving M. Copi
ferent picture. Locke characterized the (allegedly un-
knowable) real essences of things as the " ... constitu-
tion of their insensible parts; from which flow those
sensible qualities, which serve us to distinguish them
one from another... :'46 Now modern atomic theory
is directly concerned with the insensible parts of things.
Through the use of his Periodic Table, interpreted as
dealing with atomic number and valency, " ... Men-
deleev was enabled to predict the existence and proper-
ties . .." of half a dozen elements whose existence had
not been previously known or even suspected.47 And
other scientists have subsequently been able to make
similar predictions. Modern science seeks to know the
real essences of things, and its increasing successes
seem to be bringing it progressively nearer to that goal.
It must be granted that Locke's distinction between
real and nominal essence is a helpful one, even though
it is not absolute. The construction of nominal essences
is usually relative to practical interests, and the or-
dinary notion of the essence of a thing is relative to
the words used in referring to it. I think that Locke
(and Dewey and Lewis) are correct in that contention.
Surely different interests lead different people to clas-
sify or sort things in different ways, and thus to adopt
different nominal essences, the more permanently use-
ful of which receive separate names in ordinary lan-
guage. Thus it is that: "Merchants and lovers, cooks
and taylors, have words wherewithal to dispatch their
ordinary affairs ...."48
The distinction, however, is not absolute. Not every
interest is narrowly practical. The interest of the scien-
tist is in knowledge and understanding. The scientist
desires to know how things behave, and to account
46Bk. 3, ch. 3, S17.

47J. D. Main Smith, in the Encyclopaedia Britannica (14th
ed.; 1947), Vol. 17, p. 520 (my italiCS).
48Bk. 3, ch. 11, SlO.
for their behavior by means of explanatory hypotheses
or theories which permit him to predict what will oc-
cur under specified conditions. He is interested in
discovering general laws to which objects conform,
and the causal relations which obtain among them.
The scientist's sorting or classifying of objects is rel-
ative to this interest, which is not well served by clas-
sifying things on the basis of properties which are
either most obvious or most immediately practical. It
is better served by classifying things in terms of proper-
ties which are relevant to the framing of a maximum
number of causal laws and the formulation of explan-
atory theories. Thus a foodstuff and a mineral source
of aluminum, common salt and cryolite, are both clas-
sified by the chemist as sodium compounds, because
in the context of modern chemical theory it is this
common characteristic which is most significant for
predicting and understanding the behavior of these
substances. In the sphere of scientific inquiry, the dis-
tinction between real and nominal essence tends to
disappear. The scientist's classification of things is in-
tended to be in terms of their real essences. And here,
too, the process is reflected in vocabulary, not neces-
sarily or even usually in that of the man in the street,
but rather in the technical jargon of the specialist.
The essences which science seeks to discover, then,
are real essences rather than nominal ones. Since the
arguments for subjectivity or relativity to interest or
vocabulary were concerned with nominal rather than
real essences, they simply do not apply to real essences
as either Locke or Aristotle conceived them.
In one passage of his Essay, though, Locke does make
the further claim that even a real essence relates to a
sort and supposes a species.49 But on Locke's own ac-
count of real essence, the real essence of a particular
must be that set of its properties on which all of its
49Bk. 3, ch. 6, S6.
164 Irving M. Copi
other properties depend. And that can be investigated
independently of any sorting or classifying we may do
-although once its real essence is discovered, that will
determine how we should classify it scientifically if the
occasion for doing so arises.
At this point let me indicate the direction in which
I think the Aristotelian doctrine of essence and ac-
cident might well be modified. Aristotle definitely held
that there could be no scientific knowledge of acci-
dents,50 but contemporary science would admit no
such limitation. It seems to me that both Locke's and
Aristotle's views about unknowability should be re-
jected. Contrary to Locke, I should hold that real
essences are in principle knowable, and contrary to
Aristotle, I should hold that non-essential or acciden-
tal properties can also be objects of scientific knowl-
edge.
It seems to me also that neither Locke nor Aristotle
gives a satisfactory account of the relationship between
essence and accident. For Locke, all (other) properties
of a thing depend on its "real constitution" or real
essence51; but it is not clear whether the dependence
is supposed to be causal or logico-deductive. The
former is obviously the more acceptable doctrine. Aris-
totle, on the other hand, held that some properties of
a thing, namely, its accidents, do not in any way
depend upon its essence. I think that Locke's view,
understood as asserting a causal dependence of acci-
dent on essence, is the more plausible one, and that
the Aristotelian doctrine ought to be so modified as
to accord with that of Locke in this respect.
Now if both essences and accidents are scientifically
knowable, on what basis are they to be distinguished
from each other? I suggest that the epistemological or
methodological distinction is still valid. For example,
GOI064b30-I065a25.
GIBk. 3, ch. 3, SIB.
common salt has many properties, some more obvious
than others, and some more important than others
relative to different practical interests. The scientist
singles out its being a compound of equaI. parts of
sodium and chlorine as its essential nature. In doing
so he surely does not mean to imply that its chemical
constitution is more easily observed than its other
properties, or more important to either cook, tailor,
merchant, or lover. He classifies it as sodium chloride
because, within the context of his theory, that property
is fundamental. From its chemical formula more of
its properties can be inferred than could be from any
other. Since the connection is causal rather than logi-
cal, the inference from essence to accident must make
use of causal law premisses or modes of inference as
well as strictly logical ones. Hence to derive conclu-
sions about all accidental properties of a substance, we
should need to know both its real essence and all rel-
evant causal laws. That is an ideal towards which
science strives, rather than its present achievement, of
course. To the extent to which one small group of
properties of a substance can serve as a basis from
which its other properties can be causally derived, to
that extent we can be justified in identifying that group
of properties as its real essence. This view, it should
be noted, is in agreement with Aristotle's doctrine that
the definition of a thing should state its essence, 52 and
that definition is a scientific process.53
There is a certain relativity implied in this account,
although it is quite different from those previously
discussed. Our notion of what constitutes the real es-
sence of a thing is relative to the science of our day.
Centuries hence, wiser men will have radically different
and more adequate theories, and their notions will be
closer approximations than ours to the real essences
5291al, 101b21, 38.
531039b32.
166 Irving M. Copi
of things. But it will still be the real essences of things
that are destined to be known by Peirce's ultimate com-
munity of knowers.
There is one other and more radical sense of accident
that I would agree to be relative. Each separate science
is concerned with only some of the properties or aspects
of things which it studies. Those left out will be acci-
dental relative to the special science which ignores
them. They will not be derivable from what that
science considers to be the real essences of those things,
although a different special science might be much
concerned with them, and even include them in its
notion of the thing's real essence. But as (and if) the
sciences become more unified, no properties of a thing
will be wholly accidental in this sense, and all will be
causally derivable from the real essence.
In closing, I should like to refer once again to the
topic of change. If all of a thing's properties depend
on its real essence, then it would seem to follow that
every change is an essential one. In my opinion, that
unwelcome conclusion can be evaded in two ways. In
the first place, with respect to common-sense, practical
usage, our ordinary sortings will continue to be based
on nominal rather than real essences, so that changes
can continue to be classified as accidental or essential
in the traditional way. And in the second place, with
respect to scientific usage, we can say the following.
The real essence of a thing will consist· very largely of
powers or, in modern terms, dispositional properties.
An essential change in a thing will involve the replace-
ment of some of its dispositions or powers by other
dispositions or powers. But a change which is non-
essential or accidental would involve no such replace-
ment; it would rather consist in differently actualized
manifestations of the same dispositional property or
power. Unfortunately, lack of space prevents an ade-
quate development of this suggestion.
TIeENAI TA <l>AIN6MENA
G.E.L.OWEN
The first part of this paper tries to account for an

apparent discrepancy between Aristotle's preaching and
his practice on a point of method. The second part
reinforces the first by suggesting a common source for
many of the problems and methods found in the
Physics.
I
There seems to be a sharp discrepancy between the
methods of scientific reasoning recommended in the
Analytics and those actually followed in the Physics.
The difference is sometimes taken to lie in the fact
that the Posterior Analytics pictures a science as a
formal deductive system based on necessary truths
whereas the Physics is more tentative and hospitable
both in its premisses and in its methods. But this is
too simple a contrast. It is true that for much of the
Physics Aristotle is not arguing from the definitions
of his basic terms but constructing those definitions.
He sets out to clarify and harden such common ideas
as change and motion, place and time, infinity and
continuity, and in doing so he claims to be defining
From Aristote et les problemes de la methode (Symposium

Aristotelicum; Louvain, 1961), pp. 83-103. Reprinted by per-
mission of the author and Editions Nauwelaerts, S.P.R.L.,
Louvain.
168 C. E. L. Owen
his subject matter. 1 But after all the Analytics shows
interest not only in the finished state of a science but
in its essential preliminaries; it describes not only the
rigorous deduction of theorems but the setting up of
the apxa[, the set of special hypotheses and definitions,
from which the deductions proceed. And the Physics,
for its part, not only establishes the definitions of its
basic concepts but uses them to deduce further theo-
rems, notably in books VI and VIII. The discrepancy
between the two works lies rather in the fact that,
whereas the Analytics tries (though not without con-
fusion and inconsistency) to distinguish the two proc-
esses of finding and then applying the principles, the
Physics takes no pains to hold them apart. But there
seems to be a more striking disagreement than this.
It concerns the means by which the principles of the
science are reached.
In the Prior Analytics Aristotle says: "It falls to
experience to provide the principles of any subject. In
astronomy, for instance, it was astronomical experience
that provided the principles of the science, for it was
only when the phainomena were adequately grasped
that the proofs in astronomy were discovered. And the
same is true of any art or science whatever."2 Elsewhere
he draws the same Baconian picture: the phainomena
must be collected as a prelude to finding the theory
which explains them. The method is expressly associ-
ated with <pUGlK~ and the <pUGlKot;,3 and from the stock
example in these contexts-astronomy-it seems clear
lPhs. III 1, 200b12-21.

2An. Pro I 30, 46aI7-22: 5u) Tae; !lEV apxae; mp! EKaoTov
t!l1TElplae; EaT! 1Tapa50uvat, AEYUJ 5' oTov -nlV aOTpoAoYIKr,V
!lEV E!l1TElplav Tij<; aO"TpOAoYIKij<; E1TIO"TTl!lIl<; (AIl<l>6EVTUJV yap
lKavwc; ,wv <l>atvO!lEVUJV OlTrUJe; EUPE61loav at aO"TpOAoYIKai a-
1To5EIsEIC;), 6!loIUJe; 5E Ka! 1TEP! aAAllv 61Tolavouv EXEI TEXVIlV
TE Ka! E1TIO"TTl!lIlV.
3De Part. Anim. I 1, 639b5-10 with 640a13-15; De Caelo III
7, 306a5-17.
Tithenai ta Phainomena 169
that the phainomena in question are empirical observa-
tions.4 Now such a method is plainly at home in the
biological works and the meteorology;5 equally plainly
it is not at home in the Physics, where as Mgr. Mansion
observes "tout s'y n§duit en general a des analyses plus
ou moins poussees de concepts,-analyses guidees sou-
vent et illustrees par des donnees de l'experience, plu-
tot qu'appuyees sur celle-ci."6 In this sense of "phaino-
mena" it would be grossly misleading for Aristotle to
claim that he is establishing the principles of his phys-
ics upon a survey of the pl1ainomena. And there his
critics are often content to leave the matter.
But in other contexts similarly concerned with meth-
ods of enquiry "phainomena" has another sense.7 In
the Nicomachean Ethics Aristotle prefaces his discus-
sion of incontinence with the words: "Here as in other
cases we must set down the phainomena and begin by
considering the difficulties, and so go on to vindicate
if possible all the common conceptions about these
states of mind, or at any rate most of them and the
most important."8 Here Sir David Ross translates
q>ulv6f.lEVU by "observed facts," a translation evidently
designed to bring Aristotle's programme into conform-
4Cf. further An. Po. I 13, 78b39 with 79a2-6; De Caelo II 13,
293a23-30; 14, 297a2-6; Met. A 8, 1073b32-38; Bonitz, Index,
809a34 ff.
5De Part. Anim. II 1, 646a8-12, referring to Rist. Anim. I 7,
491a7-14; Meteor. III 2, 371bI8-22 with Olympiodorus' scho-
Hum (217.23-27 Stueve. Olympiodorus' reference to De Gen. et
Corr. is to I 5, not II 8 as Stueve and Ideler think).
6Introduction a la physique aristo/elicienne2 , p. 211.
7There is a temptation to distinguish this sense as what q,aive:-
Tal e:Tval by contrast with what q,aivETal 6v. But this overstates
the difference; see pp. 174-76 below. Aristotle is ready to use
q,aivEoem with the infinitive even of empirical observations, De
An. I 5, 411bI9-22.
8NE VII 1, 1145b2-6: oe:i 0', WO"Tre:p t-rrt TWV &~Ae.>V, Tl-
BEVTac; TO: q,mVO\le:va Kat -rrPWTOV Ola'ITOPtlaavTac; OUTe.> Oe:IKVU-
val \lcXAlaTa \lEV 1TCxIITa TO: Evoo~a -rre:pt Tcx\ha TO: -rraB'l, EI OE
\ltl, TO: -rr~e:iaTa Kat Kuplc:,TaTa.
170 c. E. L. Owen
ity with such passages as those already cited. But this
can hardly be its sense here. For, in the first place,
what Aristotle proceeds to set out are not the observed
facts but the Ev5ot:,a, the common conceptions on the
subject (as the collocation of cpCXlv61lEva and Ev5ot;.a
in his preface would lead us to expect). He concludes
his survey with the words ,0: IlEV oav AEy61lEva Lath'
EO,(V,9 and the AEy61lEva turn out as so often to be
partly matters of linguistic usage or, if you prefer, of
the conceptual structure revealed by language. 1o And,
secondly, after this preliminary survey Aristotle turns
to Socrates' claim that those who act against their
own conviction of what is best do so in ignorance, and
says that this is plainly in conflict with the phainome-
naP But he does not mean that, as Ross translates it,
"the view plainly contradicts the observed facts." For
he remarks later that his own conclusion about incon-
tinence seems to coincide with what Soorates wanted
to maintain,12 and in reaching it he takes care to an-
swer the question that he had named as a difficulty
for Socrates, namely what kind of ignorance must be
ascribed to the incontinent manP So Socrates' claim
conflicts not with the facts but with what would com-
monly be said on the subject, and Aristotle does not
undertake to save everything that is commonly said.
He is anxious, unlike Socrates, to leave a use for the
expression "knowing what is right but doing what is
wrong," but he is ready to show a priori that there
is no use for the expression "doing what is wrong in
the full knowledge of what is right in the given cir-
cumstances."14 It is in the same sense of the word that
9Ibid. 2, 1145b8-20.
lOEspecially Ibid. 1145blO-15, 19-20.
llNE VII 3, 1145b27-28.
12Ibid. 5, 1147b14-15.
lSIbid. 3, 1145b28-29; 5, 1147bI5-17.
14Ibid. 5, 1146b35-1147alO, 1147a24-b14. But Ross's transla-
tion of cpaIVOJ.lEVa in the two passages 1, 1145b3 and 3,
all dialectical argument can be said to start from the
pl1ainomena. 15
This ambiguity in CPCXlVOt·U:.vcx, which was seen by
Alexander,16 carries with it a corresponding distinction
in the use of various connected expressions. 'ErccxyU)y~
can be said to establish the principles of science by
starting from the data of perceptionP Yet £rccxyU)y~
is named as one of the two cardinal methods of dia-
lectic 18 and as such must begin from the EVl'>ot;CX, what
is accepted by all or most men or by the wise;19 and
in this form too it can be used to find the principles
of the sciences.2o Similarly with the d:rcOp[CXl. When
the CPCXLVO!-lEVCX are empirical data such as those col-
lected in the biology and meteorology, the d:rcOp[CXL
associated with them will tend to be questions of em-
pirical fact 21 or of the explanation of such facts,2~ or
the problem of squaring a recalcitrant fact with an
empirical hypothesis. 23 In the discussion of incon-
tinence, on the other hand, where the CPCXLVO!-lEVCX are
things that men are inclined or accustomed to say on
1145b28 is at any rate consistent and so superior to that adopted
by most scholars from Heliodorus to Cauthier.Jolif, who see that
at its first occurrence the word must mean Ev6o~o (TOU<; 60KOUVTO<;
"lTEpl olJTCilV Myou<;. Heliodorus Paraphr. 131.16 Heylbut) but
suppose that at its occurrence 25 lines later it means the unques·
tionable facts (TOI<; 'llOVEPOI<;, ibid. 137.29-30).
15An. Pr. I I, 24bl0-12; Top. VIII 5, 159b17-23. Cf. Ph. IV
I, 208a32-34, where the phai1lomenon is the theory as contrasted
with the facts (TCx wOPXOVTo). At De Caelo II 5, 288al-2;
12, 291b25; IV 1, 308a6; De Part. Anim. I 5, 645a5, it is the
speaker's owh view.
16Meteor. 33.6-9 Stueve.
17AI!. Po. II 19, 100b3-5; I 18, 81a38-b9.
18Top. I 12, 105a1O-19.
1UTop. I 1, 100b21-23.
20Top. I 2, 101a36-14.
21Meteor. II 3, 357b26-30.
22MeteoT. I 13, 349aI2-14 with a 31·b2; II 5, 362a11-13;
De long. et brev vitae 1, 464b21-30; De Cen. Anim. IV 4, 770b
28-30 with 771a14-17; Hist. Anim. VI 37, 580bl4-17.
23Meteor. II 2, 355b20-32.
172 C. E. L. Owen
the subject, the ch:op(al that Aristotle sets out are not
unexplained or recalcitrant data of observation but
logical or philosophical puzzles generated, as such
puzzles have been at all times, by exploiting some
of the things commonly said. Two of the paradoxes
are veterans, due to Socrates and the sophists.24 The
first of the set ends with the words "If so, we
shall have to say that the man of practical wis-
dom is incontinent, but no one would say this"
(not that it happens to be false, but that given
the established use of the words it is absurd) .25 The
last ends "But we say (i.e. it is a common form of
words) that some men are incontinent, without further
qualification."26
Now if the Physics is to be described as setting out
from a survey of the CPaLV6!lEva it is plainly this sec-
ond sense of the word that is more appropriate. Take
as an example the analysis of place. It opens with four
arguments for the existence of place of which the first
states what BOKEl (it appeals to established ways of
talking about physical replacement) ,27 the third states
what certain theorists ~Eyoual,28 the fourth quotes what
Hesiod and the majority VO!l(~oual,29 and the remain-
ing one relies on the doctrine of natural places which
is later taken as an EvBot;ov.30 Of the O:1tOp[al which
follow, one is due to Zeno, one is due to an equally
rich source of logical paradoxes of which I shall say
more in a later section, and all ultimately depend on
the convictions or usage of the many or the wise. Nor
are these arguments merely accessory to the main
24NE VII 3, 1145b23-27, 1146a21-31.
25Ibid. 1146a5-7.
26Ibid. 1146b4-S.
27Ph. IV 1, 208bl, 5.
28Ibid. 208b26.
29Ibid. 208b32-33.
80Ph. IV 1, 208b8-25; 4, 211a4-6 with Ross's note on 5,
212b29-H (Aristotle's Physics, p. 580).
analysis: those of the OOKOTuva which survive the
preliminary difficulties are taken over as premisses for
what follows. 31 "For if the difficulties are resolved and
the EvooE,a are left standing," as Aristotle says in both
the Physics and the Ethics, "this in itself is a sufficient
proof."32 As for E11:ay(,)y~, when it is used in the ar-
gument it proves to be not a review of observed cases
but a dialectical survey of the senses of the word "in."33
By such arguments the Physics ranks itself not with
physics, in our sense of the word, but with philosophy.
Its data are for the most part the materials not of
natural history but of dialectic, and its problems are
accordingly not questions of empirical fact but con-
ceptual puzzles. Now this reading of the work is strik-
ingly reinforced, as it seems to me, when we recognize
the influence of one other work in particular on the
argument of the Physics. In a following section of this
paper I shall try to show that in the Physics Aristotle
over and again takes his start, not from his own or
others' observations, but from a celebrated set of logical
paradoxes that may well have appeared during his own
early years in the Academy. Far more than that over-
mined quarry the Timaeus, it is the Parmenides which
supplies Aristotle in the Physics not only with many
and perhaps most of his central problems hut with the
terminology and methods of analysis that he uses to
resolve them. But before turning to this evidence let
us see whether we are yet in a position to explain the
discrepancy from which we set out.
31Ph. IV 4, 210b32-211a7. Thus for instance the common

conception of place as a container which is not part of what it
contains (1, 208bl-8; 2, 209b28-30) must be rescued from
Zeno's puzzle (1, 209a23-26; 3, 210b22-27) by a survey of the
senses of "this is in that" (3, 210a14 ff.) and can then be taken
as secure (4, 210b34-211al).
32NE VII 1, 1145b6-7; Ph. IV 4, 211a7-11. The verb for
proof in each case is OEIKVUVai.
33Ph. IV 3, 210b8-9 (E1TOKTIKWC; CTK01TOVCTIV) with 210aI4-24.
174 C. E. L. Owen
Can we appeal to this ambiguity in Aristotle's ter-
minology in order to explain how such a generaliza-
tion as that quoted from the Prior Analytics could be
taken to cover the methods of the Physics? By now
the ambiguity seems too radical for our purpose. Even
within the second sense of <P<XlVO[lEVO:, the sense in
which it is equated with gv50~O: and AEyO[lEVO:, some
essential distinctions lie concealed. For an appeal to a
AEYO[lEVOV may be an appeal either to common belief
about matters of fact 34 or to established forms of lan-
guage 35 or to a philosophical thesis claiming the fac-
tual virtues of the first and the analytic certainty of
the second. 3S And the broader ambiguity between the
two senses of the word \vas one which Aristotle him-
self had the means to expose. For when he wishes to
restrict <p<XlVO[lEVOV to its first scnse he calls it expressly
a perceptual <pO:lVO[lEVOV and distinguishes it from an
gv50~OV.37 And in the De Caelo it is this more precise
form of words that he uses to describe the criterion by
which the correctness of our principles in physics must
ultimately be assessed. 38
I think such considerations show that it is a mistake
to ask, in the hope of some quite general answer, v.. hat
function Aristotle assigns to <pCXlVO[lEVO:, or to cmop[CXl,
or to ET(O:yWy~; or they show how the function can
vary with the context and style of enquiry. But we
have pressed them too hard if they prevent us from
understanding how Aristotle could have taken the for-
mula in the Analytics to apply to the Physics as well
as to the Historia Animalium. If there is more than one
use for the expression <p<XlVO[lEVO:, the uses have a
great deal in common. Thus for example it is not a
34E.g. NE I 11, 1l01a22-24.
35E.g. Ibid. VII 2, 1145b19-20; 3, 1146b4-5.
36E.g. Ibid. I 8, l098b12-18.
37 TCJV '1>alVOJ..lEVWV KaTCl TIJV ai'0'61]0'Iv, De Caelo III 4, 303a
22-23.
38Ibid. 7, 306al6-17.
Titllenai ta Pllainomena 175
peculiarity of <palVo[lEva in the second sense that they
may fail to stand up to examination; for so may the
<palVo[lEva of perception,39 and within this latter class
Aristotle is careful to specify only the reliable mem-
bers as a touchstone for the correctness of physical
principles. 40 As for his favourite example, astronomy,
Aristotle knew (or came to realize) how inadequate
were the observations of the astronomers.41 And of
the biological "observations" many were bound to be
hearsay, AEyo[lEva, to be treated with caution. 42 Such
<palVo[lEva must be "properly established," ascertained
to be "true data."43 In the same fashion the Ev5oE,a
must pass the appropriate scrutiny, but in doing so
they too become firm data. 44 Nor, if Aristotle associates
the <palVo[lEva with E[lTrElp(a, as he does in the text
from the Analytics, must it be supposed that his words
are meant to apply only to <palVo[lEva in the first
sense. "Ev5oE,a also rest on experience, even if they
misrepresent it. 45 If they did not Aristotle could find
no place for them in his epistemology; as it is, an
Ev5oE,ov that is shared by all men is ipso facto beyond
challenge.46
Nor is it in the least surprising if Aristotle, writing
in the tradition of Parmenides and Protagoras, tended
to assimilate these different senses of <paLVo[lEva.
39De Caelo II 8, Z90a13-Z4 and esp. Met. r 5, 101Obl-11.

(On Protagoras cf. p. 176 below.)
4oDeCaeloIIL 17, 306a16-17: TO <l'OIVOIJ£vov ad KUp((.)~ KaTa
TJ1v ai'o8T]olv, "the perceptual phaihomenon that is reliable
when it occurs," not, as Tricot translates, "I' evidence toujours
souveraine de la perception sensible": for Kupi(.)<; here d. Met.
r 5, 101Ob14-19.
41De Part. Anim. I 5, 644b24-28.
42E.g. Hist. Anim. II I, 501aZ5-bl.
43An. Pr. I 30, 46a20, 25.
44Ph. IV 4, Z10b3Z-34, Zl1a7-11, NE VII 1, 1145b6-7.
45E.g. De Div. per Som. 1, 462b14-18.
46NE X Z, 1172b36-1173al; d. VII 14, 1153b27-Z8, EE I 6,
1Z16b26-35.
176 G. E. L. Owen
For Parmenides, the B6I;CXl [3pon:i:cxl include not only
the supposed evidence of the senses but the common
assumptions (and specifically the common uses of
language) which form men's picture of the physical
world.47 As for Protagoras, both Plato and Aristotle
represent his theory as applying indifferently to per-
ceptual phenomena and ~v5ol;cx, and use cpCX(VW9CXl
in describing both these applications.48 It is the same
broad use of the word that is to be found in the for-
mula from the Prior Analytics. In the De Caelo, it is
true, Aristotle observes that it is the cpcxlV6!lEVCX of
perception by which we must ultimately test the ade-
quacy of our principles in physics;49 but this is said of
CPUOlK~ as a whole, a body of science in which the
analyses of the Physics proper are preliminary to other
more empirical enquiries and consequently must be
justified, in the last resort, by their success in making
sense of the observations to which they are applied.
But this is not to say (and it does not commit Aristotle
to supposing) that in the Physics proper the analyses
either start from or are closely controlled by our in-
spections of the world. Nor in fact is he liable to con-
sider his analyses endangered by such inspections: if
his account of motion shows that any unnatural move-
ment requires an agent of motion in constant touch
with the moving body, the movement of a thrown
ball can be explained by inventing a set of unseen
agents to fill the gap.50 The phainomena to which the
Physics pays most attention are the familiar data of
dialectic, and from the context in the Prior Analytics
it seems clear that Aristotle's words there are meant
to cover the use of such data. For in concluding the
41A conflation helped by talking as though data of perception
were themselves arbitrary assumptions (B 8, 38-41 Diels-Kranz).
On the "common uses of language" see B 8, 53; B 9; B 8, 38.
48Crat. 386al; Met. r 5, 1010bl, 1009a38-b2.
49De Caelo III 7, 306al6-17.
50Ph. VIII 10, 266b27-267a20.
passage and the discussion in which it occurs Aristotle
observes that he has been talking at large about the
ways in which the premisses of deductive argument
are to be chosen; and he refers for a more detailed
treatment of the same matter to the "treatise on dia-
lectic."5l He evidently has in mind the claim made in
the Topics that the first premisses of scientific argu-
ment can be established by methods which start from
the EvDot,a. 52
II
I turn to the part played by the Parmellides, and
specifically by the arguments in which "Aristotle" is
the interlocutor, in shaping the Physics. Perhaps it is
by misreading the Physics as a confused and cross-bred
attempt at empirical science that critics have been led
to look for its antecedents elsewhere and so to make
excessive claims for its originality. So it is worth dwell-
ing on this particular Platonic influence, partly for
the light that it throws on the methods and interests
of Aristotle's work, partly to call in question the claim
that "the discussions in books III-VI ... attack a series
of problems for which there was little in Plato's teach-
ing to prepare the way,"53 and partly to establish, if
this needs establishing, that the Parmenides was not
read by the Academy either as a joke or as a primer
of fallacies. 54 What the positive aims of the dialogue
51An. Pr. I 30, 46a28-30.
52Top. I 2, 101a36-b4. Ross seems to mistake the sense of the
An. Pro text (46a28-30) when he writes: "It is of course only the
selection of premisses of dialectical reasoning that is discussed in
the Topics; the nature of the premisses of scientific reasoning is
discussed in the Posterior Analytics" (Aristotle's Prior and Pos·
terior Analytics, p. 396). But in this passage Aristotle is con·
cerned with finding the principles of scientific reasoning, and
must be thinking of the claim made in the Topics to find such
principles dialectically.
53 Ross, Aristotle's Physics, p. 9.
54 In this respect what follows can be read as complementary
178 C. E. L. Owen
may have been does not concern US; the present enquiry
is a necessary preliminary to settling such questions.
Consider the celebrated account of the point. It is
Plato in the Parmenides who argues first that what is
indivisible (viz. the One, which cannot be plural and
so has no parts) cannot have a location. For to have a
location is to have surroundings i.e. to be contained
in something; and this is to be contained either in
something other than oneself or in oneself. But to be
contained in something other than oneself is to have
a circumference and to be in contact with that other
thing at various points, and an indivisible thing can-
not have various points or a circumference distinct
from its centre. Nor can a thing without parts be con-
tained in itself, for this would entail dividing it into
container and contained, and no such division of it is
possible. "Hence it is not anywhere, since it is neither
in itself nor in another."55 This concept of place as
surroundings is normal in Greek philosophy, as the
arguments of Zeno and Gorgias show (and in ordinary
conversation, which has small use for plotting objects
by Cartesian co-ordinates, it still is so). Aristotle took
it over as an Ev50~ov and made a more sophisticated
version of it in the fourth book of the Physics. And
one problem that he raises at the start of his argument
depends on the assumption that if a point has any
location it must be its own location, an assumption
that flatly conflicts with the received view that place
is a container distinct from the thing contained. 56
Aristotle does not argue the assumption; plainly he is
drawing on Plato's argument that an indivisible can-
not be contained in something else, nor yet can there
be any distinction within it between container and
to Professor D. J. Allan's essay in Aristotle and Plato in the Mid-
fourth Century (Aristotle and the Parmenides).
55Parm. 138a2-b6 (Burnet's lineation). TIle lack of shape and
circumference is proved in 137d8-138al.
1l6Ph. IV 1, 209a7-B.
contained. And he concludes that a point cannot be
said to have a location.57
On the way to this conclusion, and as a preface to
his general account of place, he lists the different senses
in which one thing can be said to be in another, 58
and follows this with an argument to show that a thing
cannot be said to be in itself except in the loose sense
that it may be a whole having parts present in it.59
This sense is sharply distinguished from the "strictest
sense of all," that in which a thing is said to be in a
place.60 Why does he spend so much time on this?
Because of further arguments in the Parmenides. Hav-
ing maintained, in the first arm of his argument about
the One, that an indivisible cannot be contained in
itself, Plato goes on in the second arm to reduce his
subject to a whole of parts and so, by dubious steps,
to reimport the notion of place. For (a) since the
subject is in itself in the sense that all its parts are
contained in it,6! it is always "in the same thing," i.e.
in the same place and hence at rest;62 and (b) since
the subject is not in itself, in the sense that as a whole
it is not contained in any or all of its parts, it must
be always in something else03 and so never at rest.64
Among other eccentricities, the argument clearly relies
on (and I think is clearly out to expose) an ambiguity
in the form of expression "being in so-and-so": it shows
that any sense of the phrase in which a thing can be
said to be in itself cannot be the appropriate sense for
talking of location, otherwise paradoxes result. Anaxa-
goras had traded on this ambiguity,65 and no doubt
57Ph. IV 5, 212b24-25.
58Ph. IV 3, 210a14-24.
59Ibid. 210a25-b22.
60Ibid. 21Oa24.
61Parm. 145b6-c7.
62Pann. 145e7-146a3.
6SParm. 145e7-e3.
64Pann. 146a3-6.
65Ph. III 5, 205bl-5.
180 C. E. L. Owen
Plato wrote with Anaxagoras in mind; but that Aris-
totle's arguments are framed primarily with a view to
those of the Parmenides is shown by the fact that he
mentions Anaxagoras' thesis not in this context but
elsewhere and by the clear echoes of Plato's language
in his own. 66
Points, then, cannot have location. And it is Plato
who first proves the corollary, that something without
parts cannot be said to move. But his reason is not just
that what has no location cannot be said to change
location. It is that to move to a certain place is a
process, and there must be some intermediate stage of
the process at which the moving body has arrived partly
but not altogether. 61 And it is just this argument that
Aristotle in the Pl1ysics takes over and generalizes, so
that it applies to other forms of change besides locomo-
tion. 6S Again, Plato prefaces his proof that an indivisible
thing cannot change place by showing that it cannot
even rotate in one place, since rotation entails a dis-
tinction between a centre and other parts;69 and with
this in mind Aristotle prefaces his argument by notic-
ing the case in which a point might be said to move
if it were part of a rotating body, but only because the
whole body, which has a distinct centre and circum-
ference, can be said to move in the strict sense.70 Since
it is often mistakenly said that Aristotle accepted the
66E.g. Ph. IV 3, 210a25-26 = Pann. 145d7-e1; Ph. IV 3,

21Oa27-29 = Parm. 145c4-7. Notice too that by jJEpTJ here
Plato means attributes of the subject, i.e. its being and unity and
their derivatives (cf.142dl-5); and that in the corresponding con-
text of the Physics Aristotle corrects this use of the word by point-
ing out that attributes may be contained KaTO: jJEpTJ in the sub-
ject not as being IIEPTJ themselves (which he rejects, Cat. 2,
la24-25) but as being attributes of jJEPTJ(Ph. IV 3, 210a29-30).
61Pann. 138d2-e7.
68Ph. VI 10, 240b8-241a6.
69Pann. 138c7-d2.
7oPh. VI 10, 240b15-20.
definition of a line as the path of a moving point,71
it is worth stressing how thoroughly he accepts Plato's
reduction of this idea to absurdity-a reductio which
no doubt counted as part of Plato's "war against the
whole class of points."72
Again, consider the account of a connected concept,
continuity. In the Parmenides Plato defines "contact"
(a:rrtEOSat) in terms of "succession" (EcpEf,~<;) and
"neighbouring position" (EXOtlEVT] xwpa) .73 These
terms Aristotle takes up in the fifth book of the Physics.
"Contact" he defines as holding between terms whose
extremities are t()gether, i.e. in one and the same
place;74 an unhappy suggestion, since in themselves
extremities can have no magnitude and so no position.
,And then, changing Plato's order of definition, he de-
fines "neighbouring" (EX611Evov) in terms of "contact"
and "succession."75 From both accounts, it is clear,
the same implication can be derived: Plato, by defin-
ing contact in terms of neighbouring position, and
Aristotle, by defining it in terms of things having ex-
tremities, preclude the attempt to talk of a series of
points as having contact with each other and so mak-
ing up a line or any other magnitude. But this result
only follows from Plato's definition if it is coupled
with the argument that an indivisible thing cannot
have position; and no doubt it was this that determined
71E.g. by Heath, Mathematics in Aristotle, p. 117; he cites
De An, I 4, 409a4-5, where Aristotle is reporting someone
else's theory. Of other passages which seem to imply this view
Ph. IV 11, 219b16-20 can be read otherwise and Ph. V 4,
227b16-17 may represent an objector's view. But Aristotle does
inconsistently credit points with location at An. Po. I 27, 87a36;
32, 88a33-34; Met . .1 6, 1016b25-26, 30-31, and perhaps with
the possibility of being in contact at Ph. V 3, 227a27-30 (but
this seems to depend on the unaristotelian thesis in lines 27-28).
72Met. A 9, 992aI9-22.
73Parm. 148e7-1O.
74Ph. V 3, 226b23. "Together" (aflcx) is defined in 226b
21-22.
75Ph. V 3, 227a6-7.
182 G. E. L. Owen
Aristotle to reform the definition so that the conclu-
sion would follow directly from the simple premiss
that a point has no parts or extremities. This reorder-
ing of the definition would not have served Plato's
purpose, for in this particular chain of reasoning in
the Parmenides he reserves the right to treat his sub-
ject as indivisible76 without committing himself to the
conclusion that it can therefore have no location. His
definition allows him to talk of an indivisible thing as
having contact with something else, and when he
proves that it cannot have contact with itself it is on
other grounds than the mere lack of 10cation.77 As a
result his proof is valid for all things and not merely
for indivisibles. But it is plain that his definition of
contact, taken together with his denial of location to
indivisibJes, produces exactly the conclusions which
Aristotle draws from his own definitions at the be-
ginning of the sixth book of the Physics,18 namely that
there is no sense in saying that lines are collections of
points in contact. It was in the Parmenides that Aris-
totle found not only the general approach to his prob-
lem but the special ideas in terms of which he treats
it.79
There is another point in these contexts at which
Aristotle corrects Plato. For Plato, contact requires
immediate (Eu96c;) succession in the contiguous terms,
76Parm. 147a8-b2; but earlier in the same movement he has
treated it as divisible into parts and continues to do so later.
77Parm. 148el0-149a3.
78Ph. VII, 231a21-bl0.
79Another such term in the same context is XColpfC; (Parm.
149a5), taken over and defined by Aristotle. And there are other
reminiscences of Plato's treatment of these ideas. One is the
comment at Ph. I 2, 185bIl-16, which Aristotle admits to be
irrelevant to the argument in hand. Why does he introduce it?
Because he has just mentioned continuity, and this reminds him
to Plato's argument in this connexion that, since the parts can be
distinguished from the whole, the whole can have contact with
itself (Parm. 148d6-7, 148el-3).
Titl1enai ta Pl1ainomena 183
and this immediacy he explains by saying that they
must occupy neighbouring positions. 80 But a little later
he explains this requirement in turn by saying that
there must be no third thing between the two terms;81
and Aristotle is anxious to find room for this condition
too in his definitions. He cannot use it to define
"neighbouring," since he has another definition of that
concept in view; so he uses it to define "successive,"82
and in doing so he adds an important qualification:
there must be nothing between the terms of the same
kind as themselves. 83 If ABC are consecutive sections
of a straight line, C cannot follow E<jJEf,~C; after A, but
it evidently can do so if B is merely a point. In cor-
recting Plato here Aristotle may have in mind the
treatment of limits in one passage of the Parmenides
as parts of a thing, logically comparable with what lies
between them;84 but this is a treatment that Plato's
own argument enables Aristotle to reject.
There is an embarrasing ,vealth of examples of this
influence in the Physics, and I shall not bore you with
them all. But one group is too important to omit. \iVe
saw earlier that, in arguing that an indivisible thing
cannot move, Plato (and Aristotle after him) treated
movement as a process taking time and having inter-
mediate stages. As Aristotle would say, it is a con-
tinuous change, divisible into parts which are them-
selves changes taking time. But later in the Parmenides
Plato argues that if a change is construed as the pas-
sage from not-A to A the change must be instantaneous;
for there is no time in which a thing can be neither A
nor not-A, neither at rest (for instance) nor in mo-
tion. 85 And this introduction of changes which are not
80Parm. 148e7-IO.
81Parm. 149a6.
82Ph. V 3, 226b34-227a4.
83Ph. V 3, 227al; cf. VI I, 23Ib8-9.
84Parm. 137d4-5.
85Parm. 1 56c6-7 : the whole context is 155e4-157b5.
G
184 C. E. L. Owen
processes is carefully prepared by some earlier argu-
ments. Twice-once in each of the first two chains
of argument about the One-Plato discusses the logic
of growing older. In the first argument86 he considers
it as a special case of becoming different; and he argues
that if X is becoming different from Y it cannot be
the case that Y already is different from X, since other-
wise X would already be different from Y and not
merely becoming so. All that follows from "X is be-
coming different from Y" is another proposition about
becoming, "Y is becoming different from X." The
conclusion is applied forthwith to the particular case,
to show that if X is becoming older than itself it is at
the same time becoming younger. But on a later page
the same example is taken up again. 87 Now Plato argues
that at any moment during the process of growing
older the subject must be older; at any stage of becom-
ing different, the thing must already be different. For
to say that it is becoming different is to say something
about its future as well as its present; but so far as the
bare present is concerned, it must already be some-
thing that it was becoming, given that the process of
change is under way at all. Thus the argument relies
heavily on the law of excluded middle: either the
changing thing is already different, or it is not. If it is
not, the process of change is not yet under way. And
if it is, then the old conclusion, that from "X is be-
coming different from Y" we can infer only what X
and Y are becoming and not what they are, breaks
down. The old conclusion relied on inserting a tertium
quid between "X is different" and "X is not different,"
namely "X is becoming different," something tem-
porally intermediate between the first two; but such
a tertium quid is ruled out by the law of excluded
middle. Yet it is just this law that leads to the problem
86Porm. 141a6-c4.
87Porm. 152a5-e3.
of instantaneous change with which we began; for Plato
goes on to argue that, if there is no time in which a
thing can be neither A nor not-A, neither still nor
moving, it baffies us to say when it makes the change
from the one to the other. 88 When it changes from
rest to motion it cannot be either at rest (for then the
change would be still to come) or moving (for then
the change would be past) . Yet the change is not to
be talked away: "if a thing changes, it changes."89
Here then is the problem, and the whole context of
argument, taken over by Aristotle. It is generally held
that Plato's purpose was to show that there can be no
period of time during which a thing is neither A nor
not-A, and consequently that the change from one to
the other must occur at a moment of time. 90 But Aris-
totle evidently thought the puzzle more radical, and
I think he was right. For by the same law of excluded
middle not only is there no period but there is no
point of time at which a thing can be neither A nor
not-A. At any rate, whether Aristotle is enlarging or
merely preserving Plato's problem, he gives it consider-
able space in the Physics. He agrees that some changes
take no time at an.91 Among other instances he cites
the recovery of health, which is "a change to health
and to nothing else" ;92 in other words, although the
process towards recovery may take time, the actual
88Parm. 156cl-7.
89Parm. 156c7-8:' AAA' ouSe: iJ~v iJETCXj36:AAEI avw TOU IlETCX-
j36:AAEIV. Cornford (Plato and Parmenides, p. 200, n. 2) mistakes
the sense, insisting that the statement is "intelligible only if we sup·
pose that Plato shifts here from the common use of IlETCXj36:AAEIV
for 'change' in general to the stricter sense of 'transition' or pass-
ing from one state to another." What Plato means is like our
truism "business is business"-sc. it mustn't be taken for any-
thing else or explained away. He would probably regard Aristotle
as explaining such changes away.
90Comford goes so far as to call it a "businesslike account of
the instant" (Ibid., p. 203).
91Ph. VIII 3, 253b21-30; cf. I 3, 186a13-16.
92Ph. VIII 3, 253b26-28.
186 C. E. L. Owen
recovery is simply the change from not-A to A.D3 In
any process of change to a given state there will be a
similar completion of the change, and this will take
no time: 04 the argument at once recalls Plato's dis-
cussion of the transition from movement to stillness.
Later, in the eighth book, Aristotle faces the problem
squarely. It will not help, he argues, to postulate a
time-atom between the period in which something is
not white and the subsequent period in which it is
white, with a view simply to providing a time for the
change to occur from not-white to white. For one
thing, time-atoms cannot be consecutive to periods of
time or to other time-atoms, just as points cannot have
contact either with lines or with other points. More-
over the suggestion would set a regress on foot. For
when we have postulated one time-atom to house the
change from not-white to white, there will be another
change to be accommodated in the same way: the
change from changing to being white. o5 In brief, Aris-
totle takes the puzzle to show that it is a mistake to
look for a special time-reference such that the subject
is then neither white nor not-white. The primary mo-
ment at which the subject becomes (or, as Aristotle
prefers to say, has become) white is the first moment
at which it is white.o6 And, given this moment, it
becomes improper to talk of the last moment at which
93Ross explains it otherwise; but for the treatment of uyiovalC;
as the limit of a KiVT]alC; cf. Met. e 6, 1048b18-23.
94Ph. VI 5, 235b32-236a7.
95Ph. VIII 8, 263b26-264al.
96Ibid. 263b9-26, 264a2-4, d. the earlier argument in VI 5,
235b32-236a7. The solution of Plato's puzzle given in Ph. VIII
8 is more trenchant than the earlier reply in VI 9 (240a19-29):
there Aristotle suggested that even between not·A and A a ter-
tium quid could be inserted, viz. when the subject is neither
wholly not-A nor wholly A; but this is easily defeated by reformu-
lating the contradictions as "wholly A" and "not wholly A."
Just as the reply to Zeno which is given in VI 9 is admitted to be
inadequate in VIII 8 (263aI5-18), so the reply to Plato's puzzle
given in VI 9 is superseded in the same later chapter.
the subject was not white, for the two moments would
have to be consecutive.97 Equally, given a last moment
of stability there cannot be a first moment of change.9s
And Aristotle, having thus saved the situation and the
law of excluded middle, can take over without qualms
the moral of Plato's second analysis of growing older:
namely that at any time during the period in which a
thing is becoming different, it has already completed
a change and to that extent is different from what it
was. 90
His reply to Plato's puzzle has side-effects on other
discussions. To underline the paradox, Plato had called
all change from not-A to A "sudden" change
(EE,XallJlVYJ<;) .100 Aristotle restores the word to its
proper use: it is used of what departs from its previous
condition in an imperceptibly short time. 101 But all
change, he adds, involves departing from a previous
condition; and his motive for adding this is clear. He
has in mind that because of this characteristic Plato had
tried to reduce all change to sudden change, and he
implies that this was a misleading extension of the
word's use. There is nothing physically startling in
most changes and nothing logically startling in any
of them.
There is no need to go on. It might indeed be ob-
jected that the evidence does not necessarily show that
Aristotle was indebted to the Parmenides; both Plato
and Aristotle may have been drawing on a lost source.
These problems were surely discussed in the Acad-
emy,l°2 and the Academy in turn must surely have
97Ph. VIII 8, 264a3-4.
9sPh. VI 5, 236a7-27.
99Ph. VI 6, 236b32-237aI7.
100Parm. 156dl-e3.
101Ph. IV 13, 222b14-16.
102We know for instance that others had tried to define con·
tinuity (Ph. III 1, 200b18-20), though they did not make use of
the nexus of ideas common to Plato's and Aristotle's treatments
]88 C. E. L. Owen
drawn on earlier arguments, in particular those of Zeno
and Gorgias. The general purposes of this paper would
be as well served by snch a theory, but it cannot ac-
count for the intricate correspondence that we have
seen in our two texts. Gorgias' part in the matter is
guesswork: the evidence for his sole adventure into
abstract thought has been contaminated, probably be-
yond cure, by traditions to which both the Parmenides
and the Physics contributed. Of Zeno luckily we know
more; we know that Plato does echo some arguments
of Zeno, but that he transforms them radically for his
own ends. IQ3 The Parmel1ides was not an historical
anthology, and when Aristotle's words and ideas coin-
cide closely with those of the dialogue he is under
the spell of a work of astonishing brilliance and origi-
nality. A work, moreover, of logic or dialectic, not in
the least a piece of empirical science; and the Pllysics
is in great parts its successor.
This is not to say, of course, that Aristotle would
call his methods in the Physics wholly dialectical. He,
and his commentators on his behalf, have insisted on
the distinction between "physical" and "dialectical,"
or "logical," or "universal," arguments; and no doubt
of the subject; hence Aristotle C3n take over their definition at the
start of the Physics (I 2, 185b10-11) before producing his own
revision of Plato's account.
loaThe Arrow underlies Parm. 1 52b2-d2, and the argument of
Bland 2 in Diels-Kranz (the resolution of a thing into its frac-
tions without ever reaching ultimate units) underlies Parm. 164c
8-d4 and 165a5-b6. I have not been convinced by Hermann
Fraenkel's interpretation of B 3, nor therefore by his claim that
it underlies the last-mentioned passages of the dialogue ("Zeno of
Elea's Attacks on Plurality," American Journal of Philology,
LXIII [1942], pp. 6, 198-99 = Wege und Forme1l, pp. 203,
227-28). Fraenkel is also inclined to see the Arrow behind Parm.
145-46 (art. cit., p. 13, n. 33 = \Vege und Formcn, p. 210, n.
1), where others will more readily detect Anaxagoras (cf. p. 94
above); and he sees B 4 behind Parm. 156c-d ibid. pp. Il-13
= pp. 207-9). He says all that is necessary for my purpose when
he observes that in such echoes "Plato modifies the argument
and ... transfers it, as it were, to a higher order."
some of the reasoning in the Physics falls within the
first class. Yet even if the distinction were (as it sel-
dom is) sharp and fundamental in sciences where a
knowledge of particular empirical fact is in question,104
we need not expect it to be so in such an enquiry as
the Physics. This is clear from the one major example
of the contrast that is offered in the work, the dialec-
tical and physical proofs that there can be no infinite
physical body.lo5 The dialectical proof is evidently dis-
tinguished by the fact that it proves too much: starting
from a definition that applies to mathematical as well
as to physical solids, it reaches conclusions that :Ipply
to both sciences,I06 Yet immediately after his promise
to turn to physical arguments Aristotle produces a proof
that no complex body can be infinite, and this proof
shares the characteristics of its predecessor. It relies
partly on quite general definitions of "body" and "in-
finite,"107 partly on a treatment of the ratio between
finite and infinite terms which could be formulated
quite generallyl08 and which in fact is later given a
different application to speed and resistance;109 and
partly, perhaps, on the argument against an infinite
number of elements which occurs in the first book and
relies largely on quite general premisses.l1O Certainly
there are other arguments in the context which seem
I04E.g. De Gen. Anim. II 8, 747b27-748aI6.
l05Ph. III 5, 204b4-206a8. There is a second use of the same
distinction (unnoticed by Bonitz s.v. AOYIK(;)C;) at VIII 8, 264a-
7-9, and here too it proves elusive. The "logical" arguments can
hardly be marked by their generality (the MyoC; flW,AOV OiKElOC;
at 264bl-2 itself applies to kinds of change other than move-
ment) nor the "physical" by their reliance on the special theo-
rems of physics (the "logical" also may do this, 264a24).
loaph. III 5, 204b4-7, cf. Ross's notes on 204b4, 204b6.
l07Ibid. 204b20-21.
l08lbid. 204bll-19: a particularly clear case of the artificial
restriction of a general theorem of proportion so as to bring it
within "physics."
l09Ph. IV 8, 215blO-216all.
lloPh. III 5, 204b12-13; I 6, 189aI2-20.
190 C. E. L. Owen
to depend on special empirical claims, such as the
unfortunate hypothesis of natural places.111 But the
impulse throughout the work is logical, and the restric-
tion of its subject-matter to movable bodies and their
characteristics does not entail a radical difference of
method from other logical enquiries. It makes for bet-
ter understanding to recall that in Aristotle's classifica-
tion of the sciences the discussions of time and move-
ment in the Parmenides are also physics.
11lPh. III 5, 205a10-12; but for the treatment of this too as an

EVSO~OV see n. 30 above.
MATTER AND PREDICATION
IN ARISTOTLE
JOSEPH OWENS, C. SS. R.
§l INTRODUCTION
In describing the basic matter of things, Aristotle
removed from it aU determinations and so aU direct
intelligibility. Yet he regarded the basic matter just in
itself as a subject for predication. You can say things
about it. You can say, for instance, that it is ingenerable
and indestructible, and that it is the persistent sub-
strate of generation and corruption. Still more strange-
ly, Aristotle means that a substance or substantial form,
like that of a man, of a plant, of a metal, can be pre-
dicated of matter.l How can this be, if matter is in
itself whoUy undetermined and entirely unintelligible?
From The Concept of Matter in Greek and Medieval Philos·
ophy, ed. E. McMullin (Notre Dame, Ind., 1963), pp. 99-11;.
Reprinted by permission of the University of Notre Dame Press.
lSee Aristotle Met. Z 3, 1029a20-30. The technical term used
by Aristotle for matter was the Greek hyle or 'wood'. He scems to
have been the first to coin a term for this notion, though the
philosophic use of hyle for materials in general was prepared by
Plato at Ti. 69A, and Phlb. 54C. In modern times the overall
approach to the scientific notion of matter is hardly different;
e.g.: "By the building materials I mean what we call matter, ...
ordinary matter is constructed out of two types of ultimate things
called 'electrons' and 'protons.''' C. G. Darwin, The New Con-
ceptions of Matter (London, 1931), p. 8. Aristotle, however, is
approaching the question on a level that does not lead to electrons
and protons but to very different principles; cf. Appendix. For
texts, see Bonitz' Index Aristotelicus, 652b49-51, 785a5-43.
G*
192 Joseph Owens
How can matter even be indicated, if it exhibits noth-
ing that can halt the gaze of the intellect?
The above observations envisage two ways in which
characteristics may be predicated of matter. One IS es-
sential (per se) predication. Matter is of itself ingener-
able and indestructible, somewhat as man is animal and
corporeal. The other way is through added forms.
Matter is metallic, bovine, human through the forms
of a metal, a cow, a man. But these forms are substan-
tial, not accidental. Yet their predication in regard to
matter resembles accidental predication, just as the
specific differentia in the category of substance is pred-
icated of the genus as though it were a quality. As
changes within the category of substance are called by
Professor Fisk in the present volume "qualified-like
changes," this type of predication may correspondingly
be designated "quality-like" predication. It is one type
of the medieval predicatio denominativa.
In ordinary predication, as treated in Aristotelian
logic, the ultimate subject is always actual and con-
crete. The universal, from a metaphysical viewpoint, is
potential (Met. M 10, 1087a15-22). The concrete
singular always retains its actuality as its various features
are universalized and made potential. It cannot be
treated as an undetermined residue that remains after
its predicates have been removed. Logical analysis of
predication, therefore, leads ultimately in Aristotle to
the actual, and not to something wholly potential like
matter. Ultimate matter is arrived at through the rea-
soning of the Physics. So reached, it poses problems for
metaphysics. How does it have being, and how are
forms predicated of it? The Stagirite had here to grap-
ple with a refined concept attained by his scientific
thinking and established to the satisfaction of that
technical procedure itself, but which broke through the
systematized logic presupposed by him for every the-
oretical science.
Matter and Predication in Aristotle 193
The solution reached by Aristotle in this question
mayor may not provide light for other disciplines when
in the course of their reasonings they arrive at concepts
that cannot fit into the grooves of the logic they have
been using. Such new concepts may well appear self-
contradictory when stretched on the Procrustean bed
of a closed logical system. Certainly in metaphysics
pertinent help for understanding the notion of essence
can be obtained from studying Aristotle's procedure in
establishing the notion of matter. Whether or not such
help may be extended to other disciplines has to be
left a question for investigators who specialize in them.
But the contingency is one that can be encountered
when any discipline pushes its concepts far past the
experiences in which human thought commences. Con-
cepts taken from immediate experience sometimes have
to be refined in peculiar ways if they are to function
in very remote areas of inquiry. The procedure of a
first-rate thinker in meeting such a contingency belongs
to the common treasury of achievements in the history
of thought, and hardly deserves to be forgotten. Aris-
totle's method in this problem seems, then, prima facie,
a subject worthy of investigation and critique.
§2 THE SUBJECT OF PREDICATION

First, what is the basic subject of predication in
Aristotelian logic? As is well enough known, this ulti-
mate subject of predication is the highly actual con-
crete singular thing. It is the individual man, or the
individual horse, or the individual tree, according to
the examples used in the Aristotelian Categories.2 In
a logical context, the real individual thing was called
"primary substance" by the Stagirite. In Greek the
term was prote ousia, primary entity. The term charac-
terized the concrete singular thing as absolutely basic
2See Cat. V 2a13-14, 2bl3.
194 Toseph Owens
among the subjects with which logic deals, and as the
fundamental being that received the predication of all
other perfections. Secondary substances, in that logical
context, were man, animal, body, and the like, taken
universally. They were all predicated of a primary sub-
stance, of a concrete individual man like Socrates or
Plato, or of an individual horse or stone. Accidental
characteristics, like white, large, running, and so on,
were predicated of substances and ultimately of an in-
dividual substance. There was nothing more fundamen-
tal of which they could be predicated. For Aristotelian
logic the concrete individual was the basic subject of
predication. It was the primary entity upon which all
logical structure was raised. In a logical context it was
primary substance in the full sense of the expression.s
This doctrine of predication functioned without spe-
cial difficulty when applied throughout the world of
common sense thought and speech. Quite obviously
the ultimates with which ordinary conversation deals
are shoes and ships and sealing-wax and cabbages and
kings, individual pinching shoes and flat-tasting cab-
bages and uncrowned office kings, as one meets them
in the course of everyday life. These are all concrete
individual things or persons. Aristotelian logic, it should
be kept in mind, was expressly meant as a propaedeutic
to the sciences. It did not presuppose knowledge of
any theoretical science. Rather, it had to be learned
before any theoretical science could be approached.4
There should be little wonder, then, that Aristotelian
logic was not geared to function smoothly in situations
brought into being solely through the results of scien-
SCat. V 2b4-6, 15-17. In a metaphysical context, on the other
hand, the form and not the composite was primary substance, as
at Met. Z 7, 1032bl-14; 11, 1037a28. On the category mistake
occasioned by this twofold use of 'primary substance' in Aristotle,
see my article "Aristotle on Categories," The Review of Meta·
physics, XIV (1960),83-84.
4See Aristotle Met. r 3, 1005b2-5.
Matter and Predication in Aristot1e 195
tific analysis and construction. Yet those situations
have to be expressed in concepts and in language. Logic
has to be applied to them as they occur. Aristotle, as
may be expected, could not go very deeply into any
theoretical science without encountering situations that
broke through the logical norms presupposed in his
hearers. Was he prepared to meet such situations? Was
he able to adapt his logic to them as they presented
themselves in the course of his scientific investigations?
§3 THE PROBLEM OF MATTER

An instance that could hardly be avoided was that
of matter. Matter quite obviously did not come under
the notion envisaged for an ultimate subject of predica-
tion in a logic where that ultimate subject is the con-
crete singular thing. In the everyday universe of dis-
course the material or stuff out of which things are
said to be made is always of the concrete individual
stamp. The wood of which a house is constructed con-
sists of individual pieces. The bronze in which a statue
is cast is a piece of bronze in definite dimensions in a
definite place at a definite time. In the later Scholastic
vocabulary these concrete materials out of which more
complex things were made received the designation, ma-
teria secunda, or 'secondary matter: Bronze and wood
and stone were indeed matter, in the sense that things
were made out of them. But they were not the basic or
ultimate matter out of which those things were made.
That was signified by calling them secondary matter.
The designation implied that there was a still more
basic matter that was not concrete nor individual. Aris-
totle had not finished the first book of his Physica or
philosophy of nature before he had established in sen-
sible things a subject still more fundamental than the
concrete individual. A visible, tangible, or mobile thing,
the Stagirite showed, was necessarily composite. It was
196 'oseph Owens
literally a con-cretum. It was composed of more fun-
damental elements. These ultimate constituents of sen-
sible things, according to the Aristotelian reasoning,
were form and matter. Matter played the role of ulti-
mate subject, and a form was its primary characteristic.
The absolutely basic matter of the Aristotelian
Physics became known in Scholastic terminology as
materia prima, 'primary matter'. By Aristotle himself
it was simply called matter. However, Aristotle uses
the term 'matter' regularly enough to designate the
concrete materials out of which artifacts are made,
materials like bricks and stones and wood. So there
was ground for the Scholastic insistence on the use of
two expressions, 'primary matter' and 'secondary mat-
ter,' to mark the important distinction. For convenience
in the present study the term 'materials' or 'material'
will be used wherever possible to denote what the
Scholastics called 'secondary matter', and the term
'matter' without any qualification will be used regularly
for the absolutely basic substrate of things as established
in the Aristotelian Physics. By 'matter', then, will be
meant what the mediaeval vocabulary designated as
'primary matter'.
With matter in this sense established as subject, and
form as its immediate though really distinct charac-
teristic, you may readily expect to hear that the form
is considered to be predicable of matter. You will not
be disappointed; Aristotle actually does say that sub-
stance, in the sense of substantial form, is predicated
of matter: " ... for the predicates other than substance
are predicated of substance, while substance is pred-
icated of matter."5 That is his express statement.
What does it mean? At the very least, it means that
matter is the ultimate subject with which predication
is concerned. Everything other than substance you can
predicate of substance. But what is intelligible about
5Met. Z 3, l029a23-24; Oxford tr.
substance can in turn be predicated, denominative1y of
course, of matter. The principle of intelligibility in
a substance is its form, and its form is the primary
characteristic of its matter, from the "quality-like"
viewpoint.
At first sight, perhaps, nothing could seem more
natural than to predicate a form of its corresponding
matter. Characteristics are regularly predicated of sub-
jects. A new subject has been unearthed by the Aris-
totelian philosophy of nature. The substantial charac-
teristic of that subject has been isolated. \Vhat is more
normal, then, than to say that here as in other cases
the characteristic is predicable of its subject?
Yet as soon as one tries to express this type of predi-
cation in any definite instance, linguistic and concep-
tual difficulties arise. How would you word a sentence
in which a substance, or a substantial form, is predicated
of matter? The first part of Aristotle's assertion was
clear enough: "Predicates other than substance are
predicated of substance." The predicates other than
substance are the accidents. They are quantity, quali-
ties, relations, activities, time, and place. They are pred-
icated without difficulty of a concrete, individual sub-
stance. You may indicate a particular tree and say with-
out hesitation that it is large, green, near to you, grow-
ing in the yard at the present moment. Each of these
accidents is obviously predicated of a substance, the
individual tree. But, the Aristotelian text continues:
"the substance is predicated of matter." How would
you express this in the case of the tree? You would
have to say that matter is this particular tree. You would
have to say that matter is likewise Socrates, or is Plato,
or is this particular table or that particular stone. Such
predication is unusual, and requires considerable ex-
planation even to make sense.
Some light may be obtained from the way in which
for Aristotle a thing may be defined in terms of the
198 Joseph Owens
materials of which it is composed. If asked what a house
is, you may answer that it is "stones, bricks, and tim-
bers."6 If that may be called a definition, it is surely
the least perfect type of definition possible. But Aris-
totle does refer to it as a definition in terms of the
materials that are able to be made into a house. From
that viewpoint the house is the materials that constitute
it, and conversely the materials are the house insofar as
definition and thing defined are convertible. In general,
then, in the way in which a thing may be said to be
its materials, the materials themselves may be said to
be the thing. Awkward though this predication is,
what prevents it from being applied in the case of the
basic matter of which things are composed? In each
particular ease it should allow you to say that matter is
this individual man, this individual stone, this indi-
vidual tree. Substance, even the individual substance,
would in this way be predicated of matter.
The context in which the present doctrine occurs is
one of the central books of the Aristotelian Metaphys-
6Met. H 2, 1043a15; Oxford tr. On this doctrine, and Aris-

totle's use of the expression 'primary matter' in connection with
it, see W. D. Ross, Aristotle's Metaphysics (Oxford, 1924), II,
256-57. 'Primary matter' is found in various senses at Ph. II I,
193a29; GA I 20, 729a32; and Met. <1 4, 1015a7-1O. 'Matter' in
its chief or primary sense, however, meant for Aristotle the sub-
strate of generation and corruption (GC I 4, 320a2-5), even
though the designation 'primary matter' never seems to have
been limited by him to that sense. The therapy required by the
concept's genesis has to be kept applied in representing the abso-
lutely undetermined matter as that of which things are composed.
Such matter is not individual, like any of the materials of which
a house is composed. Still less is it something universal, for the
universal is subsequent to the individual in Aristotelian doctrine.
Rather, it is below the level at which individuality and universal-
ity appear. Considered just in itself, it has nothing to distinguish
it as found in one thing from itself as found in another. From
this viewpoint it parallels the common nature of Duns Scotus,
which of itself had nothing to distinguish it as found in Socrates
from itself as found in Plato (see Duns Scotus, Quaest. Metaph.,
7, 13, No. 21; ed. Vives, 7, 421b. In contrast to the Scotistic
ics. In a metaphysical context, the universal is not
substance. When in this context substance is said to
be predicated of matter, it can hardly mean just another
instance of universal predicated of particular. From the
viewpoint of logic, the secondary substance or the sub-
stance taken universally is predicated of the particular
substance. Even though present as a condition, that
logical doctrine can scarcely be what Aristotle meant
in saying in the Metaphysics that substance is predi-
cated of matter. It is not just another case of predicat-
ing universal of singular, as in the assertion: 'Socrates
is a man: Subject and predicate are really the same
when a universal substance is predicated of a particular
substance. If you say: "Matter is a man," however, you
have a different type of predication. Matter does not
coincide in reality with a man in the way Socrates does.
A really distinct principle, the form of man, is added.
From this viewpoint the predication resembles rather
the assertion of an accidental form in regard to sub-
common nature, however, the Aristotelian basic matter lacks all

formal determinations, and so not only individual determina-
tions). The absolutely undetermined matter is accordingly one
through the removal of all distinguishing characteristics. It is
wholly formless in the Physics (I 7, 191a8-12) as well as in the
Metaphysics. In this sense only, may it be regarded as common.
When actuated, it differentiates by its very nature in making
possible the spread of the same form in parts outside parts and
the multiplication of singulars in a species. In that way it is an
individuating principle without being of itself individual. As the
substrate of substantial change, it may be said-with the appropri-
ate therapy-to change from one form to another. So doing, it
shows itself to be really distinct from its forms, since it really per-
sists while the forms really replace each other. But it is not
therefore a really distinct being from the form. In the individual
there is but the one being derived from the form to the matter
and the composite. Thus any single thing is differentiated from
a "heap" (Met. Z 17, I04Ib7-31). Subsidiary forms, for instance
those indicated in water by the spectra of hydrogen and oxygen,
would accordingly be accidental forms for Aristotle, and in a
substantial change would be replaced by new though correspond-
ing accidental forms.
200 Joseph Owens
stance, as when one says that a man is pale, or fat. The
accidental form is really distinct from the substance,
as the substantial form is really distinct from its mat-
ter. Such predication will be of the "quality-like" type.
That indeed is the way in which Aristotle presents
the situation. As an accidental form, for instance quan-
tity, is predicated of substance, so substance is predi-
cated of matter. What is predicated of matter, accord-
ingly should be the substantial form or act, and not
the composite. Later in the same part of the Metaphys-
ics it is stated in exactly that manner: " ... as in sub-
stances that which is predicated of the matter is the
actuality itself, in all other definitions also it is what
most resembles full actuality.'" As accidental forms are
predicated of substances, then, so the substantial form
is what is predicated of matter within the category of
substance.
The doctrine clearly enough is that form in the
category of substance may be predicated of its matter
as of a subject. You may accordingly apply the form
of man to matter, the form of iron to matter, and so
on, and call it predication. But how can you express
this in ordinary language? It can hardly be done. Ordi-
nary language has not been developed to meet this
contingency. The best you can do, perhaps, is to say
that matter is humanized, equinized, lapidified, and
so on, as it takes on forms like those of man, horse,
and stone. To say that matter is human, equine, lapide-
ous, or that it is a man, a horse, a stone, may be true
enough in this context; but with all its linguistic od-
dity the way of speaking hardly brings out the full
import of the situation. It tends to give the impression
that matter is of itself these things. The Aristotelian
meaning, on the contrary, is that matter is not of itself
any of these things, but becomes them by receiving the
appropriate substantial forms. As their real subject it
7Met. H 2, 1043a5-7; Oxford tr.
remains really distinct from them, somewhat as a sub-
stance remains really distinct from its accidents. The
assertion that matter is humanized, equinized, lapidified
by the reception of different substantial forms expresses
the predication with less danger of being misunder-
stood, though with still less respect for linguistic usage.
The linguistic difficulties, however, tum out to be
mild in comparison with the conceptual. The im-
mediate context of the Aristotelian passage that gave
rise to this discussion is enough to cause doubts about
the very possibility of the predication. Matter had just
been defined as "that which in itself is neither a par-
ticular thing nor of a certain quantity nor assigned to
any other of the categories by which being is deter-
,mined."8 Matter is not anything definite. It is not a
particular thing. It is not a "what" nor at all an "it."
It exhibits nothing that could provide a direct answer
to the question "\Vhat is it?" It has in itself none of
the determinations by which a thing can be or be
recognized or indicated or known or understood. The
text states explicitly that it has no quantitative nor
other categorical determination. Of itself, therefore, it
has no length nor breadth nor thickness nor number
nor parts nor position. It cannot at all be conceived in
the fashion of the Cartesian concept of matter. In
this concept, matter was identified with extension.9
Nor can the Aristotelian matter be represented as any-
thing capable of detection by means of a pointer-read-
ing. There is nothing about it, in itself, that could reg-
ister in quantitative terms. It belongs to a level on which
neither quantitative nor qualitative physics has any
means of functioning. It eludes quantitative and quali-
tative and other accidental determinations, as well as
all substantial determinations.
8Met. Z 3, 1029a20-21; Oxford tr.

9See Descartes, Principia Philosophiae, 2, 4-9; A-T 8, 42.4-
45.16 (92 , 65-68).
202 Joseph Owens
Yet it cannot be expressed by negations of known
characteristics, as for instance non-being is expressed
negatively in terms of being. lO The nature of matter
cannot be represented in terms of what it is not. The
same Aristotelian text continues:
Therefore the ultimate substratum is of itself neither
a particular thing nor of a particular quantity nor
otherwise positively characterized; nor yet is it the
negations of these, for negations also will belong to
it only by accident. ll
All categorical determinations are first denied to mat-
ter. They are outside its nature, and in that sense "be-
long to it only by accident." This has been expressed
in the preceding paragraphs of the present study by
saying that the forms are really distinct from matter
somewhat as accidents are really distinct from sub-
stances. But, the Aristotelian text insists, the negations
of all the different determinations are just as accidental
to matter. None of them can express its nature, as the
term 'nature' is used of matter in the Physics (II 1,
193a28-30; 2, 194a12-13). It eludes even negations.
You can indeed say that matter is not something, or
better still, that it is a "not-something." What you say
is true. But you have not thereby expressed the nature
of matter, even negatively. Negations are just as ac-
cidental to it as are the determinations it takes on in
the actual world. You are still only skimming its ac-
cidental manifestations. You have not penetrated to its
proper nature. Its nature eludes the negations.
In a word, matter as reached by Aristotle escapes in
itself both determinative and negative characterizations.
lOFor Aristotle, predication of being is made through reference

to the primary instance of being. Even the negation of being,
namely "non-being," is asserted in this way. See Met. r 2,
1003b5-10.
llMet. Z 3, 1029a24-26; Oxford tr. 'Positively' refers here to
determination; cf. a21.
It cannot be conceived or described in any direct fash-
ion, either determinatively or negatively. It is not
even a "what" nor an "it" that is capable of being in-
dicated. In terms of modern logic, it is not the "re-
ferent" of any "demonstrative" (i.e. monstrative) sym-
bol, because it cannot be presented directly to one's
cognition. Nor can the referent be any property or set
of properties, because such determinations are lacking
to matter in itself.12
How, then, is the Aristotelian matter to be conceived
and represented? How can it be set up as a subject for
predication? Quite obviously, from the above consider-
ations, no direct method, either affirmative or negative,
is capable of grasping what Aristotle meant in this
regard. The concept will have to be that of a positive
subject, able to receive predication. No negation is
able to express the nature of matter. Yet from that
notion of positive subject every determination will have
to be removed, even, or rather especially, the determi-
nation expressed by "something." Matter is explicitly
not a "something" nor a "what" nor an "it." All deter-
mination, even the most elementary, has to be drastical-
ly eliminated from the notion of the positive in this
concept. The concept that expresses the Aristotelian
notion of matter will have to be the concept of a pos-
itive object that is wholly indeterminate. Is the human
mind able to form such a concept? If so, upon what
referents will it be based?
Presumably Aristotle could not have spoken so co-
gently about matter if he had not worked out its con-
cept to his own satisfaction. The most likely way to
12"The referend of a demonstrative symbol (i.e. a word used
demonstratively) is the ob;ect directly presented to the speaker.
The referend of a descriptive phrase is a property, or set of
properties." L. Susan Stebbing, A Modern Introduction to Logic
(6th ed.; London, 1948), p. 499. On the technical term 'refer-
end,' cf.: "We shall find it convenient to use the word 'referend'
to stand for that which is signified." Ibid., p. 13.
204 Joseph Owens
learn how the concept is formed, accordingly, should
be to follow the steps by which the originator of the
concept reasoned to the presence of matter in sensible
things. In this context, of course, the referents will be
sensible things in themselves, and not Kantian phenom-
ena.
§4 SUBSTANCE AND CHANGE

How, then, did Aristotle arrive at the notion of mat-
ter as a real subject, and as a subject denominatively
characterized by forms that remained really distinct
from it? In the first book of the Physics, the Stagirite
surveyed the teachings of his philosophic predecessors
on the basic principles of natural things. Things in the
world of nature were known by observation to be
capable of motion or change. In the course of his sur-
vey, attention is focused upon the universal require-
ments for change. Any change whatsoever needs three
principles. It has to have a subject that loses one form
and acquires another. The three principles necessarily
involved are therefore the form that is lost, the form
that is acquired, and the subject that undergoes the loss
of the old form and the acquisition of the newP
The Aristotelian examples meant to illustrate this
doctrine are clear enough. They are concrete individual
materials that lose and acquire different forms. Bronze
is the subject that becomes a statue. At first the bronze
13See Ph. I 7, 189b30-191a7. The analysis of change or motion
is made by Aristotle without dependence on the notion of time.
Rather, motion is first defined, and then the notion of time is
worked out in terms of motion, that is, as the numbering of mo·
tion in respect of prior and subsequent (Ph. IV 11, 219b1-2).
Since Kant the tendency has been first to establish the notion of
time, and then to describe motion in terms of relation to time;
e.g.: "Change thus always involves (1) a fixed entity, (2) a
three·cornered relation between this entity, another entity, and
some but not all, of the moments of time." Bertrand Russell,
Principles of Mathematics (Cambridge, Eng., 1903), 1,469.
has a nondescript form or shape. Then it is cast into
the form of a statue, say of the Greek god Hermes. It is
the subject that changes from one form or shape to
another. The notion of form in this example is readily
understood from its ordinary English use. It is the
external shape of the bronze. Another Aristotelian ex-
ample, however, uses 'form' in a more esoteric way.
A man from an uneducated state comes to be educated.
The man is the subject that changes from uneducated
to educated. 'Uneducated' describes the quality of the
man who has not had proper schooling. 'Educated'
means the quality of adequate instruction and cultural
training. Both 'educated' and 'uneducated' mean quali-
ties; and in the Aristotelian vocabulary qualities are
forms.
As can be seen in these examples, the original form
from which the subject changes is more properly re-
garded from the viewpoint of a privation of the ·form
to be acquired in the change. It is expressed in a pri-
vative way, as in the term 'uneducated.' Any of the
Aristotelian categories, like the thing's quantity, its
place, its time of occurrence, or any of its relations, is
a form in this technical Aristotelian sense. Change can
take place in any of the categories of being. 14 But in
its very notion, as has emerged from the foregoing
analysis, it involves indispensably the three principles
-a subject that changes, a form that is lost, a form
that is acquired.
This essential notion of change is reached from the
changes that are observed in the accidental categories,
like change from place to place, from size to size, from
color to color. But the analysis of the notion establishes
it as a general concept that will hold wherever change
is found, regardless of the particular category. It is
accordingly applied by Aristotle in the category of sub-
14Ph. III 1, 201a8-9.
206 Joseph Owens
stance. In all other categories the subject of the change
is observable. You can see the man who changes from
uneducated to educated. You can touch the bronze
that is cast from a nondescript form into a statue. You
can handle the wood that is made into a bed. But
with change in the category of substance you cannot
observe the subject that changes, even in principle.
This means that you cannot observe the subject chang-
ing. Change in the category of substance is accordingly
not observable, even in principle.
There need be little wonder, then, that Aristotle is
sparing in examples of change in the category of sub-
stance. Without too much enthusiasm he accepted the
tradition of the four Empedoclean elements as the
basic simple bodies, and admitted as generation the
change of anyone of these bodies into another.15 But
he is very circumspect in determining just where sub-
stance is found. Earth, air, and fire, three of the tradi-
tional elements, do not seem to him to have sufficient
unity in their composition to be recognized as sub-
stances. Living things seem to have that unity, yet just
where the unity is cannot be located too easily.16 The
one instance that he does mention definitely, though
only in a passing way, is the change to plants and ani-
mals from seedP
Today this Aristotelian example may not seem any
too happy an illustration of substantial change. With-
out having to call the fertilized ovum of a rhinoceros
a little rhinoceros, one may argue either for or against
the position that an embryo is the same substance as
the fully developed animal. To say that a tadpole is
not a frog does not commit you to the stand that the
15See Cael. I 2, 26Sb26-29; GC II 1, 329a2-S; 8, 334b3I-

335a23.
16Met. Z 16, I040b5-16.
17Ph. I 7, 190M-5. Cf. GA I IS, 722b3-5, and St Paul's
simile, I Cor. 15: 36.
one is a different substance from the other. In general,
it may be easy enough to claim that the change from
something non-living to something alive is a change
in substance. But in regard to pinpointing the change
from non-living to living substance, or even to show-
ing definitely that there was change from the truly
non-living, are we today in any noticeably further ad-
vanced position than was Aristotle? Similarly, with
modern chemical knowledge, it is easy to show defi-
nitely that air, fire, and earth are not substances in
the Aristotelian sense. We no longer share the Stagi-
rite's hesitations in that regard. vVith respect to water,
however, can a definite decision be given? In the highcr
kinds of living things, Aristotle's criterion was a unity
that distinguishes the complex organism from a heap.
It is the same criterion that enables us now to con-
sider the ant a different thing from the sandpile. In
man, consciousness adds a still more profound criterion
of unity. Every man considers himself a different be-
ing from other men, a differcnt being from the sub-
stances he absorbs in nutrition and from those into
which he will be dissolved when he dies. Apart from
preconceived positions arising out of conclusions in
metaphysics or in modern physics, and illegitimately
transferred to the domain of natural philosophy, the
difference of one being from another and the change
of one sensible being into another may in general be
admitted. Tlle evidence ot pertinent bearing eitller tor
or against, tllougll, is scarcely any greater now tllan it
was in Aristotle's day.
However, the plurality of things in the universe wiII
hardly be contested any more today in a properly
physical context than in the Stagirite's time. IS As long
as a plurality of beings in the sensible universe is ad-
mitted without subjecting the term 'beings' to intol-
I8See Ph. I 2, 184b25-185a16.

208 Joseph Owens
erable strain, the plurality of substances required for
the Aristotelian demonstration of matter is present.
'Substance' in Aristotle's terminology meant the entity
or ousia of things. Wherever you have a being, simply
stated, you have an ousia, a substance. Nor should there
be too much difficulty about the change of one thing,
macroscopically speaking, into another. Molecular com-
pounds are changed into other compounds, transmuta-
tion of the elements is no longer a dream. The one
real difficulty might lie in the proposal to locate the
individuality of things in sub-atomic particles. In that
case might not all the changes taking place in the
physical universe be merely new combinations of the
particles, as in Democritean atomism? There would be
only accidental change, not substantial change.
The denial of any unifying principle in things over
and above the sub-atomic particles would leave the
behavior of every particle wholly unrelated to that of
the others. A cosmic puppeteer would have to cause
the regularity of the world processes. A principle of
unity in each thing itself, on the other hand, would
have to be deeper than the division into particles and
into quanta, and indeed would have to be of a different
order. It would have to function on a more profound
level, in order to dominate the polarity of the sub-
atomic particles and to maintain the statistical regulari-
ty of the quanta. Such a principle would function ex-
actly as the Aristotelian substantial form. It would be
the deepest principle of unity in a thing, and so would
make a thing "a being" simply and without qualifica-
tion. It would be the principle that rendered the thing
intelligible. It would be the thing's basic determinant,
making the thing one kind of thing and not another.
It would be deeper than the entire qualitative and
quantitative or measurable orders in the thing, and so
would enable the thing to exist and function as a unit
in spite of the common patterns of atomic and sub-
atomic motion that it shares with other things. When
this formal principle gave way to its successors in
changes like nutrition or death, a radically new thing
or things would come into being, in spite of common
spectra before and after the change and in spite of the
equality of the total weight before and after. It would
enable the thing to function as a nature and not just
artificially at the hands of a cosmic puppeteer. In a
word, this principle would coincide entirely with the
Aristotelian form in the category of substance.
The argument for the change of one substance into
another, accordingly, seems neither stronger nor weaker
in any notable way than it was in fourth century
Greece. If you grant that you are a different thing or
a different being from the food you absorb in nutrition
and from the substances into which you will dissolve
in death, you have recognized the data necessary to
understand the Aristotelian demonstration. When one
substance changes into another, what disappears is the
most basic principle of determination and knowability,
the principle that most radically made food one thing
and man another thing. Without it, nothing in the
thing could be knowable or observable. It is of course
immediately succeeded by the form of the new thing.
But the change of the one thing into the other requires
a common subject, according to the very notion of
change. Such a common subject will be unobservable
both in principle and in fact, because it is what loses
and acquires the most basic of forms and so of itself
has not even the most rudimentary principle of know-
ability or observability. It has to be known in virtue of
something else. That "something else," quite naturally,
will be the observable subject in accidental change,
like the wood that becomes a bed or the bronze that
becomes a statue. Some corresponding subject has to
be present for substantial change. In that analogous
way, then, the subject of substantial change, namely
210 Joseph Owens
matter, is indirectly known. It is known as the con-
clusion of scientific reasoning in the Aristotelian sense
of 'scientific'. In Aristotle's own words:
The underlying nature is an object of scientific knowl-
edge, by an analogy. For as the bronze is to the statue,
the wood to the bed, or the matter and the formless
before receiving form to anyone thing which has
form, so is the underlying nature to substance, i.e. the
"this" or existent. 19
The presence of matter is proven stringently from
the requirements for change, while the nature of mat-
ter is established through analogy with the subject of
accidental change. The demonstration presupposes the
universal notion of change and the two terms, but not
the substrate, of substantial change.
The original referent upon which the Aristotelian
concept of matter is based is therefore the subject of
accidental change, like wood or bronze. From that
notion of "subject," however, all determinations are
removed, with the proviso that the negations as well
as the determinations are accidental to it. In its own
nature, then, this refined notion of subject remains
as positive as ever. It was a positive notion from the
start, as seen in a positive subject like wood or bronze,
and all determinations were denied it under the express
condition that none of these pertained to its own
nature. In this way the notion positive is shown to be
independent of determinate. For Aristotle, 'actual' was
a synonym for 'determinate'. What lacked actuality,
or in technical language the potential, could therefore
be positive. By establishing the concept of the poten-
tial as positive even though non-actual or indetermi-
nate, Aristotle has been able to set up matter as a
positive though entirely non-actual subject of predica-
tion. Because the potential is positive without being
19Ph. I 7, 191a7-12; Oxford tr.

determinate, this concept of matter is possible to the
human mind. Its referent is any sensible thing con-
sidered potentially as substance. It is the concept of a
principle wholly undetermined, yet necessarily posited
in reality by any form that is extended, multiplied in
singulars, or terminating substantial change.
§5 CONCLUSION
As should be clear from the foregoing considerations,
matter in the category of substance can be an object
of scientific inquiry only on the level of natural philoso-
phy. It cannot at all be reached by qualitative or quan-
titative procedures like those of chemistry and modern
physics. What is predicated of it, in itself, does not
belong to the order of the measurable or the directly
observable, even in principle. Its predicates are notions
like purely potential, unknowable ot itself, incorrupti-
ble, and so on. Its presence is still necessary to explain
substantial change, if such change is admitted. In any
case, its presence is absolutely required to account for
the extension of a formally identical characteristic in
parts outside parts, and for the multiplication of the
characteristic in a plurality of individuals, without any
formal addition whatsoever. The Aristotelian matter
has not been superseded nor even touched by the
stupendous progress of modern physics. Nothing that
is measurable can perform its function in explaining
the nature of sensible things, and by the same token
it cannot be brought forward to account for anything
that requires explanation in measurable terms. Any
type of matter dealt with by chemistry or modern
physics would in comparison be secondary matter, and
not matter that is a principle in the category of sub-
stance. "Matter" in the basic Aristotelian sense is there-
fore in no way a rival of the "matter" that can be
measured or of the mass that can be transformed into
212 Joseph Owens
energy, but is rather a very different means of explana-
tion for sensible things on another scientific level, the
level of natural philosophy.
In distinguishing his two tables, the solid one he
wrote on and the "nearly all empty space" table he
knew as a physicist, Eddington failed to stress that his
knowl~dge of his scientific table was constructed from
his knowledge of the ordinary table.20 The scientific
construct was the result of understanding the ordinary
table in quantitative terms. The same ordinary table
can also be understood scientifically (in the traditional
sense of knowledge through causes) in terms of sub-
stantial principles, form and matter, as is done in
natural philosophy. It can also be understood in terms
of entitative principles, essence and being, as is done in
metaphysics. They are all different accounts of the
same thing, given on different levels of scientific (again,
in the centuries-old meaning of "scientific") investiga-
tion. All these different accounts are necessary for a
well-rounded understanding of sensible things. None
of these accounts can afford to despise any of the others,
nor seek to substitute for any of them, nor to interfere
with any of them. Each has its own role to play, a
role that only itself can play. The Aristotelian matter
is a principle for explaining things on the level of
natural philosophy. On that level it has its own predi-
cates, predicates that still have to be used today in the
properly balanced explanation of nature.
20See Arthur Stanley Eddington, The Nature of the Physical

World (Cambridge, Eng., 1928), pp. ix-xi. Cf.: "The whole
reason for accepting the atomic model is that it helps us to
explain things we could not explain before. Cut off from these
phenomena, the model can only mislead, ..." Stephen Toulmin,
The Philosophy of Science (New York, 1953), p. 12.
APPENDIX
On the independence of these different scientific proce-

dures, see my paper "Our Knowledge of Nature," Proceed-
ings of the American Catholic Philosophical Association,
XXIX (1955), 80-86. A widely accepted view at present is
to regard natural philosophy as a sort of dialectic that pre-
pares the way for genuine physics; e.g.: " ... frontier phys-
ics, natural philosophy. It is analysis of the concept of
matter; a search for conceptual order amongst puzzling
data." Norwood Russell Hanson, Patterns of Discovery
(Cambridge, Eng., 1958), p. 119. "Not so very long ago
the subject now called physics was known as "natural phi-
losophy". The physicist is by origin a philosopher who has
specialized in a particular direction." (Arthur S. Edding-
ton, The Philosophy of Physical Science [Cambridge, Eng.,
1939], p. 8.) It is true that before physics was developed
through quantitative procedure as a special science, its
problems iKd in point of historical fact been given over to
the non-ma'hematical treatment of natural philosophy.
That way of dealing with its problems was entirely illegiti-
mate. The specific differentiae of natural things remain un-
known and impenetrable to the human mind. They cannot
be made the source for scientific knowledge of the specific
traits of corporeal things. For this reason any new attempt
to treat the experimental sciences as a continuation of nat-
ural philosophy, e.g. C. de Koninck, "Les Sciences Experi-
mentales sont-elles Distinctes de la Philosophie de la Na-
bue?" Culture, II, (1941), 465-76, cannot hope to be
successful. On the other hand, the view that natural phi-
losophy consists only in "a search for conceptual order
amongst puzzling data" seems continuous with the trend
that has given rise to the conception of philosophy in gen-
eral as linguistic analysis, concerned with words and con-
cepts and not with things. Similarly, the notion that natural
philosophy is a frontier investigation rather than a full-
fledged science in its own right, seems to stem from
Comtc's law of the three stages, in which speculative phi-
losophy in general was but an immature stage in the uni-
linear development towards positive science. In this view,
214 Joseph Owens
philosophical treatment "will naturally be expected to deal
with questions on the frontier of knowledge, as to which
comparative certainty is not yet attained." Bertrand Russell,
Introduction to Mathematical Philosophy (London, 1930),
p. v. This is nothing but a cavalier dismissal of natural'
philosophy as a science.
Theoretically, it is indifferent whether the substantial
principles (matter and form) used for the explanation of
things through natural philosophy are reached by way of
substantial change, or of extension, or of individuation. In
point of fact, the way used by Aristotle himself was through
substantial change. To show that the same two principles
are required to explain extension and multiplication of
singulars, is not a tour de force to safcguard the principles
against someone who does not admit substantial change. It
is rather a global view of the whole approach, from a theo-
retical standpoint, to the problem of matter.
May I express my thanks to Msgr. C. B. Phelan for
many helpful suggestions, and to Fr. Eman McMullin for
carefully reading the first draft of this paper and pointing
out a number of deficiencies. These I have tried to remedy
in the final draft. This draft, of course, benefits from the
other papers and the discussions at the conference, and in
particular from the clear statement of the issues in Profes-
sor Fisk's contribution. Friedrich Solmsen's recently pub-
lished work, Aristotle's System of the Physical \Vorld
(Ithaca, N.Y., 1960), with its illuminating discussion (pp.
118-26) on the historical background of Aristotle's wholly
undetermined matter, reached me too late to be of help in
preparing the present paper.
PROBLEMS IN
METAPHYSICS Z, CHAPTER 13
M.J. WOODS
The purpose of this paper is to attempt a clarification

of the theory of substance which Aristotle had devel-
oped at the time when he wrote books Z, H, and 9 of
the Metaphysics, by considering a passage where he
puts forward a number of arguments against a certain
doctrine about substance. Some of my conclusions
could, I think, be supported by an examination of cer-
tain other passages in Z. But in this paper I confine
my discussion to Z 13.
As Ross points out in his edition of the Meta-
physics, chapters 13-16 form a single section, the
main upshot of which is summed up in 1041a3-5: 1
~'tl !lEv oov -rwv Ka96Aou AEY0!lEVUlV ouoEv ouo(cx
ou-r' EO-rLV ouo(cx 0600!l(cx E~ OUOlWV, O~AOV*
The suggestion that -ro: Ka96AOU AEy6!lEVCX have
claims to be regarded as substances is raised at the
beginning of chapter 13, a discussion to which Aristotle
looks forward in chapter 3 at 1028b34. It is in the
course of these four chapters that the main, though
lLine references are all to Jaeger's Oxford text of ivfetaphysics.

[For the convenience of the reader, translations of some Greek
passages appear in footnotes. The translations used are those of
W. D. Ross and J. L. Ackrill.-Ed.]
*"Clearly, then, no universal term (TWII Kcx96AOU AEyo~EvCilv
ouBEv) is the name of a substance, and no substance is composed
of substances."
H
216 M. J. Woods
not the only, criticism of the Theory of Forms occurs
in this book.
To turn to detailed discussion of chapter 13: the
exact course of the argument is difficult, and at some
points the text is disputed. However, it is fairly dear
that the main part of the chapter ends at 10 39a 14.
an appendix. Beginning EX£L oe 1:0 oUfl~aLvov cmop[a
a dilemma is presented; it is a firmly held doctrine
that it is substance which is the object of definition;
but the earlier arguments tend to show that no actual
substance is composite (OUV9E1:OV). So substances ap-
pear to be indefinable. This dilemma is clearly before
Aristotle's mind in the succeeding three chapters.
It is important to make clear exactly what doctrine
Aristotle was attacking in these chapters. He was op-
posing the view that a correct answer to the question,
"What is oCo[a?" would be given by saying that it is
1:0 Ka90Aou. Thus he is denying that something
Ka90AOU is as such an ouo[a. To deny this is not to
claim that nothing Ka90AOU is an ouo[a, only that
being Ka90Aou is not by itself a sufficient ground for
so describing something. I wish to argue that Aristotle
was in fact maintaining that only some things properly
described as Ka90AOU are to be regarded as ouv[al;
something Ka90AOU iis an ouo[a only if it is not predi-
cated universally (Ka90Aou AEyoflEVOV). That is, the
main purpose of these chapters is to deny that anything
Ka90Aou AEyoflEVOV is an ouo[a. I will return to the
distinction between being a universal and being pred-
icated universally later. '
We notice that he begins his discussion by saying
(1038b3f.) that some have regarded 1:0 Ka90AOU as an
alnov or dPX~. * and it therefore needs investigation.
No indication of who holds this is given, but it is dif-
ficult to believe that Aristotle would not have regarded
*"Cause" or "principle."
Problems in Metaphysics Z 217
the Platonic Theory of Forms as an example of the
sort of theory he had in mind. 2 I shall refer to Aris-
totle's opponcnt for convcnience as "the Platonist,"
without intending to beg any question by the use of
this label.
He continues (1038b8-9) with the remark that it
is impossible for anything which AEYE'WL Ka86AOU
to be an ouo[a. For that which is predicated univer-
sally is necessarily something which can be common
to many things (0 TIAE[OOLV UTIO:PXELV TIE<jlUKEV)
whereas the ouo[a of something is that which is
peculiar (loLOe;) to it. To Ka86Aou cannot therefore
be the ouo[a of all the sct of objects of which it is
predicated; equally it cannot be the oUola of nothing.
But if it is the ouo[a of one of them, then all the
other members of the set will have to be identical with
this one, which is absurd; for things which share one
ouo[a and T[ ~v ElvaL are one.
According to Ross, 3 the argument has to be inter-
prcted as follows: the Platonist seeks to satisfy the
requirement that the universal be '(OlOV to a single
thing of which it is the ouo[a by supposing it to be
the ouo[a of just one of its particulars; but unfortunate-
ly it has as good a claim to be the oUo[a of the other
particulars of which it is predicated; so if the require-
ment of uniqueness is to be fulfilled, these others must
be identical with the one selected, which is absurd.
(From now on I shall call the requirement that an
ouo[a be 'loLOe; to that of which it is the ouo[a "the
uniqueness requirement.")
This interpretation, which requires us to sup-
ply a good deal, seems untenable. It is, of course,
2(;£. H 1042aI5-16: Tc';l O£ Ko90AOU Kol Tc;i YEVEI Kol oi lOEOI
OIlVO:1TTOUOIV (KoTa TOV OLlTOV yap Myov OUOla! OOKoOCl'lV ElVa!)
["And with the universal and the genus the Ideas are connected
(uuvarrTouUlv); it is in virtue of the same argument (KOTe). TOV
a(lTC~ AOYOV) that they are thought to be substances"].
3Aristotle's Metaphysics (Oxford, 1924), Vol. II, p, 210.
218 M. ,. Woods
true that if the Platonist said that the universal
animal were the oOala of only one particular
I;(;)ov (say, Socrates), not of Callias, Bucephal us,
and Fido, his position is vulnerable and indeed
absurd. Any such selection would be arbitrary. But
nothing in the text suggests that Aristotle regarded
the argument as vulnerable in just this way. Moreover,
the very absurdity and indeed lunacy of the position
put into the mouth of the Platonist may be thought
to be an argument against the view that he is toying
with the possibility of satisfying the uniqueness re-
quirement in this way. Also, as reconstructed by Ross,
the argument seems to beg the question. The Platonist
begins by supposing that a universal like I;(;)ov is the
oOala of just one animal, e.g. Socrates; he is then told
that he will be committed to regarding 'ta If'A'Aa as
identical with him, since 'ta o:''A'Aa will also have 'to
Ka9o'Aou as their oOala. But this is precisely what the
Platonist is desperately denying. Ta O:.AAa must, on
this view, refer to those particulars of which 'to
Ka96'Aou is predicated other than the one of which it
is the oOo(a; and the Platonist will not agree that they
have '[0 Ka86'Aou for their oOa(a without argument,
for the position he has just taken up is the denial of
this. As the passage stands he is represented as agreeing
without argument that all particulars of which '[0
Ka86'Aou is predicated have it for their ooo(a, and not
simply one of them. So Ross's interpretation does not
really find any argument in the passage at all.
Cherniss interprets the passage in a different way.4
According to him, the sentence hoc; 1)' el EO'tal, Kat
'[dA'Aa '[00'[ Eo'[al'" provides a reason for the dis-
junction stated in the first half of the previous sentence
4AristotZs's Criticism of Plato and the Academy (Baltimore,
1944), p. 318, n. 220.
'" "If it is to be the substance of one, this one will be the others
also."
Problems in Metaphysics z 219
("either TO Ka8oAou is the OUOla of all the particulars
falling under it or of none of them") ; while
&v yap flla ~ OUOla Kat TO 1:l ~V Etval EV, Kat a(31:a EV*
provides a reason for saying that it cannot be the
OUOla of all of them. This interpretation avoids the
difficulty mentioned just now in Ross's interpretation:
that no argument is presented which could lead the
Platonist to abandon his suggestion that 1:0 Ka8oAOU
might be the OUOlo: of one thing. On the other hand,
it is hard not to take 11, 14-15 (&v yap flla ... )
as providing a reason for what is asserted immediately
before, especially as it is readily intelligible as a reason
for it. But even if we accept Cherniss' view, the pas-
sage concerned (11, 9-15) seems best interpreted, as
I shall argue, not as an argument which is decisive as
it stands, but as one which shows what a Platonist is
committed to if he insists on regarding TO Ka80AOU as
ooola. He will be committed to accepting it as the
OUOla of all its particulars, and treating these as iden-
tical with one another.
In order to see how this and later passages in these
chapters are to be interpreted, we need to raise a major
problem which immediately faces anyone who reads
this passage in the light of Aristotle's criticism of
Plato. Aristotle's answer to the question, "What is
ouola?" is that what is OUOla in the fullest sense is
the EtiSoc; or Tt ~v EIVaL of something. This comes
out very clearly in chapter 17, but it is already present
as a doctrine in the chapters of Z before chapter 13.
It emerges from the identification of diSoc; with
1:t ~v ElVaL and of this with TIPWTT) OUota at 1032bl-2.
Thus Aristotle is presumably committed to hold-
ing that the form of the species man is a sub-
stance. But this seems incompatible with the doctrine
that nothing Ka8oAOU can be a substance: for man
""For things whose substance is one and whose essence is one
are themselves also one."
220 M. J. Woods
is surely predicated universally of Socrates, Callias, etc.
How can the species man be an ouo[a, if any 000[0:
has to belong c0e; LOLOV to that of which it is the
ouo[a? It is clearly of no use to seek to avoid this
difficulty by saying that, when discussing the suggestion
that TO: K0:86AOU are ouo[m, Aristotle has in mind only
the higher genera into which species fall. For his own
theory of substance may obviously still be open to the
objections he makes to the claims of 1:0: K0:86AOU
AEy6[lEva to be substances even if he himself only
uses these arguments against the claims of higher gen-
era to be ouo[m.
One way of escaping these difficulties is to suppose
that in Z Aristotle, where he says that ouo[a is dooe;,
has in mind not the form common to all members of
a species but something peculiar to an individual mem-
ber. Thus the ouo[a of Socrates will be peculiar to
him and will not be something predicated of anything
universally. The thesis that, as well as the form of
species, there is a fOlm peculiar to each member of a
species to be found, almost certainly, in A, and fur-
thermore, the doctrine that in the case of animate
objects their form is their soul is difficult to interpret
without presupposing some such view as this. If this
view can be found independently in Z, we shall have
some grounds for interpreting chapter 13 along
these lines. There will then be no conflict between the
requirements for being an ouo[a stated at the begin-
ning of 13 and the doctrine that ouo[a is form.
This question has been discussed in an article by
Professor R. Albritton. 5 Albritton finds clear evidence
that Aristotle believed in a distinct form for each in-
dividual substance in /\ and M of the Metaphysics
and in the De Anima; but in Z and H of the Meta-
physics the most he can find is some evidence that
5"Forms of Particular Substances in Aristotle's Metaphysics,"
Journal of Philosophy, Vol. LIV, No. 22 (October 1957).
Aristotle accepted that there was a distinct form for
each living substance. This evidence lies in the passages
in which the doctrine that the form and Quota of a
man is his soul occurs. So whether or not these passages
be regarded as evidence for the doctrine of particular
forms depends on whether we regard the doctrine that
the soul is the form of the body as intelligible only with
such a view. But, in any case, the most that these pas-
sages could be said to point to is the doctrine that there
is an individual form for each living individual. How-
ever, if Z 13, the passage with which we are immediately
concerned, be interpreted as showing that Aristotle
thought that there was a distinct form for each indi-
vidual substance, it would show that Aristotle regarded
the doctrine as holding for all individual substances,
and not merely those with a ljJUX~; for the argument
would be a general one. Since nothing KaeOA.OU
AEYOf!EVOV can be an oUota, and the form of the
species is predicated of the plurality of individuals in
the species, the QUOta of each individual substance,
whether animate or inanimate, must be a form which
is peculiar to it.
Albritton considers the suggestion that chapter 13
implies a doctrine of particular forms and argues that
there is nothing in the chapter which forces us to in-
terpret Aristotle in that way. To quote Albritton (p.
706): "One might distinguish, as Aristotle does, ways
of being 'one' and agree that things whose substance
is one need not be one in every way but only in that
of their substance. These are many in number. It fol-
lows that nothing one in number can be their sub-
stance. But the universal form of man is not one in
number .... It is only one in form. And men are one
in form. The one form of man may, therefore, be their
substance."
This interpretation is very attractive, and I agree
with the general approach. But as it stands, it seems to
222 M. J. WOlJds
allow the Platonist a very obvious rejoinder. If the
form of a species can be said to be an ouo[a, com·
patibly with the uniqueness condition, because the
particulars are tfv ELOEl, one might equally hold that
the genus could be said to be an ooo[a also, in spite
of being predicated of many things, since these things
(species and particulars) though not EV d:pl91l4'> nor
EV ElOEl are EV yEVEl.* Now it may be thought that
the sense in which a group of things which belong to
the same genus can be said to be one is weaker than
the sense in which things in the same species can be
so described, and that therefore the Platonist would
at least have to admit that genera like animal are less
fully entitled to be called ooo[al than species like man.
But at least this defense would seem to show that
"[0: Ka96AOU have some claim to be regarded as ouo[al;
and this is something which Aristotle seems firmly to
reject in this chapter. He insists roundly that nothing
Ka96AOU AEy61lEVOV is a substance, and these asser-
tions are difficult to reconcile with the position that
they are substances of a sort, though less fully than
certain other items. If he were prepared to concede
that both man and animal qualify as ouo[al though
the latter less than the fonner, we should have a posi-
tion like the one he takes up in the Categories
(2b7 f.: "[WV os. OEU"[EPCo)V OUOlWV IlCXAAOV "[0 Eloot; "[OU
yEVOUt;).1" But it seems clear that this view is now
abandoned. So we have the position that if we inter-
pret the chapter along the lines that Albritton suggests,
and make the fact that man is the ouo(a of many
individuals compatible with the uniqueness require-
ment, the Platonist will be able, apparently, to find
a comparable sense in which animal can be said to be
*"One in number," "one in form," "one in genus."

t"Of the secondary substances the species is more a substance
than the genus, since it is nearer to the primary substance."
Problems in Metaphysics z 223
an OUOtcx: and that is something Aristotle will not
accept.
Although Albritton does not consider this way out
for the Platonist, the next paragraph of his article sug-
gests an answer that Aristotle might have made to the
position. He says: "The argument would not allow
any universal of species to be the substance of its
species. The species of a genus, for example, are pre-
cisely many in form, not in number, and therefore this
genus, which is one in form cannot be their substance."
But where does Aristotle say that a genus is one in
form? In order to find the argument acceptable we need
to know that the genus animal is one in form in pre-
cisely the same way as that in which the species or
individuals of which animal is predicated are not one
in form. Only if that is so can we regard the argument
as establishing a difference in the relation between
man and the members of this species on the one hand
and the relation between animal and the species and
individuals of which it is prcdicated on the other.
Albritton does not mention any passage in which
Aristotle says that a genus is one in form. But
he draws attention a little earlier, amongst other
passages, to the beginning of Book I of the Metaphysics.
There Aristotle says that those things are one
wv ~ V6TjOlC; IIIcx, 'tOlCXU'tCX b' L)VablcxlpELOC;, ablCXtpE'tOC;
M 'tou ablCXlpE'tOU EtbEl ~ apl9114l (1052a30). *
Again, a little later he says 'to Evl ElvCXl 'to
ablcxlpE'tct> to'tlv Elvcxl. t
Since he does not in this
passage explicitly mention YEVTj as examples of things
which are EV E'(bEl we must consider whether he would
in fact have allowed a YEVOC; to be something which was
*"... the other things that are one are those whose definition
is one. Of this sort are the things the thought of which is one
(&v r1 v61']0"1t; ilia) i.e. those the thought of which is indiviso
indivisible (alilaipETOt;); and it is indivisible if the thing is
indivisible in kind (EiIiEI) or number."
t" 'To be one' means 'to be indivisible'."
n*
224 M. J. Woods
aOta[pE·rov ElOEt. It seems to me that he would not.
When he says that something is EV E'LOEl if it is
O:Ota[pE,ov ELOEl, he seems to have in mind the same
idea that he expressed Z 10 34a8 (for example). There
he says that Socrates and Callias are ,au,6 O£ E'iOEl
((hOflOV yap ,6 ELooc;). * An E'iOOC; is ch6floV in the
sense that it is not capable of further differentiation.
This is precisely what is not the case with yEVT]. If
lh6floV in this passage means the same as 'aOta[pE,ov'
in Book I, then it seems that we are forbidden to treat
a Y€VOc; as being EV ElOEt in the sense distinguished in
that passage. Moreover, even if passages can be found
in which a genus is described as EV ElOEl, this will
presumably be in a wider sense of dooc; in which it
is not contrasted with 'YEVOC;', and therefore not that
sense in which the various species of a genus are not
EV ElOEl. Again, if it is the case that a YEVOC; is correctly
described as EV ELOEl, presumably, derivatively from this,
there will a sense in which the different species of the
genus can themselves be EV E'iOEl, which could be ex-
ploited by the Platonist.
I conclude that Albritton's suggestion does not pro-
vide Aristotle with a way of meeting the argument we
have put in the Platonist's mouth: that if the form
of a species can be allowed to be a substance compatibly
with the uniqueness requirement, so can a higher genus.
The difficulty seems to be that the suggestion does
not take seriously enough Aristotle's several times
repeated statement that nothing predicated universally
(Ka96Aou AEy6flEvov, Ka'T]yopouflEVOV, unapxov) is
an ouo[o:. On the view we have just criticized, Aristotle
is saying simply that nothing that is predicated univer-
sally can be a substance, unless the plurality of objects
of which it is predicated can themselves be said to be
one in a certain sense. This runs into the difficulty
*"The same in form; for their form is indivisible."

that the Platonists can provide a sense in which the
things of which a genus is predicated can be said to
be one. Albritton's suggestion implicitly makes the
predication of a universal of many things compatible
with the satisfaction of the uniqueness requirement.
But what Aristotle says is that since to Ku90AOU is
common to many objects, it cannot be tfllOV to one.
He does not restrict himself to saying that nothing can
be a substance which is common to plurality of ob-
jects, unless the many are also one. The suggestion
under discussion waters down Aristotle's remarks con-
siderably, since it treats the doctrine that nothing
Ka90AOU A£yoll£vOV is a substance as one accepted by
Aristotle only with considerable qualification.
If we take these remarks as they stand and wish still
to hold that when Aristotle says that the £1ooe; is
ouo(u he has in mind a form common to all numbers
of a species, we must suppose that he would have de-
nied that the form man is predicated universally of
Socrates, Plato, Callias. If genera, by contrast are pred-
icated of a plurality of objects, we have found a way
of interpreting chapter 13 which allows Aristotle, con-
sistently with his own doctrine that substance is form,
to deny that anything predicated universally is a sub-
stance. For the remarks about to: Ka90AOU A£y0f.l£va
will have application only to genera. I wish to claim
that this is precisely the position that Aristotle adopts
in Metaphysics Z.
This doctrine that the form of a species is not pred-
icated universally of the members of the species in-
volves an obvious departure from the sort of view ad-
vanced in, for example, the Categories. Indeed it might
be said that the theory rejected in chapter 13 is as
much the theory held by Aristotle himself earlier as
it is that of Plato, though it is probable that in Z 13
Aristotle has Plato mainly in mind, in view of the
contents of the succeeding chapters.
226 M. J. Woods
In saying that Aristotle denied that a species may
properly be said to be predicated universally of its
members, I am not of course denying that Aristotle
might have allowed that the name of a species was
predicated of its members. Nor, perhaps, would he
have said that it was for all purposes impermissible
to speak of the form of a species as being predicated
universally of its members. What he seems to have
thought is that, for the problems with which he is
concerned in Z, it is incorrect to say that a species form
is predicated universally of a plurality of individuals.
It may be helpful at this stage to specify exactly what
Aristotle thought could properly be said about species
and what could not, on the view that I am advancing.
Firstly, he appears to be denying that a species could
be predicated universally of its members, and he ex-
presses this in various ways. No ouokx (and therefore
no species) is KaS6Aou AEy6flEVOV (1038b9): again
(1038b35) none of the things KaS6AOU urrapxOV"ra
is an ouo[a; the reason given in the next line is that
none of the things KOlVD Ka1:T]yopoflEva* is 1:60E 1:l,
earlier (1038bll) the claims of the Platonist are re-
jected because he allows to be a substance something
which is KOlV6v; it is KOlV6v because rrAE[OOlV
urrapXElv rrE<!'uKEv. At the end of the whole discussion
in these chapters (16, 104Ia4), his conclusion is re-
stated: "(WV KaS6Aou AEYOSEVWV ouoEv ouo[a. An
oOo[a is not common to a plurality of objects; it is
"(60E "(l. A genus is not "(60E 1:l but "(ol6vOE (I039al).
So there is a clear contrast between the things that
Aristotle is willing to say about species-forms and the
things that can, in his view, be said of genera.
It is interesting to note that in the Categories he
is unwilling to say that a secondary substance (in the
·"Common predicates_"
Problems in Metapllysics Z 227
tenninology he then used) properly speaking signified
't60E 'tl (3b14 f.). At that time he did not object to
the notion that a species was predicated of a plurality
of objects. A secondary substance is then described as
1toI6v 'tl, * and the reason given is that Ka'tO: 1tOAAWV
6 av9p(')1t0C; AEYE'tal Kat 'to 1;;wov. t Thus, already at
the time when he wrote the Categories, he regarded
being predicated of a plurality as strictly incompatible
with being 't60E 'tl. In Metapllysics Z, where the title
1tpw'ty] oOo[a is given to an dooc;, he rejected the idea
that an dooC; is predicated of a plurality along with the
idea that it cannot strictly be called 't60E 'tl. It is charac-
teristic of the Categories to regard the difference be-
tween genus and species, in respect of right to the name
oOo(a, as merely a matter of degree; and this, I suggest,
goes with a view which regards the relation between a
species and the particulars belonging to it as essentially
the same as that between a genus and its species. In
the Metapllysics he thought of genera as KOlVn
Ka'tY]yop06~Eva; and this is regarded as incompatible
with saying that a genus is 't60E 'tl. However, in Meta-
pllysics Z 13 he no longer says,as he does in the Cate-
gories that a genus is not 'tOOE 'tl but 1tOlOV; instead
he says (1039al): OOOEV OY]~a[vEl 'tWV KOlvTI Ka'tY]yo-
pOU~EV(')V 'tOOE 'tl &XAO: 'tOlOVOE.:j: The word 'tol6vOE
replaces 1tOlOV as the other term of the contrast with
't60E 'tl. As the use of the term 1tOlOV in connection
with something which belongs in the category of sub-
stance is obviousy potentially very misleading, it is not
difficult to see why he should have felt the need to use
another word. But I suggest that the reasons he has in
Metapllysics Z for regarding a genus as not 't60E 'tl but
saying that an dooc; is strictly not 't60E 'tl but 1tOl6v.
Another passage in Z which is worth noticing is the
*"A certain qualification."
t"For both man and animal are predicated of many things."
:j:"No common predicate indicates a 'this', but rather a 'such'."
228 M. ,. Woods
well-known passage explicitly about the Theory of
Forms in 16, 1040b27: or...r...' ot tex E'15T] MyOV't:EC; "tft
~EV cpewe; AEYOUOlV XWp[l:OV1:Ee; alJ"t(x, EL-rtEp ouo(m
£10(, "tft fl' OUK cpewe;, on "to EV E1tl 1tOAAWV E150e;
MyOUOlV.* It is the last phrase I am concerned with.
When Aristotle says that the Platonists went wrong in
describing, "to EV E1tl 1tEAr...WV as an £I5oe; he says some-
thing which it would be difficult to square with a
view that E'(5T] in one sense are properly described as
EV E1tl 1tOAAWV. Admittedly, Aristotle's conception of
£,(5T] was different from Plato's. But in a passage where
Aristotle tells us what is right about the Theory of
Forms as well as what is wrong, it would be strange
if he singled out as the fault in the theory that it calls
something which is 'one over many' an d50e; if he
himself were ready to use the word d50e; of some-
thing which was EV ETIl 1tOAAWV. I suggest that Aris-
totle would have denied that the species man was
something 'ETI1' individual men. His complaint against
Plato is that the Theory of Forms treats species and
genera as if they were alike; genera which are cor-
rectly described as EV ETIl 1tOAAWV are being treated as
if they were £loT] and therefore substances. The denial
that an £I50e; is EV ETIl TIOAAWV goes with the denial
that anything KOlvfj KatT]yopOl)~EVOV is a substance.
This interpretation fits in well with the context, when
it is clear that he still concerned himself with the
evaluation of the claim of things predicated universally
to be substances.
Before I return to a detailed discussion of chapter
13, I ought to deal with a possible objection. It may
be said that Aristotle could not have denied that the
form of a species was something KaeOAOU and there-
*"But those who say the Forms exist, in one respect are right,
in giving the Forms separate existence, if they are substances; but
in another respect they are not right, because they say the one
over many is a Form."
fore the denial that any universal is a substance is in-
compatible with the view that it is ELOT] which are
ouo[m, if the Etoll he has in mind are Etoll of species
and not E'(511 distinct for each individual member of
a species. Apart from the inherent implausibility of
denying that Etoll are Ka8oAou, Aristotle himself com-
mits himself to this in Z. At 1035b27 f., he speaks of
6 0' o:v8pwTtoC; Kal 6 tTtTroc; Kal 1:a olhwC; ETtl 1:C>v
Ka8' EKa01:a, Ka8oAou oE etc. * The context is full of
difficulties, but it is clear that what he has in mind in
this passage are the forms of the species to which men
and horses belong. Again, it is substances which are
capable of definition in the strict sense, but definition
is only of 1:0 Ka8oAou (d. 1035b34, and elsewhere).
As I have indicated earlier, my reply to this is that
a distinction has to be made between the question
whether an item is something Ka8oAOU and the ques-
tion whether it is Ka8oAOU AEyOf-lEVOV. What he is
concerned to deny is that that which is predicated un-
iversally is a substance. He is not denying that some-
thing which is Ka8oAou may be an ouo[a; he is saying
that nothing which AEynal Ka8oAOU or is KOlvTI
Ka1:11yopouf-lEvov is an ouo[a. Thus I think that when,
early in chapter 13 (1038bl1-12) he says: 't001:O yap
"-EYE'tal Ka8oAou 8 TtAE[OOlV UTtO:PXElV TtE<jluKEv, we
should regard him as saying, not that something is
called a universal if it is such as to belong to several
things, but that something is predicated univesally if
it is such as to belong to several things. Again, when
he sums up the whole discussion at the end of chapter
16, he states his conclusion by saying not simply that
nothing universal is a substance, but that nothing
predicated universally is one.
*"But man and horse and terms which are thus applied to
(Err!) individuals, but universally, are not substance but something
composed of this particular formula and this particular matter
treated as universal."
230 M. J. Woods
It is time now to return to a detailed consideration
of chapter 13. This general discussion arose out of a
difficulty in seeing how 1038b8-15 provide an argument
against the Platonist. I shall return to this passage
later. In 11, 15-16, he offers another brief argument
against his opponent. An ouo[a is something which
is not said of a ll'ltOKElflEVOV, whereas the universal is
always said of ll'ltOKElflEVOV 'n, The last statement will
have to be regarded, on the view I am defending as
claiming, not that everything universal is said of
&rtOKElflEVOV 'rl, but that everything predicated univer-
sally is said of ll'ltOKElflEVOV 'rl. If the sentence is taken
in context, this reading is, I think, possible. If this is
correct, Aristotle must deny that an dooe; MYE'tal
Ka9' lI'ltOKElflEVOU 'tlv6e; if it is to be regarded as an
eUo[a. This conflicts, apparently, with what is said in
chapter 3, at 1029a23-24: 'ta ~EV yap 6:.AA.a 'r~e;
ouo[c:c; KaTIJyopEl'ral, athT] Of. 'r~c; UAT]C;.* But I think
that the conflict need not worry us greatly. Since the
word lI'ltOKElflEVOV is used in several senses, the ques-
tion whether something MYE'ral KaS' lI'ltOKElflEVOU
'rlv6c; will also have a number of senses, and when
asked about a given item, will have different senses
according to how it is taken. The statement in chapter
13 that no ouo[a MYE"tal K<XS' lI'ltOKElflEVOU 'rlVOC;
is to be regarded as equivalent to the denial that an
eUo[a is, properly speaking, predicated of the individu-
als which belong to it. This will not prevent Aristotle
from saying that an ELOOe; is, in a sense, predicated of
matter.
Before returning to the argument with which we
started, I will now consider the next section, 1038bl6-
30. These are taken by Ross and others to be a series
of further arguments offered by Aristotle against the
view that any universal is a substance. The first argu-
*"For the predicates other than substance are predicated of
substance, while substance is predicated of matter."
ment is rather obscure; in 1, 19 Eon is corrected by
Jaeger in the Oxford Classical Text to 'E01:W', to bring
it into line with E01:W in 1, 22. This passage, 11, 16-23,
is interpreted by Ross as follows: The Platonist con-
cedes that something K0:90AOU cannot be regarded as
an OUo[o: in the way a 1:[ ~v. ElVW is; none the less a
genus is an element in a 1:[ ~v E{VW, and thus ought to
have some claim to be regarded as an 060[0:. But it
itself will have to be the V1.E. of something in just the
way that the V].E. in which it occurs as an element is.
Thus the amended suggestion of the Platonist amounts
to the same as the already rejected suggestion that
something K0:90AOU can be an ouo[o: through being a
V1.E.; it is subject to the objections raised earlier in
the chapter.
At first sight this is plain sailing. The Platonist
amends his position to meet the difficulties raised by
Aristotle; Aristotle then shows that the amended posi-
tion is really the same as the original one. But if we
look at the actual reasons for saying that the new
position is the same as the old, the passage becomes
more and more puzzling. The reason given is that it
will be £KElVOU OUo[o: EV 0 ElOEl Q<; tblOV UTIeXPXEl*
(if we follow Jaeger and excise 'olov 1:0 I:;wov'). But
surely if it is conceded that it is the OUo[o: of some-
thing to which it belong Q<; tblOv, then this under-
mines the reasons given earlier for saying that nothing
K0:90AOU AEyoflEVOV can be said to be an OUo[o:. The
argument given earlier was that nothing which did not
satisfy what I called the uniqueness requirement can
be regarded as an oua[o:. But now Aristotle appears to
be saying that it does after all satisfy the uniqueness
requirement. It is no use to appeal to the earlier estab-
lished conclusion that nothing K0:90AOU AEY. is an
oua[o: in the way that a V].E. as if the reason for saying
* "The substance of that in which it is present as something

peculiar to it."
232 M. J. Woods
that it is after all a T.11.£. is that it belongs to a unique
ElOoc;; it was precisely because this condition is not
fulfilled in the case of Ka90AOU AEyo[lEva that it was
denied that anything Ka9. AEY. can be an ouota in the
earlier argument. If it now turns out that something
Ka9. AEY. can be the ouota of a single Elooc;, the
earlier argument is called into question. The way out
of this difficulty that I suggest is that 1038b22-23 not
only does, but is intended to, undermine the earlier
argument. The words in question are not used by
Aristotle propria persona but are part of a set of argu-
ments which he represents the Platonist as advancing
in favour of his position. I think we should regard the
whole passage from lines 16 to 30 as put into the mouth
of Aristotle's opponent. Aristotle's own objections to
the position begin at line 30: 'OA(;)C; OE oU[1~atvEl .. .'
On the usual interpretation, only 11. 16---18 are attrib-
utable to the Platonist, and from 'OUKOUV OfjAOV'
onwards there are three arguments against his amended
position. But it seems to me that much better sense
can be made of the whole passage down to line 30
if we suppose that it is the Platonist who is speaking.
The first argument, in 11. 16---23, can then be inter-
preted as follows: The Platonist concedes that it is
not an ouota in the way that '(0 ,(l ~v Elvm is. This
concessio'n is temporary; what he means is that it is
not an ooola in the way that what are agreed to be
ouolm, namely species, are; it will nevertheless be an
element in a substance-species. It will have a MyoC;;
the fact that the MyoC; of swov is not a complete
MyoC; of the OUOla man makes no difference; since it
has a MyoC;, this MyoC; will define a unique class of
species of which it is the OUOla in just the way that
o:v9p(;)TIoC; is the OUOla of individual men. In other
words, the Platonist is insisting that the very same
reasons that enable one to regard o:v8p(;)TIoC; as an
QUCJla consistently with the uniqueness requirement
Problems in Metapbysics Z 233
can also enable I:;wov to be regarded as an 060 [a:. He
insists that species and genera be treated on the same
footing. It is noticeable that in 1038b23 he says,
~v 4'> Ei:OEl w<; 'if>lOV lJTIeXPXEl. This would be an odd
remark if Aristotle were talking in his own person; the
plurality of species of which I:;wov is predicated do not
fall into a single Etoo<; but a single yEVO<;. But the
universalist is naturally happy to say that the various
species of animal fall into a single EIf>O<; just as much
as single men do, since he wishes to assimilate the re-
lation of a genus to its species to the relation of a
species to the individuals belonging to it.
Finally, on this section, one point of translation:
Ross has to translate 'Eo'n' both in 1. 19 and in 1. 20
existentially: "It makes no difference if not all the
elements in the substance have a definition." He thinks
that Aristotle has in mind the difficulty that if the
elements of a substance revealed in a definition are
themselves substances, we shall be forced to allow in-
definable substances, when the process of analysis is
carried to its limit, if an infinite regress is to be avoided.
On the view I am taking, 'Eon' is to be construed in
1. 20 as a copula: "It makes no difference if it (sc. the
Myo<; of I:;wov) is not a Myo<; of everything in the
substance (sc. man)." Cherniss6 takes 'Eon' in 1. 19 and
in 1. 20 as a copula; on my interpretation, it has to be
taken existentially in 1. 19, and as a copula in 1. 20.
The next argument, beginning with En in 1. 23 also
seems to make more sense if it is treated as an argu-
ment advanced by the Platonist. The point is that if
the elements of a substance which are revealed by a
definition are prior to the whole in which they occur,
then, if they are not substances, they must be (e.g.)
qualities; so quality will be prior to substance. Regarded
as an argument in the mouth of the Platonist the
60p. cit.
234 M. J. Woods
argument makes good sense. No doubt the way in
which the dilemma is presented is a little simple-
minded; it does not follow from the fact that genera
are not, properly speaking, oOOlOl that they belong in
some other category. Aristotle's own solution, which
is admittedly not easy to understand, is to invoke
the notion of potentiality. He says later in the
chapter that aouvaTOV yap OUO[OV £f, OUOlWV EtVaL
£vuTI:apxouowv £VTAEXEla (1039a4). * Still later he in-
vokes the matter-form model to explain the relation
of genus and differentia (1045a33 f.). However, the
objection that if the elements of substance are prior
to the whole, they must be substances is exactly the
sort of objection that might occur to someone who
had read Z 10, where Aristotle is quite happy about
saying that some of the parts of a A6yCX; (by which,
presumably, he means the dooe;) are prior to the
whole; and this might readily make someone wonder
how the parts of a substance can fail to be substances
themselves.
To turn briefly to the third argument in this section,
at 1038b29-30, he says that T4'> ~(')Kp6:TEl EVUTI:apf,El
ouo[a, WOTE OUOlV EOTal ouo[a. Ross7 glosses this as
follows: "'In Socrates, himself a substance, there will
be present as an element a substance (sc. animal),
which will therefore be a substance of two things (sc.
of the class of animals and also of Socrates)'." I think
this is correct; but it makes sense in the context only
if we regard it as an argument advanced by the Pla-
tonist. It cannot be an objection to regarding something
Ka96AOU AEY0I:lEVOV as a substance that it would have
to be the ouo[a of two things; for in tIlat sense the
species man is the 000(0 of two things, viz. the class
of men and also Socrates. Regarded as an argument
*"A substance cannot be composed of actually existing sub·
stances ...."
70p. cit., p. 211.
used by the Platonist, it makes sense. Just as the Et!)OC;
man is the OOOlCX of the class of men and, derivatively
from that, of Socrates, so the genus animal is the oOOla
of the class of animals and also, derivatively from that,
of Socrates. Considered as member of the class of men,
the oOo[a of Socrates is the species man; considered
as a member of the class of animals, his 000[a is the
genus I:;wov.
\Ve are now in a position to explain the earlier ar-
gument in chapter 13 with which we began. It is not
intended as a refutation of the Platonist's position;
what Aristotle does is to draw out a consequence of
it, which makes his opponent formulate it more clcarly.
Starting with the class of men, it is suggested that,
despite appearances, if the species man can be said
to be the oOo(a of them, so can I:;wov; Aristotle points
out that if it is to be the OOo[a of the class of man,
since animal is predicated of non-men, then other ani-
mals must be shown in some way to be identical with
men. This leads the Platonist to formulate his position
more carefully. Although animal, in relation to the
class of men is not an oOo[a in quite the way the
species man is, it is the T.Tj.E., and thereforp. ooo[CX of
a larger class. There is an EIEloc; tv 4'> Qc; ,[EllOV l)'Jt(:XPXEl.
Aristotle's own refutation of the Platonist begins at
1. 30. He says, in effect, that if we allow that species
are substances, then no element of a species revealed
by a definition is. His reasons for this are, firstly, that no
I:;wov exists xcuplC;, TIcxpa Ta TlVO::. (As Ross remarks,S
when he says that there is no I:;wov TIapa a TlVO::, he has
in mind the point that there is no animal apart from
the particular species of animal, not that there is no
1;wov apart from the individual animals.) Secondly,
nothing predicated in common is a T6BE 'n but only
'[ol6vBE. I have already discussed this passage. If I am
BOp. cit.
236 M. J. Woods
right, he is concerned here to deny that I:;wov is genu-
inely predicated of a plurality of objects.
He ends by saying that if these propositions are not
accepted, aAAa 1:£ nOAAO: oU[l~a[VEl Kat 6 1:p[1:o<;
(iv9pcvno<;.* I end my discussion of this chapter by
considering how he thought that the denial of the
doctrines led to the Third Man. We noticed a little
before that in order to refute the Platonist he makes
use of the fact that no genus exists apart from its
species. This doctrine is not argued for in chapter 13
but in chapter 12. (This, incidentally, explains why
chapter 12 comes where it does in Z; he needed to
establish this in order to disallow the claim of things
Ka90Aou AEyo[lEva to be substances.) The idea that
no genus exists apart from its species raises many prob-
lems of interpretation, but I take it that part of what
he is saying can be stated quite simply; he is saying
that nothing can be an animal without being a partic-
ular species of animal. This may seem platitudinous but
it is certainly un-Platonic; for in our version of the Theo-
ry of Forms, Forms are construed as paradigms or ex-
emplars with only one characteristic; so the form ani-
mal would have to be an animal without being a
particular kind of animal, and would thus be a counter-
example to the doctrine that there is no I:;wov napo:
1:0: 1:lVa in a sense which can be precisely stated. VVe
are by now familiar with the arguments by which
Aristotle sought to derive the Third-Man regress from
the Theory of Forms thus construed. Can we find
anything analogous relevant to the present passage?
I think we can. The Platonist, who maintains that
the genus which is an element in a species like man is
an ouola, is committed to holding that the word I:;wov
is predicated in the same sense both of the species Illan
and the supposed ooola which is an element in the
* "If not, many difficulties follow and especially the 'third
man' ."
Problems in Metapl1ysics Z 237
species man. But the common element which is present
in all species of animal cannot of course itself be re-
garded as belonging to any animal-species. So he has
to reject the doctrine that there is no l;wov TIapa TO:
T[Va. It is not difficult to see that Aristotle could have
developed a regress similar to the traditional Third-
Man regress against this position; the first step would
be to ask how the word l;wov came to be used in the
same sense of the species man and the supposed com-
mon element it shared with other animal species. Some-
thing very like this argument seems to be adumbrated
in 14, 1039a30 f.
I have argued that the arguments of chapter 13
against those who say that things Ka96AOU AEy6[lEva
may be substances are not incompatible with Aristotle's
own doctrine that EH">11 are substances. E'l511 are things
which are not AEy6[lEva KaTO: TIOAAWV. However, it
may be thought that this remains no more than a
verbal maneuver unless some justification is offered for
treating species differently from genera. What justifica-
tion is there for Aristotle's denial that man KaTO:
TIOAAWV MYETm? The answer, I think, is along the
following lines: It is the species-form man which sup-
plies us with a principle for individuation for man:
it is only in virtue of possessing the form man that
bits of matter which constitute men are marked off
from one another. To speak of a plurality of objects
I need some means of marking off each member of the
set from other things; I do this, according to Aristotle,
by recognizing occurrences of a certain form in matter.
Thus I must already regard things as possessing the
form before I can think of objects as a genuine plurali-
ty. In so far as the statement that the form of a species
is predicated universally of its members implies the
contrary of this, it is incorrect. Aristotle refused to say
that aV9pCilTIo<; was Ka96AOU AEy6[lEVOV because that
would suggest that you could distinguish men inde-
238 M. ,. Woods
pendently of their possession of the form-as if you
could first distinguish individual substances and then
notice that the predicate applied to them which sup-
plied a basis for distinguishing them in the first place.
With species in relation to genera, on the other hand, it
is the other way round. The genus does not itself supply
a basis for distinguishing species; the species are dis-
tinguished by appropriate differentia; in Z a species is
virtually equated with its differentia (e.g. at 1038aI9),
so that species are in a certain sense self-individuating;
hence the genus to which they belong is predicated
universally of them.
IV. ETHICS
THE MEANING OF 'ArAeON
IN THE ETHICS OF ARISTOTLE
H. A. PRICHARD
I have for some time found it increasingly difficult

to resist a conclusion so heretical that the mere ac-
ceptance of it may seem a proof of lunacy. Yet the
failure of a recent attempt to resist it has led me to
want to confess the heresy. And at any rate a statement
of my reasons may provoke a refutation.
The heresy, in brief, is that Aristotle (in the Nico-
machean Ethics, except in the two discussions of plea-
sure-where ayo:66v is opposed to q>o:Gt-.ov and
fl oX6'lp6v) really meant by ayo:66v conducive to our
happiness, and maintained that when a man does an
action deliberately, as distinct from impulsively, he
does it simply in order to, i.e. from the desire to, be-
come happy, this being so even when he does what is
virtuous or speculates. Of this heresy a corollary is
the view that Aristotle, being anxious to persuade men
first and foremost to practise speculation and second-
arily to do what is virtuous, conceived that, to succeed,
what he had to prove was that this was the action
necessary to make a man happy. This corollary, how-
From Moral Obligation (Oxford, 1949), pp. 40-53. Reprinted

by permission of the Clarendon Press, Oxford. Originally pub·
Iished in Philosophy, Vol. X, No. 37 (January 1935). For the
convenience of the reader, the editor of this volume has added
some translations in footnotes, using the Ross version current at
the time the essay was written.
242 H. A. Prichard
ever, which may seem only a further heresy, I propose
to ignore. The heresy, in my opinion, is equally at-
tributable to Plato, and for much the same reasons.
But for simplicity's sake I propose to confine con-
sideration to Aristotle, with, however, the suggestion
that the same argument can be applied to Plato.
In attributing this view to Aristotle I do not mean
to imply that he does not repeatedly make statements
inconsistent with it. Nor do I mean to imply that the
question of the consistency of these statements with
the view simply escapes him; it seems to me that it
does not, but that owing to a mistake he thought
they were consistent with it. Nor do I mean to imply
that his acceptance of this view appears on the surface;
but rather that it becomes evident once we lay bare
certain misleading elements in his account of the mo-
tive of deliberate action.
The first two chapters of the Ethics, and especially
its opening sentence, are undoubtedly puzzling. Aris-
totle begins by saying: TCaoa 'tEXVT] Kat TCaoa ~E905Q(;,
6~otu)C; 5E TCpat;tc; 't£ Kat TCpoa[p£oL<;, aya900 'tlVOC;
Eq>[w9al 50K£l 510 KaAWC; aTC£q>~vaV't:o 'taya96v 05
mxV'£' Eq>lE'taL. 'Every art and every inquiry, and similar-
ly every action and purpose, is thought to aim at some
good; and for this reason the good has rightly been
declared to be that at which all things aim'. Then after
pointing out that certain aims or ends are subordinate
to others, he contends that there must be one final end
to which all others are subordinate, and that this will be
Taya96v, the good, and that, consequently, knowledge
of this final end will have great influence on our lives,
since if we have it, we shall have a definite mark or goal
to aim at. And he goes on to say that, this being so, his
object in the Ethics is to discover what this final end is.
Here, as the rest of the first book shows, Aristotle,
in his first sentence, is not simply stating a common
opinion, but stating it with approval and on the as-
TIle Meaning of AgatIlOn 243
sumption that it is an opinion which his hearers will
accept and so which can be used as a basis for his sub-
sequent argument. And, so regarded, it is very sweeping.
Even if he had said that in every deliberate action
we have an aim or are aiming at something, we should
have regarded the statement, put forward as expressing
a fact obvious to everyone, and so as needing neither
elucidation nor discussion, as sufficiently sweeping. But
what he does say is more sweeping. In effect, taking
for granted that there is always something at which we
are aiming, he commits himself to a general statement
about its nature, stating that it is always aya96v 1:l,
or, as we may translate the phrase, a good.
But besides being sweeping it is obscure. Even if
Aristotle had said that in all action we are aiming at
something, we should have felt that the statement
needed elucidation. But saying as he does that we are
aiming at something good, we have an additional puz-
zle. If, instead, he had said that we are always aiming
at a pleasure, or at an honour, or at doing some good
action, then we should have at least suspected we knew
what he meant, whether or not we agreed. But the
meaning of aya96v is not clear.
Consequently to discover his meaning we have to
find out not only what he means when he speaks of
us in a deliberate action as aiming at something
(~<p(Eo9a( 1:lVOC;) or as having a 1:EAOC; or end, but also
what he means by aya96v. And of these tasks, plainly
the former has to be accomplished first.
The idea, which of course underlies the Ethics, that
in all deliberate action we have an end or aim, is one
the truth of which we are all likely to maintain when
we first consider action, 'action' being a term which,
for shortness' sake, I propose to use for deliberate ac-
tion. The idea goes back to Plato; and Mill expresses
it when he says that all action is for the sake of an
end. We take for granted tllat in doing some action
244 H. A. Prichard
there must be some desire leading or moving us to do
the action, i.e. forming what we call our motive, since,
as we should say, otherwise we should not be doing
the action; or, for this is only to express the same idea
in other words, we take for granted that in doing the
action we have a purpose, i.e. something the desire of
which moves us to do the action. And, taking this for
granted, we are apt to maintain that our purpose in
doing the action always consists in something, other
than the action, which we think the action likely to
cause, directly or indirectly, such as an improvement
in our health which we expect from taking a dose of
medicine.
Further, taking this view of the motive of action,
we are apt to express it metaphorically by saying that
in any action we have an aim or that there is something
at which we are aiming. For when we consider, e.g.,
taking a drug from the desire to become healthy, we
are apt to think of the thing desired, viz. our health,
as that by reference to which we have devised the
action as what is likely to cause it, and so as similar
to the target by reference to the position of which a
shooter arranges his weapon before shooting. We are
also apt to speak of our purpose metaphorically as our
end, as being something which we think will come
into existence at the end of the action. In either case,
however, it is to be noticed that the terms 'end' and
'aim' are merely metaphorical expressions for our pur-
pose, i.e. for that the desire of which is moving us to
act. No doubt further consideration may afterwards
lead us to abandon this view. For certain actions and
notably acts of gratitude or revenge seem prompted by
the desire to do the action we at least hope we are
doing, such as the desire to inflict an injury on another
equal to that to which he has done us. Yet we may
not reflect sufficiently to notice this, or even if we do
we may fail to notice that such actions require us to
The Meaning of Agathon 245
modify the view, or may even think, as Aristotle did,
that the doctrine may be made to apply to them.
Plato, it may be noticed, expressly formulates this
view in the Gorgias. In trying to show that orators
and tyrants have the least power in States, he lays
down generallyl that a man in doing what he does
wishes not for the action but for that for the sake of
which he does it, this being implied to be some result
of the action. And in support he urges that a man
who takes a drug wishes not for taking the drug but
for health, and that a man who takes a voyage wishes
not for the sailing and the incurring of dangers but
for the wealth for the sake of which he takes the voyage.
He is, however, here obviously going too far in asserting
that the man does not want to do the action itself,
for if the man did not want to do the action, he would
not be doing it. What Plato should have said and
what would express the view accurately is this:
A man undoubtedly wants to do what he does, and
this desire is moving him. But the desire is always
derivative or dependent. His having it depends on
his having another desire, viz. the desire of something
to which he thinks the action will lead, and that is
why this latter desire should be represented as what
is moving him, since it is in consequence of having
this latter desire that he has the desire to do the
action.
The view, therefore, implies the idea that the desire
to do some action is always a dependent desire, depend-
ing on the desire of something to which we think the
action will lead. But, as we soon notice, this latter
desire must either be itself an independent desire, i.e.
a desire which does not depend on any other, or else
imply such a desire, since otherwise, as Aristotle put
it, desire would be empty and vain. We are therefore
lGOTgiaS, p. 467.
246 H. A. Prichard
led to draw a distinction between an independent desire
and a desire depending on a desire of something which
we think the thing desired will cause. Aristotle, of
course, recognized and even emphasized the distinction,
but unfortunately he formulated it with a certain inac-
curacy. He implies that it should be expressed as that
between TO ~OUAEaea[ Tl fn' alJTO, or Kae' aUTO, and
1:0 ~OUAEOea[ 1:l fll' ETEpOV. But the latter phrase must
be short for 1:0 ~OUAEOea[ 1:l flu): 1:0 ~OUAEOeaL
ETEpOV 1:l, and, this being so, the former phrase must
be short for 1:0 ~OUAEOea[ Tl fllCx TO l3ouAEOeal aUTO,
which, meaning wishing for something in consequence
of wishing for itself, is not sense. The distinction
should have been expressed as that between TO
~OUAEOea[ Tl !l~ fllCx TO ~OUAEOeal ETEPOV 1:l and TO
~ouAEOea[ 1:l EllCx 1:0 130UAEOeaL ihEpOV 1:l or, to be
more accurate, between desiring something not in con-
sequence of desiring something else, and desiring some-
thing in consequence of desiring something else to
which we think it will lead. And in this connexion it
should be noticed that the English phrase for an in-
dependent desire, viz. the desire of something for its
own sake, which is the equivalent of Aristotle's
l3ouAEOea[ Tl fll' aUTO, has really only the negative
meaning of a desire which is not dependent on any
other desire.
Further, having reached this distinction, we are soon
led, as of course Aristotle was, to hold that in every
action we must have some ultimate or final aim, con-
sisting of the object of some independent desire, and
to distinguish from this aims which we have but which
are not ultimate.
Having drawn this distinction we do not ask: 'Of
what sort or sorts are our non-ultimate aims?' since
obviously anything may be such an aim. But we do
raise the question: 'Of what sort or sorts is our ulti-
mate aim in various actions?'
The Meaning of AgatllOn 247
To this question Aristotle's answer is aya90v 'tl,
since his opening statement covers ultimate as well as
non-ultimate aims. And the most obvious way to as-
certain what Aristotle considers our ultimate aim is,
of course, simply to find out what he means by aya96v.
But, as should now be obvious, there is also another
way. Like ourselves, he must really mean by our ulti-
mate or final aim that the independent desire of which,
or, as he would put it, that the desire of which KaS'
au'to, is moving us to act. Consequently, if he says of
certain things that we desire and pursue, i.e. aim at,
them Ka9' a6't6:, we are entitled to conclude that he
considers that in certain instances they are our ultimate
aim. Now in chapter 6 of Book I he maintains that
there are certain kinds of things, viz. 'tlllT), q>poVTJou;,
and ~50vT), which are 5lCUK61lEVa Kal ayallwllEva
Ka9' au't6:;" and to these he adds in chapter 7, § 5,
voU<; and llaoa apE'tT),t of which, together with 'tt1lT)
and ~50vT), he says that though we choose them for
the sake of happiness, we also choose them 5l' au't6:,
i.e. as being what they severally are, since we should
choose them even if nothing resulted from them. And
to say this is only to say in other words that in some
instances our ultimate end is an honour, in others it
is a pleasure, in others our being q>p6vlllo<;, and so on.
Consequently, if we hold him to this, the only pos-
sible conclusion for us to draw is that he considers
(1) that in ,such cases our ultimate end is not aya96v
'tt, whatever he means by aya96v, and also (2) that
our ultimate end is not always of the same sort, so that
no single term could describe it. We thus reach the
astonishing conclusion that Aristotle, in insisting as
he does that we pursue these things for their own sake,
is really ruling out the possibility of maintaining that
""Honour, wisdom, and pleasure, which are pursued and loved
for themselves"-Ed.
t "Reason and every virtue"-Ed.
248 H. A. Prichard
our end is always ayaeov 'n, or indeed anything else,
so that we are in a position to maintain that he has
no right to assert that our ultimate end is always an
ayaeov, even before we have attempted to elucidate
what he means by ayaeov.
Further, if we ncxt endeavour, as we obviously
should, to do this, we get another surprise. Aristotle's
nearest approach to an elucidation is to be found in
chapter 6, §§ 7-11, and chapter 7, §§ 1-5. There he
speaks of TeX Ka9' aLn:o: OLCilKollEva Kat ayaTIwllEva
as called ayaeo: in one sense, and gives as illustrations
'nll~, cppovTJaL<;, and ~oov~; and he speaks of TO:
TIOLTJ'nKO: TOUTCilV fj tpUAaK'nKa TICil<;* as called aya8a
in another sense, and he implies that these latter are
OLCilKTO: Kat atpETO: OL' ETEpOV2 and that 11AO[31:0<; is
an illustration. 3 Further, he appears to consider that
the difference of meaning is elucidated by referring to
the former as ayaeo: Ka8' aUTO: and to the latter as
ayaeo: oux wOw, i.e. aya90: OLO: ayaeo: Ka8' a6TO:.
But this unfortunately is no elucidation, since to state
a difference of reason for calling two things ciyaeov is
not to state a difference of meaning of ayaeov, and
indeed is to imply that the meaning in both cases is
the same. Nevertheless, these statements seem intended
as an elucidation of the meaning of ayaeov. And the
cause for surprise lies in this, that if they are taken
seriously as an elucidation, the conclusion can only
be that ayaeov includes 'being desired' in its meaning,
and indeed simply means TEAO<; or end. For if they
are so understood, Aristotle must be intending to say
(1) that when we say of something that it is aya80v
Kae' aUTO what we mean is that it is OLCilKollEVOV Kat
ayaTIWf-lEVOV Kae' aUTO, i.e. simply that it is an ulti-
*"Those which tend to produce or to preserve them somehow."

2Ethics, I 7, 4. ["Pursued and desirable for the sake of some-
thing else."-Ed.]
3Ibid. I 5, 8. ["Wealth"-Ed.]
The Meaning at Agathon 249
mate end, and (2) that when we say of something that
it is o:ya80v Ot' E1:EpOV, what we mean is that it is
otWK6[.lEVOV KaL O:yaTIw[.lEvoV Ot' E1:EpOV, i.e. simply
that it is a non-ultimate end. In other words, if here
he is interpreted strictly, he is explaining that eXya86v
means 1:£Aoc;" and by the distinction between an
o:ya80v Ka8' aL)"[6 and an o:ya80v Ot' E1:EpOV he
means merely the distinction between an ultimate and
a non-ultimate end. Yet if anything is certain, it is
that when Aristotle says of something, e.g. TIAo[l'toc;"
that it is an eXya86v he does not mean that it is a
1:£Aoc;" i_e. that it is something at which someone is
aiming, and that when he says of something, e.g. 1:t[.l~
or cpp6vTJOtc;" that it is an o:ya80v Ka8' al)"[6, he does
not mean that it is someone's ultimate end, i.e. what
he speaks of in VI 9, 7 as 1:0 1:£11.0<; 1:0 cXTIAwc;,. Apart
from other considerations, if he did, then for him to
say, as he in effect does, that we always aim at o:ya86v
"[t would be to say nothing, and for him to speak, as
he does, of the object of ~ouATJOtc;, as 1:o:ya86v would

be absurd.
But this being so, what does Aristotle mean by
o:ya86v? Here there is at least one statement which
can be made with certainty. Aristotle unquestionably
would have said that where we are pursuing something
of a certain kind, say, an honour, Ka8' aU1:o, we are
pursuing it ci>c;, eXya80v, i.e. as a good. Otherwise there
would not even have been verbal consistency between
his statements, that we pursue, i.e. aim at, things of
certain stated kinds, and that we always aim at o:ya86v
"[t. Again, unless we allow that he would have said
this, we cannot make head or tail either of his puzzling

statement in Book I, chapter 2 that since, as there must
be, there is some end which we desire for its own
sake, this end must be 1:o:ya80v, or, again, of its sequel
in chapter 7, where he proceeds to consider what that
is to which the term 1:o:ya80v is applicable by con-
250 H. A. Prichard
sidering which of our various ends is a final end. For
we are entitled to ask: 'Why does Aristotle think that
if we discover something to be desired and pursued
for its own sake, we shall be entitled to say that it is
-raycx96v?' And no answer is possible unless we allow
that he thought that in desiring and pursuing some-
thing for its own sake we are desiring and pursuing it
ci>C; d:ycx96v.
But Aristotle in saying, as he would have said, that
in pursuing, e.g., an honour, we are pursuing it
ci>C; d:ycx96v could only have meant that we are pur-
suing it in virtue of thinking that it would possess a
certain character to which he refers by the term
d:ycx96v, so that by d:ycx96v he must mean to indi-
cate some character which certain things would have.
Further, this being so, in implying as he does that in
pursuing things of certain different kinds Kcx9' CXUl:CX
we are pursuing them ci>C; aycx9a, he must be implying
that these things of different kinds have, nevertheless,
a common character, viz. that indicated by the term
ayo:96v. It will, of course, be objected that he expressly
denies that they have a common character. For he says:
-rl!l~C; 5E. Ko:t <J>POV~(JEQC; Kcxt ~f>o~C; ~-rEpOl KCXt EHO:<J>E-
pOV-rEC; Ol A6yOl -rcx6-rTI n d:ycx9a.4 But the answer is
simple; viz. that this is merely an inconsistency into
which he is driven by his inability to find in these
things the common character which his theory requires
him to find, and that if he is to succeed in maintain-
ing that we pursue these things of various kinds
ci>C; d:ycx9a, he has to maintain that in spite of ap-
pearances to the contrary they have a common char-
acter.
Nevertheless, though we have to insist that Aristotle
in fact holds that in pursuing any of these things
4Ethics I 6, 11. "But of honour, wisdom, and pleasure, just in
respect of their goodness, the accounts are distinct and diverse."
-Ed.]
Ka8' alJTO, i.e., as we should say, for its own sake, we
are pursuing it we; ayaeOV, we cannot escape the ad-
mission that in doing so he is being inconsistent. For
to maintain that in pursuing, e.g., an honour, we are
pursuing it Ka8' aLn:o, or, as we should say, for its
own sake, is reallv to maintain that the desire of an
honour moving us' is an independent desire, i.e. a desire
depcnding on no other. And, on the other hand, to
maintain that in pursuing an honour, we are pursuing it
we; ayaeov, or as a good, is really to maintain that the
desire of an honour moving us is a dependent desire, viz.
a desire depending on the desue of something which
will possess the character indicated by the word d:yaeov,
i.e. that we desire an honour only in consequence of
desiring something which will possess that character
and of thinking that an honour will possess it. It is,
in fact, really to maintain that in pursuing an honour,
our ultimate aim, i.c. that the independent desire of
which is moving us, or what Aristotle would call that
which we are pursuing Kae' aLn:o, is not an honour
but a good, i.e. something having the character, what-
ever it may be, which is indicated by the word ayaeov,
i.e. that we desire an honour only in consequence of
desiring a good. The principle involved will become
clearer, if we take a different illustration. In chapter 6
Aristotle speaks of DpaV as one of the things which
are pursued for their own sake; and if he had said that
we pursue DpaV we; aloe6:vwew he would in consist-
ency ha\'e had to maintain that what we are pursuing
Kae' auto is not DpaV but aloe6:vwew, * and that the
desire of DpaV moving us is only a dependent desire
depending on our dcsiring something else which we
think DpaV will be.
lt will be objected that there is really no inconsisten-
cy, since Aristotle conceives the charactelistic referred
""Sight ... perception."

252 H. A. Pricllard
to by &ya8ov as a characteristic of an honour and of
anything else which he would say we pursue Ka8' aLY[O,
and that to speak of us as desiring something in respect
of some character which it would have is not to rep-
resent our desire of it as dependent. In illustration
it may be urged that to speak of us as, in desiring to
do a courageous action, desiring it as a worthy or vir-
tuous action is not to represent our desire to do a
courageous action as dependent. But the objection can-
not be sustained. For if we desire to do a courageous
action, as something which would be a virtuous action,
i.e., really, a something which we think would be' a
virtuous action, although our desire does not depend
on a desire of something which we think a courageous
action would cause, it does depend on the desire of
something which we think it would be. And as a proof
of this dependence we can point to the fact that if,
while having this desire, we were to do a good action
of another sort, e.g. a generous action, the desire would
disappear.
What is in the end plain is that Aristotle cannot
succeed in maintaining that our ultimate end is always
aya90v "[l without abandoning his view that we pur-
sue such things as "[l~~ and O:pE"[~ Ka9' aL)"[(x, or, as
we should say, for their own sake, and maintaining
instead that we pursue them as things which we think
will have the character to which the term &ya96v
refers. Nevertheless, in spite of having to allow that
we are thereby attributing inconsistency to Aristotle,
we have to admit that he, in fact, holds that in desiring
and pursuing certain things for their own sakes we are
desiring and pursuing them in respect of their having
a certain character, viz. whatever it be to which he
refers by the term &ya90v.
So far the only clue reached to the meaning of
aya80v is the idea that Aristotle used it to refer to
a certain character possessed by certain things, the
thought of the possession of which arouses desire for
them, and indeed is the only thing which arouses de-
sire for anything, except where our desire depends on
another desire.
We have now to try to get to closer quarters with
the question of its meaning. The question is really:
'What is the character which Aristotle considered we
must think would be possessed by something if we
are to desire it, independently of desiring something
else to which we think it will lead, that character being
what Aristotle used the word d::ya96v to refer to?'
Here it seems hardly necessary to point out that the
answer cannot be 'goodness'. To rule out this answer
it is only necessary to point out two things. First, if
Aristotle had meant by d::ya96v good, he would have
had to represent us as desiring for its own sake any
good activity, whether ours or another's, whereas he
always implies that a good activity which we desire is
an activity of our own, and in addition he would have
had to drop, as he never does, the idea of a connexion
between a good activity and our own happiness. And
second, Aristotle's term d::ya96v is always d::yo::90v 'nv[,
as appears most obviously in the phrase d::v9pomvov
d::ya96v and in the statement in IX 8, 8-9, where he says
that reason always chooses what is best for itself-TIae;
yap voGe; o::tpEltaL TO ~EATlOTOV EauTt;)-and goes on to
add that the man who gives wealth to a friend assigns
the greater good, the having done what is noble
(TO KO:AOV), to himself. Once, however, we regard this
answer as having to be excluded once for all, there
seems to be no alternative to attributing to Aristotle
a familiar turn of thought to which we are all very
prone and which is exemplified in Mill and T. H.
Green.
When we consider what we desire we soon come to
the conclusion, as of course Aristotle did, that there
are things of certain kinds which we desire, not in
254 H. A. Prichard
consequence of thinking that they will have an effect
which we desire, but for themselves, such as seeing a
beautiful landscape, being in a position of Rower, help-
ing another, and doing a good action. We then are
apt to ask, 'What is the condition of our desiring such
things?' and if we do, we are apt to answer-and the
tendency is almost irresistible-'It is impossible for us
to desire any such thing unless we think of it as some-
thing which we should like, since, if we do not think
of it thus, we remain simply indifferent to its realiza-
tion.' Then, if asked what we mean by its being some-
thing we should like, we reply: 'Something which
would give us enjoyment, or, alternatively, gratification,
or, to use a term which will cover either, pleasure.'
The tendency is one to which Mill gives expression
when he says that desiring a thing -and finding it pleas-
ant are two parts of the same phenomenon; and Green
exhibits it when he maintains, as in effect he does,
that we can desire something only if we think of it
as something which will give us satisfaction, i.e. gratifi-
cation. In maintaining this we are really maintaining
that the thing which we at first thought we desired
for its own sake, such as seeing a beautiful landscape,
or doing a good action, is really only being desired for
the sake of a feeling of enjoyment or gratification, or,
to put it generally, pleasure, which we think it will
cause in us. And correspondingly, where we think of
the desire as moving us to act, we are really maintain-
ing that what we at first thought our ultimate end is
really only our penultimate end or the proximate
means, and that our ultimate end is really a pleasure
which we think this will cause. We are, however, apt
to think of a thing's giving us enjoyment, or alterna-
tively gratification, as if it were a quality of the thing,
just as we think of the loudness of a noise as a quality
of the noise. And our tendency to do this is strength-
ened by the fact that the ordinary way of stating the
fact that something X excites a feeling of pleasure, or
of gratification, is to say that X is pleasant or gratifying,
a way of speaking which suggests that what is in fact
a property possessed by X of causing a certain feeling
is a quality of X. The tendency is mistaken, since, as
anyone must allow in the end, something's giving us
enjoyment is not a quality of it, and when we say that
something is pleasant, we are not attributing to it a
certain quality but stating that it has a certain effect.
Nevertheless, the tendency exists. And when it is oper-
ative in us, we state our original contention by saying
that in desiring to see a beautiful landscape for its own
sake, we desire it as something which will be pleasant,
and that when we are acting on the desire, our ulti-
mate end is the seeing a beautiful landscape as some-
thing which will be pleasant, thereby representing
what on our view is really the proximate means to our
end as our end.
This being the line of thought to which I referred,
it remains for me to try to show that it was taken by
Aristotle. Before we consider details we can find two
general considerations which are in favour of thinking
that he took it. In the first place, if we assume it to
be indisputable that he thought that there are things
of a certain sort which we desire for their own sake,
but that in desiring them we desire them in respect of
having a certain character to which he refers by the
term aya86v, and then ask 'What can be the character
of which he is thinking?' the only possible answer
seems to be: 'That of exciting either enjoyment or
gratificati.on.' And in particular two things are in fa-
vour of this answer. First, it is easy, from lack of con-
sideration, to think of exciting pleasure as a quality
of the thing desired-as indeed Aristotle appears to do
when he speaks of virtuous actions (at KaT' O:pE1~V
'ITpaE,ElC;) as q>60El ~5Ea and as ~5ELal Ka8' alJ'ra:c;,5
5Ethics I 8, 11.
256 H. A. Prichard
i.e. as pleasant in virtue of their own nature; and sec-
ond, the perplexity in which he finds himself in chap-
ter 6 whcn trying to elucidate the meaning of aya80v
would be accounted for if what he was referring to was
something which is not in fact a character common
to the various things said to be aya8eX, although he
tended to think of it as if it was. In the second place
he applies the term aya80v not only to the things
which we desire for themselves, but also to the things
which produce or preserve them, and it is difficult to
see how he can apply the term to the latter unless
aya80v means productive of pleasure, whether directly
or not. In fact, only given this meaning is it possible
to understand how Aristotle can speak not merely of
nll~ but also of TIAOU'rOC; as an ay<x8ov.
To pass, however, to special considerations, we seem
to find evidence, and decisive evidence, in a quarter
in which we at first should least expect it. At the be-
ginning of chapter 4 he directs his hearers' attention
to the question: 'r[ (Eon) 'ro TIeXV'rC0V aKpo'r<X'"[oV
'rwv TIpaK'rwv aya8wv, i.e. 'What is it that is the
greatest of all achievable goods?' and he proceeds to
say that while there is general agreement about the
name for it, since both the many and the educatcd
say that it is happiness, yet they differ about what
happiness is, the many considering it something the
nature of which is clear and obvious, such as pleasure,
wealth, or honour, whereas, he implies, the educated
consider it something else of which the nature is not
obvious. Then in the next chapter he proceeds to state
what, to judge from the three most prominent types
of life, that of enjoyment, the political life, and that
of contemplation, various men consider that the good
or happiness is, viz. enjoyment, honour, and contem-
plation. And later he gives his own view, contending,
with the help of an argument based on the idea that
man has a function, that happiness is tjJuX~<; ~VEPYEl(X
1:l<; KCX1:' O:pE1:~V 1:EAdcxv.6
Here it has to be admitted that Aristotle is express-
ing himself in a misleading way. His question '\Vhat
is the greatest of goods?' can be treated as if it had
been the question 'What is a man's ultimate end?' i.e.
1:0 1:EAO<; 1:0 CX7tM'><;. For as I 2, 1 and I 7 show, he
considers that to find what is the greatest good, or the
good, we must find a man's final end, i.e. that which
he desires and aims at for its own sake, and in I 5 he
judges what men consider the good from what their
lives show to be their ultimate aim. And his answer
to this question, if taken as it stands, is undeniably
absurd. For, so understood, it is to the effect that,
though all men, when asked 'What is the ultimate
end?', answer by using the same word, viz. EUOcnflov[cx,
yet, as they differ about what Eul'>CXlflovla: is, i.e., really,
about the thing for which they are using the word
Eul'>CXlflov[cx to stand, some using it to designate pleas-
ure, others wealth, and so on, they are in substance
giving different answers, some meaning by the word
Eul'>CXlflov[cx pleasure, others wealth, and so on. But of
course this is not what Aristotle meant. He certainly
did not think that anyone ever meant by Eul'>CXlflov[cx
either 1:lfl~ or 'TtAOlho<;; and he certainly did not him-
self mean by it tjJuX~<; ~VEPYEl<X n<; KCX1:' O:pE1:~V
1:EAEkxv. What he undoubtedly meant and thought
others meant by the word Eul'>CXlflov[cx is happiness.
Plainly, too, what he thought men differed about was
not the nature of happiness but the conditions of its
realization, and when he says that Eul'>CXlflov(cx is
tjJuX~<; EVEpYEl<X n<; KCX1:' O:pE1:~V 1:EAE[CXV, what he re-
ally means is that the latter is what is required for the
realization of happiness. Consideration of the Ethics
by itself should be enough to convince us of this, but
6Ethics I 13, 1 ["an activity of the soul in accordance with
perfect virtue."-Ed.)
258 H. A. Prichard
if it is not, we need only take into account his elucida-
tion of the meaning of the question ',[ Eonv;' to be
sure that when he asks ',[ Eon ~ EUOalf!0v[a';
his meaning is similar to that of the man who, when
he asks 'What is colour?' or '\Vhat is sound?' really
means 'What are the conditions necessary for its reali-
zation?' We must therefore understand Aristotle in
chapter 4 to be in effect contending that while it is
universally admitted that our ultimate aim is happi-
ness, there is great divergence of view about the con-
ditions, or, more precisely, the proximate conditions,
of its realization.
But, this conclusion reached, we can plainly take
one step farther and conclude that Aristotle himself
is in agreement with the view that our ultimate end
is happiness, and that, taking its truth for granted, his
Ethics is concerned first to prove that it is by virtuous
action that it will be realized, and then to work out
in detail the character of virtuous action, so that we
shall be better able to obtain our aim. In other words,
we can conclude that his real answer to the question,
'\Vhat is ,0 TEAO<; TO O'ITAw<;, i.e. our ultimate aim?'
is not, as we may at first think, tjJuX~<; EVEPYEL<X n<;
Ka,' d:PEulv ,EAdav but EuoaqlOv[a, i.e. happiness.
Putting this otherwise, we can say that the accurate
statement of his own view is to be found in I 12,
where he gives as a reason why Euoaqlov[a is ,If!LOV,
whereas d:pE'~ is merely E'ITalVE,OV, that it is for the
sake of Euoaqlov[a that we all do everything.7
Now, if by thus going behind Aristotle's terminology
we are driven to conclude that Aristotle really con-
sidered our ultimate end to be always our happiness,
or alternatively some particular state of happiness on
our part-for sometimes he seems to imply the one
view and sometimes the other-we are also driven to
7haUTTJ<; (Le. EOSaI110via<;) yap xaplv TO: Aama mwra TraVTE
TrpaTTOI1EV, Ethics, I 12, 8.]
conclude that, though he at times makes statements
to the contrary, he also holds that where we are said
to have as our ultimate end nfl~ or EVEpYElCx nc; KCX't'
d:pE't~V or anything else of a kind which we consider
a condition of happiness, the thing in question is really
according to him only our penultimate end, and the
desire of it is only a derivative desire depending on our
desire of happiness. And then it becomes obvious that
when he implies, as he always does, that in desiring
one of these things we desire it as an dya:96v, what he
means by dycx96v is 'producti~e of a state, or rather a
feeling, of happiness', i.e., as I think we may say in this
context, a feeling of pleasure. Further, this being so,
we have to allow that he fundamentally misrepresents
his own problem. Assuming that we all always have
either a single ultimate aim, or at least, alternatively,
an aim of one sort, what he ostensibly maintains is
that we are uncertain about its nature, and that there-
fore he has to discover its nature in order to help us
to achieve it. But, as we must now conclude, what he
is really maintaining is that though the nature of our
ultimate aim, happiness, is known to us, for we all
know the nature of that for which the word 'happi-
ness' stands, we are doubtful about the proximate
means to it, and that consequently he has to discover
the proximate means. In other words, in maintaining
l\'ux~c; EVEpYElCx 'tlC; KCX't' d:pE't~V 'tEAElCXV to be our ul-
timate end of the nature of which we are uncertain,
he is putting what on his view is really the proximate
means to our end in the place of what on his view is
really our end. And if we ask 'How can he have come
to misrepresent his own view so fundamentally?', then,
if the contentions already advanced are true, we have
at hand a satisfactory answer. We can reply that the
misrepresentation is due to his making two mistakes
to which we are all prone: first, that of thinking of the
property of causing happiness as a quality of what
260 H. A. Prichard
causes it, and secondly, that of thinking that where
we are aiming at something of a certain kind for its
own sake, and so having it as our ultimate end, we are
nevertheless aiming at it in respect of its having a
certain character.
By way of conclusion it may be well to refer to an
objection which will inevitably be raised, viz. that I
have been, in effect, representing Aristotle as a psycho-
logical hedonist, and that to do this is absurd. I admit
the charge, but do not consider the representation
absurd. It seems not only possible, but common, to
hold that there are a number of things other than
pleasure which we desire for their own sake, and then
when the question is raised, 'How is it that we desire
these things?', to reply: 'Only because we think they
will give us pleasure.' In my opinion, the reply is mis-
taken, and is made only because we are apt to think of
the gratification necessarily consequent on the thought
that something which we have desired is realized as
that the thought of which excites the desire. But the
mistake is a very insidious one, as, if I am right, is
shown by the fact that Green, in spite of all the trouble
he takes to point out that Mill falls into it, falls into
it himself.
'ArAeON AND EY~AIMONIA IN THE
ETHICS OF ARISTOTLE *
J. L. AUSTIN
This article takes its start from an article by Professor

H. A. Prichard (Philosophy, X [1935], 27-39 [241-
60 above]) on "The Meaning of o:ya66v in the Ethics
of Aristotle." It will be seen that I disagree with
him, but I think his article has the great merit of
raising serious questions.
Statement ot Prot. Prichard's conclusions. Prof.
Prichard begins by stating his "heretical" conclusions,
as follows:
(1) Aristotle really meant by o:ya66v "conducive to
our happiness."
(2) Aristotle maintained that when a man does an
action deliberately, as distinct from impulsively, he
does it simply in order to, i.e. from the desire to,
become happy, this being so even when he does
what is virtuous or speculates. 1
*This paper was composed by the late Professor John L. Austin
before World War II, while Professor Prichard was still alive.
Some changes and additions were made after World War II. Its
publication was made possible by the kind consent of Mrs. J. L.
Austin. The author's footnotes are numbered. For the conven·
ience of the reader some translations have been supplied by the
editor of this volume. The translation employed is that of W. D.
Ross, since his was the version current at the time this paper was
written.
lThis distinction between "speculating" and "doing what is
virtuous" is not strictly Aristotelian: 6Ecupia is EVEP¥EIO Ka1"a
Tilv 1"EAEI01"01"T)v apE1"~v.
262 J. L. Austin
(2.1) A corollary: Aristotle, being anxious to per-
suade men first and foremost to practice speculation
and secondarily to do what is virtuous, l conceived
that, to succeed, what he had to prove was that this
was the action necessary to make a man happy.
(The corollary Prof. Prichard ignores, and, at least for
the present, I shall do the same.)
His reason for excluding certain passages from con-
sideration invalid. We must first direct our attention
to a curious and important reservation, which Prof.
Prichard makes in stating his view. Aristotle, he says,
means by aya86v conducive to our happiness "in the
Nicomachean Ethics [abbreviated NE hereafter-Ed.]
except in the two discussions of pleasure-where
aya86v is opposed to cpauAov and lloX8T)p6v." We are
not here concerned with the restriction to the NE,
but it is necessary to examine the further restriction,
by which we are precluded from using NE VII xi-xiv
and X i-v.
The argument implied in Prof. Prichard's words
seems to be as follows:
(a) In these passages aya86v means something dif-
ferent from what it means in the rest of the NE.
(b) This is shown by the fact that it is, in these
two passages opposed to cpauAov and lloX8T]p6v.
With regard to (a), it is most unfortunate that Prof.
Prichard does not tell us what aya86v does mean in
these two passages.
As to (b), we clearly need further explanation. I
hope that the following is a correct expansion of Prof.
Prichard's argument.
(1) Throughout the NE, with the exception of
these two passages, aya86v is never opposed to
q>auAov or lloX8T)p6v, but to something else, pre-
sumably KaK6v.
(2) In these two passages alone, aya86v is opposed,
not to KaK6v, but to cpauAov and lloX8T]p6v.
Agathon and Eudaimonia 263
(3) Since we know on independent grounds that
KaK6v has a different meaning from cpaGAov or
!-'o)(ElTJp6v, it follows that &ya96v, in these two pas-
sages, must have a meaning different from that
which it has throughout the rest of the NE.
(To take a parallel case. Suppose that I do not know
the meaning of the adjective "green": and that through-
out a certain work I find it opposed to the adjective
"experienced," except in two passages where it is op-
posed to "red" and "yellow." Then if I know on other
grounds that "experienced" means something sufficient-
ly different from "red" or "yellow," I can infer that
"green" must, in these two passages, have a meaning
different from that which it has throughout the rest
of the work.)
If this is actually the sort of argument on which
Prof. Prichard is relying, I think there are considera-
tions which will lead him to abandon it.
(1) &ya:96v is opposed to [loX9TJp6v elsewhere in
the NE-e.g. in IX viii 7-where pleasure is not un-
der discussion. (Not to mention passages in other
works, e.g. Met. 1020b21). I have not found a case
of &ya96v being explicitly opposed to q>aGAov else-
where, but cpo (3) intra.
(2) &ya96v is constantly opposed to KaK6v in the
two discussions of pleasure: VII i 1-2, xiii 1 and 7,
xiv 2 and 9, X ii 5. In VII in particular, the discus-
sion is introduced and terminated by an opposition
between &ya86v and KaK6v, and in xiv 2 we read:
KaKctl yap &ya:86v tvavrlOV.
( 3) I do not know of any clear distinction between
the meaning of KaK6v and the meanings of q>aGAov
and !-'0X811P6v (or TIovTJp6v) any more than I can
clearly distinguish between &ya96v, tTIlElKEC;, and
oTIou5ai:ov. The words seem to be used almost in-
differently, or at least for "species" of one another
which would be equivalent in certain contexts.
264 J. L. Austin
Very many passages in the NE, such as III v 3, are
evidence of this. But the point seems clear even from
the two discussions of pleasure. OTIOU5alOv seems equiv-
alent to aya80v in VII xiv 4, ETIlElKEC; to aya80v
in X ii 1, floX8Tjpov seems equivalent to or a species
of KaKOV in VII xiv 2, cpaUAOV to KaKOV in X i 2.
Pleasures are called aya8al, OTIOU5alat, ETIlElKElC;
(also KaAal, etc.) in apparently the same sense or
senses not sufficiently distinguished: and similarly
OUK aya8al, flOX8T] pal,. cpaUAal (also a[oxpal, etc.)
in apparently the same sense. And these adjectives
seem opposed to one another indifferently, cpo e.g.
X v 6. (It must be admitted that Aristotle does not
use the expression KaKal ~50val: if this requires some
explanation, I think one could easily be found.)
It is, of course possible that some distinction can
be drawn between KaKOV on the one hand and cpauAov
and floX8T]pov on the other. But (a) It is clearly in-
cumbent on Prof. Prichard to draw it-which he does
not do. (b) Even so he would by no means be out of
the wood, for (i) it does not seem to be true that
aya80v is, in these passages on pleasures more com-
monly opposed to cpauAov and floX'8Tjpov than to
KaKOv. floX8T]pov only occurs once in each book, and
is only used as opposite of aya80v in VII xiv 2: and
there it is only so opposed because it is equated with
KaKOv. cpauAov is opposed to aya80v only once in X i
2, a rather popular passage: and there, as section 5 of
the same chapter shows, it is equivalent to KaKOv.
(ii) Actually aya80v is in these same passages much
more commonly opposed to KaKov, and so presumably
has its 'normal' sense. (But we shall see that it is
vital for Prof. Prichard that, when Aristotle says "~50viJ
is an aya8ov," aya80v should never have its 'normal'
sense of "conducive to our happiness.") (iii) In a
most important passage, X ii 1, exactly the same re-
marks are made about aya80v as in IiI, a passage
on which Prof. Prichard relies in arriving at his inter-
pretation of it as "conducive to our happiness." Here
then dycx86v must presumably have that meaning:
but this is one of the places where ~oovi) is said to be an
dycx86v, which, on Prof. Prichard's interpretation of
dycx86v, does not make sense. v.i.
It would seem then that Prof. Prichard's ostensible
argument for excluding these two discussions of pleas-
ure from consideration will not bear examination. And
it is in any case so very recondite that we may be
tempted to think he would never have chanced upon
it, unless he had been searching for some reason to
justify the exclusion of these passages.
Why does he wish to exclude these passages? His
interpretation of EUOCXlflov[cx. Why then, we must ask,
should it be important for Prof. Prichard to secure the
exclusion from consideration of the two discussions of
pleasure? In order to understand this, we must first
understand that the whole argument of Prof. Prichard's
paper is based upon a premise which is never expressed,
no doubt because it seems to him obvious; namely,
that "happiness" (his translation of EUOa:lfloV[cx) means
a state of feeling pleased. This may be shown as fol-
lows:
(1) On p. 39, line 16 [po 259, 1. 37 above] "causing
happiness" is substituted without remark for "causing
pleasure," which was the expression used in the par-
allel argumentation on e.g. pp. 35-36.
(2) On p. 38, [po 259)] at the bottom we read "what
he means by dycx86v is productive of a state, or rather
a feeling, of happiness, i.e. as I think we may say in
this context, a feeling of pleasure." This remark is in-
teresting, since it seems to imply that "in other con-
texts" being happy does not mean feeling pleased.
However, we need not worry, I think, about these
other contexts, for it is quite essential to Prof. Pri-
chard's argument at a crucial point that the word
266 J. L. Austin
EUl:iat~OV(a:, or in English "happiness," should be and
should be known by all to be, at least for purposes of
Ethics, entirely clear and unambiguous in meaning:
so that, if being happy means feeling pleased in some
contexts in the NE, it clearly must do so in all thought
that concerns the moral philosopher.2 (The crucial
point referred to is found on p. 37 [po 257] at the
bottom, and repeated on p. 39 [po 259] at the top.
Prof. Prichard there maintains that Aristotle cannot
be really asking the question he "ostensibly" asks, viz.
"What is the nature of happiness?" for there is no
uncertainty about that-"the nature of ... happiness
is known to us, for we all know the nature of that
for which the word 'happiness' stands.")
(3) On p. 39 [po 260], in summing up his own con-
tentions, Prof. Prichard says that he has in effect rep-
resented Aristotle as a psychological hedonist; i.e., I
understand, he claims that, according to Aristotle, all
deliberate action is done from a desire to produce in
ourselves feelings of pleasure. Now on p. 27 [po 241],
quoted above, Prof. Prichard said that, according to
Aristotle, all deliberate action is done from a desire to
become happy. Hence being happy and feeling pleased
are evidently, for Prof. Prichard, equivalent expressions.
(And this enables us to see why he considers his view,
as he says on p. 27 [po 241], "heretical," although, at
first sight and as here stated, it does not appear very
extraordinary: if we realize that "become happy" means
"feel pleased" the view certainly is very strange.)
It is, therefore, clear that Prof. Prichard does not
distinguish what we call "being happy" from what we
2Prichard's reservation, "in this context" whatever it means, is
not important here; nor is the contrast between state ("disposi·
tion") and feeling. Similarly, on p. 38 [po 258], a distinction is
drawn between "happiness" and "some particular state of happi-
ness": but I do not think that concerns us here.
Agathon and Eudairnonia 267
call "feeling pleased": and he has in fact been good
enough to tell me that that is so.
lt now becomes evident why the two discussions
of pleasure in the NE should have a peculiar interest
for him.
For if cXyaS6v means "conducive to our happiness,"
and if "happiness" is equivalent to "pleasure"-then
how can we ask, as Aristotle does in these passages,
whether ~5ovTj is an aycx86v? For ~5ovTj must pre-
sumably be translated "pleasure": so that the question
we are asking becomes "Is (our) pleasure conducive
to our pleasure?" which is absurd, or at least absurdly
limited. Further similar difficulties arise, if we ask, for
instance, what could be meant by saying that some
~50vcx( are aycx,8al, and, odder still, that some are not
cXycx8cxL
Hence it is essential for Prof. Prichard to maintain
that ayaS6v has in these passages a meaning different
from "conducive to our happiness": but, as we have
seen, his reason for saying ayaS6v has a new meaning
in these two passages is invalid (nor does he explain
what cXycx86v does mean in them).
But Prof. Prichard has another and quite radical
difficulty to face in connection with Aristotle's dis-
cussions of pleasure, which in his paper he appears
not to appreciate. For Aristotle there discusses the
relation of ~5ovTj to Eu5CXl!lOVlCX in such a manner as
to make it quite plain that these two Greek words do
not mean 'the same,s \Vhereas Prof. Prichard's whole
argument depends on translating Eu5CXl!l0V[CX "happi-
ness," and taking "happiness" to be equivalent to
"pleasure" which must (we assume) be the translation
of ~5ovTj : so that ~50v~ and Eu5CXlfloVlCX ought to mean
the same. Hence Prof. Prichard must hold that, in
these discussions, not merely cXycx86v, but also either
3EIi50llJOvio is TO aplO'TOV, 1i6ovft is notTo aplO'Tov: and so
on.
268 J. L. Austin
Euf>CXlflov[a or ~f>0vfJ changes its meaning from the
normal. Otherwise his view is untenable. As to which
alternative he would choose, I do not know: but both
are very difficult. He can scarcely hold that EuooqlOv[a
changes its meaning, for, as we shall see, much of his
argumentation depends on his view that the meaning
of EUoolflov[a was clear and unambiguous. As to
~f>ovfJ, he does hold, as we shall see, that the word
is sometimes used in a special restricted sense to in-
clude only the OQflaTlKai ~f>ova[: but it is obviously
quite impossible to hold that it has only this restricted
sense throughout the two full dress discussions of pleas-
ure. Unfortunately, Prof. Prichard does not notice
this additional difficulty.
Even the exclusion of these passages would not suf-
fice to save Prof. Prichard's view from refutation. It
would still be open to Prof. Prichard to maintain (and
in view of his low opinion of Aristotle's Ethics, I think
it possible he might do so) that, even apart from
other arguments such as that about q>auAov and
floXBT]p6v, these very facts which I have just men-
tioned are themselves sufficient to show that the two
discussions of pleasure are inconsistent with the rest
of the NE and may therefore be neglected. It need
scarcely be pointed out how dangerous this would be:
for we are trying to discover the meanings of ayaB6v
(and Euoolflov[a) and it is scarcely permissible to elim-
inate a large part of the evidence, not otherwise
known to be incompatible with the rest, on the ground
that it will not square with our interpretation of those
meanings. At least it would be necessary to prove that
EUoolflov[a must elsewhere mean "pleasure": but, it
seems to me, Prof. Prichard does not prove this, he
assumes it.
However, it would in any case be of no use to ex-
clude from consideration the "two discussions of pleas-
ure." For pleasure is mentioned in many other parts
of the NE, and precisely the same difficulties for Prof.
Prichard's view are to be found in them also.
Let us confine ourselves to Book I, since it is upon
that book that Prof. Prichard principally relies, in-
cluding chapters v and xii, which he cites.
In I v we are told that ot TIOAAOL KOt CPOP'LlKc0'LO:'Wl
maintain that EUO<Xlflov[o is Tjoov~ and that Aristotle
himself rejects this view. According to Prof. Prichard's
interpretation, it would seem that he ought to accept
it, as tautological. Prof. Prichard did reply, when faced
with this, that Tjoov~ here, being the end of the
aTIOAOUO'LlKOC; ~[oc;, has a special restricted meaning
which includes only the OCUflO'LlKOi Tjoovo[ (cp. VII
xiii 6).4 This is scarcely obvious, and we should have
expected Aristotle's rejection to take a rather different
form, if he himself held that our end is ~oov~, although
not merely the OG.lflO'LlKOl TjoovoL However, we need
not insist on this passage; others are plainer.
In I xii 5, what Eudoxus said about Tjoov~ is com-
pared with what Aristotle himself says about EUO<XlflO-
v[o:: clearly, then, ~oov~ and EUO<Xlflov[o: are distinct.
And, whatever may be true of Sardanapallus, there is
no reason whatever to suppose that Eudoxus meant by
~oov~ merely the OG.lflO'LlKOi ~oovoL
Finally, in I viii the relation of Tjoov~ to EUO<Xlflov[o
is discussed very much as in books VII and X:
EUO<Xlflov[o: is flE8' Tjoov~c; ~ OUK <XVEU ~oov~C;, * but
quite clearly it is distinct from it. (Even if they are
necessarily connected, we must not confuse one with
the other: cpo EE I ii 5.)
We do not, naturally enough, find in Book I, a
discussion as to whether ~oov~ is an ayo:86v. But the
4Just as, when Aristotle says ~6ovtl is aya66v, Prof. Prichard

says aya66v has a meaning different from the ordinary, so, when
Aristotle distinguishes ~6ovtl from Eu6aq.lOvla, Prof. Prichard
says ~6ovtl has a meaning different from the ordinary.
*"Accompanied by pleasure or not without pleasure."-Ed.
270 J. L. Austin
views that it is '1:0 ayo:96v or '1:0 aplO'l:OV are men-
tioned, and, though rejected, not rejected as absurdities
(I v and xii). Moreover, as was pointed out above,
Eudoxus' views are mentioned in I i and xii in pretty
much the same words as in X ii, so that it would seem
that the meaning of ayo:96v ought to be the same in
each case.
We see then, that Prof. Prichard cannot exclude
VII xi-xiv and X i-v from consideration, and that they
are fatal to his view. But even if we do exclude them,
other passages, equally fatal, can be produced even
from Book I, on which he relies. So that, if he still
maintains the view, it would seem that he must be
prepared to attribute to Aristotle even more and graver
inconsistencies and oversights than those, already so
numerous, which he attributes to him in his article.
Myself, I am not yet prepared to do this: though I am
only too well aware how imperfect the Ethics is in
these respects.
Present state of the problem. So far our results are
negative. EOl'lO:lf.L0v[a does not mean a state or feeling
of pleasure and cXya96v does not mean conducive to
our pleasure. (It is, however, still possible that cXyo:96v
may mean conducive to our happiness, if "happiness"
is not equivalent to "pleasure.") It certainly is im-
portant to discover, therefore, what these two words
do mean. Of the two, EOl'lCXlf.L0v[a is, for reasons which
will appear, considerably the easier to elucidate, and
accordingly I shall consider it first.
n EO't:lV EOl'lO:lf.L0V[O:; the meaning of the question.
Once again I shall take my start from Prof. Prichard's
article. In a passage extending from p. 36 at the bottom
to p. 38 [pp. 256-58 above], he argues that evidence
for his view is to be found in NE I iv. I must quote,
I am afraid, at some length.
At the beginning of Chapter IV he [Aristotle]
directs his he.arers' attention to the question ...
"What is it that is the greatest of all achievable
goods?" and he proceeds to say that while there is
general agreement about the name for it, since both
the many and the educated say that it is happiness,
yet they differ about what happiness is, the many
considering it something the nature of which is
clear and obvious, such as pleasure, wealth, or
honour, whereas, he implies, the educated consider
it something else, of which the nature is not ob-
vious. Then in the next chapter he proceeds to state
what, to judge from the most prominent types of
life, that of enjoyment, the political life, and that
of contemplation, various men consider that the
good or happiness is, viz. enjoyment, honour, and
contemplation. And later he gives his own view ...
that happiness is tjJuX~c; EVEpY£leX nc; Ka't' 6:pE't~V
'tEA.E(av.
Here it has to be admitted that Aristotle is express-
ing himself in a misleading way. His question,
"What is the greatest of goods?" can be treated as
if it had been the question, "What is man's ultimate
end?" ... And his answer to this question, if taken
as it stands, is undeniably absurd. For, so under-
stood, it is to the effect that, though all men, when
asked """hat is the ultimate end?" answer by using
the same word, viz. EuBaq.J.ov[a, yet, as they differ
about what EuBaL!10v[a is, i.e. really, about the
thing for which they are using the word EuBal!10v[a
to stand, some using it to designate pleasure, others
wealth, and so on, they are in substance giving dif-
ferent answers, some meaning by the word EuBal!10-
v[a pleasure, others wealth, and so on. But of course
this is not what Aristotle meant. He certainly did
not think that anyone ever meant by EuBaL!10v[a
either 'tl!1~ or 'ltAo(hoC;; and he certainly did not
himself mean by it tjJuX~c; EVEPY£leX nc; Ka't' 6:pE't~V
'tEAE(av. What he undoubtedly meant and thought
272 J. L. Austin
others meant by the word Euoaq.LOv[a is happiness.
Plainly too, what he thought men differed about was
not the nature of happiness but the conditions of
its realization, and when he says that EUow[l0v[a is
tjJuX~<; EVEPYCL<X ne; KaT' cXpE'T~V TEAElaV, what he
really means is that the latter is what is required for
the realization of happiness ... this meaning is sim-
ilar to that of the man who, when he asks "What is
colour?" or "What is sound?" really means "What
are the conditions necessary for its realization?"
Here is the passage, I iv 1-3: MY(,)[lEV .•• 'tL 'to
TIaVT(,)V cXKPOLaTOV 'tWV TIpaK'twv aya8wv, 6vo[lan
[lEV OUV O)(EOOV lmo TWV TIAE[OT(,)V 0[l0AOYELLW' 't~V
yap EUOW[lov[av Kat OL TIOAAOt Kat OL Xap[EVTEe; M-
youow, TO 0' EU I:;~v Kat TO 'to TIpCXL'tElV 'taUTOV lmo-
Aa[l~avouOl 'tip Euow[l0vElv. TIEpl OE T~e; EUow[l0v[ae;,
T[ Eonv, cX[ltjllO~~'tOUOl Kat OUX O[lo[(,)e; OL TIOAAOt 'tOLe;
ootjlol<; cXTIoolMaOlv. Ol [lEV yap TWV Evapywv n Kat
tjlaVEpwv, olov ~oov~v ~ TIAo{hov n[l~v K'tA.*
vVhence does Prof. Prichard derive his confidence
that Aristotle is misrepresenting his own problem? If
he is, it must be admitted that the misrepresentation
is pretty consistent. Right through to Book X Aristotle
always purports to be telling us "what EUOW[lOVla is."
(He summarizes the present passage in almost the same
words again in I vii 9.)5 Moreover, Aristotle is, of course,
aware of the very kind of misrepresentation of which
Prof. Prichard accuses him: compare what he says about
*"Let us state ... what is the highest of all goods achievable

by action. Verbally there is very general agreement; for both the
general run of men and people of superior refinement say that it is
happiness, and identify living well and ,doing well with being
happy; but with regard to what happiness is they differ, and the
many do not give the same account as the wise. For the former
think it is some plain and obvious thing, like pleasure, wealth, or
honour; etc."-Ed.
5The only reasonable alternative i; to hold that in both pass·
ages it is really TO aplOTOV av(lpc;l1T4) which is being elucidated.
AgatllOn and Euclaimonia 273
pleasure in VII xii 3 and X iii 6, rebuking those who
maintained that pleasure is a YEVEOLC;, when they really
meant that a certain YEVEOLC; is the condition of the
realization of pleasure.
The real reason for Prof. Prichard's confidence is to
be found in his unquestioned assumption that EUOaL[l0-
v[a means pleasure. This assumption is stated in the
passage above: "\Vhat Aristotle undoubtedly meant
and thought others meant by EUOaL[l0v[a is happiness."
(This is, of course, rather odd as it stands, since Aris-
totle did not know English: it would lose its plausibili-
ty if we substituted ~oov~ for "happiness.") For if
EUOaL[l0Vla were the Greek word for 'pleasure', it might
well be contended that to ask .r[ £OllV EUO<'.(l[lov[a;
must be misleading: for it might very well be held 6
that 'pleasure' stands for something unanalyzable and
sui generis, which we either know, and know with entire
adequacy, from experiencing it, or do not know at
all. Pleasure might, on these lines, very well be con-
sidered to be in the same case as colour and sound, to
which, accordingly, Prof. Prichard without hesitation
compares EUOaL[l0v[a. In such cases, one who asks
"what is so-and-so?" will very probably be found to
be asking "what are the conditions for its realization?"
And whether or not 'pleasure' could in any sense be
analyzed, at least e.g. the person who said pleasure is
honour or wealth would obviously be suspect of only
really intending to give conditions of realization.
However, EUOaL[l0v[a docs not mean pleasure, as we
have seen. And this very passage proves as well as an-
other that it does not. This is not merely because the
theory that EUOaL[l0v[a means or "is" ~oov~ is rejected,
but also because in a most important clause omitted in
Prof. Prichard's paraphrase, EUOaL[l0v[a is said to be
equivalent to 1:0 EO ~~v Kat EO TIpanElV which can-
61 do not enquire whether this would be correct.

274 J. L. Austin
not mean "feeling pleased." To this we shall return
shortly.
Prof. Prichard seems to make out that, apart from
the fact that the Greeks did not disagree about what
EU!)alflov[a stands for, Aristotle's actual presentation
of his question makes it in general an absurd one. Cer-
tainly, Prof. Prichard's ostensible argument for main-
taining that Aristotle here misrepresents his own prob-
lem is not very explicit. According to him, Aristotle
says that men agree only on the name for the 'rEA.O<;,
viz. EUf>alflovla, but disagree about what it is used to
stand for, or to designate. This, he says, is undeniably
absurd. Now why? Has Prof. Prichard any other reason
for saying so, except his belief that there could, in
fact (owing to the unanalyzable nature of what
EU!)<Xlflovkx does stand for), be no disagreement about
what EU!)<Xlflov[a stands for?
He seems to suppose that, according to Aristotle,
(I) men agree only on the name; (2) that is a sub-
stantial measure of agreement; (3) what is being (mis-
takenly) asked for is some synonym for EU!)<Xlflov[a,
in the simplest sense-some other word or phrase
which stands for precisely the same as EU!)<Xlflov[a
stands for. (Somewhat as though, when asked for the
answer to a mathematical problem, all should agree
that the name for the answer was 'k' while disagreeing
as to the number which 'k' stands for.) I do not know
whether even this would be undeniably absurd (al-
ways assuming that in fact all did not know EU!)<Xlfl<:>v[a
to stand for 'pleasure') : but in any case, Aristotle does
none of these things.
(1) According to Aristotle, men agree, not merely
on the name Eu!)aq..LOv[a:, but also that Eu!)al!l0v[a
is equivalent to .0 EO l;~v Kat EO 'I1:pcXTTElV. (This
statement Prof. Prichard omits in his paraphrase.)
Moreover, it transpires later that they also agree on
a number of other propositions about the characteristics
of EUOCXl[.lovkx, which are listed in chapters Vlll and
ix-xii.
(2) As is shown by I vii 9, Aristotle does not think
that the agreement on the name alone is very sub-
stantial. And indeed it is clear even from chapter iv
that this agreement could cover most radical disagree-
ments.
( 3) Aristotle is not, I believe, searching for some
simple synonym for Euoal[.lov[a, but rather for an
'analysis' of its meaning. While satisfactory as a pre-
liminary statement, this does not make it sufficiently
clear what exactly Aristotle is doing. * All men know
more or less vaguely what is meant by Euoal[.l0v[a or
TO EO 1:;~v Kat EO npaTrElV, and agree on many prop-
ositions about it: but when they attempt to
clarify that meaning, they disagree. Cpo I vii 9:
cr.AA' faCile; T~V [.lEv EUOCXl[.l0v[av TO aplaTov MYElV
0[.lOAOYOU[.lEV6v Tl q><x[VETCXl, 1!09ElTal 0' £vapYEOTE-
pov Tl £anv En AEx9~val.t
To search thus for an analysis of the meaning of
Euoal[.l0v[a does not seem to me absurd, except on the
false assumption that its meaning was, and was known
to be, simple and unanalyzable. We might, to take
a similar case, agree that the aim of the statesman is
"liberty" or "justice," and yet, in a perfectly intel-
ligible sense, disagree about "what liberty is" or "what
justice is."
**There is no doubt, however, that this account of
what Aristotle is asking when he asks TL EOnv
Euoal[.l0v[a; is far from entirely satisfying. For we need
to distinguish from the analysis of the meaning of
Euoal[.l0v(a another procedure altogether, namely the
.. At this point, in view of the state of the manuscript, some

editing by Mr. Urmson was rcquired.-Ed.
+"Presumably, however, to say that happiness is the chief good
seems a platitude, and a clearer account of what it is is stiIl
desired."-Ed.
**See below under the same sign.
276 ,. L. Austin
"discovery" of those things, or that life, as he would
rather ordinarily say, which satisfy the definition of
EUOalflOVla when that has been discovered. As Moore
has insisted, in the case of 'good' (which will occupy
us later), it is important to distinguish the discovery
of what a word means from the discovery of those
things in which the characters meant by the word in
fact reside. (Of course this latter procedure is still not
what Prichard means when he speaks of discovering
'the conditions for the realization' of something.) This
is too simple a view for modern times, since few will
accept that goodness is a character in this simple sense.
But we still would distinguish the meaning of
Euocxlflov[cx-the best life for man, etc.-from what
we may call the specification of the good life: what
the good life allegedly consists in concretely. The
whole problem arises over the connection between
these two.
There is justice in Prichard's remark that Aristotle
"certainly does not think that anyone ever meant by
EUOalflov[cx either 'tlfl~ or 'ITAoihoc;." They were capable
of having said that they meant this, but would have
more plausibly claimed that 'tlfl~ or 'ITAoihoC; was what
satisfied the specification of EUOCXlflov[cx. But it is not
so clear that "he certainly did not himself mean by it
tjIuXOc; ~VEPYElCX 'tlc; KCX't' apE't~v 'tEAE[CXV," at least
~V P[ep 'tEAE[ep. It is hard to discover, especially in I vii,
where the analysis ends and the other process begins.
It is perhaps impossible to judge how much is meant
to be analysis. Certainly EUOCXlflov[cx is analyzed ('to:u'tov
l)'f[OACXfl~o:vOUOl) as 'to EO ~~V Kcxl EO 'ITpO:'t'tElV. Then
~V P[ep 'tEAElep (ambiguous phrase!) also seems clearly
part of the meaning of EUOCXlflov[cx. I believe that the
whole of I vii 9-16 is intended to be an analysis of
that meaning. And it results, as it should, in a clear
and full definition, referred to as 6 A6yoC; of EtX>CXlflov[cx
or of 'to apLO'tov. Moreover EE II i 10 says that
AgatllOn and Eudaimonia 277
ljJuxue; EVEpYElCX KaT' cl:pE'r~V is 'r0 YEVOe; Kat 'rOY opOV
Euoaq.lOVlae;. If it had been said that Aristotle did not
mean by EUoCXlflovla 9EWplcx, that would, I think, be
certainly true: and it is 9Ewpla which, in Aristotle's
theory, occupies the place of ~oov~ and TIAOU'rOe;
in rival theories, as chapter v plainly shows.
But when Aristotle discusses what are in fact the
special virtues, and which is the most perfect, he can-
not be said any longer to be analyzing the meaning of
EUOCXlflovlcx. He is asking 'rl Eonv EUoCXlflovla; in a dif-
ferent sense: what are the virtues that fill the bill. Even
in X viii 8, where the conclusion is reached:
&Ot' ELY] av ~ EUOCXlflovlcx 9EWplcx ne;, it is evident that
EUoCXlflovla does not mean 9EWplcx-E<jl' ooov o~
OlCXtElVEl ~ HEwpla Kat ~ EtloCXlflovla, Kat Ole; flO:AlOtCX
UTIO:PXEl to 9EWPElV Kat 'rO EUOCXlfloVElV OU KaTO:
OUfl~E~Y]KOe; cl:AAcl: KaTO: 'r~V 9EWplcxv.*"
So Aristotle's distinction between analysis and speci-
fication is most unclear. But there is some excuse for
Aristotle perhaps, in that EUOalflovlcx does not stand
for some character, such as goodness might be, but for
a certain kind of life, or EVEPYElCX (Aristotle is un-
clear as to which): in such a case it is not so easy
clearly to observe Prof. Moore's distinction, or even
one such as Hare's or Urmson's between meaning and
criteria. Suppose we were to ask, for instance, "What is
golfing?" But there is this finally to be said. If Aris-
totle had thought that EUoCXlflovla, like golfing, resided
in fact in only OIle activity of one kind of creature,
there would have been more excuse for him than is
* *The material included between the double stars required

extensive editing by Mr. Urmson because the manuscript on these
pages was complicated by notes, corrections, and revisions. Ross's
translation of this Greek passage reads: "Happiness extends, then,
just so far as contemplation does, and those to whom contem-
plation more fully belongs are more truly happy, not as mere
concomitant but in virtue of the contemplation; ... Happiness,
therefore, must be some form of contemplation."-Ed.
278 ,. L. Austin
actually the case. For actually he does think that
EUoolllov[a is achieved, in different ways, by gods and
by men: hence EUoolllov[a cannot mean those activi-
ties in which human EUoolllov[a is found. (Unfortu-
nately, of course, his statements on divine EUoolllovla
are rudimentary, and it is very doubtful, e.g. how
£v [3lct:» 1:EAElct:» can be a part of the meaning of EU-
5alllov[a if the gods are also Eu5alllovEC;!)
Some distinctions, however, though not this requisite
one, Aristotle does draw: in a way he was perhaps on
his way to it. In Rhet. I v ad init. he says we must ask:
1:( £01:lV Eu5aLIloVla Kai 1:a Ilopla aU1:~C;; and this dis-
tinction is common, though not in the NE. In EE I v
13-14, he calls the particular virtues Ilopla 1:~<;
aycxe~<; l;cu~<;, which OUV1:elVOUOl npo<; Eu5alllov[av.
(This does not, of course, mean that they are "the
conditions for the realization" of Eu5aLllov[a: in NE
1129b18, where the Ilopla EUoaLllov[ac; are mentioned,
they are distinguished from 1:a nOlT]1:lKa Kai
<puAaK1:lKa Euoalllov[ac; [cpo VI xii 5], a distinction
insisted on in e.g. EE I ii 5. There is a similar dis-
tinction in EE I i of EV 1:[Ol 1:0 EO ~~v from
nw<; K1:T]1:OV; compare also MM I 11 9-11:
ou yap EOTlV O:AAO Tl Xc.ilplC; 1:0ll1:CUV ~ Eu5aLllov[a
ana mU1:a.) 7 Now in effect, the discovery of the
Ilopla EUOaLllov[ac; is the discovery of the activities
which together make up the life which, for man, satis-
fies the definition of EUoolllov[a; and we can in a sense
say, as Aristotle does, that this life is Eu5aLllov[a.
Nevertheless, EUoolllov[a does not mean that life, and
its discovery is posterior to the analysis of the meaning
of EUoolllov[a. Two erroneous presuppositions were,
however, encouraged by Aristotle's failure to be clear
what question he is asking when he writes 1:( £01:lV
7It is important to remember this when interpreting sueh a
passage as I vii 5. Compare EE II i 12, a better statement.
[Austin's MM is a reference to Magna Moralia-Ed.]
EUOCXlflOV[CX; first, the presupposition that EUOCXlfloV[cx-
the ideal life-is not a will-O'-the-wisp, and that there
is only one possible ideal life; second, that the question
of what fills the bill is throughout purely factual.
Summing up then, the question 't[ EO'tlV EUOCXlflov[cx;
is sensible but ambiguous. Aristotle means to ask
firstly: what is the analysis or definition of EUOaLflov[cx?
and secondly: what life, in particular for man, satisfies
that definition or specification? (A subsidiary and dis-
tinct question is: \Vhat are the conditions WV OUK
CiVEU, and the methods necessary for the realization of
such a life?)
\Vhat does EUOCXlflov[cx mean? Some general con-
siderations. \Ve must now concern ourselves with the
first of these questions: vVhat docs EUOaLflov[cx mean
in Greek and incidentally what is its translation in
English? But I cannot enter into details of Aristotle's
own analysis: I shall only attempt to show what, ac-
cording to him, was the vague and common notion
of EUOaLflov[CX from which his analysis starts.
According to I iv 2, 'to E00CXlfloVElV was admittedly
equivalent to 'to IOU s~v KCXt IOU TIpO:'t'tElV. (In I vii 4 there
is a weaker statement: it is admitted "that the Euocx[flC0V
lives and acts well" -but, as Aristotle remarks, on his
own account EUOCXlflov[cx is living and acting well: and
in both the other Ethics the statement is given as in
I iv 2; cpo MM I iii 3; EE i 10 and I i 7.) EUOCXlflov[cx,
then, means living a life of a certain kind-of what
kind can of course only be discovered by analyzing the
word IOU, and hence aycx86v (so that the full analysis
of EUOCXlflov[cx includes that of aycx86v). EUOCXlflC0V,
I suppose, means literally or, what is often the same,
etymologically, "prospered by a deity": and what the
deities prosper is lives or careers or activities or parts
of these. Aristotle insists on two further points-
EUoCXlflovla means a complete life of activity of a cer-
tain kind. On the latter point, that the reference is to
K
280 J. L. Austin
EVEpYEla not to EE,lC;, he is always firm (cp. also I xii).8
On the former point he IS not so happy: not only is
!3[OC; 'rEAELOC; hopelessly ambiguous (cp. MM I iv 5),
but Aristotle often omits to remember this qualifica-
tion. So much so that in the end he never explains how
the !3[oc; is made up-only the apE'ral which predomi-
nate it.
At any rate, what is important in all this is, that,
though of course we can speak of a man as Eu!)a[flUlv,
the substantive with which Eu!)a[flUlv naturally goes is
!3[oc; or a similar word: a man is only called Eu!)a(flUlv
because his life is so. Hence the discussion in chapter
v of the various l3(ol which lay claim to being
Eu!)alflov[a.9 And hence the saying "call no man
Eu!)a[flUlv until he is dead" (I xi).
Similarly the forms Eu!)alflov[l;ElvandEu!)alflOVlOfl6c;,
which seem to mean "to congratulate," "congratula-
tion." That whereon I congratulate someone is an
achievement, an activity and normally a completed
one (though normally also, of course, of less extent
than his whole career). With reference to this point
consider I v 6 and viii 9.
These considerations show conclusively that EUOOl-
flov[a could not mean "pleasure": pleasure is a teeling,
not a life of a certain kind nor an achievement: nor do
I congratulate someone on his feeling pleased: and it
80n p. 38 [po 258], Professor Prichard makes what seems to me

serious misstatements about I xii. "In other words ... everything
else." The contrast in the first sentence I, of course, consider
mistaken. As for the rest, the quotation given is actually fom
section viii where EUOa!lIovia is not contrasted with O:PET~ at
all. The contrast with O:PET~ is in ii and vi: O:PET~ is E1TalvETclV
qua TIl 1Toiov Kal 1TPOC; Tl Elva! - viz. the EVEpyEia! (1Tpa~EIC;,
Epya) which are EV6alllovia: the EVEpyEia! are Tillia and are
Evliaillovia. Professor Prichard talks as though in xii EVEpyEial
KaT' 6:PET~V were ETTalvETO:: they are Tillla!
9And when in I iv the suggestion is made that Ev6a!llovia is
li6ov~ or 1T~OilTOC;, that is loose language (as V shows) for
the life in which most pleasure or most wealth is gained.
would be silly to say "call no man pleased until he
is dead."
There is, however, Aristotle's own remark in I iv
that many people do maintain that EUool!lov[a is
~50v~, as others TCAO(ho<; or "[L!l~. This is a loose re-
mark-as Prof. Prichard claims, though not in his way.
His explanation, that Aristotle meant that some main-
tain that EU5al!lov(a is produced by ~50v~, also TCAOtho<;.
etc., is incorrect. Aristotle himself shows what he meant
more fully in chapter v. The view was that Eu5aq,.lOv[a
is the aTCoAauo-rlKo<; ~[o<;, the life in which most pleas-
ure is felt. Likewise the identification with TCAOU-rO<;
should be taken as "the life in which most wealth is
gained" (not, of course "is by definition" but "in fact
resides in").
If we want a translation of EUool!lov(a which will
not mislead, as "happiness" appears to mislead Prof.
Prichard, we might use as a prophylactic "success."
"Success" is at least a word of the same type, so to
say, as Eu5aL!l0v[a. "Success" does mean living and act-
ing well: a life or a part of it is "successful": and with
some hesitations I do congratulate on "success"; it
might well be said "call no man successful until he is
dead." Furthermore, success demands just that fortu-
nate supply of Eu-rux(a, EU1')!lEp[a, which Aristotle ad-
mits in I viii 15-17 is also a necesS<U)' ronJition for
Eu5al!l0v[a.
It is true, however, that "success" is 'not a moral
notion for us. Perhaps this is no great disadvantage, for
it is doubtful how far a pagan ethic, such as the Creek
(or the Chinese) gains by a translation which imports
our own moral notions: Eu5aL!l0v[a is certainly quite
an unchristian ideal. Still, we do require a word to
import some form of commendation, as Euoa[!lwv
did, and to certain non-pcrsonal standards.
**That Eu5aL!l0v[a did mean life of activity of a
certain kind is almost certainly the correct analysis;
282 ,. L. Austin
and that it did further mean life of apETu[ seems equal-
ly correct. So we may say that the analysis in I vii 8-16
is correct and is supported by I viii. However we must
also say: (a) That roundabout way of bringing into
the discussion ljJuxT'] and EpyOV is a piece of unneces-
sary Aristotelian metaphysics. It is not really made use
of until the very end, in the argument for the su-
premacy of SEC0p[U; the argument proceeds straight on
to the apETal. (b) The whole discussion here is not
purely factual; its nature is disguised by the trans-
ference of commendation to the apETUl. (c) Of the
three lives suggested none accords with the popular
view which was that of Tellus the Athenian.* His
ideal is omitted altogether except when allowed in by
the back door suddenly in Book X. (d) It might be
argued that he omits to give sufficient consideration to
[lE!f ~oov~c; ~ OUK aVEU ~oov~c; and it might be ar-
gued further that not enough deference is shown to
ljJuXlKov-people may have meant by this something
genuinely internal, as did later the Stoics. Quite pos-
sibly EUOal[l0V[U did have this "meaning" too. * *
Nevertheless, "happiness" is probably after all to be
preferred as a translation, partly because it is traditional,
and still more because it is fairly colourless.1° It seems
to me very rash to assume that in common English
"happiness" obviously means feeling pleased: probably
it has several more or less vague meanings. Take the
lines:
*Tellus the Athenian is mentioned by Plutarch in his Life of

Solon, where Solon is said to describe him as a happy man on
account of his honesty, his having good children, his having a
competent estate, and his having died bravely in battle for his
country. Tellus appears to have been a' plain man whose happi-
ness, nevertheless, Solon holds up to Croesus.-Ed.
* * ... * * This passage appears to be a later addition by Austin,
which we inserted here in a somewhat expanded form.-Ed.
lOSuccess also does not always or usually refer to life as a whole:
and it has perhaps a nuance of competitiveness.
AgatllOn and Eudaimonia 283
"This is the happy warrior, this is He
That every Man at arms should wish to be."
I do not think Wordsworth meant by that: "This
is the warrior who feels pleased." Indeed, he is
"Doomed to go in company with Pain
And fear and bloodshed, miserable train."
(Though no doubt his life is !lEe' ~bovilc; ~ OLlK avEU
~bovilc; as Aristotle likes to assume or feebly to argue
his own chosen one must be.) \Vhat every man at
arms is being incited to wish, is not so much to get
for himself feelings comparable to those of the paragon,
as to imitate his lite. I think, then, that if we are on
our guard against misleading nuances, "happiness" is
still the best translation for EUool!l0v[a.
The question of the relation of EUOaL!lov[a to
~bOV~ has not, it should be noticed, yet been cleared
up. It to a great extent coincides with the equally
difficult problem of the relation of ,0 aya86v to
,0 ~M. Both must be reserved for a separate discus-
sion later.
6:ya86v-Does Aristotle tell us what its meaning is?
Prof. Prichard's contention is, it will be remembered,
that Aristotle really means by 6:yae6v "conducive to
our happiness." Not concerning ourselves for the mo-
ment with the precise interpretation given, we may
notice, firstly, that he is purporting to give Aristotle's
answer to a question which Aristotle himself expressly
declines to answer, viz. what is the meaning of ayae6v?
And secondly, that the answer given, implying as it does
that ayae6v does have a single meaning, is of a kind
which Aristotle himself is at pains to prove impossible.
Let us take the translation of 6:yae6v as "good"
here for granted, and let us once more, following Prof.
Moore, distinguish between two very different sorts
of question, which are commonly asked in books on
ethics. We are investigating, we may say, the Good:
284 J. L. Austin
but we may intend to ask: (1) What does the word
'good' mean?l1 or (2) What things are good, and in
what degrees? Of course, these two questions may be
formulated in a variety of ways. For (1) we may sub-
stitute: What is the nature of goodness? or: In saying
of anything that it is good, what am I saying about
it? even: What sort of a predicate is 'good'? and so
on. For (2) we may substitute: Of what things may
it be truly said that they are good? And which is the
best of them? and so on.
To these two different sorts of question, as Moore
claimed, we shall get two correspondingly different
sorts of answer. To (1) the answer might be: Good-
ness is a simple unanalyzable quality like yellow, or:
"Goodness" means "approved by me," or: To say of
anything that it is good, is to say that it is conducive
to happiness, or: 'Good' is an evaluative word. Whereas
to (2), the answer might be: Friendship is good, or:
Violence is better than justice, and so on. Note here
that it is assumed that the only sense in which ayo96v
has a "meaning" is in some "factual" sense. Aristotle
assumes this too.
Now we must ask, on this distinction does Aristotle
concern himself with both these questions or with
one only? And if the latter, with which? And I think
that we must answer, that he concerns himself pro-
fessedly with the second only,12
For Aristotle himself is aware of and draws this dis-
tinction between the two questions, and says that, in
the Etllics, he is concerned only with the second. This
he does in the celebrated chapter I vi. He there con-
llWhich is itself probably an ambiguous question, but we may
let this pass for the moment. (We also have to distinguish (a)
how to translate a word? (b) what does someone say about the
analysis of the meaning or of the definition of a word? and both
of these from (c) possibly different senses of 'mean'.)
12And with that only in the more special form: "What par-
ticular things are good, and in what degrees, fOT manl"
futes those who had supposed that the word ayae6v
always stands for a single identical predicate. 13 But in
proving this, he docs not tell us what are the various
meanings of ayae6v-the furthest he goes in that direc-
tion is to give us a hint as to how the various meanings
may be related to one another, i.e. how the variety
yet forms a unity. He then dismisses the matter,
1096b 30: aAA' lO(0<; 1:a01:a llcV a<jJ2:1:COV 1:0 vOv.
Et:aKpl~oOV yap UTIEp atnwv aAAl']<; av c'il'] <jJlAo-
oO<jJ[a<; OlKcl61:cpov. * He then turns to his present
problem: \Vhat is that good which is TIpaK1:0V Kat
K1:l']1:0V avepC:my, that is: What particular things are
good for man, (and in ,,,hat degrees)? Unfortunately,
as is well known, he does not in fact discuss "the
meaning of c\yae6v" elsewhere.
It is clear, then, that Aristotle declines in general
to discuss the meanings of aya66v but argues that it
has no single meaning. In both respects Prof. Prichard's
view conflicts, prima facie, with Aristotle's statements.
Nevertheless, it may be urged, both that Aristotle is
unjustified in declining to explain the meaning of
ayae6v and that he must himself attach some meaning
to it in using in throughout the Ethics, which we may
be able to discover. And further, that meaning might
be identical in all important cases (this will be
explained later).
If asked to justify himself, there is no doubt how
Aristotle would reply.I! The NE is only intended as
13Compare [H.\V.B.] Joseph, Some Problems in Ethics [Ox-
ford, 1931], p. 75.: "That goodness is not a quality is the burden
of Aristotle's argument in NE I vi.": but this is not quite correct.
Aristotle is anxious to say thatayex80v has no single meaning-
whether a quality, a relation, or anything else. As a matter of fact,
he says it does sometimes stand for a quality_
*But perhaps these subjects had bctter be dismissed for the
present; for perfect precision about them would be more appro-
priate to another branch of phiiosophy."-Ecl.
14Aristotle does to some extent reply to this objection in I vi
14-16.
286 J. L. Austin
a guide for politicians, and they are only concerned
to know what is good, not what goodness means. Prob-
ably Platonists would have said that I cannot dis-
cover what is good until I have found the definition
of goodness, but Aristotle would claim that the defini-
tion, if possible at all, is only necessary if we wish to
demonstrate scientifically that certain things are good.
So much o.Kpt~Ela is not called for in the NE, and
in any case one can know what things are good without
knowing the analysis of 'good'.15 Whether this reply
is satisfactory is doubtful. Certainly there is much to
be said on Aristotle's side. Firstly, as Moore pointed
out, we can know something to be true without know-
ing its analysis. Secondly, as Aristotle pointed out in
I vi 4, if goodness is an isolable, definable property
we should be able to study it in isolation from all
subject matters, whereas in fact there is no such study.
Thirdly, even those who, like Moore, find goodness
"unanalyzable" still go on to discuss what is good.
But in any case it must be admitted that his method
has its dangers; and whether they are serious can best
be judged by its results in the body of the NE. (In at
least two cases, those of ~50v~ and <illAta and perhaps
also in that of Tlfl~ the lack of a clearer account of
the meaning or meanings of aya96v is in fact most
serious. )
The extent to which Aristotle does in fact attach a
discoverable, and even single, meaning to aya96v will,
of course, concern us largely. But before proceeding,
it is worth considering an example of the extreme
lengths to which Prof. Prichard is prepared to go in
imputing inconsistency to Aristotle. His view, we said,
implies that aya96v does have a 'single meaning, i.e.
does always stand for an identical common character
in the subjects of which it is predicated. (Note that
151 do not, however, accept Burnet's exaggerated view of the
"dialectical" method of the NE.
Agatbon and Eudaimonia 287
we need 'character' here, not 'quality'; for, of course,
one of Prof. Prichard's main contention is that Aris-
totle mistakes "being conducive to our happiness" for
a quality of that which is so conducive, whereas it is
not.) And this, we said, conflicts with Aristotle's own
statement that ayaS6v does not do this.
Now it may be thought that I have been wasting
unnecessary words over this, since Prof. Prichard him-
self notices this objection and answers it on pp. 32-33
[po 250.]
But Aristotle in saying, as he would have said, that
in pursuing, e.g. an honour, we are pursuing it
we; ayaS6v, could only have meant that we are
pursuing it in virtue of thinking that it would possess
a certain character to which he refers by the term
ayaS6v, so that by aya86v he must mean to indi-
cate some character which certain things would
have. Further, this being so, in implying as he does
that in pursuing things of certain different kinds
Ka8' alrro: we are pursuing them we; ayaS6:, he
must be implying that these things of different kinds
have, nevertheless, a common character, viz. that
indicated by the term ayaS6v. It will, of course,
be objected that he expressly denies that they have
a common character. For he says: llfl~e; bE Kat
cppov~aEC.)e; Kat ~bOV~e; ELEpOl Kat blacpEpOVLEe; ot
MyOl LauLD n ayaS6: (Etbics I vi 11).* But the
answer is simple; viz. that this is merely an incon-
sistency into which he is driven by his inability to
find in these things the com1110n character which
his theory requires him to find, and that if he is to
succeed in maintaining that we pursue these things
of various kinds we; ayaS6:, he bas to maintain that
in spite of appearances to the contrary they have a
common character.
*"But of honour, wisdom, and pleasure, just in respect of their
goodness, the accounts are distinct and divcrse."-Ed.
288 J. L. Austin
We are not concerned for the moment with this
doctrine about "pursuing things <:0<; aya86:."16 But what
we must notice is that Prof. Prichard quotes the sen-
tence about tl[l~, <i'p6VT]at<;, and ~oov~, as though it
were an unwilling admission, into which Aristotle's
honesty drives him: as though Aristotle is admitting
that in certain cases he cannot find the common char-
acter, which he must find. Yet a glance at the context
will show that, on the contrary, the whole passage is
designed to prove that there is no such common char-
acter: and the sentence quoted clinches the argumentP
It is not too much to say that, if TL[l~ and the rest
are possessed of a cornman character denoted by
aya86v, the whole argument against the Platonists is
undermined: and Prof. Prichard must accuse Aristotle
not so much of inconsistency as of bad faith.
In any case, this passage cannot be thus lightly dis-
missed. The objection retains, it seems to me, very
considerable weight.
aya86v-Does it mean "that which is desired"? We
may now turn to the main problem, the meaning of
dya86v. It must, I think, be agreed, that there is some
chance of our being able to discover what Aristotle
does in fact believe about the meaning of the word:
but, since Aristotle appears to decline to assist us, that
chance is small.
It is on p. 31 [po 248] that Prof. Prichard begins his
attempt to elucidate the meaning of dya86v: and he
says that "Aristotle's nearest approach to an elucida-
l6 W'; aycx8a does not in Greek mean "in virtue of their possess-
ing a certain characteristic, 'goodness.' .. It means rather "pursuing
them in the way in which we pursue things we say are good."
Thus the phrase is noncommittal as to whether there is a com-
mon quality or not. Cpo in English 'stating as a fact: 'stating as
a matter of opinion.'
l7Aristotle does not say he "cannot find" a common character
denoted by aycx86v in these instances: he says that he knows
uycx86v stands for different characters.
tion" is to be found in I vi 7-1118 and vii 1-5. Certain
statements there made, he says, "seem intended as an
elucidation of the meaning of aya96v"; nevertheless
they fail, even formally, to constitute such an elucida-
tion: and moreover, if they are "taken seriously as an
elucidation," the result is to render absurd other well-
known doctrines in the NE.
I suppose it cannot be denied that Aristotle is capa-
ble of getting himself into pretty tortuous confusions;
but this time, at least, I think he can be exonerated.
In the first place, it seems to me that the statements
in question are not intended at all as an "elucidation
of the meaning of aya96v" in Prof. Prichard's sense. 19
It is, consequently, not surprising that they fail for-
mally to constitute such an elucidation: what is sur-
prising, I think, is Prof. Prichard's reason for saying
that they so fail. In the second place, if they are "taken
seriously as an elucidation," they do not so much
render absurd the doctrines mentioned by Prof. Pri-
chard as others which it is easy to find.
In these passages, says Prof. Prichard, Aristotle
speaks of "ta Ka9' au"ta f>lc.>K6~Eva Kat ayancb~Eva
as called aya9a in one sense . . . and he speaks of
"ta nOlll"tlKa "tou"tc.>v ~ q>uAexK"tlKa nc.>c; as called
aya9a in another sense, and he implies that these
latter are f>lc.>K"ta Kat atpE"ta fa' E"tEpOV (Ethics I
vii 4) ... Further he appears to consider that the
difference of meaning is elucidated (sic) by referring
to the former as aya9a Kex9' au"ta and to the lat-
ter as aya9a f>la "texO"tex, i.e. &yex9a f>la &ya9a
Kex9' exu"ta. But this unfortunately is no elucidation,
since to state a difference of reason for caIling two
things &ya96v is not to state a difference of mean-
18For "7-11" we should read "8-11."
191£ any passage deserves the description "Aristotle's nearest
approach to an elucidation," it is probably I vii 10 (not men·
tioned by Professor Prichard); on this v.i.
290 J. L. Austin
ing of ayaS6v, and indeed is to imply that the
meaning in both cases is the same. Nevertheless,
these statements seem intended as an elucidation
of the meaning of ayaS6v. And the cause for sur-
prise lies in this, that if they are taken seriously as
an elucidation, the conclusion can only be that
ayaS6v includes 'being desired' in its meaning, and
indeed simply means TEAoc; or end.
It will, I am afraid, necessarily take some considerable
space to clear this matter up.
Let us, as a preliminary, remark that the passage vii
1-5 is not concerned with at all the same matter as vi
8-11. In vii 1 he asserts that TO d:ya86v of any activity
whatsoever is TO TEAOC; (incidentally this passage im-
plies that TO ayaS6v and To TEAOC; do not mean the
same): he then proceeds to ask, what is the TEAoc;
TWV ll:paKTWV a1'[(XVTwv? And the word ayaS6v
is not so much as mentioned again. Moreover, though
he does proceed to distinguish a TEAOC; KaS' aUTO
blWKT6v (atpET6v) from a TEAOC; bl' ETEpov blWKT6v
(atpET6v) he does not assert that the word TEAoc; has
(in any sense) two meanings, only that there are some
TEAT] which are TEAEla (complete or final), others
which are not. Still less, of course, does he assert that
d:ya96v has two meanings, or endeavour to elucidate
its meaning: he says nothing about it at all; just as, in
vi 8-11, no mention is made of TEAOC; at all.
We may confine ourselves, therefore, to vi 8-11.
Upon what is Aristotle there engaged? In the earlier
part of the chapter he has been arguing that d:yaS6v
cannot have always an identical meaning. But now he
produces an attempted answer to his own objections,
which is to be understood by referring to his own logic
(Categories I al). A word (Bvolla) may be used either
auvwvullwC;, 0llwvullwC;, or ltapwvullwC;. If, on each
occasion of its use, its connotation (6 KaTtl: Touvolla
MyoC; T~C; oualac;) is identical, then the word is used
Agathon and Eudairnonia 291
auvcuvuf-lCUC;. If, on different occasions of its use, its
connotations are different, then the word is used
of-lcuvuf-lcuc;: e.g. KAdc; may be used to mean "key" or
"collarbone," t,yov "animal" or "picture." But there
is a third possibility: on different occasions of its use,
the word may possess connotations which are partly
identical and partly different, in which case the word
is said to be used TICXPCUVUf-lCUC;.20 There are evidently
many ways in which a word can be used paronymously:
Aristotle names some of them, and gives examples
(,(0 UYlElV6v in Met. 1003a33 '(0 lalplK6v and '(0
UYlElV6v in Met. 1060b37. Also notoriously '(0 QV with
which ci:ya86v here is again compared, and '(0 EV. For
more details, the reader may refer to those passages, or
to Burnct's notes on p. 29 of his edition of the
Et11ics) .21
One such type of paronymity is known as the "npoc;
EV." \Vhen we speak of a "healthy exercise" the word
'healthy' has a connotation which is only partly the
same as that which it has in the phrase "a healthy bo-
dy": a healthy exercise is an exercise which produces or
preserves healthiness in bodies. Hence healthiness a ,
when predicated of an exercise, means "productive or
preservative of healthiness b " i.e. of healthiness in the
sense in which it is predicated of bodies. Thus "healthi-
ness b " and "healthiness a" have connotations which are
partly identical and partly different.
Now, in our present passage, vi 8-11, Aristotle is
producing as an objection against himself the fact (ad-
mitted elsewhere, cpo Rhet. 1362a27) that ci:ya86v is
paronymous in this way. His opponents are supposed
20Compare the traditional classification of terms as univocal,
equivocal, and analogous.-Joseph, An Introduction to Logic (Ox.
ford, 1916), p. 46: "analogous" is unsatisfactory, since KaT'
CevaJo-oYlav is only one form of paronymity.
"This supplements and amends Cat. I. a bit; homonymy and
parOll\·my are defined by very limited examples there. Consider
also Rhet. 1362a21 If. with regard to this point.
292 ,. L. Austin
to claim that, although aya90v does not always have
an identical meaning, that is merely because it is paron-
ymous in the above manner. Sometimes it means "x,"
sometimes "productive, etc. of x," etc.; and clearly it
is only the "nuclear" meaning of "x" which is common
to both, with which they are concerned. And it is al-
ways identical. To which Aristotle replies as follows.
First, let us distinguish aya80v as used in the one
sense, from aya86v as used in the other, by substituting
for it, in the one case aya80v Ka8' aLrro, and in the
other aya80v fHO: aya80v 'tl Ka8' a6'to.22 Then let
us disregard all cases where aya80v is used in the latter
sense. And finally observe that, even in cases where it
is used in the former sense only, it still has not always
an identical meaning.
Now let us ask, with reference to Prof. Prichard's
argument quoted above, does Aristotle intend here to
"elucidate the meaning" of aya86v? It is clear that
there is no simple answer, Yes or No, to this question.
He does intend to point out that it has at least two
meanings, at least in a reasonably strict sense, and to
explain how those two meanings are related, and to
show how they are in part identical. But he does not
intend to elucidate that identical part. Moreover, he
proceeds to assert that in fact aya86v, used in the
nuclear way, has at least three different meanings (so
that there may be at least six meanings of aya86v in
general): but he does not mean to elucidate anyone
of the three, nor even to explain how they are related
to each other. It is not, then, surprising that Prof.
Prichard should feel dissatisfied; if we ask his over-
simplified question, the passage must "seem intended
as an elucidation" of the meaning of aya96v, and yet
fail to be one-for plainly, Prof. Prichard would only
count as a real elucidation, an elucidation of the mean-
2271'pOC; might have been expected rather thanBlCic: but Buicis
more general.
ing of d:yaS6v in the nuclear sense of senses, which
Aristotle is not concerned to give him.
Prof. Prichard's remark, that "to state a difference
of reason for calling two things d:yaS6v is not to state
a difference of meaning of d:ya86v, and indeed is to
imply that the meaning in both cases is the same,"
remains to me obscure. But it geems clear that he has
not appreciated the doctrine of 'ITapc0vu!la, since he
uses the rigid dichotomy "same meaning-different
meaning" whereas 'ITapc0vu!la are words which have
meanings partly the same and partly different. His
"difference of reason" might be the point about, say,
healthy. Then Aristotle would accept the claim, and
say that what he offers is not an elucidation and is not
intended to be one. vVe have seen, however, that
Prof. Prichard is prepared to believe that the passage
may be seriously intended as an elucidation of d:ya96v
in the "nuclear" sense. Seeing that he has already dis-
missed the distinction between d:yaSa KaS' aU1:a
and d:yaSa Ola 1:aU1:a as "no elucidation," and
seeing that Aristotle makes no attempt to elucidate the
meaning of d:yaS6v in the nuclear sense, and further
states that it has at least three nuclear senses, it is
clearly bound to be difficult for Prof. Prichard to find
anything in the passage which could be described as
"an elucidation of the meaning of d:ya86v." He fastens
upon the words OlUlK6!lEva Kat d:ya'ITc0!lEva, Olc0KE1:at,
Olc0K0!lEv and concludes that, according to Aristotle
here, o:ya86v "simply means 1:£AO<; or end."23
Let us proceed to consider first Prof. Prichard's ar-
guments to prove that d:ya86v does not have that
meaning. After asserting that it quite certainly is not
used by Aristotle with that meaning [which in a way
is true: but surely then, it is unlikely that he intends
23It seems to me doubtful whether TEAOC; means simply "some-
thing desired": but for present purposes it will do no harm to
suppose it does so.
294 J. L. Austin
to ascribe that meaning to it in this passage?] he pro-
ceeds as follows: "Apart from other considerations, if
he did [mean by d:ya8ov simply TEAOC;], then for him
to say, as he in effect does, that we always aim at
d:ya8ov n would be to say nothing, and for him to
speak, as he does, of the object of ~OUA1l0lC; as TCxya80v
would be absurd."
The first of these arguments is perhaps too concise
to be clear. "We always aim at aya80v n" may per-
haps mean
(1) whenever we act, we aim at aya80v Tl, or
(2) whenever we aim, we aim at d:ya8ov Tl
If (I) is meant, I very much doubt whether Aristotle
would maintain it; for we sometimes act from E11l8ufl[a
or 8ufloC;, and then we aim at ~ou n or at aVnAUTIllolC;.
But he might say24 that in all deliberate action we
aim at d:ya8ov n, as he implies in the first sentence of
the Ethics. 25 However, this qualification need not con-
cern us here (though, since he would certainly have
said that every action aims at some TEAOC;, it is clear
enough that d:ya8ov does not mean TEAOC;). If, then,
(1) is meant, and if d:ya8ov n means TEAOe; n, Aris-
totle will be saying "vVhenever we act, we aim at some
end." But to say that is not to say nothing: indeed,
it is to say something which Prof. Prichard declares,
on p. 29 [po 243], to be false. 2G At most the sentence is
pleonastic: for we could write TlVOe; for TEAOUe; nvoc;
(or d:ya8oQ nvoe;) without loss of meaning. But that
is a very minor matter. And this would serve, e.g., to
define d:ya8ov.
If (2) is meant, the sentence is certainly emptier:
"whenever we aim, we aim at some end." It might be
said to be a matter of definition. But it might well be
24Wrongly.
25The importance of which should not be exaggerated.
26Because he understands by TEAO<; only what Aristotle calls a
TEAO<; m:xpa TtlV 1Tp6:~IV, I confess that his criticism at this point
seems to me perverse.
said, for emphasis, in explaining the notion of "aiming,"
or in otherwise explaining the use of words. However,
for reasons similar to those given in the preceding
paragraph, I do not consider that Aristotle would ever
have made such a remark about ayu86v, though he
might well have made it about TEAoe;. If we change it
to "whenever we aim deliberately, and only then, we
aim at ayu86v n," Aristotle might agree: but tIlen,
if we substitute TEAoe; n for aya86v n, the statement
would simply be false. Thus proving, though not by
Prof. Prichard's argument, that aya86v does not mean
TEAoe;.
Now take Prichard's second argument. Here, no such
objection applies as in the case of the first argument,
for Aristotle certainly would say that the object of
~06AY]0le; is Taya86v. If aya86v mcans TEAoe;, the
statement becomes: "the object of ~06A1']ale; is TO
TEAOe;." Now this is not, as a matter of fact, as it stands,
absurd-it is in fact a remark which Aristotle often
makes, with good sense (cp. e.g. 1111 b26, or 1113a15,
l'j ~06A1']ale; TOG TEAoue; EaTlV). However, TO TEAOe; is
there being contrasted with Ta npoe; TO TEAOe;, and
~OUA1']Ole; with npOalpWle;: whereas it is clear that
Prof. Prichard means, and fairly enough, to refer to a
context where ~06A1']0le; is being contrasted with
Em8u[1la and 81'] [16e;, and Taya86v with TO l'jM and
avnAUny] au:;. In such a context, it is clear enough that
we cannot substitute TEAOe; for aya86v. For in the
case of the other two types of 6pEE,ElC; we are also aim-
ing at a TEAOe;, and we need ayu86v not 'rEAoe; to
provide the contrast. Prof. Prichard's arguments, then,
reduce to one, which I think is sound: ayu86v cannot
mean "that which is desired," because Aristotle holds
that there are other objects of desire besides TO aya86v.
Apart from this argument, it is absolutely clear from
I vi itself that aya86v docs not, for Aristotle, mean
(or at any rate merely mean?) "that which is desired."
296 J. L. Austin
It is not merely that he says that aya96v has no single
meaning, and hence a fortiori, not that suggested. But
further, he says of "t:l!l~ cpp6VT]OlC; and ~oov~ that they
are all OlC0K6!l€va Ka9' aLno:, so that, presumably,
their MYOl n OlC0K6!lEVa Ka9' a61:0:* are identical:
whereas their MYOl mU1:TI n cl::ya90: are different.
So that being aya96v cannot mean being desired.
I expect the reader has found the proof of this
tedious. But it is important to establish the point,
since the relation between "being cl::ya96v" and "being
desired" is one of the most baffling puzzles in Aris-
totle's, or for that matter Plato's, ethical theory. This
puzzle, like that of the relation between cl::yae6v and
~06, must be reserved for a separate discussion.
"Their accounts as things pursued for their own sake." (Ed.
trans!.) .
THE FINAL GOOD IN
ARISTOTLE'S ETHICS
w. F. R. HARDIE
Aristotle maintains that every man has, or should

have, a single end (-rEA.OC;), a target at which he aims.
The doctrine is stated in NE I 2. 'If, then, there is
some end of the things we do which we desire for its
own sake (everything else being desired for the sake
of this), and if we do not choose everything for the
sake of something else (for at that rate the process
would go on to infinity, so that our desire would be
empty and vain), clearly this must be the good and
the chief good. Will not the knowledge of it, then,
have a great influence on life? Shall we not, like archers
who have a mark to aim at, be more likely to hit upon
what is right?'l (1094a18-24). Aristotle does not here
prove, nor need we understand him as claiming to
prove, that there is only one end which is desired for
itself. He points out correctly that, if there are objects
which are desired but not desired for themselves, there
must be some object which is desired for itself. The
passage further suggests that, if there were one such
object and one only, this fact would be important and
helpful for the conduct of life.
From PhilosoPhy, XL (1965), 277-95. Reprinted by permis·
sion of the author and the editors of Philosophy.
1 Here, and in quoting other passages, I have reproduced the
Oxford translation. I refer to the Nicomachean Ethics as NE
and to the Eudemian Ethics as EE.
298 w. F. R. Hardie
The same doctrine is stated in EE A 2. But, whereas
in the NE the emphasis is on the concern of political
science, statesmanship, with the human good conceived
as a single end, the EE speaks only of the planning by
the individual of his own life. 'Everyone who has the
power to live according to his own choice (TIpoa[pEOL<;)
should dwell on these points and set up for himself
some object for the good life to aim at, whether honour
or reputation or wealth or culture, by reference to which
he will do all that he does, since not to have one's
life organised in view of some end is a sign of great
folly. Now above all we must first define to ourselves
without hurry or carelessness in which of our posses-
sions the good life consists, and what for men are the
conditions of its attainment' (1214b6-14). Here, then,
we are told that lack of practical wisdom is shown in
a man's failure to plan and organise his life for the
attainment of a single end. Aristotle omits to say, but
says elsewhere, that lack of practical wisdom is shown
also in a man's preference for a bad or inadequate end,
say pleasure or money. We learn in NE VI 9 that the
man of practical wisdom has a true conception of the
end which is best for him as well as the capacity to
plan effectively for its realisation (1l41b31-33).
How far do men in fact plan their lives, as Aristotle
suggests they should, for the attainment of a single
end? As soon as we ask this question, we see that there
is a confusion in Aristotle's conception of the single
end. For the question confuses two questions: first,
how far do men plan their lives; and, secondly, so far
as they do, how far do they, in their plans, give a
central and dominating place to a single desired object,
money or fame or science? To both these questions
the answer that first suggests itself is that some men
do and some do not. Take the second question first.
It is exceptional for a life to be organised to achieve
the satisfaction of one ruling passion. If asked for ex-
The Final Good in Aristotle's Ethics 299
amples we might think of Disraeli's political ambition
or of Henry James' self-dedication to the art of the
novel. But exceptional genius is not incompatible with
a wide variety of interests. It seems plain that very few
men can be said, even roughly, to live their lives under
the domination of a single end. Consider now the
first question. How far do men plan their lives? Clearly
some do so who have no single dominant aim. It is
possible to have a plan based on priorities, or on equal
consideration, as between a number of objects. It is
even possible to plan not to plan, to resolve never to
cross bridges in advance. Hobbes remarked that there
is no 'finis ultimus, utmost aim, nor summum bonum,
greatest good, as is spoken of in the books of the old
moral philosophers .... Felicity is a continual progress
of the desire, from one object to another, the attaining
of the former being still but the way to the latter'
(Leviathan, ch. xi). But even such a progress may be
planned, although the plan may not be wise. Every
man has, and knows that he has, a number of inde-
pendent desires, i.e. desires which are not dependent
on other desires in the way in which desire for a means
is dependent on desire for an end. Every man is capable,
from time to time, of telling himself that, if he pur-
sues one particular object too ardently, he may lose or
imperil other objects also dear to him. So it may be
argued that every man capable, as all men are, of
reflection is, even if only occasionally and implicitly,
a planner of his own life.
vVe can now distinguish the two conceptions which
are confused or conflated in Aristotle's exposition of
the doctrine of the single end. One of them is the
conception of what might be called the inclusive end.
A man, reflecting on his various desires and interests,
notes that some mean more to him than others, that
some are more, some less, difficult and costly to achieve,
that the attainment of one may, in different degrees,
300 W. F. R. Hardie
promote or hinder the attainment of others. By such
reflection he is moved to plan to achieve at least his
most important objectives as fully as possible. The
following of such a plan is roughly what is sometimes
meant by the pursuit of happiness. The desire for hap-
piness, so understood, is the desire for the orderly and
harmonious gratification of desires. Aristotle sometimes,
when he speaks of the final end, seems to be fumbling
for the idea of an inclusive end, or comprehensive
plan, in this sense. Thus in NE I 2 he speaks of the
end of politics as 'embracing' other ends (1094b6-7).
The aim of a science which is 'architectonic' (1094a26-
27; ct. NE VI 8, 1141b24-26) is a second-order aim.
Again in NE I 7 he says that happiness must be 'most
desirable of all things, without being counted as one
good thing among others since, if it were so counted,
it would be made more desirable by the addition of
even the least of goods .. .' (1097b 16-20). Such con-
siderations ought to lead Aristotle to define happiness as
a secondary end, the full and harmonious achievement
of primary ends. This is what he ought to say. It is
not what he says. His explicit view, as opposcd to his
occasional insight, makes the supreme cnd not inclusive
but dominant, the object of one prime desire, philoso-
phy. This is so even when, as in NE I 7, he has in
mind that, prima facie, there is not only one final
end: ' ... if there are more than one, the most final
of these will be what we are seeking' (1097a30). Aris-
totle's mistake and confusion are implicit in his for-
mulation in EE A 2 of the question in which of our
possessions does the good life consist (1214b12-13).
For to put the question thus is to rule out the obvious
and correct reply; that the life which is best for a man
cannot lie in gaining only one of his objects at the
cost of losing all the rest. This would be too high a
price to pay even for philosophy.
The ambiguity which we have found in Aristotle's
conception of the final good shows itself also in his
attempt to use the notion of a 'function' (EPYOV)
which is 'peculiar' to man as a clue to the definition
of happiness. The notion of function cannot be de-
fended and should not be pressed, since a man is not
designed for a purpose. The notion which Aristotle in
fact uses is that of the specific nature of man, the
characteristics which primarily distinguish him from
other living things. This notion can be given a wider
interpretation which corresponds to the inclusive end
or a narrower interpretation which corresponds to the
dominant end. In NE I 7, seeking what is peculiar to
man (1097b33-34), Aristotle rejects first the life of
nutrition and growth and secondly the life of percep-
tion which is common to 'the horse, the ox and every
animal' (1098a2-3). What remains is 'an active life
of the element that has a rational principle' (1098a3-
4). This expression need not, as commentators point
out, be understood as excluding theoretical activity.
'Action' can be used in a wide sense, as in the Politics
VII 3 (1325bl6--23), to include contemplative think-
ing. But what the phrase specifies as the proper func-
tion of man is clearly wider than theoretical activity
and includes activities which manifest practical intel-
ligence and moral virtue. But the narrower conception
is suggested by a phrase used later in the same chapter.
'The good for man turns out to be the activity of soul
in accordance with virtue, and if there are more than
one virtue in accordance with the best and most com-
plete' (1098al6--18). The most complete virtue must
be theoretical wisdom, although this is not made clear
in NE I.
The doctrine that only in theoretical activity is man
really happy is stated and defended explicitly in X 7
and 8. Theoretical reason, the divine element in man,
ffl<)re than anything else is man (1l77b27-28, 1178a6--
7). 'It would be strange, then, if he were to choose
302 w. F. R. Hardie
not the life of his self but that of something else.
And what we said before will apply now; that which
is proper to each thing is by nature best and most
pleasant for each thing' (1l78a3-6). Man is truly hu-
man only when he is more than human, godlike. 'None
of the other animals is happy, since they in no way
share in contemplation' (1178b27-28). This statement
makes obvious the mistake involved in the conception
of the end as dominant rather than inclusive. It is no
doubt true that man is the only theoretical animal.
But the capacity of some men for theory is very small.
And theory is not the only activity in respect of which
man is rational as no other animal is rational. There
is no logic which leads from the principle that happi-
ness is to be found in a way of living which is common
and peculiar to men to the narrow view of the final
good as a dominant end. What is common and peculiar
to men is rationality in a general sense, not theoretical
insight which is a specialised way of being rational.
A man differs from other animals not primarily in be-
ing a natural metaphysician, but rather in being able
to plan his life consciously for the attainment of an
inclusive end.
The confusion between an end which is final because
it is inclusive and an end which is final because it is
supreme or dominant accounts for much that critics
have rightly found unsatisfactory in Aristotle's account
of the thought which leads to practical decisions. It is
connected with his failure to make explicit the fact
that practical thinking is not always or only the finding
of means to ends. Thought is needed also for the setting
up of an inclusive end. But, as we have seen, Aristotle
fails to make explicit the concept of an inclusive end.
This inadequacy both confuses his statement in NE
I 1 and 2 of the relation of politics to subordinate arts
and leads to his giving an incomplete account of delib-
eration.
I have represented Aristotle's doctrine as primarily
a doctrine about the individual's pursuit of his own
good, his own welfare (Eu5alflov[a). But something
should be said at this point about the relation between
the end of the individual and the 'greater and more
complete' end of the state. 'While it is worth while to
attain the end merely for one man, it is finer and more
godlike to attain it for a nation or for city-states' NE
I 2, 1094b7-1O). This does not mean more than it
says: if it is good that Smith should be happy, it is
even better that Brown and Robinson should be hap-
py too.
What makes it inevitable that planning for the at-
tainment of the good for man should be political is
the simple fact that a man needs and desires social
community with others. This is made clear in NE I 7
where Aristotle says that the final good must be suf-
ficient by itself. 'Now by self-sufficient we do not mean
that which is sufficient for a man by himself, for one
who lives a solitary life, but also for parents, children,
wife and in general for his friends and fellow-citizens,
since man is born for citizenship' (1097b7-11). That
individual end-seeking is primary, that the state exists
for its citizens, is stated in eh. 8 of NE VI, one of
the books common to both treatises. 'The man who
knows and concerns himself with his own interests is
thought to have practical wisdom, while politicians
are thought to be busybodies .... Yet perhaps one's
own good cannot exist without household manage-
ment, nor without a form of government' (1142a1-1O).
The family and the state, and other forms of associa-
tion as well, are necessary for the full realisation of
any man's capacity for living well.
The statesman aims, to speak roughly, at the greatest
happiness of the greatest number. He finds his own
happiness in bringing about the happiness of others
(NE X 7, 1177b14), especially, if Aristotle is right, the
304 W. F. R. Hardie
happiness of those capable of theoretical activity.
Speaking in terms of the end as dominant Aristotle,
in NE VI 13, sets a limit to the authority of political
wisdom. 'But again it is not supreme over philosophic
wisdom, i.e. over the superior part of us, any more
than the art of medicine is over health; for it does not
use it but provides for its coming into being; it issues
orders, then, for its sake but not to it' (1145a6--9).
This suggestion that science and philosophy are insu-
lated in principle from political interference cannot be
accepted. The statesman promotes science but also
uses it, and may have to restrict the resources to be
made available for it. If the secondary and inclusive
end is the harmonisation and integration of primary
ends, no primary end can be sacrosanct. But, even if
Aristotle had held consistently the extravagant view
that theoretical activity is desired only for itself and
is the only end desired for itself, he would not have
been right to conclude that there could be no occasion
for the political regulation of theoretical studies. For
the unrestricted pursuit of philosophy might hinder
measures needed to make an environment in which
philosophy could flourish. It might be necessary to
order an astronomer to leave his observatory, or a
philosopher his school, in order that they should play
their parts in the state. Similarly the individual who
plans his life so as to give as large a place as possible
to a single supremely desired activity must be ready to
restrain, not only desires which conflict with his ruling
passion, but the ruling passion itself when it is mani-
fested in ways which would frustrate its own object.
In NE I 1 and 2 Aristotle expounds the doctrine that
statesmanship has authority over the arts and sciences
which fall under it, are subordinate to it. An art, A,
is under another art, B, if there is a relation of means
to end between A and B. If A is a productive art, like
bridle-making, its product may be used by a superior
art, riding. Riding is not a productive activity, but it
falls under generalship in so far as generals use cavalry,
and generalship in turn falls under the art of the states-
man, the art which is in the highest degree architec-
tonic (1094a27; cf. VI 8, 1141b23-25). Thus the man
of practical wisdom, the statesman or legislator, is com-
pared by Aristotle to a foreman, or clerk of the works,
in charge of technicians and workmen of various kinds,
all engaged in building an observatory to enable the
man of theoretical wisdom to contemplate the starry
heavens. In the Magna Moralia the function of prac-
tical wisdom is said to be like that of a steward whose
business it is so to arrange. things that his master has
leisure for his high vocation (A 34, 1198b12-17). Per-
haps the closest parallel to the function of the states-
man as conceived by Aristotle is the office of the Bursar
in a college at Oxford or Cambridge.
This account of statesmanship as aiming at the ex-
ercise of theoretical wisdom by those capable of it is
an extreme expression of the conception of the end as
dominant and not inclusive. The account, as it stands,
is a gross over-simplification of the facts. \Vhen he
speaks of a subordinate art as pursued 'for the sake
of' a superordinate or architectonic art (1094a15-16),
Aristotle should make explicit the fact that the subor-
dinate activity, in addition to serving other objects,
may be pursued for its own sake. Riding, for example,
has non-~ilitary uses and can be a source of enjoyment.
Again two arts, or two kinds of activity, may each be
subordinate, in Aristotle's sense, to the other. Riders
use bridles, and bridle-makers may ride to their work.
The engineer uses techniques invented by the mathe-
matician, but also promotes the wealth and leisure in
which pure science can flourish. Aristotle does not fail
to see and mention the fact that an object may be
desired both independently for itself and dependently
for its effects (NE I 6, 1097a30-34). He was aware also
306 W. F. R. Hardie
that theoretical activity is not the only kind of activity
which is independently desired. But he evidently
thought that an activity which was never desired ex-
cept for itself would be intrinsically desirable in a
higher degree than an activity which, in addition to
being desired for itself, was also useful. It is, so to say,
beneath the dignity of the most godlike activities that
they should be useful. Aristotle is led in this way, and
also by other routes, to give a narrow and exclusive
account of the final good, to conceive of the supreme
end as dominant and not inclusive.
Aristotle describes deliberation, the thinking of the
wise man, as a process which starts from the concep-
tion of an end and works back, in a direction which
reverses the order of causality, to the discovery of a
means. Men do not, he asserts, deliberate about ends.
'They assume the end and consider how and by what
means it is to be attained; and if it seems to be pro-
duced by several means they consider by which it is
most easily and best produced, while, if it is achieved
by one only, they consider how it will be achieved by
this and by what means this will be achieved, till they
come to the first cause, which in the order of discovery
is last' (NE III 3, 1112b15-20). Such an investigation
is compared to the method of discovering by analysis
the solution of a geometrical problem. Again in VI 2
practical wisdom is said to be shown in finding means
to a good end. 'For the syllogisms which deal with acts
to be done are things which involve a starting-point,
viz. "since the end, i.e. what is best, is of such and
such a nature" .. .' (1l44a31-33).
This is Aristotle's official account of deliberation.
But here again, as in his account of the relation be-
tween political science and subordinate sciences, a too
narrow and rigid doctrine is to some extent corrected
elsewhere, although not explicitly, by the recognition
of facts which do not fit into the prescribed pattern.
Joseph, in Essays in Ancient and Modern Philosophy,
pointed out that the process of deciding between al-
ternative means, by considering which is easiest and
best, involves deliberation which is not comparable to
the geometer's search (pp. 180-81). But he remarks
that Aristotle does not 'appear to see' this. What the
passage suggests is that the agent may have to consider
the intrinsic goodness, or badness, of the proposed
means as well as its effectiveness in promoting a good
end. A less incidental admission that there is more in
deliberation than the finding of means is involved in
Aristotle's account of 'mixed actions' in NE III 1. Aris-
totle recognises that, if the means are discreditable,
the end may not be important enough to justify them.
'To endure the greatest indignities for no noble end
or for a trifling end is the mark of an inferior person'
(11lOa22-23). 'It is difficult sometimes to determine
what should be chosen at what cost, and what should
be endured in return for what gain' (1110a29-30).
Alcmaeon's decision to kill his mother, on his father's
instruction, rather than face death himself is given as
an example of a patently wrong answer to a question
of this kind. luis kind of deliberation is clearly not
the regressive or analytic discovery of means to a pre-
conceived end. It is rather the determination of an
ideal pattern of behaviour, a system of priorities, from
which the agent is not prepared to depart. It is what
we described earlier as the setting up of an inclusive
end. It is a kind of practical thinking which Aristotle
cannot have had in his mind when he asserted in NE
III 3 that 'we deliberate not about ends but about
means' (1l12bl1-12).
I have argued that Aristotle's doctrine of the final
human good is vitiated by his representation of it as
dominant rather than inclusive, and that this mistake
underlies his too narrow account of practical thinking
as the search for means. But to say that the final good
308 W. F. R. Hardie
is inclusive is not to deny that within it there are
certain dominant ends corresponding to the major in-
terests of developed human nature. One of th~se major
interests is the interest in theoretical sciences. Of these,
according to Aristotle, there are three; theology or first
philosophy, mathematics and physics (Metaphysics E
1, 1026a18-19, cf. NE VI 8, 1142al6-18). His account
of contemplation in the Ethics, based on the doctrine
of reason as the divine or godlike element in man (NE
X 7, 1177al3-17; 8, 1178a20-23), exalts the first and
makes only casual mention of the other two. Elsewhere,
in the De Partibus Animalium I 5, he admits that
physics has attractions which compensate for the rel-
atively low status of the objects studied. 'The scanty
conceptions to which we can attain of celestial things
give us, from their excellence, more pleasure than all
our knowledge of the world in which we live; just as
a half-glimpse of persons that we love is more delight-
ful than a leisurely view of other things, whatever their
number and dimensions. On the other hand, in cer-
titude and in completeness our knowledge of terrestrial
things has the advantage. Moreover their greater near-
ness and affinity to us balances somewhat the loftier
interest of the heavenly things that are the object of
the higher philosophy' (644b31-645a4).
I cannot here discuss the theological doctrines which
led Aristotle to place 'the higher philosophy' on the
summit of human felicity. But there is an aspect of
his account of the theoretic life which has an immediate
connection with my main topic. He remarks in NE
VII 14 that 'there is not only an activity of movement
but an activity of immobility, and pleasure is found
more in rest than in movement' (1154b26-28). This
doctrine that there is no 'movement' in theoretical
contemplation, and the implication that its immobility
is a mark of its excellence, is determined primarily by
Aristotle's conception of the divine nature. The latest
The Final Good in Aristotle's Etllics 309
commentators on the NE, Gauthier and Jolif, say, with
justification, that he here excludes discovery from the
contemplative life. 'On pourrait meme dire que !'ideal,
pour Ie contemplatif aristotelicien-et cet ideal Ie Dieu
d' Aristote Ie realise-ce serait de ne jamais etudier et
de ne jamais decouvrir .. .' (855-56). In NE X 7 we
are told that 'philosophy is thought to offer pleasures
marvellous for their purity and their enduringness' and
that it is 'reasonable to suppose that those who know
will pass their time more pleasantly than those who
enquire' (1177a25-27). It is not reasonable at all. It
is a startling paradox. I shall now suggest that Aristotle's
apparent readiness to accept this paradox, like his con-
fusion between the dominant and the inclusive end, is
to be explained, at least in part, by his failure to give
any explicit or adequate analysis of the concept of end
and means.
Aristotle states in NE I that an end may be either
an activity or the product of an activity. 'But a certain
difference is to be found among ends; some are activi-
ties, others are products apart from the activities that
produce them. Where there are ends apart from the
actions, it is the nature of the products to be better
than the activities' (1094a 3-(6). The suggestion here is
that, when an activity leads to a desired result, as
medicine produces health or shipbuilding a ship or
enquiry knowledge, the end-seeking activity is not it-
self desired. As he says (untruly) in the Metaphysics,
'of the actIons which have a limit none is an end' (e 6,
1048b18). But an activity which aims at producing a
result may be an object either of aversion or of indif-
ference or of a positive desire which may be less or
greater than the desire for its product. It is necessary
to distinguish between 'end' in the sense of a result
intended and planned and 'end' in the sense of a result,
or expected result, which, in addition to being in-
tended and planned, is also desired for itself while
310 w. F. R. Hardie
the process of reaching it is not. It is true that
travel may be unattractive, but it may also be more
attr~ctive that arrival. A golfer plays to win. But,
if he loses, he does not feel that his day has
been wasted, that he has laboured in vain, as he
would if his only object in playing were to win a prize
or to mortify his opponent or just to win. Doing cross-
word puzzles may be a waste of time, but what makes
it a waste of time is not the fact that we rarely get one
out. It would be a greater waste of time if we never
failed to finish them. In short, the fact that an activity
is progressive towards a planned result leaves quite
open the question whether it is the process or the result
which is desired, and, if both, which primarily. If Aris-
totle had seen and said this, he might have found it
more difficult than he does to suggest that the pleasures
of discovery are not an essential element in science as
a major human interest. Philosophy would be less at-
tractive than it is if only results mattered. God's per-
fection requires that his thinking should be unpro-
gressive. But men, who fall short of perfect simplicity,
need, to make them happy, the pleasures of solving
problems and of learning something new and of being
surprised. For them the best way of life leads, in the
words of Meredith,
'through widening chambers of surprise to where
throbs rapture near an end that aye recedes'.
We have seen that Aristotle's doctrine of the final
human good needs clarification in terms of a distinction
between an end which is inclusive, a plan of life, and
an end which is dominant as the satisfaction of theo-
retical curiosity may be dominant in the life of a phi-
losopher. No man has only one interest. Hence an end
which is to function as a target, as a criterion for decid-
ing what to do and how to live, must be inclusive. But
some men have ruling passions. Hence some inclusive
ends will include a dominant end. I shall now try to
look more closely at these Aristotelian notions, and to
suggest some estimate of their relevance and value in
moral philosophy.
It will be best to face at once and consider a natural
and common criticism of Aristotle; the criticism that
his virtuous man is not moral at all but a calculating
egoist whose guiding principle is not duty but prudence,
Bishop Butler's 'cool self-love'. Aristotle is in good
company as claiming that rationality is what makes a
man ideally good. But his considered view, apart from
incidental insights, admits, it is said, only the rational-
ity of prudent self-interest and not the rationality of
moral principle. Thus Professor D. J. Allan, in The
Philosophy of Aristotle, tells us that Aristotle "takes
little or no account of the motive of moral obligation"
and that "self-interest, more or less enlightened, is as-
sumed to be the motive of all conduct and choice"
(p. 189). Similarly the late Professor Field, a fair and
sympathetic critic of Aristotle, remarked that, whereas
morality is 'essentially unselfish', Aristotle's idea of the
final end or good makes morality 'ultimately selfish'
(Moral Theory, pp. 109, lll).
When a man is described as selfish what is meant
primarily is that he is moved to act, more often and
more strongly than most men, by desires which are
selfish. The word 'selfish' is also applied to a disposi-
tion so to plan one's life as to give a larger place than
is usual or right to the gratification of selfish desires.
But what is it for a desire to be selfish? Professor Broad,
in his essay 'Egoism as a theory of human motives' (in
Ethics and the History of Philosophy), makes an im-
portant distinction between two main kinds of 'self-
regarding' desires. There are first desires which are
'self-confined', which a man could have even if he were
alone in the world, e.g. desires for certain experiences,
the desire to preserve his own life, the desire to feel
L
312 w. F. R. Hardie
respect for himself. Secondly there are self-regarding
desires which nevertheless presuppose that a man is
not alone in the world, e.g. desires to own property,
to assert or display oneself, to inspire affection. Broad
further points out that desires which are 'other-regard-
ing' may also be 'self-referential', e.g. desires for the
welfare of one's own family, friends, school, college,
club, nation.
A man might perhaps be called selfish if his other-
regarding motives were conspicuously and exclusively
self-referential, if he showed no interest in the welfare
of anyone with whom he was not personally connected.
But usually 'selfish' refers to the prominence of self-
regarding motives, and different kinds of selfishness
correspond to different self-regarding desires. The word,
being pejorative, is more readily applied to the less
reputable of the self-regarding desires. Thus a man
strongly addicted to the pursuit of his own pleasures
might be called selfish even if his other-regarding mo-
tives were not conspicuously weak. A man whose rul-
ing passion was science or music would not naturally
be described as selfish unless to convey that there was
in him a reprehensible absence or failure of other-
regarding motives, as shown, say, by his neglect of his
family or of his pupils.
The classification of desires which I have quoted
from Broad assumes that their nature is correctly rep-
resented by what we ordinarily think and say about
them. Prima facie some of our desires are self-regarding;
and, of the other-regarding desires, some are and some
are not self-referential. But there have been philoso-
phers who have questioned or denied the reality of
these apparent differences. One doctrine, psychological
egoism, asserts in its most extreme form that the only
possible objects of a man's first-order independent de-
sires are experiences, occurrent states of his own con-
sciousness. Thus my desire to be liked is really a desire
to know that I am liked; and my desire that my chil-
dren should be happy when I am dead is really a desire
for my present expectation that they will be happy.
The obvious criticism of this doctrine is that it is pre-
posterous and self-defeating: I must first desire populari-
ty and the happiness of my children if I am to find
gratifying my thought that I am popular and that my
children will be happy. To most of us it seems that
introspective self-scrutiny supports the validity of this
dialectic. We can, therefore, reject psychological ego-
ism. A fortiori we can reject psychological hedonism
which asserts that the only experiences which can be
independently desired are pleasures, feelings of enjoy-
ment. This further doctrine was stated as follows by
the late Professor Prichard. 'For the enjoyment of
something which we enjoy, e.g. the enjoyment of seeing
a beautiful landscape, is related to the thing we enjoy,
not as a quality but as an effect, being something ex-
cited by the thing we enjoy, so that, if it be said that
we desire some enjoyment for its own sake, the correct
statement must be that we desire the experience, e.g.
the seeing of some beautiful landscape, for the sake of
the feeling of enjoyment which we think it will cause,
this feeling being really what we are desiring for its
own sake' (Moral Obligation, p. 116). Surely most of
us would be inclined to say that we can desire for its
own sake 'the seeing of some beautiful landscape' and
that we do not detect a distinct 'feeling of enjoyment'.
Was Aristotle a psychological egoist or a psychologi-
cal hedonist? A crisp answer would have been possible
only if Aristotle had explicitly formulated these doc-
trines as I have defined them. So far as I can see, he
did not do so even in his long, but not always lucid,
treatment of friendship and self-love in NE IX. This
being so, he cannot be classed as a psychological egoist
in respect of his account of first-order desires. When
Aristotle confronts the fact of altruism, he does not
314 w. F. R. Hardie
refuse to accept benevolent desires at their face value
(NE VIII 2, llSSb31; 3, llS6b9-1O; 7, llS9a8-12).
But he shows acuteness in detecting self-referential
elements in benevolence. Thus he compares the feelings
of benefactors to beneficiaries with those of parents
for their children and of artists for their creations. 'For
that which they have treated well is their handiwork,
and therefore they love this more than the handiwork
does its maker' (NE IX 7, 1167b31-1168aS).
The nearest approach which Aristotle makes to the
formulation of psychological hedonism is, perhaps, in
the following passage in NE II 3. 'There being three
objects of choice and three of avoidance, the noble,
the advantageous, the pleasant, and their contraries,
the base, the injurious, the painful, about all of these
the good man tends to go right and the bad man to go
wrong, and especially about pleasure; for this is com-
mon to the animals, and also it accompanies all ob-
jects of choice; for even the noble and the advantage-
ous appear pleasant' (l104b30-110Sal). But there are
passages in his discussion of pleasure in NE X which
show that, even if he had accepted psychological ego-
ism, he would not have accepted psychological hedo-
nism. 'And there are many things we should be keen
about even if they brought no pleasure, e.g. seeing,
remembering, knowing, possessing the virtues. If pleas-
ures necessarily do accompany these, that makes no
odds; we should choose them even if no pleasure re-
sulted' (1174a4-8). This reads like a direct repudia-
tion of the doctrine in my quotation from Prichard. In
NE X 4 he asks, without answering, the question
whether we choose activity for the sake of the attendant
pleasure or vice versa (1l7Sal8-21). The answer which
his doctrine requires is surely that neither alternative
can be accepted, since both the activity and the at-
tendant pleasure are desired for their own sake. But it
•
is open to question whether, when we speak of a state
or activity, such as 'the seeing of some beautiful land-
scape', as pleasant, we are referring to a feeling distinct
from the state or activity itself.
The charge against Aristotle that his morality is a
moraliy of self-interest is directed primarily against his
doctrine of the final good, the doctrine which I have
interpreted as a conflation of the distinct notions of
the 'inclusive end' and the 'dominant end'. But the
critic may also wish to suggest that Aristotle overstates
the efficacy of self-regarding desires in the determina-
tion of human conduct. To this the first answer might
well be that it is not easy to overstate their efficacy.
The term 'self-regarding' applies, as we have seen, to
a wide variety of motives; and there is a 'self-referential'
factor in the most potent of the other-regarding mo-
tives. Altruism which is pure, not in any way self-re-
garding or self-referential, is a rarity. The facts support
the assertion that man is a selfish animal. But the
criticism can be met directly. Aristotle does not ignore
other-regarding motives. Thus, while he points out that
the philosopher, unlike those who exercise practical
virtue, does not need other men 'towards whom and
with whom he shall act', he admits that the pleasures
of philosophy are enhanced by interest in the work of
colleagues. 'He perhaps does better if he has fellow-
workers, but still he is the most self-sufficient' (NE
X 7, 11 77a27-b 1 ). When, in the EE, Aristotle speaks
of philosophy as the service of God, he seems to im-
ply that the love of wisdom is not directed merely
to the lover's own conscious states (1249b20). Again,
in NE IX 8, he can attribute to the 'lover of self' con-
duct which is, in the highest degree, altruistic and
self-sacrificing. 'For reason always chooses what is best
for itself, and the good man obeys his reason. It is true
of the good man too that he does many acts for the
sake of his friends and his country, and, if necessary,
316 W. F. R. Hardie
dies for them; for he will throwaway both wealth and
honours and in general the goods that are objects of
competition, gaining for himself nobility (TO KuMv);
since he would prefer a short period of intense pleasure
to a long one of mild enjoyment, a twelvemonth of
noble life to many years of humdrum existence, and
one great and noble action to many trivial ones. Now
those who die for others doubtless attain this result;
it is, therefore, a great prize that they choose for them-
selves' (1169aI7-26).
But it is not enough, if we are to do justice to the
criticism that Aristotle makes morality selfish, to quote
this passage, or the passage in NE I 10 where Aristotle
speaks of the shining beauty of the virtue shown in
bearing disasters which impair happiness (1100b30-33).
Such passages, it may be said, show Aristotle's moral
sensibility and moral insight. But the question can still
be asked whether their commendation of the ultimate
self-sacrifice, and of endurance in suffering, is consistent
with Aristotle's doctrine of the final human good. Per-
haps he is speaking more consistently with his own
considered views when, again in NE IX 8, he makes
the suggestion (or is it a joke?) that a man may show
the finest self-sacrifice, the truest love, by surrendering
to his friend the opportunity of virtuous action (1169a
33-34). Perhaps Aristotle's commendation of the sur-
render, in a noble cause, of life itself needs to be
qualified, from his own point of view, as it was qualified
by Oscar Wilde:
And yet, and yet
Those Christs that die upon the barricades,
God knows it I am with them, in some ways.
To this question I now turn. My answer must and can
be brief.
\Ve have found two main elements in Aristotle's
doctrine of the final good for man. There is, first, the
suggestion that, as he says in EE A 2, it is a sign of
'great folly' not to 'have one's life organised in view
of some end'. Perhaps it would be better to say that
it is impossible not to live according to some plan, and
that it is folly not to try to make the plan a good one.
The inevitability of a plan arises from the fact that a
man both has, and knows that he has, a number of
desires and interests which can be adopted as motives
either casually and indiscriminately or in accordance
with priorities determined by the aim of living the
kind of life which he thinks proper for a man like
himself. But in an agent naturally reflective the omis-
sion to make such a plan is not completely undesigned:
the minimal plan is a plan not to plan. To this side of
Aristotle's doctrine I have applied the term 'inclusive
end', inclusive because there is no desire or interest
which should not be regarded as a candidate, however
unpromising, for a place in the pattern of life. Wisdom
finds a place even for folly. The second element which
we have found in Aristotle's doctrine is his own answer
to the question what plan will be followed by a man
who is most fully a man, as high as a man can get on
the scale from beast to god. Aristotle's answer is that
such a man will make theoretical knowledge, his most
godlike attribute, his main object. At a lower level, as
a man among men, he will find a place for the happi-
ness which comes from being a citizen, from marriage
and from the society of those who share his interests.
I have called this the doctrine of the dominant end.
The question whether Aristotle's doctrine of the final
good can be reconciled with the morality of altruism
and self-sacrifice must be asked with reference both to
the inclusive end and to the dominant end.
To say that a man acts, or fails to act, with a view to
an inclusive end is to say nothing at all about the com-
parative degrees of importance which he will ascribe
to his various aims. His devotion to his own good, in
318 w. F. R. Hardie
the sense of his inclusive end, need not require him
to prefer self-regarding desires to other-regarding de-
sires, or one kind of self-regarding desire to another.
All desires have to be considered impartially as can-
didates for places in the inclusive plan. To aim at a
long life in which pleasures, so far as possible, are en-
joyed and pains avoided it is a possible plan, but not
the only possible plan. That a man seeks an inclusive
end leaves open the question whether he is an egoist
or an altruist, selfish or unselfish in the popular sense.2
While a man seeking his inclusive end need not be
selfish, he can be described as self-centred in at least
three different ways. First and trivially his desire to
follow his inclusive plan is his own desire; it is self-
owned. Secondly, a man can think of a plan as being
for his own good only if he thinks about himself,
thinks of himself as the one owner of many desires.
His second-order desire for his own good is self-reflec-
tive. Thirdly, this second-order desire, being a desire
about desires, an interest in interests, can be gratified
only through the gratification of his first-order desires.
Even the martyr plans to do what he wants to do.
We can express this by saying that the pursuit of the
final good is self-indulging as well as self-reflective. But
'self-indulgence' as applied to a way of life in which
pleasures may be despised and safety put last carries
no pejorative sense. That action in pursuit of an in-
clusive end is self-centred in these ways does not mean
21 owe this point, and less directly much else in my discussion
of the criticism of Aristotle's ethical system as egoistic, to Profes-
sor C. A. Campbell's British Academy Lecture (1948), "Moral
Intuition and the Principle of Self-Realisation" (especially pp.
17-25). Professor Campbell's lecture discusses the ethical theory
of T. H. Green and F. H. Bradley, and 1 do not know whether he
would think of his arguments as being relevant to the interpreta·
tion of Aristotle. But 1 have found his defence of 'self-realisation'
as a moral principle helpful in my attempt to separate the strands
of thought in Aristotle's doctrine of the final good.
that the agent is self-regarding or self-seeking in any
sense inconsistent with the most heroic or saintly self-
sacrifice.
To the question whether the pursuit of the human
good, understood in terms of Aristotle's conception of
the dominant end, can be reconciled with the morality
of altruism, and in particular the extreme altruism of
the man who gives his life for his friends or his country,
a different answer must be given. Here reconciliation
is not possible. In order to see this it is necessary only
to reflect on Aristotle's definition in NE I 7 of the
dominant end, which he calls happiness, and to com-
pare this definition with what is said about the self-love
of the man who nobly gives up his own life. 'Human
good turns out to be activity of soul in accordance
with virtue, and if there are more than one virtue, in
accordance with the best and most complete. But we
must add "in a complete life." For one swallow does
not make a summer nor does one day; and so too one
day or a short time does not make a man blessed and
happy' (1098al6-20). How then can the man who, to
gain nobility ('to KaMv) for himself, gives his life for
his friends or his country be said to achieve happiness?
Aristotle's answer, as we have seen, is that such a man
prefers 'a short period of intense pleasure to a long
one of mild enjoyment, a twelvemonth of noble life to
many years of humdrum existence, and one great and
noble action to many trivial ones' (1169a22-25). But
the scales are being loaded. For why should it be sup-
posed that the man who declines to live the final, if
crowded, hour of glorious life will survive to gain only
'mild' enjoyments and a 'humdrum' or 'trivial' exist-
ence? If such existence is, or seems, humdrum because
the 'intense pleasure' of self-sacrifice has been missed,
then Aristotle's thought here is circular and self-stulti-
fying. The intensity of the brief encounter, it is sug-
gested, is such that by contrast the remainder of life
L*
320 w. F. R. Hardie
would be humdrum. But, unless the alternative would
be humdrum in its own right, the encounter would
not be intense enough to compensate for the curtail-
ment of life and happiness. A 'complete life' either is, or
is not, a necessary condition of happiness. Aristotle as a
theorist cannot justify the admiration which, as a man,
he no doubt feels for the 'one great and noble action',
Confronted with the facts he would have to admit that
the man who, whether by good fortune or design, sur-
vives a revolution or a war may live to experience in-
tense enjoyments and to perform activities in accord-
ance with the best and most complete virtue. He may
become a professor of philosophy or at least a prime
minister. We must conclude, therefore, that Professor
Field was right: the doctrine of the good for man, as
developed by Aristotle in his account of the dominant
end, does make morality 'ultimately selfish' (Moral
Theory, pp. 109, HI).
Aristotle offers us in his Ethics a handbook on how
to be happy though human. To some it may seem that
a treatise on conduct with an aim so practical and so
prudential can do little to clarify the concepts with
which moral philosophy is mainly concerned, the con-
cepts of duty and of moral worth. 'He takes little or
no account', Professor Allan tells us, 'of the motive of
moral obligation' (The Philosophy of Aristotle, p.
189), Perhaps not. The topic is too large for a con-
cluding paragraph. Certainly most men feel moral ob-
ligations which cannot be subsumed under the obliga-
tion, if there is one, to pursue their own happiness by
planning for the orderly satisfaction of their self-regard-
ing desires. But 'obligation' and 'duty' are words with
many meanings, meanings variously related to the con-
cept of moral worth. Perhaps Aristotle is not wrong,
as he is not alone, in connecting the concept of moral
worth with the fact that man is not just the plaything
of circumstance and his own irrational nature but also
the responsible planner of his own life. This aspect of
Aristotle's teaching is what I have called his doctrine
of the 'inclusive end', and I have argued that there
is no necessity for the doctrine to be specified and
developed as a recommendation of calculated egoism.
Aristotle himself, as we have seen, does not adhere
consistently to his own exaltation of self·regarding aims.
He is, indeed, always ready to notice facts which are
awkward for his own theories. Thus in NE I 10 he
recognises that the actual achievement of happiness,
virtuous activity, is largely outside a man's control. 'A
multitude of great events if they tum out well will
make life happier ... while if they tum out ill they
crush and maim happiness; for they both bring pain
with them and hinder many activities' (1l00b2S-30).
He adds that, even when disaster strikes, 'nobility
shines through, when a man bears with resignation
many great misfortunes, not tbrough insensibility to
pain but through nobility and greatness of soul'
(1l00b30-33). 'The man who is truly good and wise',
he goes on to say, 'bears all the chances of life be-
comingly and always makes the best of circumstances
as a good shoemaker makes the best shoes out of the
hides that are given to him' (1l00b3S-1101aS). The
suggestion of this passage is that a man's worth lies
not in his actual achievement, which may be frustrated
by factors outside his own control, but in his striving
towards achievement. In an earlier chapter (5) of NE
I he speaks of the good as something which 'we divine
to be proper to a man and not easily taken from him'
(1095b25-26). Aristotle's doctrine of the final good is
a doctrine about what is 'proper' to a man, the power
to reflect on his own abilities and desires and to con-
ceive and choose for himself a satisfactory way of life.
What 'cannot be taken from him' is his power to keep
on trying to live up to such a conception. Self-respect,
322 ~.17.lt. IIardie
thus interpreted, is a principle of duty. If moral philoso-
phy must seek one comprehensive principle of duty,
what other principle has a stronger claim to be regarded
as the principle of duty?
ARISTOTLE ON PLEASURE
J. O. URMSON
Aristotle's most mature and careful account of pleas-

ure or enjoyment-he uses the noun l'JoovfJ and its
cognates and the verb XalPELV without any apparent
discrimination-is to be found in Book X of the Nico-
machean Ethics (1174a13 ff.). I propose to summarize
this very acute account and then to discuss some of
the problems arising out of it.
Like sight, Aristotle holds, pleasure is not a process
(KlVTJOL<;) but an activity. As such, it is complete at
any time. When you build a temple, you begin, con-
tinue, and perhaps finish it, and until you have finished
it, you have not built a temple; but if at any moment
you see or are enjoying something, then you have seen
or enjoyed it. Sight and enjoyment cannot be left
half-finished.
Perception, of which sight is an example, thought,
and contemplation have objects. When the perception
or thought is high-grade and its object is worthwhile,
then the perception or thought or contemplation is
enjoyable (pleasant). The higher grade the perception
and the more valuable its object, the more pleasant
and more perfect the activity. But the excellence of
the perception or thought and its object make the
activity perfect in a different way from that in which
enjoyment perfects it; the former are constitutive of
its perfection while the enjoyment is a manifestation
of it as the grace of youth is a manifestation of physical
324 J. O. Urmson
prime. We might say that it adds zest to the activity.
When, through tiredness or illness, the activity is im-
paired, so is the enjoyment of it. Again, novelty com-
mands effortless attention and therefore enjoyment,
but when novelty wears off, the attention wanders and
enjoyment wanes. So activity and life itself are bound
up with enjoyment (pleasure); there is no enjoyment
without activity enjoyed and enjoyment is the mark
of activity at peak performance.
Different activities (1175a21) are differently en-
joyable. Just as perception and thought are different
species of activity, so the pleasures of perception are
different in species from the pleasures of thought.
Every activity has its own 'proper' (otKElO) pleasure;
one could not chance to get the pleasure of, say, read-
ing poetry from stamp collecting. The enjoyment
proper to an activity promotes that activity, whereas
the enjoyment of something else impedes it. If we are
doing two things at once, perhaps arguing and listen-
ing to music, the more enjoyable gets in the way of
the other. So when we enjoy something very much, we
do not do anything else at the same time; we only eat
sweets at the theater when we are not enjoying the
plot very much. So, if two enjoyments can conflict,
they must be distinguishable. Just as the enjoyment
of something else maims the enjoyment of an activity,
so does the 'proper pain', that is, the dislike of that
activity. \Ve are disinclined to draw or reason if we
find it disagreeable.
As activities differ in worth (1175b23) so do their
enjoyments (proper pleasures). Even desires for the
worthwhile or the base are praised and blamed, al-
though the appetite is not so intimately bound up with
the activity of pursuing them; how much more so are
the enjoyments of these activities which are scarcely
distinguishable from the activities themselves.
This very acute analysis of the enjoyment of activi-
Aristotle on Pleasure 325
ties can be supplemented from elsewhere in the NE.
As a prologue to his discussion of temperance in Book
III, Aristotle says that the pleasure of the soul need
be distinguished from the bodily pleasures, instancing
the love of honour and love of learning. He adds that
the lover of each of these "enjoys what he loves, his
body being in no way affected, but rather his intel-
ligence." This remark puzzled the commentator As-
pas ius to the point of exasperation: "\Vhat does he
mean," he protests, "when he says that the enjoyment
of lovers of learning or honour involves a condition
of the intelligence? For enjoyment and the pleasures
are not in the intelligence but in the affective
(rra8rrrlKcp) part of the soul." Aspasius never under-
stands Aristotle's view that enjoyment of learning is
exhibited in the effortless concentration of the intel-
ligence on its problems rather than in getting some
feeling as a result or concomitant of one's study.
It has, indeed, been commonplace in the history of
philosophy to regard doing something for the pleasure
of it as doing something for the sake of gaining some
sensation or feeling named 'pleasure'. Psychological
hedonists have typically assumed this analysis and
concluded therefrom that mankind has only one goal
of action which they call pleasure. Sometimes their
opponents have seemed to think that to reject psycho-
logical hedonism, they must say that we do not do
things for pleasure but for their own sake, thus con-
senting to the analysis; they take the pleasure of an
activity to be a feeling resulting from it but claim
that the pleasure is a mere bonus resulting from gain-
ing what one wanted rather than the end for the sake
of which the activity was performed. It is not surpris-
ing, perhaps, that the fascination of this analysis should
have made Aspasius and many of his successors among
the commentators on Aristotle's ethics unable to see
that Aristotle has no inclination to accept it. The en-
326 J. O. Urmson
joyment of an activity is for him not a result of it but
something barely distinguishable from the activity; it
is more like the effortless zest with which the activity
is performed than a result or concomitant of it. His
account is, indeed, very similar to the account of en-
joyment given by C. Ryle in his Concept of Mind
(London, 1949), a fact which should surprise nobody.
Ironically enough, we shall find that Aristotle is at
his weakest in discussing just those aspects of pleasure
to which the 'traditional' analysis is most plausibly
applicable. He is in the end most interested in the
enjoyment of intellectual activities, and it is here that
his view is most obviously attractive.
But what of the bodily pleasures? It is clear that in
Book X Aristotle wishes to give the same general ac-
count of them as of intellectual enjoyment. Thus,
bodily pleasures (0UJ!lan Ka[ TjE>ova:[) are construed
as the enjoyment of activities of sense-perception. Aris-
totle gives the same account of them in general in
Book III. He makes it clear that he is thinking of the
enjoyment of colors and shapes and pictures (sight),
of music (hearing), and smells. But the objects of
taste and touch have a special position among the ob-
jects of bodily pleasure, so that usually when Aristotle
speaks of bodily pleasures (e.g. at 1104b5, 1153b33,
1154a8, 1154a1O, 1154a26, 1177a7) he seems to have
them alone in mind. They also are the only objects
of bodily pleasure which fall within the sphere of tem-
perance.
Aristotle, we have said, allots a special position to
the objects of taste and touch; at times he was little
inclined to distinguish between them, as when he
treats XU!l0C; as a amov in De Anima 414b6-11. But
in a remarkable passage (NE 1118a26-33) he makes
it clear that touch is the true sphere of temperance:
"They [the intemperate] seem to make little use of
taste. For to taste belongs the judgment of flavors, as
is done by wine-tasters and cooks. But it is not these
that people enjoy, intemperate ones at any rate, but
the experience which is all a matter of touch, whether
in the field of food or drink or sex." Here, it seems
to me that Aristotle is on the brink of making a dig.
tinction far more important than his distinction of
the enjoyment of perceptual and intellectual activities.
To this further distinction, that he never quite makes
clear even, I think, to himself, we must now tum.
The account of the enjoyment of activity, whether
it involves the exercise of thought or the senses, given
in Book X seems to me to be remarkably acute; it is
surely, at least in general, correct. But Aristotle appears
to offer it as an account of pleasure in general and as
such it raises problems, of which we shall immediately
raise the most acute.
One is inclined to say that it is a correct account of
the enjoyment of an activity in contrast to those ac-
counts of it which make such enjoyment the obtaining
of pleasant experiences which accrue from the activity,
accounts such as, apparently, Aspasius and Prichard in
his essay in this volume [pp. 241-60] would be in-
clined to give. But surely there is such a thing as gain-
ing pleasant experiences as a result of activity, even if
this is not what it is to enjoy that activity? Let us
suppose that I am engaged in geometrical thinking and
on the verge of completing an important proof. Cer-
tainly such a situation as this· could result in a glow
of excitement and elation welling up within me, though
this is not the enjoyment of the activity, which Aris-
totle held to be scarcely distinguishable from the activi-
ty itself. Moreover, this excitement might well inter-
cept and impede my geometrical thought in the way
that 'foreign pleasures' are said by Aristotle to do;
I might have to light a pipe and pace about until the
pleasurable excitement died down and I could again
become absorbed in my geometry. Yet in his account
328 f. O. Urmson
of pleasure Aristotle seems to leave no room for this
sort of pleasurable feeling. Even the bodily pleasures
are regarded as the enjoyment of activities of sense
perception; they are less clear or pure than the pleas-
ures of intellectual activity and arise from painful
conditions such as thirst, but Aristotle seems to think
that they can be subsumed under the same general
analysis.
But it is pretty clear that the pain of thirst, which
is alleged to be the opposite of the pleasure of drink-
ing, is not like the disagreeableness of doing geometry,
which is also in some way the opposite of finding
geometry enjoyable. When we find geometry disagree-
able, Aristotle tells us, we tend to avoid geometry and
can attend to it only with difficulty. But Aristotle can-
not wish to claim that the pain of thirst is such that
we find it hard to pay attention to our thirst, nor does
he speak of the 'proper pleasure' of thirst or the
'proper pain' of drinking. And the reason for these
facts is clear; thirst is not an activity, and the pangs
of thirst are an unpleasant sensation unconhected with
activity; similarly when Aristotle refers to the pleasure
of drinking he is not referring to the enjoyment of the
activity of drinking but the pleasant feeling engendered
by drinking when thirsty.
Certainly Aristotle's account of enjoyment can be
applied to activities essentially involving sense percep-
tion. One can, as he points out, be interested in tastes
(enjoy tasting things) in the way that an expert cook
or connoisseur of wines is likely to be; one might be
interested in tastes as one is interested in geometry
and enjoy tasting things as one enjoys doing geometry.
Or, again, watching plays and looking at colors, shapes,
and pictures are activities which essentially involve
sense perception, as also does listening to music. If
these are to be called bodily pleasures, as involving
the senses, we do not have to exclude them from the
field happily covered by Aristotle's account.
But we should notice that the activities are looking
at (or watching), listening to, tasting (as understood
in wine tasting), and smelling (understood as sniffing
at things), not merely seeing, hearing, having a taste
in one's mouth or an odor in one's nostrils. There
could be a comparable pleasant activity of active touch-
ing, a quasi-aesthetic enjoyment of textures; touch is
not essentially different here from the other senses.
Now surely if we look attentively at Aristotle's dis-
cussion of temperance and intemperance in Book 111,
especially the passage from it 'quoted earlier, it is clear
that when the intemperate man eats, drinks, and in-
dulges in sexual activity, he is characterized not as
enjoying these activities, but as performing them for
the sake of the pleasant feelings they produce. It is
the feelings that are enjoyed by the intemperate, not
the activities that engender them. But Aristotle fails
to make explicitly clear to himself the central point;
he persuades himself that the intemperate pleasures
are to be distinguished from the pleasure of looking
at pictures because they involve the sense of touch
and thus fails to see that he has really made a distinc-
tion between enjoying activities essentially involving
use of the senses and doing things which produce a
pleasant feeFng. Pleasant bodily feelings such as those
engendered by food, drink, and sex are merely con-
spicuous examples of the latter. We must surely rec-
ognize that one might smell roses not because one en-
joyed smelling roses but because one enjoyed the smell
-a passive experience to which the activity of smelling
is but a means. Similarly, one might, though the pos-
sibility is remote, do geometry not because of an in-
terest in geometry and enjoyment of geometrical work,
but in order to produce the feeling of exhilaration
330 ,. O. Unnson
\yhich successful work can sometimes give rise to, the
feeling being what was enjoyed.
In the Topics (106a36 ff.) Aristotle remarks that
"the pain from thirst is opposite to the pleasure from
drinking, but there is none opposite to the pleasure
from contemplating the incommensurability of the
side and the diagonal. So the term 'pleasure' is used
ambiguously." But in the NE, he has rightly noticed
that there can be a 'proper pain' of thinking; I might
find the contemplation of this incommensurability in-
finitely tedious. In the same way, I might find the
activity of drinking tedious. Moreover, abstinence from
geometrizing, like abstinellce' from drinking, might
produce pangs of deprivation akin in principle to the
pangs of thirst. What again we need is what Aristotle
only half sees, a distinction between the pleasantness
(welcomeness) and unpleasantness (unwelcomeness)
of things, in particular feelings, which may be produced
by or be otherwise concomitants of an activity and the
pleasantness or unpleasantness of the activity itself.
No doubt the pleasantness of the feelings produced by
drinking is causally dependent on their being preceded
by unpleasant feelings of thirst, whereas the pleasure
of geometrizing is not so closely bound to the un-
pleasantness of anything. But these are contingent
facts, for we might conceivably have pleasant sensa-
tions from drinking until we burst and enjoyed doing
geometry only after a toothache.
So it would seem that Aristotle, having adumbrated
a distinction between the enjoyment of perceptual ac-
tivity (bodily pleasure in the wider sense), and the
enjoyment of feelings produced by the bodily, but
scarcely perceptual, activities of eating, drinking, and
sex; fails to recognize its proper significance, making
it, wrongly, a question of what sense is employed. It
is remarkable that while most philosophers have wrong-
ly assimilated the enjoyment of activity to the enjoy-
ment of feelings, a mistake for which Aristotle duly
castigates them, he himself makes the uncommon error
of assimilating the enjoyment of feelings to the en-
joyment of activity.
Yet at times, it seems that Aristotle effectively has
the distinction clear, particularly in Book VII of the
NE. There he is attacking the views of those who
denigrate pleasure and it seems that he has two main
contentions to make. First, he sees in such views a
mere excess of asceticism; the activities from which
pleasant feelings arise are essential features of human
life and thus have their place if not cultivated in ex-
cess. Further, if one wishes to engage in theoretical
work a neutral state is no doubt preferable to either
pleasant or unpleasant bodily feeling; but no one would
wish to regard this as a reason for a general attack on
pleasure who did not think of pleasant bodily feeling
as the whole of pleasure so that (1153b33) "the bodily
pleasures have gained a proprietory right to the name."
As he duly points out (1153a22), the enjoyment of
speculative thought can hardly be claimed to impede
speculative thought. Throughout these discussions in
Book VII, one has to take the term 'bodily pleasure'
as meaning 'pleasant feelings' rather than 'enjoyment of
activities involving use of the senses'. And yet even in
Book VII, he gives a summary definition of pleasure
as 'unimpeded activity', which I take to be essentially
similar to that of Book X; that is, I take him to mean
that an activity is enjoyable when it is unimpeded (by,
for example, "foreign pleasures").
To sum up, then, I should contend that Aristotle
gives us a clear and valuable account of the enjoy-
ableness of an activity. What leads to the obscurity
and insufficiency in his position is an ambiguity in his
use of the term 'bodily pleasure' of which he him-
self is unaware. At times, he uses it to refer to the
enjoyment of any activity essentially involving the use
332 ,. O. Urmson
of any of the senses, including listening to music and
watching plays. At other times, he uses it to refer,
sometimes quite clearly, to the pleasant sensations
which can be gained by activity. But, since he draws
his examples from localized bodily feelings resulting
from contact, and since there are no familiar examples
of enjoying active employment of the sense of touch,
he gives a mistaken formal account of the matter as
if there were no ambiguity but the generic case of
enjoyable use of the senses and the specific case of
enjoyable use of the sense of touch. Thus, in Book X
of the NE, he gives a final account of pleasure which
is appropriate as an account of enjoyable activity in
general, including those involving the senses, but, fail-
ing to see the ambiguity of 'bodily pleasure', he falsely
believes that he has given a general. account of pleasure
in giving an account of the enjoyment of activity.
I would like to add a few remarks about the ter-
minology employed by Aristotle and its translation. It
is plain that at times, particularly outside his official
discussions of 'pleasure', Aristotle is referring to pains
and other very unpleasant bodily feelings when he uses
the word MTCT]; obvious examples can be found in Book
VII, chapter 7. But the verb AUTCta9al frequently can-
not be idiomatically translated 'be pained', but means
rather 'find distasteful' as when Aristotle says that the
musician AUTCEL-ral by bad songs-he is repelled by
them, finds them distasteful. It is abundantly clear that
to speak of the 'proper pain' of an activity is highly
suggestive of just that analysis of enjoyment and its
opposite which Aristotle rejects; to perform an activity
AUTCOU\lEVO<; is not to perform it with pain but finding
it disagreeable, irksome, boring, or repulsive. Moreover,
while MTCT] is used by Aristotle as the opposite of
~f>OvfJ, pain is not the opposite of pleasure. It is not
even the opposite of pleasant feeling, a position oc-
cupied by unpleasant feeling (there are many un-
pleasant feelings which are not pains). It is hard to
translate M'TCT) because there is no obvious opposite in
English to 'enjoyment' or 'pleasure'; 'displeasure' will
only do in certain rather special contexts; but we
ought to abandon the word 'pain' as a translation. Pains
have no opposite; in this respect they resemble tickles.
It is also notable that while the word 1'j5ovTJ is com-
mon, and sometimes refers to a pleasant feeling, the
verb ~5EOeal is far less common in the NE than
Xa[pELV, which is the usual opposite of AU'ltElOeal;
I have noticed no case where ~5EOeal could mean
'have a pleasant feeling', and I suppose nobody would
suggest that Xa[pElV has such a meaning, though one
can Xa[pElV OCUf.La'rlKU 0:'ltoAa6oEl. It would seem that
for Aristotle Xa[pElV is the verb most commonly cor-
responding to the noun Tj50vTJ rather than ~5£Oeat
and if this is so, it is surely suggestive for our under-
standing of the word Tj50vTJ in Aristotle.
334
NOTES ON CONTRIBUTORS
John L. Ackrill is Professor of Ancient Philosophy at
Oxford University and Fellow of Brasenose College,
Oxford.
Miss G. E. M. Anscombe is Lecturer in Philosophy and
sometime Research Fellow in Somerville College,
Oxford.
John L. Austin was White's Professor of Moral Phi-
losophy at Oxford University and Fellow of Corpus
Christi College from 1952 until his death in 1960.
John Cook Wilson was Wy~eham Professor of Logic
at Oxford University and Fellow of New College,
Oxford. He died in 1915.
Irving M. Copi is Professor of Philosophy at the Uni-
versity of Michigan.
W. F. R. Hardie is President of Corpus Christi Col-
lege, Oxford.
K. J. Hintikka is Professor of Philosophy at the Univer-
sity of Helsinki and at Stanford University.
G. E. L. Owen is Professor of Philosophy at Harvard
University.
Father Joseph Owens teaches at the Pontifical Institute
of Medieval Studies at Toronto.
H. A. Prichard was White's Professor of Moral Phi-
losophy at Oxford University. He died in 1947.
Manley Thompson is Professor of Philosophy at the
University of Chicago.
J. O. Urmson is Fellow of Corpus Christi College,
Oxford.
M. J. Woods is Fellow of Brasenose College, Oxford.
335
SELECTED BIBLIOGRAPHY
This is a bibliography relevant to the topics discussed

in this anthology. The selections are designed primarily
for the student of philosophy.
(A) TRANSLATIONS
Oxford Translation under the editorship of J. A. Smith

and W. D. Ross. 12 vols. Oxford: Oxford University,
19°8-52.
Clarendon Aristotle Series, ed. by J. L. Ackrill. Oxford:
Clarendon, 1963- (translations and commentaries of
selected portions of the Organon, Metaphysics, and
Physics, with most volumes still under preparation).
(B) GENERAL
Allan, D. J., The Philosophy of Aristotle. Oxford: Oxford

University, 1952.
Anscombe, G. E. M., and Geach, P. T., Three Philoso-
phers. Oxford: Blackwell, 1961.
Bambrough, R. (ed.), New Essays on Plato and Aristotle.
London: Routledge & Kegan Paul, 1965.
Bonitz, H., Aristotelische Studien, I-V. Vienna, repub-
lished 1962-67.
Cherniss, H. F., Aristotle's Criticism of Plato and the
Academy. Baltimore: Johns Hopkins, 1944.
During, I., and Owen, G. E. L. (eds.), Aristotle and Plato
336
in the Mid-fourth Century. Goteborg: Almquist & Wik·
sell, 1957.
Jaeger, W., Aristotle. 2d ed.; Oxford: Clarendon, 1950.
Mansion, A., (ed.), Aristote et les problemes de methode.
Louvain: Publications Universitaires de Louvain, 1961.
Mure, G. R., Aristotle. London: E. Benn, 1932.
Ross, W. D., Aristotle. 5th ed.; London: Methuen, 1949.
(C) LOGIC
Bochensky, I. M., Ancient Formal Logic. Amsterdam:

North Holland Publishing Company, 1951.
Hintikka, K. J., "An Aristotelian Dilemma," Ajatus, XXII
(1959), 87-<)2.
- - , "Aristotle and the Ambiguity of Ambiguity," In-
quiry, II (1959), 139-51.
- - , "Necessity, Universality, and Time in Aristotle,"
Ajatus, XX (1957), 65-90.
Joseph, H. W. B., An Introduction to Logic. 2d ed.; Ox·
ford: Clarendon, 1916.
Kneale, W. C. and M., The Development of Logic. Ox·
ford: Clarendon, 1962.
Lukasiewicz, J., Aristotle's Syllogistic, 2d ed.; Oxford:
Oxford University, 1957.
review of Lukasiewicz (1st ed.) by J. L. Austin, Mind,
LXI (1952), 395-404.
McCall, S., Aristotle's Modal Syllogisms. Amsterdam:
North Holland Publishing Company, 1963.
Patzig G., Aristotelische Syllogistik. 2te Aufl.; Gottingen:
Vandenhoeck & Ruprecht, 1963.
review of Patzig (1St ed.) by J. L. Ackrill, Mind, LXXI
(19 62 ), 107-17'
(D) "CATEGORIES" AND

"DE INTERPRETATIONE"
Albritton, R., "Present Truth and Future Contingency,"

The Philosophical Review, LXVI (1957), 29-46.
337
Bonitz, H., Vber die Kategorien des Aristoteles. Vienna:
Staatsdruck, 1853.
Butler, R. J., "Aristotle's Sea Fight and Three-Valued
Logic." The Philosophical Review, LXIV (1955),
26 4-74'
De Pater, W. A., Les Topiques d'Aristote et la dialec-
tique platonicienne. Etudes Thomistes, X. Fribourg,
19 65.
Hintikka. K. J., "The Once and Future Sea Fight," The
Philosophical Review, LXXIII (1964), 461-<)2.
Husik, J., "The Categories of Aristotle," in Philosophical
Essays, Oxford: Blackwell, 1952.
Owen, G. E. L., "Inherence," Phronesis, X (1965), 97-
10 5.
Owens, J., "Aristotle on Categories," Review of Meta-
physics. XIV (1960-61), 73-90.
Saunders, J. T., "A Sea Fight Tomorrow," The Philosoph-
ical Review, LXVII (1958), 367-78.
SteinthaI, H., Geschichte der Sprachwissenschaft. 2d ed.;
Hildesheim: Olms, 1890 (1961).
Strang, C., "Aristotle and the Sea Battle," Mind, LXIX
(19 60 ),447-6 5.
Taylor, R., "The Problem of Future Contingencies," The
Philosophical Review, LXVI (1957), 1-28.
Trendelenburg, A., Geschichte der Kategorienlehre. Hil-
desheim: Olms, 1846 (1963).
(E) METAPHYSICS
Albritton, R., "Fonns of Particular Substances in Aristot-

le's Metaphysics," Journal of Philosophy, LIV (1957),
699-708 .
Cook Wilson, J., "Aristotle Metaphysics 1048a30 sqq.,"
Journal of Philology, XXXII (1913), 300-1.
Cousin, D. R., "Aristotle's Doctrine of Substance," Mind,
XLII (1933), 319-37; XLIV (1935), 168-85.
Eslick, L., "What Is the Starting Point of Metaphysics?"
Modern Schoolman, XXXIV (1957),247-63.
338
Guthrie, W. K. C., "The Development of Aristotle's The-
ology," Classical Quarterly, XXVII (1933), 162-71;
XXVIII (1934), 90-98.
Harring, E. S., "Substantial Form in Aristotle's Meta-
physics Z 1," Review of Metaphysics, X (1956), 308-
32; XI (1957), 482-501, 698-7 13.
Heidel, W. A., The Necessary and the Contingent in the
Aristotelian System. Chicago: Chicago University, 1896.
Lacey, A. R., "OUSIA and Form in Aristotle," Phrone-
sis, X (1965), 54-69.
Mansion, S., "La doctrine aristotelicienne de la substance
et Ie traite des Categories," Library of the Tenth In-
ternational Congress of Philosophy. Amsterdam: North
Holland Publishing Company, 1949.
McMullin, E. (ed.), The Concept of Matter in Greek
and Medieval Philosophy. Notre Dame: University of
Notre Dame, 1963.
Mure, G. R., "Aristotle's Doctrine of Secondary Sub-
stances," Mind, LVIII (1949), 82-83.
Owen, G. E. L., "A Proof in the PERI IDEOON," Tour-
nal of Hellenic Studies, LXXVII (1957), Pt. I, 103-11.
- - , "Aristotle on the Snares of Ontology," in R. Bam-
brough (ed.), op. cit.
- - , "Logic and Metaphysics in Some Earlier Works of
Aristotle," in I. During and G. E. L. Owen (eds.),
op. cit.
Owens, J., The Doctrine of Being in the Aristotelian Met-
aphysics. 2d ed.; Toronto: Pontifical Institute of Medi-
eval Studies, 1963.
Sachs, D., "Does Aristotle Have a Doctrine of Secondary
Substances?" Mind, LVIII (1948), 221-25.
Sellars, W., "Substance and Form in Aristotle," Journal
of Philosophy, LIV (1957), 688-<)9·
Solmsen, F., Aristotle's System of the Physical World.
Ithaca: Cornell University, 1960..
Sorabji, R., "Function," Philosophical Quarterly, XIV
(1964), 289-302.
Tugendhat, E., TI KATA TINOS. Freiburg: K. Alber,
1958.
339
(F) ETHICS
Adkins, A. W. H., Merit and Responsibility. Oxford:

Clarendon, 1960.
Allan, D. J., "The Practical Syllogism," Autour d'Aris-
tote. Louvain: Publications Universitaires, 1955.
Anscombe, G. E. M., Intention. Oxford: Blackwell, 1958.
Burnet, J., The Ethics of Aristotle. London: Methuen,
19°0.
Cook Wilson, J., On the Structure of the Seventh Book
of the Nicomachean Ethics. Oxford: Clarendon, 1912.
Frankena, W. K., Three Historical Philosophies of Edu-
cation. Chicago: Scott, Foresman & Co., 1965.
Gauthier, R., La morale d'Aristote. Paris: Presses Uni-
versitaires de France, 1958.
Grant, C. K., "AKRASIA and the Criteria of Assent to
Practical Principles," Mind, LXV (1956), 400-7.
Horsburgh, H., "The Criteria of Assent to a Moral Rule,"
Mind, LXIII (1954), 345-58.
Jackson, R., "Rationalism and Intellectualism in the Eth-
ics of Aristotle," Mind, LI (1942), 343-60.
Joachim, H. H., The Nicomachean Ethics. Oxford: Clar-
endon, 1951.
Mothersill, M., "Anscombe's Account of the Practical
Syllogism," The Philosophical Review, LXXI (1962),
44 8- 61.
Owens, J., "The Ethical Universal in Aristotle," Studia
Moralia III (Rome: Desclee, 1965), 27-47
Ritchie, O. G., "Aristotle's Explanation of AKRASIA,"
Mind, VI (1897), 536-41.
Robinson, R., "L'acrasie selon Aristote," Revue Philoso-
phique, CXLV (1955), 261-80.
Walsh, J., Aristotle's Conception of Moral Weakness.
New York: Columbia University, 1963.
Williams, B. A. 0., "Aristotle on the Good; A Formal
Sketch," Philosophical Quarterly, XII (1962), 289-96.
340
(G) CONTEMPORARY PHILOSOPHY
RELATED TO ARISTOTLE
Austin, J. L., "A Plea For Excuses," Proceedings of the

Aristotelian Society, LVII (1956), 1-30.
Bennett, J., "Substance, Reality, and Primary Qualities,"
American Philosophical Quarterly, II (1965), 1-17.
Copi, I., and Gould, J. (eds.), Readings on Logic. New
York: Macmillan, 1964.
Frankena, W. K., Ethics. Englewood Cliffs: Prentice-Hall,
19 63.
Gauthier, D., Practical Reasoning. Oxford: Clarendon,
1963.
Geach, P. T., Reference and Generality. Ithaca: Cornell
University, 196z.
- - , "Subject and Predicate," Mind, LlX (1950), 6z-
83·
Grice, H. P., "Some Remarks about the Senses," Ana-
lytical Philosophy, ed. by R. Butler. Oxford: Blackwell,
196z.
Hamlyn. D. W., "Categories, Formal Concepts, and
Metaphysics," Philosophy, XXXIV (1959), l11-z4.
Hampshire, S., Thought and Action. London: Chatto and
Windus, 1959.
Harrison, B., "Category Mistakes and Rules of Language,"
Mind, LXXIV (1965), 30<)-z5.
Hart, H. L. A., "A Logician's Fairy Tale," The Philosoph-
ical Review, LX (1951) 198-ZlZ.
Jarvis, J., "Practical Reasoning," Philosophical Quarterly,
XII (196z), 316-z8.
Ryle, G., "Categories," Logic and Language (zd ser.); ed.
by Flew. Oxford: Blackwell, 1955.
- - , "Pleasure," Dilemmas. Cambridge: Cambridge Uni-
versity. 1954.
Strawson, P. F., Individuals. Londdn: Methuen, 1959.
- - , Introduction to Logical Theory. London: Methuen,
195 z .
von Wright, G. H., "Practical Inference," The Philosoph-
ical Review, LXXII (1963), 159-79.
341
- - , The Varieties of Goodness. London: Routledge &
Kegan Paul, 1963.
Williams, D. C., "'The Sea Fight Tomorrow," Structure
Method; and Meaning, ed. by P. Henle. New York.
Liberal Arts, 1951.
New Studies in Ethics
Edited by W. D. Hudson
This major series of

monographs has been
planned to cover the whole
range of ethical theory from
the Greek philosophers to
the latest developments in
contemporary philosophy.
Each study is complete in
itself, and the series provides
a unique treatment of the
main philosophical
problems in ethics.
"The writers of the New

Studies in Ethics are all
somewhat iconoclastic; no
cows are sacred. They are
seeking to stimulate really
fresh discussion in clear
precise English."
-A. C. Adcock
Hibbert Journal
Modern Studies in
Philosophy
3001 Greek Ethics Edited by Amelie Rorty
Pamela Huby
3025 Hume
3002 Ethical Intuitionism Edited by V. C. Chappell
W. D. Hudson
3027 Aristotle
3003 Contemporary Moral Edited by
Philosophy J. M. E. Moravscik
G. J. Warnock
3028 Wittgenstein : The
3004 Existentialist Ethics Philosophical
Mary Warnock Investigations
Edited by
3005 Evolutionary Ethics George Pitcher
Antony Flew
3029 Kant
3006 Aquinas and Natural Edited by
Law D. J. O'Connor Robert Paul Wolff
Other titles in preparation Other titles in preparation

Aristotle PDF

Uploaded by

Copyright:

Available Formats

Aristotle PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Aristotle PDF

Uploaded by

Copyright:

Available Formats

Papermacs are sold subject

to the condition that they

First published in the United States of America 1967

ISBN 978-0-333-00436-4 ISBN 978-1-349-15267-4 (eBook)

What is it for us to understand Aristotle?l In order

lThe views expressed in this introduction about the nature of

2Some of what Aristotle said is likely to be true or false a priori,

SThe Hedgehog and the Fox. An Essay on Tolstoy's View of

1 For what is and for what has come about, then,

From Mind, LXV (1956), 1-15, Newly revised by the author.

AN ELUCIDATION OF THE FOREGOING

A. The Vice Chancellor will either be run over next

1. The interrelations of modal notions in Aristotle.

necessary that p not nccessary that p

From Inquiry, III (1960), 18-28. Reprinted by permission of

necessary that p = not necessary that p =

2. The two notions of possibility. The problem is to

not possible that not p = possible that not p =

necessmy that p = possible that p = impossible that p =

It is seen that (iii) and iv) differ in that in (iii)

necessary contingent impossible

The distinction between the different applications

1Aristotle's awareness of the fact that a case of a concept may

Arguments have recently been offered that purport-

lOCo S. Peirce remarks, "It is probable that Kant also under·

11The verb used to express this relation is akolouthein, which

12It is assumed here that for Aristotle simple conversion (as

l3The Greek here is tauton semainein as opposed to akolou·

§ 438. If our reasoning is correct, the universalities

§§ 438-50. Translation of principal passages referred to

6J. C. W. translated: 'It is not possible for One or Being to be

From Aristotle's Categories and De Interpretatione, tr. with

lRead TWY tTipc.JY YEVWV.

NOTES ON THE CATEGORIES

lal. The word translated 'animal' originally meant just

1316. What does Aristotle mean here by 'combination'

In several of his writings Aristotle presents what

6Trendelenburg, op. cit., pp. 24 If. and H. Stein thaI, Ge·

7Steinthal, op. cit., pp. 260 ff.

10Aristotle, like Plato, did not distinguish between these.

14For a different view of the relation between Aristotle and

18G. Ryle, "Categories," Logic and Language, second ser., cd.

The notions of essence and accident play important

7A Modern Introduction to Logic, p. 433.

12Bk. 2, ch. 27, 53.

21Bk. 3, ch. 4, 519. But d. Bk. 3, ch. 6, S4, S5.

32Bk. 2, eh. 23, S7.

39Ibid. 42Bk. 3, ch. 11, SlO.

46Bk. 3, ch. 3, S17.

The first part of this paper tries to account for an

From Aristote et les problemes de la methode (Symposium

lPhs. III 1, 200b12-21.

31Ph. IV 4, 210b32-211a7. Thus for instance the common

39De Caelo II 8, Z90a13-Z4 and esp. Met. r 5, 101Obl-11.

66E.g. Ph. IV 3, 210a25-26 = Pann. 145d7-e1; Ph. IV 3,

11lPh. III 5, 205a10-12; but for the treatment of this too as an

§2 THE SUBJECT OF PREDICATION

§3 THE PROBLEM OF MATTER

6Met. H 2, 1043a15; Oxford tr. On this doctrine, and Aris-