THE LOGIC OF BEING

A PALLAS PAPERBACK

~p~

\]Q] DaDerbaCkS

THE LOGIC
OF BEING
Historical Studies

Edited by
SIMO KNUUTTILA
Dept. of Systematic Theology, University of Helsinki

and
JAAKKO HINTIKKA
Dept. of Philosophy, Florida State University

D. REIDEL PUBLISHING COMPANY

A MEMBER OFTHE KLUWER

ACADEMIC PUBLISHERS GROUP

DORDRECHT ,I BOSTON / LANCASTER / TOK YO

Library of Co ngress Cataloging in Publica tion Data

Main entry under title:
The Logic of Bting.
(Synthese historitallibrary; v. 28)
Includes indexes.
Contents: Introduction - Retrospttt on the verb ' to be' and the concept of
being I Charles H. Kahn - Identity and predication in Plato I Benson Mates Aristotle and existence I Russell M. Dancy - letc.] I. Ontology - Histo ry Addresses. essays, lectures. I. Knuuttila. Simo, 1946II . Hintikka,
Jaakko. 1929Ill. Series.
111
85- 19622
BD3)I.L826 1985
ISBN _13: 978_90_211_237 1_0
001 : 10.1007/978-94-009-4780-1

Published by D. Reidel Pu blishing Company

P.O . Box i7, 3300 AA Dordrecht . Holland
Sold and dist ributed in the U .S.A . and Canada
by Kluwer Academic Pu blishers,
101 Philip Drive. As~in ippi Park, Norwell. MA 02061. U.S.A.
In all other cou ntries. sold and dist ributed
by Kluwer Acade mic Pu blishers Group.
P.O. Box 322. 3300A H Do rdrecht , Holland
Also published in /986 in hordbound edilion by Reidel in Ihe Synthese
Historicol Library, Volume 18

All Ri ghts Reserved

@ 1986 by D. Reidel Publishing Company, Dordrecht, Holland

No pan of the material protected by this copyright notice may be reproduced or

utilized in any form or by any means, electronic or mechanical,
including photocopying, reco rding or by any information storage and
retrieval system, without written pcrmission from the copy right owner

TABLE OF CONTENTS

ACKNOWLEDGEMENTS

vii

INTRODUCTION

ix

CHARLES H. KAHN / Retrospect on the Verb 'To Be' and the

Concept of Being
BENSON MATES / Identity and Predication in Plato

29

RUSSELL M. DANCY / Aristotle and Existence

49

JAAKKO HINTIKKA / The Varieties of Being in Aristotle

81

STEN EBBESEN / The Chimera's Diary

115

KLAUS JACoBI/Peter Abelard's Investigations into the Meaning

and Functions of the Speech Sign 'Est'

145

HERMANN WEIDEMANN / The Logic of Being in Thomas Aquinas

181

SIMO KNUUTTILA / Being qua Being in Thomas Aquinas and John

Duns Scotus

201

LILLI ALANEN / On Descartes's Argument for Dualism and the

Distinction Between Different Kinds of Beings

223

JAAKKO HINTIKKA / Kant on Existence, Predication, and the

Ontological Argument

249

LEILA HAAPARANTA / On Frege's Concept of Being

269

INDEX OF NAMES

291

INDEX OF SUBJECTS

297

ACKNOWLEDGEMENTS

.Tllakko Hintikka's work on 'The Varieties of Being in Aristotle' was

supported by NSF Grant No. BNS 8119033.
Btnson Mates's paper 'Identity and Predication in Plato' first appeared in Phronesis 24 (1979), 211 - 229.
Russell M. Dancy's paper, 'Aristotle and Existence', appeared in Synthese 54 (1983), 409 - 442.
laakko Hintikka's paper 'Kant on Existence, Predication, and the Ontological Argument' first appeared in Dialectica 35 (1981), 127 -146.
Leila Haaparanta's paper 'On Frege's Concept of Being' contains
some material which has also appeared in her dissertation Frege's Doctrine of Being (Acta Philosophica Fennica, vol. 39), Societas
Philosophica Fennica, Helsinki 1985.
All the previously published material appears with the permission of the
author and of the editor or publisher, which the editors thus gratefully
acknowledge.

vii

INTRODUCTION

The last twenty years have seen remarkable developments in our

understanding of how the ancient Greek thinkers handled the general
concept of being and its several varieties. The most general examination
of the meaning of the Greek verb 'esti'/'einai'/'on' both in common
usage and in the philosophical literature has been presented by Charles
H. Kahn, most extensively in his 1973 book The Verb 'Be' in Ancient
Greek. These discussions are summarized in Kahn's contribution to this
volume. By and large, they show that conceptual schemes by means of
which philosophers have recently approached Greek thought have not
been very well suited to the way the concept of being was actually used
by the ancients. For one thing, being in the sense of existence played a
very small role in Greek thinking according to Kahn.
Even more importantly, Kahn has argued that Frege and Russell's
thesis that verbs for being, such as 'esti', are multiply ambiguous is ill
suited for the purpose of appreciating the actual conceptual assumptions
of the Greek thinkers. Frege and Russell claimed that a verb like 'is'
or'esti' is ambiguous between the 'is' of identity, the 'is' of existence, the
copulative 'is', and the generic 'is' (the 'is' of class-inclusion). At least
a couple of generations of scholars have relied on this thesis and frequently criticized sundry ancients for confusing these different senses of
'esti' with each other. Others have found the distinction between the different Fregean senses in this or that major Greek philosopher, or otherwise used the distinction as an integral part of their interpretative
framework. Kahn's results show that all these lines of argument are
highly suspect.
Independently of Kahn, Michael Frede (in his Habilitationsschrijt
published in 1967 under the title Priidikation und Existenzaussage)
reached the conclusion that Plato did not - at least not in the Sophist
- accept anything like the Frege - Russell distinc~ion, thus striking
another blow against the received views. We hoped to include excerpts
of Frede's little classic here. Unfortunately, for reasons beyond our help
this turned out to be impossible.
IX

S. Knuullila and 1. Hintikka (eds.), The Logic of Being, ix-xvi.

Co, 1986 by D. Reidel Publishing Company.

INTRODUCTION

Most philosophers and most classicists were not immediately convinced by Kahn and Frede. It seemed to them that to say that Plato (or
any other Greek philosopher) did not show the Frege - Russell distinction amounts to accusing him of a logical howler. Isn't that distinction
a purely logical one, an indispensable part of our logical apparatus? This
question seems to have haunted even those students of Plato and Aristotle who had found the Frege - Russell distinction less a research tool than
a Procrustean bed. They seem to have been reluctant to claim in so many
words that the greatest Greek philosophers did not really operate with the
Frege - Russell distinction. In the case of Plato, it took a scholar like
Benson Mates, who combines a high degree of logical expertise with
historical scholarship, to come out of the closet and argue expressly that
to acknowledge that Plato Ia,cked the distinction is not to accuse him of
any mistake; for there are alternative logical treatments of 'is' (or 'esti')
which do not presuppose any irreducible ambiguity in these verbs between the allegedly different Frege - Russell senses of 'is'. Mates's article
'Identity and Predication in Plato' appeared in 1979, and is reprinted in
this book. He argues there that it is a radical mistake to try to project
the distinction onto Plato.
In the same spirit as Mates, but with an entirely different alternative
logical framework in mind, laakko Hintikka argued in his paper' "Is",
Semantical Games and Semantical Relativity' (1979) that the
Frege - Russell distinction is not only dispensable but indeed a mistaken
representation of the logic of natural language. In the first half of Hintikka's paper 'The Varieties of Being in Aristotle', appearing for the first
time here, he argues that Aristotle did not operate with the
Frege - Russell distinction any more than Plato did (according to Frede
and Mates), and traces some of the consequences of this insight.
In some ways, these developments are only the tip of an iceberg. Much
of the best work on Greek philosophy in the last twenty years has been
inspired or otherwise influenced by the late G. E. L. Owen. He made it
entirely clear that Aristotle handled the concept of being in a way essentially different from what the Frege - Russell logic leads us to expect.
One of the many stimulating suggestions Owen made was that the purely
existential uses of 'esti' in sentences of the type 'Homer is', in the sense
'Homer exists', are construed by Aristotle as being in the last analysis
elliptical for 'Homer is a man', more generally 'Homer is what he essentially is'. Owen never argued for his position, however, and he may never
have adopted it. In his paper, 'Aristotle and Existence', reprinted here,
Russell M. Dancy sets Owen's idea against the background of Aristotle's

INTRODUCTION

xi

treatment of the concept of being in general and defends it at length.

Dancy's paper serves at the same time as a useful introduction to the problems posed by Aristotle's use of the concept of being.
The alleged Frege - Russell ambiguity does not exhaust the varieties of
being that are relevant here. Although Aristotle fails to postulate the
Frege - Russell distinction, he does assume a distinction between the different senses of 'esti' used in the different categories. Aristotle's doctrine
of categories has recently been the subject of a great deal of interesting
discussion. Much of it was prompted by G. E. L. Owen, who in 1960
published a paper entitled 'Logic and Metaphysics in Some Early Works
of Aristotle'. The paper has since become a modern classic. It contains
a discussion of how Aristotle sought to overcome the distinction of different senses of 'esti' in the different categories by means of the idea of
pros hen multiplicity of uses. The contrast between these uses is therefore
something less than a full ambiguity (or homonymy, in Aristotle's
terminology).
It is nevertheless far from clear what Aristotle's distinction between
different categories really amounts to. In the second part of his paper
'The Varieties of Being in Aristotle' (see pp. 96 - 112) laakko Hintikka
presents a systematic analysis of logical quantification in natural
languages which yields as a by-product a rational reconstruction of
Aristotle's theory of categories. Among other things, Hintikka offers an
explanation of how the different aspects of Aristotle's theory go together
- in particular, how Aristotle can consider the distinction between different categories sometimes as a distinction between classes of simple
predicates, sometimes as a distinction between the different largest
classes of entities one can speak of together, sometimes as a distinction
between different senses of 'esti', and sometimes (especially in choosing
his names for the different categories) as going together with the different kinds of wh-questions in Greek. It turns out that the topical imperfections of this rational reconstruction naturally lead us to some of
the same conceptualizations as were used by Aristotle in further developing his theory of categories, especially to the matter-form contrast. Here
some extremely interesting possibilities of cooperation between
systematic and historical work seem to be opening up.
From this survey, it is apparent that the first half of our volume is
calculated to present a reader with the basic materials documenting the
new perspective on the ways in which the major Greek philosophers dealt
with the concept of being. The second half of our volume discusses some
important aspects of the subsequent history of what we have called 'the

xii

INTRODUCTION

logic of being'. As shown in the papers of Klaus Jacobi and Sten Ebbesen, early medieval inquiries into the logic and semantics of 'is' were
a part of an investigation of the nature of predication. It was usually
thought that the standard logical form of an affirmative proposition
could be thought of as a three-part form consisting of a subject term, a
predicate term, and an interposed copula. In the Aristotelian manner,
the three-part form was conceived of as an explanatory reformulation of
a two-part form, in which a noun in the nominative case is combined with
an inflected verb. On this approach, the question of the properties of the
copulative 'is' became the main problem of the theory of predication.
One of the difficulties was to understand the relation between 'is' as tertium adiacens, that is, as a copula, and 'is' as secundum adiacens, that
is, without additions.
Abelard's attempts to solve the question and his reports of the theories
of his contemporaries are discussed by Klaus Jacobi in his contribution
to this volume. In Abelard's time there were two main alternative positions. According to one theory, the function of 'is' as a copula is to join
the semantic content of the predicate term to that of the subject term.
It can exercise this function only when it has no semantic content of its
own. No connection was seen between the copula and the 'is' used as
secundum adiacens. The opponents of this equivocation theory argued
that 'is' used as secundum adiacens expresses that the thing under discussion exists, but that, when 'is' is used copulatively, the predicate serves
to determinate the manner in which the subject exists. According to this
view, the actual multiplicity of uses if 'is' is not accompanied by any genuine multiplicity of meaning. The non-copulative use was thought of as
an existential one, and it was suggested that the copulative propositions
with non-existent subject terms can always be translated into forms
having existential import.
In many places Abelard seems to hesitate between these alternatives.
This hesitation was connected with the fact, Jacobi argues, that both
alternatives were based on the three-part analysis of a proposition,
whereas Abelard's main interest was to develop a theory of predication
in which a two-part form is preferred. From this vantage point he tried
to interpret the copula as an auxiliary verb, which in conjunction with
a predicate noun does duty for verbs which often are not yet invented.
Abelard's ideas did not win any adherents, and as stated by Sten Ebbesen, in the thirteenth century the equivocation theory met with a
general disapproval. In his paper Ebbesen delineates the ancient and
medieval discussion of the problems of non-existent things, as ex-

INTRODUCTION

xiii

emplified by statements about chimera. One of the much discussed

medieval examples was the consequence: Chimera is opinable, therefore,
chimera is. This is, according to Aristotle, a fallacy secundum quid and
simpliciter, not a fallacy of equivocation (Soph. EI. 167al - 6). Similarly, most of the mid-thirteenth century writers thought that 'is' as secundum adiacens signifies actual being simpliciter and 'is' as tertium
adiacens signifies a special sort of being. This doctrine, influenced by
Aristotle's logical writings, was already known in Abelard's time. Some
developments can be seen from new attempts to define those special sorts
of being which do not entail actual existence.
It is historically interesting that the question of the existential import
of 'is', when not used as secundum adiacens, was usually discussed in
terms of the examples like 'chimera is opinable' on the one hand, and
'chimera is chimera', on the other hand. Although later medieval
philosophers of being were more interested in other kinds of distinctions,
this particular problem continued to be treated, too, and thus the question of the relations between something like the 'is' of existence, the 'is'
of predication, and the 'is' of identity remained one of the live topics in
the logic and semantics of 'is'. In the fourteenth century, this triple
distinction also attracted attention because, in the logical analysis of
complex propositions into immediately verifiable ones, the three uses of
'is' are found in the basic propositions. The fully analyzed form of
'Socrates is a man', for example, was presented as follows: 'This is & this
is_ Socrates & this is a man'.
In the second half of the thirteenth century, the completed reception
of Aristotle's works in the Latin West gave rise to a new systematic approach to questions of epistemology, ontology, and meaning. Traditional problems in the semantics of 'is' were discussed on the basis of a
general philosophical theory of modes of being, intellection, and
signification. 'Is' was called an analogous term whose different syntactical uses and various contextual senses were taken to be bound together
by a theory largely inspired by Aristotle's remarks on the pros hen
multiplicity of the different senses of 'is'.
This approach was part of the theoretical framework of the
philosophy of language of the modistae, as mentioned by Ebbesen. Its
role in Thomas Aquinas' thought is discussed by Hermann Weidemann
and Simo Knuuttila. Weidemann first treats Aquinas' basic distinction
between, on the one hand, the use of the verb 'be' to express the being
of something which falls under one of the categories, and, on the other
hand, the use of the verb to express the truth of a proposition. This

xiv

INTRODUCTION

dichotomy, based on Aristotle's Metaphysics (1 017a22 - 35), amounts to

distinguishing between two different existential uses of 'is', which are
called the use of the verb in an actuality sense and in a there-is sense,
respectively. By taking this distinction as a starting point, Weidemann
shows how one could find the parts of the Frege trichotomy of the uses
of 'is' in Aquinas. He nevertheless also argues that, on the basis of
Aristotelian metaphysics and some ideas of his own, Thomas Aquinas
thinks that all uses of 'is' are based on the primary existential use in an
actuality sense.
Simo Knuuttila discusses the ontological and epistemological ideas on
which the analogical focal meaning theory of 'is' is built in Aquinas and
many other medieval thinkers. Peter Aureoli wrote in 1316 that, according to the common view, when something is spoken of as a being, it is
immediately conceived as a substance or quality or a quantity and not as
falling under some one common definition. The famous Scotist doctrine
of the univocal metaphysical concept of being is discussed by Knuuttila
against the background of this disjunctive concept of being. According
to Duns Scotus, even in the disjunctive approach it is in fact presupposed
that, when we speak of something as being, it is eo ipso conceived as having some kind of identity, although we cannot always actually identify
it. This virtual identifiability is, according to Duns Scotus, the most
primary constituent of the positive nature of whatever can be and as such
the content of the univocal metaphysical concept of being. Virtual identifiability does not presuppose actual existence; it can be said of anything
the existence of which is logically possible. In this new metaphysics, all
possible individuals are considered as metaphysical beings, divided into
possible worlds, of which the actual world is one. In Scotus's theory, the
same individual can be considered as actual in several alternative worlds
at the same time, although it has existence in the actual world only.
Scotus's views on individual identity and existence were connected with
his modal theory which influenced the non-Aristotelian trends in
fourteenth-century modal logic and introduced a modal semantics having similarities with the modern possible worlds semantics.
Scotus's doctrine of the univocity of the metaphysical concept of being
was partially adopted by the fourteenth-century nominalists, because it
suggested a univocal notion of existence, which was an important part
of the nominalistic omology. The nominalists contested essential parts
of the metaphysical framework on which the general theory of being of
their predecessors was built. It seems that in their search for a new
semantics of 'is' some of them were led to think that the basic types of

INTRODUCTION

xv

the uses of 'is' are fully inderendent. Sten Ebbesen mentions John
Buridan as a representative of this trend. However, fourteenth-century
discussions of these questions (and the later significance of these discussions) are still only partially known.
One of the medieval conceptual tools which has also been used in the
modern period is the doctrine of the different kinds of ontological
distinctions, e.g., real, mental, and formal distinctions. These are
discussed by Lilli Alanen. She shows that Descartes's well-known argument for the mind-body distinction is based on a specific interpretation
of this traditional doctrine, and that the early discussions of the argument were largely concentrated on the peculiarities of Descartes's way of
drawing the ontological distinctions just mentioned. The role of these
issues concerning the logic of being has not received its due attention in
the earlier literature.
It is sometimes said, or implied, that the Frege - Russell distinction
between the allegedly different senses of 'is' goes back to Kant and to
Kant's idea that 'existence is not a predicate'. In his paper, 'Kant on
Existence, Predication, and the Ontological Argument', Jaakko Hintikka argues that Kant does not in the last analysis assume anything like
the contemporary Frege - Russell distinction. All we find in Kant is a
contrast between relative and absolute uses of the concept of being.
As a part of Hintikka's argument, he offers a largely new diagnosis of
the fallacy in the ontological argument, denying Kant's dictum that 'existence is not a predicate'. Instead, he finds the crucial flaw in the ontological argument elsewhere, viz. in its tacit dependence on our being
able to identify God (the being of whose existence is to be proved) between the different possible worlds presupposed in the argument. This
provides a new perspective on historical as well as contemporary discussions of Anselm's argument.
All these different investigations naturally raise the question: What is
the origin of the Frege - Russell distinction? What is its background? In
her paper, 'On Frege's Concept of Being', Leila Haaparanta discusses
Frege's treatment of being in its historical setting. One of the crucial ingredients in Frege's treatment of being is his idea that existence is a
second-level concept (property of a concept). Haaparanta sees the foundation of this assumption in Frege's ideas about the identification (individuality) and existence of individuals (objects), incorporated in
Frege's treatment of the senses by means of which we can grasp an individual object. These were according to her inspired by Kant's ideas,
especially by Kant's distinction between the predicative and existential

xvi

INTRODUCTION

uses of 'is'. Even though Kant did not subscribe to or even anticipate the
Frege - Russell Jistinction, he thus seems to have inspired it.
SIMO KNUUTIILA
JAAKKO HINTIKKA

CHARLES H. KAHN

RETROSPECT ON THE VERB 'TO BE' AND THE

CONCEPT OF BEING

When I began work on the Greek verb to be in 1963, in the project that
took shape in the article 'The Greek Verb "to be" and the Concept of
Being,l and eventually resulted in a book on the Greek verb 'to be' in
1973, 2 my aim was to provide a kind of grammatical prolegomena to
the study of Greek ontology. I wanted to give a description of the
linguistic facts concerning the ordinary use and meaning of the verb,
apart from its special use by the philosophers, in order to clarify the pretheoretical point of departure for the doctrines of Being developed by
Parmenides, Plato, and Aristotle. I thought at the time (and still think)
that the ancient use of the verb estileinail on was poorly understood, and
that much of the modern discussion is vitiated by false assumptions, in
particular by an uncritical application of the notions of existence and
copula to the interpretation of ancient texts. I take the present occasion
to summarize the results of my work both for the theory of the verb and
for the interpretation of some of the early philosophical texts, referring
to earlier publications for more detailed exposition and defense of the
views outlined here.
I. THE DISTORTING INFLUENCE OF THE TRADITIONAL VIEW

As I see it, confusion reigns both in the traditional account of the verb
given by linguists and philologists, and also in much of the philosophical
exegesis of ancient theories of being. The two lines of confusion have infected one another, since the linguists have borrowed their notions of existence and the copula from philosophy (and from rather superficial
philosophy at that), while philosophers have in turn made use of
linguistic doctrine as a basis for their own account of Greek ontology,
and in some cases as the weapon for a general attack on the Greek notion
of Being. I begin by stating what I take to be the principal errors in the
standard view, by which I mean the views prevailing twenty years ago and
still to be found in many handbooks and commentaries.
(1) It was generally assumed that the uses of einai could be classified
either as (a) meaning 'exists', or (b) serving only as copula. 3 But this
S. Knuullila and 1. Hintikka (eds.), The Logic oj Being, 1-28.
1986 by D. Reidel Publishing Company.

CHARLES H. KAHN

dichotomy is theoretically unsound and descriptively inadequate. It is

theoretically unsound because (a) is a semantic and (b) is a syntactic notion. A rational theory would contrast 'exists' with other meanings of the
verb, and the copula syntax with other constructions. The dichotomy
could be justified only if there were a one-to-one correlation between
sense and syntax, so that copula uses were all meaningless and the verb
in absolute (non-copulative) construction always meant 'exists'. But
both assumptions are false.
(2) The traditional account of einai and its Indo-European root*es- in
comparative grammar (which goes back to Brugmann and Meillet, and
is reflected in some accounts of the copula in English) takes the existencecopula distinction for granted and proceeds as follows. The verb be
(einai, *es-) was originally a verb like other verbs, with concrete meaning.
The original meaning was to exist, or perhaps something even more concrete like to be present or to be alive. Predicate nouns and adjectives were
originally expressed without any verb, in the so-called "nominal
sentence" familiar from Russian and many other languages: John is wise
was simply John/wise; John is a man was John/(a) man, and so forth.
But in the course of time it became useful to introduce a verb into the
nominal sentence in order to express the tense, person, mood, and other
modalities carried by the finite verb in Indo-European. Hence the verb
be (meaning exist) was introduced into the nominal sentence, where it
gradually lost its original meaning and degenerated into an "empty"
verb or "mere copula," a syntactic device which serves to satisfy the requirement that every sentence must contain a verb.
This historical-sounding theory is enshrined in the textbooks,4 but
there is really no evidence to support it. What looks like evidence is a
misleading parallel to other verbs that take a predicate construction (like
turn pale, grow tall) and that clearly had an independent meaning, but
which in the course of time came to be used as substitutes or suppletives
for be, and even provide forms that are now integrated into the conjugation. (Thus am, are, and is in English are derived from *es-, but be comes
from I.-E. *bheu- 'to grow', 'become'; was comes from *wes- 'to dwell,
stay in a place'; eta is, he in French, stato in Italian, estar in Spanish all
come from Latin sto/stare 'to stand'.) However, these parallels prove
nothing to the point, since in every case the known historical development presupposes the existence of a basic copula verb in Indo-European. 5
And there is no doubt that the original copula was *es-, our verb 'to be'.
The notion of a prehistorical state of Indo-European without a copula
verb is a pure figment of the imagination. !n Greek at any rate

THE VERB 'TO BE' AND THE CONCEPT OF BFINC;

the copula uses of einai are overwhelmingly more frequent than any
other use in the earliest texts. The idea that the existential uses are
somehow more fundamental or more primitive seems to be a mere prejudice, a prejudice based in part upon a mistaken view of existence as an
ordinary predicate, taken together with an empiricist theory of meaning
which assumes that the original sense of any word must have been
something concrete and vivid, something like what Hume calls an
"impression" .
Hence I propose a modest Copernican revolution: to reinstate the
copula at the center of the system of uses of einai. I do not claim that
the copula uses are older, since for that claim also there is no evidence.
In purely synchronic terms I propose that the copula uses must be regarded as more fundamental in three respects: (1) they are statistically
predominant, (2) they are syntactically elementary, whereas other uses
(existential, veridical, potential) are grammatically "second-order",
operating as functors on a more elementary sentence, and (3) they are
conceptually prior and central to the whole system of uses of the verb,
in a sense that remains to be clarified, but which bears some analogy to
the unifying role of a central term in Aristotle's scheme of "focal meaning" or pros hen legomena. Thus if we take the copula uses as given, we
can see why the same verb may serve in other ways, for existence, truth,
possession and the rest. But if we take any of the other uses as primary,
the way back to the copula becomes difficult, if not impossible [Q
understand. 6
(3) It was correctly noted by a number of linguists that the existence
of a verb to be in our sense, which is at once a verb of predication, location and existence (to name only three of its functions), is a peculiarity
of Indo-European. 7 As we can see from the monograph series on "the
verb 'be' and its synonyms" edited by .I, W. M, Verhaar, the topic of
be can itself be defined only by reference to Indo-European verbs from
the root *es-. But why should a historical peculiarity of this kind be of
any general significance, and how can a concept based upon the
parochial usage of an Indo-Eurorcan verb provide a genuine topic for
philosophical theory? Thus A. C. ~:(2harn has claimed, a rropos of the
very different situation in Chinese. thJ.t
there is no concert of Being which languages ar~ wei! Of iii cquiprcd to
tipn." or 'to ht" as ('opld3 depend lIPon a grarnmatical rule t'O!' thl...'
sentence. and it would b::.- merely a :elDcid'.:r~'(' ~r one found an\'thi:,,(~

ir-tr!!!uagc without this rule. S

l'-C",'flL:

tnr rune
,)\" the

i:-" ,~-:atinn

Icscnl~)l;-:,'

11

in (:

Such arguments from linguistic relativism tend to rei:lforce the

CHARLES H. KAHN

philosophical complaints which Mill and others have directed against

"the frivolous speculations concerning the nature of Being", which
Mill thought had arisen from overlooking the distinction between the
verb of existence and the diverse uses of the copula, and from supposing
"that a meaning must be found for it (namely, to be) which shall suit
all these cases.,,9 So Mill's godson Russell insisted that "the word is is
terribly ambiguous" and proposed that instead of a unified notion of Being we need to distinguish various "senses" of to be, including existence,
identity, and predication. Both logical and linguistic criticisms thus tend
to converge in a general suspicion that doctrines of Being in traditional
ontology reflect a projection onto the universe of the linguistic structure
of Greek or of Indo-European.
I do not intend to do battle here against a general thesis of linguistic
relativism, and I shall certainly not deny that the union of predicative,
locative, existential and veridical functions in a single verb is a striking
peculiarity of Indo-European. Whether this diversity is properly regarded as a case of ambiguity of meaning is a question on which I do not
propose to take a stand. \0 What I do deny is that this cumulation of functions in the verb to be was necessarily a philosophical disadvantage. On
the contrary, I want to suggest that the absence of a separate verb "to
exist" and the expression of existence and truth (plus reality) by a verb
whose primary function is predicative will have provided an unusually
favorable and fruitful starting-point for philosophical reflection on the
concept of truth and the nature of reality as an object for knowledge.
This was due in part to the puzzling con verge'.lce of so many fundamental
notions and functions in a single linguistic form. It was Parmenides who
first introduced "what is" (to on) as a central topic for philosophical
discussion, and the paradoxical argument by which he developed his
thesis turned out to be one of the most creative innovations in the history
of Western thought. In the first place he elaborated the stative-durative
aspect of the verb into a systematic claim that what is must be
ungenerated and incorruptible. And this claim provoked all the element
theories of the fifth century, including the theory of atoms, as an account
of how the most fundamental realities could remain immune to change.
(If the doctrine of indestructible atoms is no. longer with us, we may
perhaps recognize its Parmenidean shadow in the conservation laws of
modern physics.) And it was the same Parmenidean thesis of unchanging
reality that provided Plato with the ontological resources for his own account of immutable Forms. On the other hand, because esti is not only
a verb of truth and reality but also the sign of predication, Parmenides'

THE VERB 'TO BE' AND THE CONCEPT OF BEING

paradoxical insistence upon the monolithic unity of what is stimulated

both Plato and Aristotle into working out theories of predication, and
eventually led Aristotle to propose his scheme of categories as an account of how "what is is said in many ways". Finally, the Aristotelian
doctrines of matter and form, potency and act were also formulated as
a response to Parmenidean paradoxes about the concept of change (as
we can see from the structure of his argument in Physics I). Thus Western
physics, logic, and metaphysics have all been the beneficiaries of a fierce
century and a half of philosophical dispute generated by Parmenides'
bold attempt to fuse into a single entity the diverse features of the verb
to be.
I hope to shed a little light on some of the earlier phases of this momentous episode. But first we need a realistic description of the prephilosophical uses of the verb.
II. THE LINGUISTIC THEORY OF TO BE

1. The Copula

In order to see this as the central focus of the system of uses of einai, we
must have a more adequate account of the copula itself. Linguists often
speak of the copula as a dummy verb, as a merely formal "bearer" of
the verbal marks of tense, person, and mood. 11 Hence a recent author
can describe the entire function of the copula as "simply to act as a verbalizer," to convert "an adjective like 'cunning' into a verb-phrase 'is
cunning', which is of the same category as 'snores' " or of any finite
verb. 12 Abelard, who was either the inventor or at least the codifier of
the classical theory of the copula, rightly associated the copulative function (vis copulativa) with all finite verbs. He saw is as distinctive in that
it provides only the predicative link and not also the predicate (copulat
tan tum et non copulatur); other finite verbs do both.13 Abelard's theory
has the merit of focussing attention on the verbal function as such, and
not simply on is as the verb for nominal predicates. What he saw is that
the copula separates out the specific function of the verb, which is
obscured in the case of other verbs such as runs, sleeps, just because they
combine the information content of the predicate (running, sleeping)
with the verbal form. This general verbal function is what one might
identify as the propositional tie or mark of assertion; what we, following
Abelard, call the copula is simply the canonical expression of this function in a sentence of the form X is Y.

CHARLES H. KAHN

In sentences of this form, the copula expresses what is otherwise indicated by the verb-ending: in either case we have a "sign of predication". What is meant by predication here has both a syntactic and a
semantic aspect. (I) Syntactically, the copula or verb-ending serves to
make a grammatical sentence out of two terms that would otherwise only
form a list: John and running (or to run). It thus indicates not only tense,
mood, and person but something more fundamental: sentencehood. In
many cases the finite verb does this alone, without a noun, both in impersonal verbs like UH (pluit, "it is raining") and where the subject is
understood from the context: 7f!x.H (currit, "he is running"). (2)
Semantically, if we take the indicative-declarative form as basic, the verb
or copula gives formal expression to the truth claim of the sentence. (We
need not attend here to the ways in which this truth claim may be
modified, by interrogative or conditional sentence structure or by the
various moods; in some but not all of these cases, the modification will
be reflected in the verb ending.) Limiting ourselves to indicative forms
used in declarative sentences we can say: tbe semantic function of is as
copula or sign of predication is to bear the mark of sentential truth Claim,
to serve as focus for the claim of the whole sentence. (The truth claim
of a sentence corresponds roughly to the fact that it can have a truth
value because it does have truth conditions.)14 This basic assertive function of the copula - more precisely, the intimate connection between the
copula and the assertive function of the sentence - shows up when we
stress the verb in pronunciation: "Margaret is clever, I tell you!" "The
cat is on the mat after all." This semantic role of the copula as sign of
sentential truth claim permits us to understand one of the most important
special uses of einai in Greek, the so-called veridical, where the verb by
itself (both in the third person indicative and in the participle) expresses
the notions of truth and reality. If we lose sight of this connection between the copula and the truth claim that is fundamental for all
declarative discourse, the fact that "being" in (;reek (to on) may mean
reality will become a mysterious anomaly, quite independent of the
predicative function of the verb. Hence philolog.ists have often tended to
overlook the veridical use or to conflate it with the existential, despite the
fundamental differences in sentence structure. This is a principal support
for my claim that the copula use is fundamental; neither veridical nor existential use can be explained on the basis of the other, but both can be
understood on the basis of the copula.
Before considering these special uses of einai, I call attention to two
features of the copula verb which are often on:rlooked.

THE VERB 'TO BE' AND THE CONCEPT OF BEING

(A) The verb einai, whether used as copula or in other constructions,

has only durative (present-imperfect) forms, and no forms in the aorist,
that is, in the punctual or non-durative aspect. (The future forms lie outside this aspectual contrast.) This formal peculiarity of einai (and of I.-E.
*es-) is reflected in its semantic value as stative copula, in contrast to the
mutative copula becomes (gignesthai in classical Greek). 15 The antithesis
of Being and Becoming, opposed to one another as stability to change,
was deeply built into the system of copula verbs, long before it was exploited by Parmenides and Plato.
(B) Among the copula uses of be in a broad sense are what we may call
ldcative uses, where the complement or predicate expression is not a
noun or adjective but a local adverb (here, there) or a prepositional
phrase of place (at home, in the marketplace). Some linguists may prefer
not to count these as copula uses, since in this construction is cannot be
replaced in English by the mutative copula becomes (though in Greek
gignesthai may be used in locative sentences); whereas locative is functions (a) in contrast to a verb of motion (goes there, arrives in the
marketplace) and (b) in parallel to a wide variety of verbs (John works
at home, Socrates talks in the marketplace). This use of is, which I will
call the locative copula, shows a special affinity with a small group of
verbs of posture, which may serve as static replacers for copula is with
predicate nouns and adjectives as well as with locative phrases: sits,
stands, lies. 16 Because of these connections, and because of the more
vivid or concrete sense that seems to attach to einai in locative sentences,
and which (as we will see) often suggests an existential nuance that gets
rendered in English by there is, some scholars have suggested that the
locative or locative-existential use of einai represents the basic sense and
function of the verb. l ? But although the locative uses are certainly important for understanding the intuitive force of the verb to be in Greek as
a verb of state or station, I do not believe that they are more fundamental
than the copula use with predicate nouns and adjectives. For one thing,
it is possible to add locativt modifiers to copula sentences, as to many
others, and then derive the locative copula by zeroing the nominal
predicate: Athens is a city -+ Athens is a city in Greece -+ Athens is in
Greece; John is busy -+ John is busy at home -+ John is at home. But
there does not seem to be any plausible derivation in the opposite direction. Whereas locative phrases are optional modifiers for a wide variety
of sentences, the introduction of a predicate noun or adjective presupposes the basic construction with a copula verb, as I have argued
elsewhere. 18 Even more important, in my opinion, is that it is only the

CHARLES H. KAHN

general copula function as "verbalizer" and mark of sentencehood,

and not the locative use, that can explain the deep connections of einai
with the notions of truth and fact.
I turn now to the two principle classes of non-copulative uses of the
verb: the veridical and the existential. For present purposes we can ignore
other non-copula constructions, such as the possessive (esti moi, "I
have") and the potential (esti + infinitive, "it is possible to").
2. The Veridical Uses

The Lexicon recognizes, what every good Hellenist knows, that in many
cases the verb esti and its participle on must be translated by "is true" ,
"is so", "is the case" or by some equivalent phrase: esti tauta "these
things are so" (cf. French c'est cela "that's right"), legein to onto "tell
the truth", "state the facts" .19 Comparative grammar shows that this is
a pre-historic use of*es- in Indo-European 2o ; and Aristotle himself notes
this as one of four basically different uses of einai (Met. 117). All I have
done is give this use a name, "the veridical," and correlate it with a
definite sentence form. The veridical construction proper is characterized by three syntactic features: (i) the understood grammatical subject
of esti is not a noun form (like man or hunting) but a sentential structure,
as represented in English by a that-clause (that the man was hunting); (ii)
the construction of the verb is absolute, i.e. there is no nominal, locative
or other adverbial complement, except the comparative "so" (houtos),
which introduces (iii) a comparative clause of saying or thinking. which
is expressed in the full veridical construction and implied in every case:
These things are as you say, fCTn raimx ovrws 'WCT1I"EQ CTU Af'YHS.
Without claiming that the veridical construction is derived from the
copula use, either historically or transformationally, I nevertheless
believe that it is easy to see how the two uses logically and naturally
belong together, as long as we keep in view the semantic function of the
copula as the mark of truth claim. A properly veridical use of to be (as
in "Tell it like it is") simply makes general and explicit the truth claim
that is particularized and implicit in every declarative use of the copula
("The cat is on the mat"), as in every declarative sentence. Because the
great flexibility of the copula construction makes it possible to produce
an Sis P sentence that is roughly equivalent to every noun-verb sentence,
the copula tends to serve as verb par excellence, that is, as representative
for the finite verb as such and for its predicative force, what Abelard
called its vis copulativa. It is because of this very general function of the
copula as sign of predication and sentencehood that the very same form

THE VERB 'TO BE' AND THE CONCEPT OF BEING

(esti, "is") can serve to express the veridical idea as such: to bring out
the implicit truth claim ("This is how I say things are") and the corresponding notion of reality ("This is how things really are").
Despite my general reluctance to decide when a different use becomes
a different sense, I am inclined to speak here of a veridical meaning or
connotation of einai, in cases where the Greek verb cannot be adequately
translated by the copula or by an idiomatic use of is alone. This is most
conspicuous when the participle (to) on is used to mean "truth" or "the
fact of the matter", and when it may be replaced in Greek by a word like
aletheia or (to) alethes. 22 And there are clear cases of the veridical connotation attached to a copula construction, as in the example which
Aristotle cites of einai meaning "is true": esti S6krates mousikos, "that
Socrates is musical, i.e. that this is true". 23 In English as in Greek, this
force of the verb is typically brought out by a contrast between Being and
Seeming: "He wants not to seem (dokein) but to be (einOl) the noblest"
(Aeschylus Septem 592). Here again a basic philosophical contrast - between appearance and reality - is fully prepared in the pre-philosophical
usage of the verb.
3. Existential Uses
I briefly describe three uses of einai that we intuitively recognize as "existential" and are inclined to render by there is or even (in the third case)
by exists.24
(1) The existential copula:
(a)

"There is a city (esti polis) Ephyre in a corner of horsenourishing Argos". (Iliad VI. 152)

(b)

"There is a certain Socrates (estin tis Sokrates), a wise man,

student of things aloft ... who makes the weaker argument
the stronger". (Apology 18B 6)

Perhaps the most common of all "existential" uses of the verb in Greek
are sentences such as these, where esti seems to functions twice: to assert
the existence of a subject ("There is a city ... ") and then to say
something about it: "The cityEphyre is in the corner of Argos". In most
instances the predicative use will be locative, as in (a); (b) is one of the
rare examples where a purely nominal copula takes an existential force.
It is clear that the underlying syntax of the verb in such sentences is that
of the copula, but that this construction has been overlaid with a secondary function, which I would analyze as introducing a subject for further

10

CHARLES H. KAHN

predication, and which accounts for the existential nuance we render by

"there is". In most cases, where the construal is locative, the verb serves
to introduce (posit, assert existence for) its subject by locating it in a
definite place or context. But as (b) shows, esti may perform this function
even without the support of a locative construction. In either case the
verb typically occurs in the emphatic initial position (like "there is",
which always begins its clause); but initial position is neither necessary
nor sufficient for the Greek verb to play this role. 25
What do we mean by classifying the verb in such sentences as "existential"? It is in fact false t.J the intuitive force of the sentence to say that
it asserts the existence of Ephyre or Socrates. The verb simply introduces
its subject into the narrative, or into the stream of discourse, as a subject
for further predication or, more often, as a local point of reference for
the episode that follows. We may be inclined to connect the verb with the
existential quantifier of formal logic, since the sentence does imply that
the set of objects specified by the following predicates is not empty:
"There is something, not nothing, which is a city in Ephyre, etc.". But
the trouble with this analysis of the "existential" copula in sentences of
type (1) is that it applies equally well to straight copula versions of the
same sentences: "Ephyre is a city in Argos", "Socrates is a wise man".
What is logically implied is one thing, what is expressed is another.
Somehow, in virtue of initial position, locative function or more general
contextual features (introducing an item for future reference), the verb
in (a) and (b) gathers to itself the existential claim that properly belongs
to the sentence as whole. How can the copula verb assume this function?
Perhaps because the verb itself provides no lexical content but agrees formally (in person and number) with a subject term that does have content,
emphasis on the copula serves to focus attention on the subject, to present it in an emphatic way (usually by localizing it), and thus to focus on
the subject as such, as subject of "is" and hence precisely as a subject
for predication.
Whatever the explanation, the secondary existential force of the
copula verb in type (1) points the way to a purely existential use, with no
copula construction, in type (2).
(2) The existential sentence operator:
"There is someone (no one) who does such-and-such" (ouk)

esti has tis + relative clause. 26

In typical examples, sentences of this type refer to persons, but variants

occur which give this form the full generality of the logical scheme (3x)Fx

THE VERB 'TO BE' AND THE CONCEPT OF BEING

11

where Fx stands for the relative clause and an initial esti functions as existential quantifier: "there is someone/something such that ... ". The
verb esti serves precisely to affirm or deny (the existence ot) a subject for
the following clause, to assert that the set specified by the following formula is or is not empty. There is no trace of the copula construction, nor
any way to derive this form logically or syntactically from the copula construction. But there is a logical overlap with type (1), which can be seen
as a copula construction overlaid with the existential function of type (2).
This properly existential use is relatively rare: I found only 4 examples
out of 562 occurrences of einai in the first twelve books of the Iliad. I
think it would be unreasonable to suppose that (2) somehow represents
the original, prehistoric value of*es-. Can we offer a historical explanation of this use of esti? My suggestion is that it arose out of the copula
use by way of sentences of type (1), where the copula acquires "existential" connotations in virtue of its locative association and its rhetorical
function of introducing a subject for predication. Given these connotations, it is natural that the existential function of (2) becomes one of the
values esti can have when used alone, without nominal br locative complements. (The veridical, possessive, and potential uses represent other
values esti may possess when it appears outside of the copula
construction.)
It is on the basis of the existential force of the verb in (1), where this
force is secondary, and in (2) where it is primary but serves directly as
the basis for ensuing predication, that we can understand the appearance
of a new sentence type, in which esti itself becomes the grammatical
predicate.
(3) The existential predicate:
(a)
(b)

"There are (no) gods" (auk) eisi theoi,

"Zeus does not even exist" oud' esti Zeus.

Type (2) is rare in Homer, but type (3) does not occur at all. My earliest
specimens are from Melissus, Protagoras, and Aristophanes in the middle and second half of the fifth century B.C., and they clearly show the
influence of philosophical speculation. 27 Sentences of this kind are
sometimes cited as exhibiting the oldest meaning of *es- in IndoEuropean. On the contrary I regard this as a fifth-century innovation,
based upon the existential force of the verb in the older types (1) and (2),
but focussing attention on existence as such (i.e. on the question whether
or not there is such a thing), as a result of philosophical speculation,

12

CHARLES H. KAHN

theological sceptIcIsm, and the general disputes about "Being" that

begin with Parmenides.
Whatever its date and origin, the verb in this sentence type is best
understood as an abridgement from type (2), "There is someone (no one)
who ... ". Thus "Zeus is not" and "Ther~jlre gods" are to be construed
as generalized versions of "There is no Zeus who ... " and "There are
gods who ... ", where the effect of generality is achieved by dropping
the relative clause with its particular content and thus presenting (or rejecting) a subject for any and all unspecified predication. But the idea
that in such sentences the existential verb would itself constitute the
predicate is an illusion to which the Greeks seem never to have fallen
prey.28 The functions of einai as instrument of predication were so fundamental that the same verb could not easily be seen as forming a selfsufficient predicate. 29 In Greek linguistic intuition, "There is no Zeus"
(ouk esti Zeus) means that Zeus is not a subject for any predication, that
there is nothing true to be said about him. The Greeks are thus untroubled by the modern puzzle of negative existentials, whicli arises from
the temptation to assume that "Zeus does not exist" says something
which is true of Zeus.
III. THE EARLY PHILOSOPHIC USE OF TO BE

The outcome of my linguistic survey has been to underline the fundamental role of einai as copula verb and at the same time as verb of state and
station, characterized both by locative and by durative-stative values.
Among the non-copula uses I have called attention to the veridical expression for truth and fact, and I have insisted upon the very limited
range of early existential uses, bound to a specific locative or predicative
context. Thus while existential sentences of type 1 and 2 are well attested
in Homer, the stripped-down "absolute" use of type 3, in which einai
appears alone as existential predicate, is not found before the fifth century, and then only in contexts where philosophic or sophistic influence
is clear. 30 My suggestion is that for understanding the early philosophical
usage, both in Parmenides and in Plato, the veridical notion (whether or
not it is the case that p) turns out to be more important than the idea of
existence (whether or not there is such a thing as X), although both notions are present.
To illustrate the new, quasi-technical use of the verb as a pure existential (my type 3), we may cite what is probably the earliest unambiguous
example, the welI-known statement of Protagoras that, concerning the

THE VERB 'TO BE' AND THE CONCEPT OF BEING

13

gods, he does not know "either that they are (has eisin) 0;- that they are
not (hos ouk eisin), or what they are like in form" (fr. 4). The contrast
furnished by the last clause guarantees that einai here refers to the question of the gods' existence; and the verb itself might properly be
translated as "exists". 31
However, in another even more famous quotation from Protagoras
the natural reading of the verb to be is veridical: "Man is the measure
of all things, of the things that are (ton onton), that they are (has estin) ,
and of the things that are not, that they are-not (has ouk estin)" (fr. 1).
Here we have prefigured, in a slight modification of the old idiom for
truth ("tell it like it is"), the distinction between ~he intentional being-so
of judgment and statement (hOs estin) and the objective being-so of the
way things stand in the world (Tex 11m). This intuitive distinction between
the ways things are and the way they are judged to be, which Protagoras
recognizes only to deny its validity, is precisely what we find in the two
terms of Aristotle's definition of truth, where the participle (ta onta) is
used for the facts of the case, as in Protagoras fr. 1, while the finite
verb in Protagoras' formula is replaced in Aristotle by the infinitive, for
the asserted einai of thought and statement. 32 The parallel being so exact,
it is no accident that Protagoras' book was called "Truth". 33
Another early example from the philosophical literature shows how
veridical and existential values can intersect in a single occurrence, or
how an author can oscillate between the two. Melissus is conditionally
assuming what he wants to deny: the reality of phenomenal diversity.

If there really is (ei eSlI) earth and water and air and fire and iron and gold, and living and
dead, and black and white and all the other things men say are true or real (hosa phasin
... einai alethe), if these things really are (ei tauta estl), and we see and hear correctly ...
(Melissus fro 8.2).

Melissus is actually insisting upon a radical discrepancy between the way

things are and the way they seem to us. He is stating in an extreme form
the distinction which Protagoras is attacking, and Protagoras' formula
for truth could reasonably be read as a direct response to Melissus. Now
in Melissus' own statement the first esti is pretty clearly existential: it
looks like a pure existential of type 3. But the veridical undertone makes
itself heard in the summary clause "all the things men say are true (einai
alethe),', and the last occurrence of the verb in ei tauta esti is neatly ambiguous between "if these things exist" and "if these claims are true".
And in what follows, the thesis about reality or existence is repeated in
the veridical mode with a copula cC'nstruction: "these things would not
change if they were true" (or "real": ei alethe en).

14

CHARLES H. KAHN

This instability of the type 3 existential use is characteristic not only

of the early texts but also of the verb in Plato and Aristotle. 34 On the one
hand there is no doubt that, for Melissus as for Protagoras and for all
later writers of Greek, esti used alone can mean "exists" or "there is such
a thing". 35 On the other hand, the verb performs so many other functions, and its copula role is so prominent, that there is rarely any
system:ltic reliance upon the fixed sense of the verb as "exists", except
in certain special contexts such as the existence of gods and mythological
creatures. Bur when Plato wants an unambiguous expression for an
assertion of existence, he has recourse to a copula construction: dva,{ TL
"to be something" rather than P-T]OEV dVaL "to be nothing". 36 Furthermore, although Plato and Aristotle both use sentences of type 3 to assert
and deny existence, neither philosopher mentions existence as one of the
basic notions of einai. When Aristotle applies his scheme of categories
to show how Being (it is) "is said in many ways", we may prefer to
describe his various modes of being as so many different kinds of existence, or even as so many different senses of "exists". But Aristotle
does not speak in such terms, and he regularly illustrates his categorial
distinctions by copula uses of to be: "Socrates is a man", "is wise",
"is 6 feet tall", "is in the marketplace". For Aristotle as for Plato, existence is always dva,{ TL, being something or other, being something
definite. There is no concept of existence as such, for subjects of an indeterminate nature. 37 Thus the limited literary use and unstable semantic
value of type 3 existentials is reflected in the explicit philosophical doctrine that being (and a fortiori existence) is not a genos, not a definite
kind of thing.
So much by way of caveats before we turn to Parmenides, where I shall
urge that the veridical use gives us a better initial grip on the argument
than the existential does, although both are needed together with the
copula construction in order to give a complete exegesis. In a deliberate
challenge to what seems to be the prevailing interpretation, I want to
claim that for Parmenides, as for Plato and Aristotle and also in the prephilosophic usage of the verb, existence is a subordinate and not a
primary component in the concept of being. The notion of existence (or
the use of the verb meaning "to exist") must be included in our account
of Parmenides' argument, since "what is" (to on) is contrasted with
"nothing" on the one hand and with coming to be (genesis) and perishing
(olethros) on ;he 0ther. To sustain these contrasts, to on must be (a)
something rather than nothing, (b) something that is already there, and
(c) something that continues to be there, something that persists. These

THE VERB 'TO BE' AND THE CONCEPT OF BEING

15

contrasts define the sense in which for Parmenides esti means "it exists",
where the durative and locative values of the verb give some definite
shape to the claim of existence. But there is no reason, neither in the prephilosophic usage nor in the context of the poem nor in the later echoes
of the Parmenidean thesis in Protagoras (fr. 1, cited above) and Plato
(Rep. V, 476E - 477 A, cited below), to suppose that in the initial presentation of his thesis in the one-word sentence esti in fr. 2, the sense of
Parmenides' claim can be adequately captured by the translation "it
exists" . 38
How are we to construe this claim? The contrasts just cited require
that, if the argument is to be coherent, the content of what is claimed
must be such that it is (a) something rather than nothing, (b) already present, and (c) guaranteed to endure. But that gives us no clue as to where
the argument begins, or how we are to understand Parmenides' initial
presentation of the thesis so as to provide him with a plausible startingpoint. For this we must look at the context in the poem and above all at
the preceding context: the allegorical proem. 39
Parmenides' thesis (that it is and that it cannot not-be)is introduced as
the acceptable member of a pair of alternative "ways of inquiry" for rational cognition (noesal) to travel on (fr. 2.2). Where is this inquiry supposed to lead? Obviously, to knowledge and to truth, as is clear both
from the proem and from the words immediately following the thesis
(2.4: "it is the path of Persuasion, who follows on Truth"). In the
allegorical proem the voyager on the right road is a "knowing mortal",
transported by wise horses and clever escorts, the daughters of the Sun
who are leading him "to the light" (fr. 1.10). When he arrives, a goddess
promises to instruct him in everything, but first of all in "the unshaken
heart of persuasive Truth" .40
This is what Parmenides gives us as a background for understanding
his thesis that it is, and that it cannot not-be. To interpret the thesis we
must be able to say: what is the subject of the claim esti? And what is
the content of that claim? I think the subject can be specified with some
confidence, on the basis of clues from the proem and the immediate context. With these clues Parmenides makes quite clear that what the goddess holds out is a promise of knowledge, and that the path of it is must
lead to truth. Hence the understood "it" which the goddess is referring
to in the thesis must be located in the region of knowledge and truth; it
can also be identified as the goal of inquiry and the object of that quest
that began in the first verse of proem, where the horses are said to carry
the youth "as far as his desire can reach". So we C:in ciescribe the subject

16

CHARLES H. KAHN

referred to by the goddess as what our youth has come to find out, and
what will be made hIOwn to him in the revelation of persuasive truth.
Such an initial characterization of the subject as "the object of inquiry"
or "the knowable" can be taken for granted before the thesis is articulated, though any fuller characterization remains to be spelled out in
the course of the argument. 41
What does esti say about the object of inquiry that (a) can be taken
for granted as a point of departure, and (b) can justify the immediate,
categorical rejection of the negative way, that it is not? Recent interpreters, looking ahead to see what content is given to the thesis later in
the argument, propose to read esti as "it exists". But in addition to the
dubious procedure of reading a poem backwards, this view has the disadvantage of saddling Parmenides from the outset with an essentially
anachronistic notion of to be. If my interpretation of the linguistic
evidence sUPlmarized above is even approximately correct, then it is
highly unlikely that either Parmenides or his readers would understand
a bare unadorned esti as meaning primarily or predominantly "it exists".
Of course the parallel to sentence types 1 and 2 guarantees that "there
is such a thing" will be there as a background meaning for Parmenides
to rely upon. But the primary idiomatic sense of an unqualified esti in
the early fifth century can only be the veridical, in this case taken objectively for the reality as known: "it is so" or "this is how things stand".
And the logic in support of the initial thesis then becomes unassailable:
what is known or knowable must be the case and cannot not be so. "For
you could not know what is not (so)" (fr. 2.7).42 Thus Plato, when he
echoes this argument in Rep. Y, 476E-477A, has Socrates ask: "Does
a knower know something or nothing? ... Something which is (on) or
which is not (ouk on)?" To which the interlocutor replies: "Something
which is (on); for how could anything which is not (me on tl) be known?"
Plato adopts Parmenides' starting-point here precisely because he wants
to make the premises of his argument as plausible as possible. For the
ancients as for the moderns, knowledge entails truth: what is known
must be really SO.43
It is the veridical use, then, which not only provides the idiomatic
background for understanding Parmenides' stark initial esti, but also
provides the conceptual grounds for granting it,; necessary truth (best
understood as necessity of the consequence: if p is knowable, then,
necessarily, p is the case). Once this starting point has been granted, on
the basis of a veridical what is (so), Parmenides will go on to unfold the
richer implications of an esti whose full content will depend on other uses

THE VERB 'TO BE' AND THE CONCEPT OF BEING

17

of the verb, including the locative associations which justify his assumption that to eon is spatially continuous, indivisible, and sphere-like if not
spherical. Among these other properties of Parmenides' being there will
surely be some that depend upon the copula use (Being is unchanging,
for if it is F, for any F, then it can never be not-F without falling into
Not-being), and some that reflect the existential use in the sense specified
above: if Being is, it is not nothing; if it is ungenerated it is already there;
if it is imperishable it will persist. Whether Parmenides' move to these
richer senses of einai necessarily involves him in a fallacy of equivocation
is not entirely clear. We might well suspect something of this sort, in view
of the astonishing nature of his conclusions. Plato is at pains to show,
against Parmenides, that something can be X and also not be Y without
falling into nonentity; whereas Aristotle distinguishes being not only in
terms of the categories but also in terms of potency-act and substrateprivation-form (in Physics I) in order to avoid the conclusions which
Parmenides draws by taking to on as univocal. Here I suggest that
Plato's diagnosis cuts deeper into the actual structure of Parmenides'
argument. Some unclarity but no radical incoherence results if
Parmenides takes to on (I) veridically, as the objective state of affairs required for truth and knowledge, then (2) existentially, as a real, enduring
object which is the "subject" of this state, and also (3) copulatively as
being F for various F's, as well as (4) locative, i.e. spatially extended. 44
Fallacy enters only with negation, and the assumption that what-is-not
in any respect must be a Non-being pure and simple. The inference from
(I), "there is something which is the case, which is determinately so" to
(2) and (3) "there is'something which exists as an enduring subject, and
which is F" requires for its validity only the reasonable (if not inevitable)
assumption that for a state of affairs to be definitely so there must be a
definite subject with definite properties. The undeniable category-shift
from a propositional entity that is the implied subject of esti in (I) to a
substantial or thing-like entity for (2) and (3) is precisely parallel to the
shift between "if these things are true (einai alethe), ' and "if these things
exist" (ei tauta estl) in the text of Melissus cited above, and parallel also
to the cat ego rial ambiguity of einai alethe in the same text: "if they are
true" and "if they are real". Since similar shifts and ambiguities between
propositional and substantial entities occur in Plato and Aristotle too,45
it would be surprising indeed if the paradoxical esti of the earliest Greek
ontology were quite unequivocal in this respect. 46
This interpretation of Parmenides' thesis also provides a natural
historical explanation for the paradox of false statement and false belief,

18

CHARLES H. KAHN

which seems to have been popular with some sophists and which persistently recurs in Plato's Cratylus (4290), Theaetetus (189A 10-12)
and Sophist (236E, 237E), with an early variant in the Euthydemus
(283E - 284C). If speaking falsely is saying what is not (the case), and
what is not is nothing at all, then speaking falsely is saying nothing and
hence not speaking at all. There may be other dimensions to this
paradox, but the crucial move is clearly the slide from what is/what is
not as the object of true and false statement and belief to what is not as
that which is nothing at all, the non-existent - a slide precisely parallel
to the one we have identified in Parmenides' argument, and which has
its counterpart in Melissus' oscillation between doubts that the multitude
of phenomenal things really exist, on the one hand, and claims that we
do not see or hear correctly, on the other hand, or that there are not as
many things "as men say are true". Thus "true being" (to on alethinon)
for Melissus (in fro 8.5) is both (a) what really exists and (b) what is true,
as the content of true statement and belief. This ambiguity is relatively
harmless in the affirmative case, where (a) and (b) coincide (given the
failure to distinguish between facts and things). But the corresponding
negation leads to fallacy and paradox, if the denial of (b) in a reference
to the object of falsehood is also taken as a denial of (a).
Since I have treated Plato's use of esti and to on at length in a recent
article, I will here simply list my chief conclusions concerning Plato's ontological vocabulary in the preliminary and mature statements of his
theory of Forms.
(1) In the so-called Socratic dialogues, the first philosophically relevant use of einai is its occurrence in connection with the What-is-X?
question of Socratic definition. Examples: Laches 190B7-C6: "we ought
to possess knowledge of what virtue is" eidenai hoti pot' estin arete); "we
say we know what it is" (eidenai auto hoti estin); "But if we know, then
we can say what it is" (ti estin). In the Euthyphro we find the contrast
between such a "whatness" and other attributes of a thing hardening into a terminological distinction between ousia, "essence", i.e. the content
or correlate of a true answer to the what-is-it? question, and pathos, any
other property or attribute (Euthyphro IlA7-8). Here ousia is simply a
nominalization for the verb estin in the what-is-it? question (to hosion
hoti pot' estin, IlA7). In such contexts the verb is syntactically the
copula, but logically or epistemically strengthened by the context of use
into what we might call the definitional copula or the is of whatness,
which aims at locating the true, proper, deep or essential nature of the
thing under investigation.

THE VERB 'TO BE' AND THE CONCEPT OF BEING

19

(2) The first trace of a more technical use, growing directly out of the
definitional copula, appears in the Lysis, where the attempt to explain
what makes something dear, friendly, or beloved (phi/on) leads to the notion of "that which is primarily dear (ekeino ho esti prOton phi/on), for
the sake of which we say that all other things are dear" (219C 7), these
other things being potentially deceptive "images" (eidola) of "that
primary thing, which is truly dear" (ekeino to prOton, ho hos alethos esti
philon (219D 4). What is new here is (a) the use of the definitional copula
as a kind of proper name for the concept under discussion, or for its
primary instance: what is (truly, primarily) X, prefiguring the canonical
reference to the forms in later dialogues as to ho esti X, and (b) the
veridical strengthening of the copula in "what is truly (alethos) dear" cf.
tpC)..OIl Of T~ olin at 220 BI and B4), by contrast with the "images" which
are only "verbally" dear (220B 1), i.e., said to be dear because of their
relation to the primary case (219D 1). Just as the Euthyphro adds precision to the is of whatness by a version of the essence-accident distinction,
so the Lysis reinforces the metaphysical import of a privileged use of this
formula by introducing a contrast between Reality and Appearance, between what really is F and what is only an image or a putative instance
of F.
(3) In Plato's first explicit statement of his mature doctrine, in the
speech of Diotima in the Symposium, the Beautiful itself is announced
as the goal of a process culminating in a final study "which is (ho estin)
the study of nothing but that Beautiful itself", where the student will end
by knowing "that which itself is beautiful (auto . .. ho esti kalon, 211 C
8). This is the formula of the Lysis, with its veridical force (' 'he will know
what is truly beautiful") again underscored by contrast with appearance
and images (2IIA 5, 212A 3). But in this case the formula unmistakably
refers to the Form. For here we have a new (or newly formulated) doctrine in which, for the first time, Plato provides his specimen Form the
Beautiful with a definite ontological status, based upon the Eleatic opposition between eternal, unchanging Being (aei on) and inconstant,
perishable Becoming (211A I - 5). In this context the participle on is used
both existentially ("it is forever") and as copula ("it is not beautiful in
one respect, ugly in another,,).47 It is precisely in such Parmenidean contexts, where Being is contrasted with Becoming, that it seems most
natural to regard to on in Plato as existential, though the aspectual value
is that of the stative copula.
(4) In the Phaedo and Republic, where the doctrine for Forms is
systematically developed, the philosophical uses of einai become too

20

CHARLES H. KAHN

diverse for cataloguing here. I would emphasize only that (a) the veridical
overtones of to on, used roughly as a synonym for "truth", are predominant in the initial presentation of the Forms in both dialogues: in Phaedo
65B - 67B, to on and ta onta occur together with aletheia and to alethes
for the "true reality" which the philosopher's soul desires and pursues
(and which is identified as Forms at 650); in Republic V, 476A the Forms
are introduced by the contrast of Reality and Appearance: each of them
is one but appears (phainesthal) as many (A7); and their ontological
status is again expressed by a use of to on for the object of knowledge;
that which is wholly real (pantelos on) is wholly knowable; that which
is in no way real (me on medame) is in every way unknowable" (477A
3). The veridical-epistemic contrast between Being and Seeming
(phainesthal) serves to distinguish the Forms and "the many"
throughout this passage (cf. 479A 7 - BIO).
(b) The formula auto to ho esti (ison), "that itself which is (equal)",
familiar from the Lysis and the Symposium, is gradually developed in the
Phaedo from idiomatic phrases into a semi-technical designation for the
Forms (notably at 7502 and 7804, recalled at 9308), with a parallel use
of ousia for the distinctive being, essence, or reality of the Forms. 48 The
same designation is used to reintroduce the Forms into the central
epistemological passage of the Republic: the Beautiful itself and the
Good itself and the other unique entities, "each of which we call what
it is" (ho estin hekaston prosagoreuomen, 507B 7). In this designation
the predicative form is Fis taken separately, independent of all subjects,
and made itself the target of the question 'what is it?' Thus ho esti serves
in Plato, like to ti en einai in Aristotle, for the objective essence or definitional content given in a correct answer to the question "what is Fl" for
a given predicate F. The syntax of the verb is still that of the copula, but
its predicative role is reinforced now not only by the definitional search
for the true nature of a thing but also by the ontological dualism of
Plato's neo-Parmenidean opposition between Being and Becoming, the
One and the Many, the Intelligible and the Visible. The specifically
Platonic use of einai in the doctrine of the middle dialogues thus consists
in a convergence between (i) the definitional copula from the what-is-it?
question, (ii) the veridical Being that contrasts with Seeming, and (iii) the
stative-invariant Being that contrasts with Becoming and Perishing. An
unqualified use of to on, einai, or ousia may bear any and all of these
connotations. 47 The predicative syntax is always latent if not manifest.
The existential value appears above all in (iii), but even here the copula
use, on which the stative-mutative contrast of Being-Becoming is found-

THE VERB 'TO BE' AND THE CONCEPT OF BEING

21

ed, may reappear at any moment. so The Platonic concept of Being is constituted not by a fusion of copula and existence but by the union of
timeless-invariant Being (in contrast with Becoming) and cognitivelyreliable, veridical Being (in contrast with Appearance), both of them expressed or expressible in copula predications, but most rigorously distilled in the frozen auto ho esti version of the definitional is of whatness
in application to the Forms.
(c) In the Eleatic introduction to the doctrine of Forms at the end of
Republic V, Plato has moved beyond Par men ides in a number of interesting ways. First, by accepting an intermediate "mind" reality between Being and Not-Being as object for the cognitive state of opinion
(doxa) between knowledge and ignorance, Plato has provided an ontological basis for change and becoming, which was simply the domain
of error and falsehood for Parmenides: Plato thus accounts for the
possibility of true opinion short of knowledge by giving it an object of
its own. In the second place, by his development of the copula use for
parallel designations of Forms (as "what is F") and particulars (as "what
is and is not F"), Plato opens the way to a philosophical analysis of
predication and the diverse uses of to be which he will pursue in the
Sophist and elsewhere, and which will lead eventually to Aristotle's
theory of categories and his distinction between essential and accidental
predication. On the other hand, where Plato in the Republic has not
moved substantially beyond Parmenides is in his conception of the negation of Being as what is not in any way (to medame on); for this is indescribable and unintelligible, as Parmenides had insisted and as Plato
in the Sophist will finally agree. The paradox of false statement and false
belief will haunt Plato until he works out a way to negate the "being"
of truth without falling into this hopeless region of blank non-entity.
This is far enough to pursue a project that began as linguistic prolegomena to Greek ontology and not as a history of the subject. In conclusion, I want to say a word against the charge of linguistic relativism,
in so far as it claims that ancient ontology was vitiated or distorted by
the accidental possession of a verb that combines the functions of existence and predication. It is certainly true that the verb einai serves a
multitude of functions that are rarely combined in languages outside of
Indo-European. And if Greek-ontology had begun with a radical confusion between existence and the copula, then its first task should have been
to distinguish the two, a task that neither Plato nor Aristotle undertook.
On the contrary, both of them systematically subordinate the notion of
existence to predication; and both tend to express the former by means

22

CHARLES H. KAHN

of the latter. In their view to be is always to be a definite kind of thing:

for a man to exist is to be human and alive, for a dog to exist is to be
enjoying a canine life. Instead of existence, which is a tricky notion at
best, it was another use of to be that gave Parmenides and Plato their
philosophical starting point: the veridical use of esti and on for "the
facts" that a true statement must convey. Thus the Greek concept of Being takes its rise from that naive, pre-philosophic notion of "reality" as
whatever it is in the world that makes some statements true and others
false, some opinions correct and others mistaken. But this notion of what
is as whatever distinguishes truth from falsehood, reliable information
from idle rumor, is surely not peculiar to Indo-European. Some such notion will be functioning in any language in which questions can arise concerning what is true and what is false, what is knowledge and what is error. This notion is so essential to the basic descriptive or informative use
of language that it is bound to be in some sense a linguistic universal.
What is peculiar to Greek (and to Indo-European) is that a locution for
"reality" in this sense should be provided by a verb whose primary function is to express predication and sentencehood for statements of the
form X is Y. So doctrines of Being first arose in Greece in connection
with the question: what must reality be like for knowledge and informative discourse to be possible, and for statements and beliefs of the
form X is Yto be true? In principle, the question concerning knowledge
and informative discourse is one that might have been posed in any
language; the question about the sentential form X is Y reflects a point
of view more specifically Greek. If anyone believes that it was a disaster
for Greek theories of knowledge and reality to be concerned, from the
beginning, with problems of predication and with the propositional
structure of language and thought, let him blame the verb to be.
NOTES
I wish to dedicate this review of my own work on einai to the memory of G. E. L. Owen.
Rereading since his death his major articles on Greek ontology I see more clearly than
before how he was a powerful ally in my campaign against the uncritically "existential"
interpretation of is in Plato and Aristotle. In many cases we came by different routes to
similar conclusions; in some cases I have been echoing his formulation without realizing
it. Like all workers in this vineyard, lowe him a great debt of inspiration and
encouragement.
I Foundations of Language 2 (\966), 245 - 65.
2 The Verb 'Be' in Ancient Greek (Reidel, 1973), The Verb 'Be ' and Its Synonyms, Part
6, ed. by J . W . M. Vtrhaar (= Foundations of Language, Supplementary Series Vol. \6).
This will be cited below simple as Be' .

THE VERB 'TO BE' AND THE CONCEPT OF BEING

23

The earliest clear statement of the dichotomy known to me is that of J. S. Mill in his
System of Logic (IB43), I. iv. I, who attributes it to his father, James Mill, in the Analysis
of the Human Mind{IB29). But the philologists were already using this dichotomy as early
as G. Hermann in IBOI. (See the quotation in 'Be', p. 420, Note I.) Hermann in turn appeals to "what logicians call the copula", and is apparently dependent on the logic of
Christian Wolff, ('Be', p. 423 with Note 5).
4 References to Brugmann, Delbriick, Meillet, Kiihner-Gerth, and Schwyzer-Debrunner in
'Be', p. 199, Note 21. Compare John Lyons, Introduction to Theoretical Linguistics
(196B), p. 322: "Even in the Indo-European languages the copulative function of 'the verb
to be' appears to be of secondary development".
5 For the argument see 'Be', pp. 199 - 207.
6 Thus to explain the stative value of copula*es- we would have to posit an original sense
"to stay, remain" which is unattested, and turns out to be only a projection of the 'be''become' contrast for the copula. The priority of the copula uses is partially clarified
below; for fuller discussions see Chapter VIII of 'Be', especially pp. 395 - 402,407 -409.
For methodological remarks on the claim of priority here, see 'On the Theory of the Verb
To Be', in Logic and Ontology ed. by M. K. Munitz (New York, 1973), pp. 17 - 20.
7 See in particular E. Benveniste, 'Categories de pensee et categories de langues' and
, "~tre" et "avoir" dans leurs fonctions linguistique', in Problemes de linguistique
generale, pp. 63 -74 and IB7 -193.
8 , "Being" in Classical Chinese', The Verb 'be' and Its Synonyms, ed. by J. W. M.
Verhaar, Part I (Reidel, 1967), p. 15. Compare Lyons, Introduction to Theoretical
Linguistics, pp. 322f.
9 See the reference to Mill's Logic in Note 3, above.
10 There has recently been a noticeable trend away from the Mill- Russell view that "is"
has different senses, which the Greek philosophers should have distinguished. See, e.g.,
Benson Mates' suggestion that Plato's different uses of "is" can all be understood on the
basis of a single, univocal use of the copula: 'Identity and Predication in Plato',
Phronesis 24 (1979), 211- 229. And compare Jaakko Hintikka's paper in this volume. In
my opinion, the question whether "is" has different meanings or only different uses cannot
be answered without confronting certain very deep problems in the theory of meaning,
which is ultimately a part of the theory of knowledge. For example, are senses of a word
distinguishable as a matter of logical form and conceptual truth, independently of any factual question as to the kinds and natures of the things to which the word is applied? Up
to a point, linguistics can settle questions of syntax and sentence structure. But epistemology and metaphysics must be called in to decide how linguistic "meanings" are related to
the nature of things or to our "conceptual scheme".
II See, e.g., Lyons, Introduction to Theoretical Linguistics, pp. 322f.
12 C. J. F Williams, What is Existence? (Oxford, 19BI), echoing Quine, Word and Object, pp. 96f. This view of the copula ignores the distinction between an 'is' of identity and
of predication, a logical distinction which is not reflected in the syntax of the verb and is
not plausibly regarded as a difference in meaning for 'is'. My argument for this view (in
'Be' p. 372, Note I and p. 400, Note 33) is defended by C. J. F. Williams, op. cit., pp.
10-12. For criticism of this view on philosophical rather than linguistic grounds, see Ernst
Tugendhat, 'Die Seinsfrage und ihre sprachliche Grundlage', in Philosophische Rundschau 24 (I977), 164.
13 Logica 'Ingredientibus', ed. by Geyer, p. 351, cited with other passages from Abelard
in 'On the Terminology for Copula and Existence', in Islamic Philosophy and the Classical
Tradition. Essays presented . .. to Richard Walzer (Cassirer, 1972), pp. 146 - 149.
3

24

CHARLES H. KAHN

14 For the notion of truth claim, see 'Be', pp. 186f; 'Theory of the Verb', pp. Ilf. Compare
Quine's statement: "Predication joins a general term and a singular term to form a
sentence that is true or false according as the general term is true or false of the object, if
any, to which the singular term refers" (Word and Object, p. %). This makes clt:ar the
sense in which predication is more than a syntactic notion.
15 For the stative-mutative contrast see' Be' , pp. 194 - 198, following Lyons, Introduction
to Theoretical Linguistics, pp. 397ff. Compare Benveniste, Problemes de linguistique
generale, p. 198: "etre' ... est ... un verbe d'etat, ... est meme par excellence Ie verbe
d'etat" .
The durative aspect emerges as a distinct "sense" of the verb in the Type I ("vital") use
with persons, where einai means "continue (in life), survive": eti eisi "they are still alive",
theoi aiei eontes "the gods who live forever". See 'Be', 241 ff.
16 For the verbs of posture as static be-replacer, see' Be', pp. 217 - 219.
17 See my own exposition of this view in 'Be', pp. 225f, 375 - 379, with the work of J.
Klowski cited there (p. 375n.) from Archiv fur Geschichte der Philosophie 49 (1967), 138ff.
18 'Be', pp. 201ff.
19 Cf. Thucydides VII. 8.2: tpo{3ov,uvos 010 1'17 oi 1rtl'1rO,uVOL ... ou Ta DVTCl Cx1ra-y-yi>V.w(lIv
"(Nicias) fearing that this messengers might not report the facts (sent a written letter)".
For additional examples, see 'Be', pp. 335 - 355. The veridical "is" appears in
Shakespeare, e.g., King Lear IV.vi.l41: "I would not take this from report: it is,/ And my
heart breaks at it". The idiom is still alive and well in contemporary speech: "Tell it like

it is" .

'Be', p. 332, Note 2.

For the distinction between the intentional it-is-so of judgment and statement and the
objective being-so of things in the world, see my article in Phronesis 26 (1981) 126f. This
corresponds to the distinction between the roles of infinitives and participles, respectively,
in Aristotle's definition of truth: "to say of what is (to on, objectively) that it is (einai, intentionally), to say of what is not (to me on) that it is not ... " (Met. r, 7, 101lb26).
22 See the passages from Phaedo 65B - 66C cited in Phronesis (1981), 109. Cf. the example
from Thucydides in Note 19 above, and passages where 0 fWV M-yos means "the true
report" (in' Be', p. 354).
23 Met. t. 7, discussed in Phronesis (1981), 106f. Other examples of veridical copula in
'Be', pp. 356 - 360.
24 Thus I ignore here two types (the "vital" use in Type I and the verb of occurrence in
Type V) counted as existential in 'Be', pp. 239ff, 282ff. For Type III, see the next note.
25 In my existential Type III, which represents the plural of (I) above, instead of the verb
in initial position we often have a kind of quantifier-word like "many" or "others":
1ro>v'al -yae Cxva UTeaTov tiUL xi>-.tVIJOL "For there are many paths up and down the encampment" (Iliad X.66). Further examples in 'Be', 261ff.
26 For examples, see 'Be~, 277ff.
27 Examples ln 'Be', 300ff.
28 See 'Why Existence Does Not Emerge as a Distinct Concept in Greek Philosophy', Archiv fur Geschichte der Philosophie 58 (1976), 323 - 34.
29 The one case where einai provides .m independent predicate is my Type I "vital" use
for persons, where ouketi esti means "he is no longer alive". See Note 15 above.
30 Further discussion in 'Be', 301, 303 - 6, 320 - 323, 326 - 330.
31 With this "pure existential" use contrast a typical non-technical existential in
Herodotus, with a locative restriction as in type I above: "There is no stag or wild boar
in all of Libya" (IV, 192.2, cited in 'Be', p. 327).

20
21

THE VERB 'TO BE' AND THE CONCEPT OF BEING

25

See Note 21 above.

For this interpretation of Protagoras' homo mensura formula, see Phronesis (1981),
117 -119.
34 See the star example of this instability in Post. An. I!. I - 2, where the question ei esti
is initially existential ("is there or is there not a centaur or a god"), but then gets divided
into "particular" (epi merous) and "unqualified" (haplos) cases, where the unqualified
cases are still more or less existential (fi fonv ~ /l~ OfA~VT/ ij vu~) "whether or not there is
a moon or whether it is (?) night "), but the particular cases are not: "is the moon being
eclipsed? or is it waxing?" (90" 1 - 5). This problem has been much discussed. (See Ross'
commentary, pp. 610 - 612; A. Gomez-Lobo in Review oj Metaphysics 34 (1980), 71 - 89.)
Barnes (Aristotle's 'Posterior Analytics, p. 194) takes the particular or "partial" (epi
merous) question ei esti in Chapter 2 to be a reformulation of the hoti question in Chapter
1: " 'X is Y' says that X is 'partially' because 'X is' is a part of 'X is Y' ". Even if this
is right, it would show that Aristotle did not consistently read ei esti (or hoti est!) as existential. The copula construction X is Y (fi rae ion rt ij /l~ fan n at 90" 4) is treated as a
special case of the unqualified ("existential") esti. For what seems to me the most plausible
explanation of this rather baffling fact, see the notion of predicative complex borrowed
from Mohan Matthen in Note 46 below. According to this suggestion, Aristotle thinks of
X is Yas equivalent to the YX exists.
3S Aside from the problematic example from Parmenides to be discussed in the text, the
only type 3 existentials older than Melissus and Protagoras seem to be in the fragments of
Zeno, where the syntax i~ uncertain. In frs. I and 3, fi 1rOMa fonv can be read either as
"if (the) many exist" or as "if (beings) are many". In fro 3 aft rae fUeCX /lfrcx~V
rwv ovrwv iort "there are always other things between the beings" the locative qualification recalls my Homeric Type III, the plural of type I above. (Compare the locativeexistential from Herodotus in Note 31 above.) But fiT/ and d Of fonv at the beginning of
fro I seem to be straight-forward cases of type 3 existentials.
16 See passages cited in Phronesis 1981, p. 130, Note 17.
37 Here I am agreeing with, and in part echoing, a series of studies by G. E. L. Owen.
Compare: "There is [for Aristotle) no general sense to the claim that something exists over
and above one of the particular senses" ('Logic and Metaphysics in Some Earlier Works
of Aristotle', in Articles on Aristotle Vol. 3, ed. by Barnes, Schofield, Sorabji = p. 165
in Aristotle and Plato in Mid-Fourth Century, ed. by DUring and Owen); "To be, then,
is always to be something or other" (,Aristotle on the Snares of Ontology', in New Essays
on Plato and Aristotle ed. by R. Bambrough, pp. 76ff); "The concept of being that he [viz.
Plato) takes himself to he elucidating here [in the Sophist) is not that of existence" ('Plato
on Not-Being' in Plato. A Collection oj Critical Essays I. Metaphysics and Epistemology,
ed. by G. Vlastos, pp. 240f).
38 The view that I am opposing is defended by D. Gallop in ' "Is" or "Is not"?', The
Monist 62 (1979),61 ff and J. Barnes, The Presocratic Philosophers, Vol. I, pp. 161 ff. BOlh
Gallop and Barnes are following G. E. L. Owen, 'Eleatic Questions', Classical Quarterly
10 (1960), 84 - 102. It is ironicallhat these scholars should not have recognized the extent
to which Owen's own work on einai in Plato and Aristotle (cited in the preceding note) has
succeeded in undermining their assumption that the modern notion of existence is an appropriate instrument for capturing the sense of 'to be' in Greek philosophy. Owen himself
later expressed some qualms about his use of the "conventional choice" between copula
and existence to decide in favor of the traditional, uncriticized notion of existence for intepreting esti in Parmenides' thesis. (See 'Plato on Not-Being', p. 225.)
39 My interpretation here follows the main lines of the view developed in greater detail in
32

33

26

CHARLES H. KAHN

'The Thesis of Parmenides' and 'More on Parmenides', Review of Metaphysics 22 (1969),

700- 24 and 23 (1969),333 -40. But my earlier interpretation of esti and ouk esti in the
thesis now seems to me too schematic.
40 Despite its recent reprinting by Tanin and other editors, I think that' A>4IEL'15 fU}(V}(Af05
in fro 1.29 is indefensible, both in terms of MSS. evidence and the rules for nounformation. (On this see Gnomon 40 (1968), 124.) The reading firndlfos is better attested
and is guaranteed by the context: "persuasive truth" answers, with chiastic reversal, to
pistis alethes in the next verse . The thought ("trust only in the truth") is fundamental, and
recurs in 2.4 (the Way of Persuasion, who follows Truth) and 8.50 (the end of the pistos
logos concerning truth).
41 G. E. L. Owen identified the subject as "what can be talked or thought about" ('Eleatic
Questions', p. 95) without reference to the proem or the context. Gallop first suggests the
vaguer subject "a thing", but ends by following Owen (Monist (1979),68 and 71). Barnes
(The Presocratics I, 163) approximates to my identification by taking the subject of esti to
be "the implicit object of dizesios", i.e., the object of inquiry .
42 "Nor can you point it out" (oute phrasais, 2.8), i.e., you can give no reliable information about what is not the case. Of course you can say what is not so, and hence some
stronger notion than mere statement seems to be implied by Parmenides' claim "it is not
sayable (ouphaton) . .. that it is not'" in 8.9. (The same problem arises for the rival interpretation of esti as "exists", for of course we can talk about what does not exist.) It was
perhaps to strengthen this side of Parmenides' thesis that the paradox of false statement
was first formulated.
43 Thus J . Hintikka rightly suggested that a rule like" 'm knows that p' entails 'p' " might
be called "Parmenides' law", on the basis of fro 2.7. See his Knowledge and Belief, p. 22,
Note 7.
44 Except for the locative-spatial implication, which does not hold for the Forms, Plato's
adaptation of Parmenides' argument in Rep. V, 476 Eff follows the same steps, from the
veridical on as object of knowledge (to on gnonai hos echei at 478 A 6) to the existential
whose negation is nothing (meden at 478 B 12) and the copula at 479 B 9: "is there any
one of these many things which is any more than it is not what one says it to be?"
In view of this close affinity between copula and veridical uses (since the simpler examples of facts or states of affairs can always be framed in an X is Y construction), there
is no real incompatibility between my reading of the Parmenidean esti and Mourelatos'
proposal to regard the thesis as a predicative sentence frame: .. - - - is - - -" (with
.. - - - is not - - -" for the negation). The basic copula function of esti in Greek will
assure that, on any reading of the thesis, this sentence form is immediately felt to be implied. And I would not exclude Mourelatos' notion of "speculative predication" (as he
says, this is a narrower concept which "falls within the range" of the veridical). But I do
not think there is enough. early evidence for "quiddity" uses of einai to justify such a
restriction on the primary reading of esti. See A. P. D . Mourelatos, The Route of
Parmenides, pp. 55 - 59.
45 For Plato, see the preceding note; for Aristotle see Note 34 above. These remarks represent my answer to the second of Gallop's three objections to my interpretation, namely that
a state of affairs is "of the wrong logical type to serve as the bearer of such attributes as
'ungenerable' ... and 'immovable' ", which belong rather to a thing-like entity (Monist
(1979),66). I agree, and this might stand as an objection to Parmenides' argument. But
it counts as an objection to my interpretation only if one assumes that Parmenides is in-.
capable of overlooking such a distinction of logical type (as Plato and Aristotle certainly
were not!). Much the same holds for Gallop's first objection, that a premise which presup-

THE VERB 'TO BE' AND THE CONCEPT OF BEING

27

poses that knowledge of the truth is possible is too weak to support Parmenides' argument,
since "a sceptic might well respond, that no one knows anything, precisely because there
is no truth to be known" (ibid). This objection seems to me to mistake Parmenides for
Descartes. Why should Parmenides be thought of as arguing against a sceptic? His argument is about how to get to the truth and what one will find there, not about whether there
is any such thing.
Gallop's third objection is more substantial: if we start by rejecting "what is not the
case" as an object for knowledge, how do we get to the rejection of non-existence that is
required to disprove generation and perishing? (ibid., pp. 67 and 72). I agree that we must
find in to me eon a sense of not-being which is equivalent to "nothing at ail", and if this
is what is meant by an existential eon, then Parmenides' Being must be existential. But, as
suggested above, if to eon as a determinate state of affairs is understood to contain or imply
a real ("existing") subject and definite attributes, and if the negation (to me eon) is
understood as denying everything contained in or implied by to eon, then "what is not"
must be construed not as a well-defined, unrealized state of affairs but rather as a blank
non-entity: no subject ("what does not exist") with no attributes ("is not P' for every Fl.
It is obvious that Not-being so understood must turn out to be not only unknowable but
indescribable. Plato will defuse Parmenides' argument precisely by distinguishing this
hopelessly unqualified Not-being from the more precisely defined not-being-F for various
values of F.
46 The most enlightening explanation known to me for the easy shift from propositional
to existential and copulative construals of einai in Greek 'philosophy is the notion of a
"predicative complex" proposed by Mohan Matthen in an unpublished paper. Matthen
defines a predicative complex as "an entity formed from a universal and a particular when
that particular instantiates the universar'. Thus artistic Coriscus is such an entity, which
"exists when and only when Coriscus is artistic". In grammatical terms, a predicative complex (or rather, its linguistic expression) is the attributive transform of an ordinary copula
sentence: corresponding to X is Y we may assume the existence of a logically equivalent
predicative complex, the YX exists. Thus for (I) Socrates is healthy we have the corresponding (2) The healthy Socrates exists, where the truth conditions for (I) and (2) are assumed
to be identical. Furthermore, truth conditions will also be the same for the veridical
transform of (I), namely (lA): It is the case that Socrates is healthy. Aristotle in Met. 1!t..7
slides effortlessly between (1) and (lA). (See 1017" 33 - 35, as interpreted in Phronesis
(l9SI), 106f) Now if (I) is transformed as (2), we see how the copula-veridical-existential
slide can seem so natural in Greek, since all three formulations are logically equivalent. I
believe this construal (following Matlhen) captures something quite deep, and quite strange
to us, in the use of einai by the Greek philosophers. And it shows why our conventional
dichotomy between existence and copula imposes a choice upon the interprett:r which corresponds to nothing in the Greek data. Also, our difference in "logical form" between propositional (fact-like) and substantial (thing-like) entities as subject of einai will reappear
in this conception simply as a difference in formulation between (I A) and (2).
47 For the double construal of on at Symposium 21 IA I, see Phronesis (l9SI), lOR.
48 First in an idiomatic variant at 65D 13 - E I, then progressively from 74B 2 to 75D and
7SD. See Phronesis (1981), 109- Ill.
49 For a convergence of veridical (Being versus Seeming) and "existential" values (Being
versus Becoming), see Rep. VI, 50SD 5 - 9, where to on is first paIred with atetheia, as
radiating the light of rational cognition, and then contrasted with to gignomenon te kai
apollumenon, the source of darkness and inconstant opinion. This is of course compatible
with a slightly different contrast in the following section (50SE - 509B), where the role of

28

CHARLES H. KAHN

the Good as cause of truth and knowledge is distinguished from its role as source of Being
(to einai te kai ten ousian) for the objects known. There Being for intelligible things is
presented as parallel to generation (genesis) and growth for visible things and must refer
to the stable existence of the Forms as appropriate objects for knowledge. (Shorey renders
einai ... ousia as "existence and essence", thus recognizing that both ideas are in play;
but it would be a mistake to look for any fundamental difference in sense between the verb
and the noun. At 479C 7 he renders ousia alone as "existence or essence", again rightly.)
so See, for example, the double syntax of aei on cited in Note 47. The stative, and hence
potentially predicative rather than strictly existential force of expressions like (TO) ov ad
is clearly indicated by the alternative formula with hELv: e.g. TO aft Xo/Ta
TO/UTa WUo/VTWS fXOV for Forms as object of knowledge at Rep. VI, 484B 4 (cf. Phaedo 79A
9, 80B 2, etc.), which is immediately picked up by TO OV with veridical overtones (T~ ovn)
at 484C 6: "those who are veritably deprived of the knowledge of the veritable being of
things" (Shorey). Thus at 485B;2 ExfLVT/ ;, ouuLO/I;, ad oJu<x is first of all the object of the
philosophical eros for knowledge, and at the same time contrasted with "ousia" that is made
to wander by coming-to-be and passing-away". Of course this convergence of veridical and
stative-existential values for einai is systematically motivated by Plato's theory: the Forms
are reliable objects for knowledge and truth just because they are eternally invariant. (And
in technical contexts this invariant being-what-they-are will be expressed by strong copula
uses such as to ho esti.) It is just this union of" true reality" plus "eternally stable reality"
that is conveyed by to on and ousia throughout Rep. VI - VII, e.g. (following on the
passages just cited) "the spectacle of all time and all ousia" at 486A 8, thought which is
naturally led t1f't rilv TOU OVTOS iOi<xv lxaurov at 486D 10, the soul which is to have an adeQuate and complete grasp of to on at 486E 2.

Dept. of Philosophy,
University of Pennsylvania,
Philadelphia, PA 19104, U.S.A.

BENSON MATES

IDENTITY AND PREDICATION IN PLATO

Among the Platonic statements that have most agitated his commentators, from Aristotle's time down to the present, are those in which he
seems to be saying (and with great confidence, too, as though there were
no question about it) that beauty itself is beautiful, justice itself is just,
largeness is large, piety is pious, and the like. On the one hand, these
statements are considered by many to involve some sort of categorymistake or serious ambiguity: beauty itself, they say, is not the sort of
thing that can be beautiful, at least not in the same sense in which people,
statues, paintings, or pieces of music are beautiful. And likewise with
justice itself, largeness itself, and the other Ideas. On the other hand,
though, there is the awkward fact that these so-called "selfpredications" cannot be lightly dismissed as mere lapsus linguae on the
part of our author, for they seem essentially related to his doctrine that
each Idea is a paradigm or perfect exemplar for the particulars that fall
under it; beauty itself is said not only to be beautiful, but to be the most
beautiful thing of all.
In recent times this situation has been analyzed on the basis of the
assumption that the verb "to be" has at least two senses, viz., the
predicative sense, as in "Socrates is human", and the identity sense, as
in "Socrates is the husband of Xanthippe". Plato's critics castigate him
for being unaware of the distinction, while his defenders believe that he
was perfectly well aware of it and that the allegedly self-predicative
statements are to be understood as assertions of identity. In this paper
I wish to investigate the possibility that the assumption is false, and that
consequently neither the attacks nor the defenses that are based upon it
are well-founded. J
1. THE THIRD MAN ARGUMENT

A convenient point of entry to the matter is the notorious Third Man

argument, which, though it has been discussed in the literature over and
over again, still has a few things to teach us. One version of this argument
occurs at Parmenides 132Al - B2, and it is upon this that I wish to focus

29
s. Knuuttila and J. Hintikka (eds.), The Logic oj Being,
1979 by Van Gorcum, Assen, The Netherlands.

29-47.

30

BENSON MATES

attention. 2 Parmenides addresses young Socrates:

"This, I suppose, is why you consider that each form is one: whenever a number of things
seem to you to be large, some one idea no doubt seems to you, as you view them, to be
the same in all of them; whence you think that the large (TO p.f-ya) is one".
"What you say is true", he replied.
"But what about the large itself and the other large things; if in the same way you mentally view all of them, will not some one large (Ell TL p.f-ya) again appear, by which (~) all these
appear large?"
"Evidently" .
"Therefore another form of largeness (p.f-yfIJoVS) will show up besides that largeness that
was already there and the things participating in it; and on top of these yet another one,
by which (~) all these will be large. And no longer will each of the forms be one for you,
but infinite in number".

Now the first thing to observe here is that the point of Parmenides'
argument is not, as has often been erroneously said, that certain assumptions lead to an 'infinite regress' (for there is nothing per se wrong with
an infinite regress, anyway), but simply that Socrates' admissions are inconsistent with the principle:
(1)

Each of the forms is one.

(Let us postpone for a moment the question of what this principle

means).
Indeed, Parmenides attacks (I) throughout this portion of the
dialogue. Thus in the section immediately following the quoted passage
Socrates seeks to escape the net by raising the possibility that each of the
forms may be a thought, existing only in a mind, and in this way "each
would be one and would no longer be subject to the consequences just
now mentioned" (emphasis supplied by Plato with the particle ')'f). 2 And
in the section immediately preceding our passage it is likewise clear that
denials of statements like (1) are what Parmenides is endeavoring to
prove and Socrates cannot accept:
"Do you think that the whole form, being one, is in each of the many, or what?"
"Why not, Parmenides?" said Socrates.
"Well, being one and the same it will be separate from itself".
"Not if," he said, "just as day, which is one and the same, is simultaneously in many
places and is nevertheless not separate from itself, so each of the forms were one and the
same in all at once."
"You are quite ready. 0 Socrates," he replied, "to make what is one and the same to
be in many places at once, as if spreading a sail over a number of people you sh')uld say
that one thing as a whole was over many. Is not that the sort of thing you intend?"
.. Perhaps", he said.
"But would the whole sail be over each person, or only a part over one, another part
over another?"

IDENTITY AND PREDICATION IN PLATO

31

"Only a part."
"Then the forms themselves would consist of parts, 0 Socrates, and the things participating in them would participate in parts, and in each of them there would no longer
be the whole but only a part of each form."
"So it seems."
"Are you willing, then, Socrates, to say that our one form really has parts and still is
one?"
"Not at all", he said.

Thus the question at issue throughout is, in Platonic language,

whether each idea is one. This point is important, because the Greek
sentence I have translated as (I) could perhaps also be rendered as
(2)

There is one form in each case,

which is how it was formerly taken by Professor Vlastos, following

Cornford. 4 But in the context surrounding the passage under discussion
there are several occurrences of the same Greek phraseology that cannot
be taken in the manner of (2).5 Therefore, if we are to preserve the form
and coherence of Plato's argumentation it seems that we must prefn (I)
to (2). Maybe it can be shown that sometimes (1) means the same as (2);
if so, so much the better. But to vacillate between the two in translation,
as Cornford does, is to obscure the structure of the argument.
As I interpret the passage, then, Parmenides takes Socrates to be
holding (I), and he purports to prove (on the basis of other premises to
which Socrates agrees) the negation of this. His argument proceeds by
considering a particular instance, the form of largeness, i.e., the large,
and by deducing (in effect) the negation of
(3)

The large is one.

And, of course, the negation of (3) immediately implies the negation of

(I), which is what is explicitly stated.
What do statements like (I) and (3), and, in general, statements of the

form "A is one", mean for Plato? In my opinion this is a very difficult
question, which has never been satisfactorily answered and can be approached only through a careful study in which one would notice, among
other things, what conclusions Plato is willing to draw from such
statements and what other statements he regards as implying them. I
presume, with most others who have considered the passage we are
studying, that when Plato says that each form is one, he does not intend
merely to express the apparent triviality that each form is one form. In
particular, I presume that (3) does not mean that the form of largeness
is one form. On the other hand, the best I have to offer toward an ac-

32

BENSON MATES

count of what these statements do mean is the claim that part of the
meaning of, e.g. (3), is this:
(4)

It is not the case that there are two different forms Fand F',
such that something is large by virtue of F and something is
large by virtue of F' .6

By refuting (4) Parmenides considers himself to have refuted (3) and

hence also to have refuted (1), which is the fundamental principle under
attack. Thus, as the argument appears to me, the little pronoun [.J ('by
virtue of which') is of crucial importance; for while it would be trivially
true that any form by which large things were large would be one form,
the Platonic view at issue, which is expressed by (3), implies that there
is only one such form.
Why not accept (2) as a statement of the thesis Parmenides purports
to refute? We would then need to explain the sense of "in each case".
I cannot go along with those who find in Plato a distinction between the
forms, on one hand, and so-called "characters" or "properties", on the
other, and who then explicate (2) as "there is exactly one form corresponding to each character". 7 For insofar as I understand these various
terms, the forms are characters or properties; that is what Plato's
idealism is all about; he believes that the properties of things have an existence apart from the things of which they are properties. There are indeed certain places in which Plato appears to be distinguishing between
e.g. largeness itself (auTo TO p.i-yd)o~) and "the largeness in us"
(TO EV ~p.iv p.i'Yf:8o~), 8 'but these passages, properly read, do not require
us to add anything to the basic Platonic ontology of particulars and the
forms in which they participate and by which they are what they are. 9
In sum, the argument of Parmenides may be paraphrased as follows:
This, Socrates, is the sort of consideration that makes you think each form is one. You
think that whenever a numbq of things are large, there is a form (to be called "the large"
or "largeness") that is the same in all of them and by which they are all large; whence you
think that this form, "the large", is one. But now this form itself is also large. So there
will be another form that is the same in this form and the other large things, and by which
they are all large. Hence there will be more than one form by which large things are large,
and therefore the large will not be one after all.

2. SOUNDNESS

Is the argument, as thus stated, sound? 10 Some scholars have held that
there must be a gap in it, from Plato's point of view at least, or else he
would have given up his Theory of Ideas then and there. 11 However that

IDENTITY AND PREDICATION IN PLATO

33

may be, it is clear that there is indeed a fairly conspicuous gap, which
cannot be filled by the addition, as a "suppressed premise", of any thesis
for which Plato argues elsewhere. This gap is at the point where it is concluded that "then there is another form by which all these are large";
there is no apparent reason why the first form cannot be one of the things
that are large by virtue of it, i.e., cannot be large by virtue of itself. Of
course, everything here depends on the sense of "by virtue of", or, more
precisely, on that of the datives we thus translate. If, to mention but one
possibility, 'x is <p by virtue of ex' is analyzed along the lines of 'x is <p and
if there were no such thing as ex it would be impossible for x to be <p',
Platonic doctrine would justify not only such assertions as "The Mona
Lisa is beautiful by virtue of beauty itself" but also "Beauty itself is
beautiful by virtue of beauty itself". 12
The jump from "there is again a form by which ... " to "there is
another form by which ... " is, in my opinion, the only gap in
Parmenides' argument. All other aspects of it are consonant with Plato's
views and should pass inspection by logicians. 13
Now many influential commentators, from ancient times down to the
present, have in effect located the difficulty at a different place, namely,
at the point where it is assumed that largeness is large. Aristotle says that
the proof that there is a "third man" distinct from Man and from individual men rests on the fallacious assumption that "Man", like the
proper name "Callias", denotes an individual substance, whereas in fact
every such general term denotes either a quality, or a relation, or a quantity, or something of that kind. 14 Applied to the argument as given in our
passage, this evidently amounts to the claim that largeness is not the sori
of thing (i.e., an individual substance) that can be large.
In model n times Russell has made essentially the same point, using
much more drastic language:
In the first place, PlalO has no understanding of philosophical syntax. I can say "Socrates
is human," "Plato is human." and so on. In all these statements, it may be assumed that
the word "human" has e.xactly the same meaning. But whatever it means, it means
something which is not of the same kind as Socrates, Plato, and the rest of the individuals
who compose the human race. "Human" is an adjective; it would be nonsense to say
"human is human". Plato makes a mistake analogous 10 saying "human is human". He
thinks that beauty is beautiful ... He fails altogether to realize how great is the gap
between universals and particulars ... He himself, at a later date, began to see this difficulty, as appears in the Parmenides, which ~ontains one of the most remarkable cases in
history of self-criticism by a philosopher.1 5

And not only Plato's critics but also the more sympathetic commentators have problems with his assertions that largeness is large, beauty is

34

BENSON MA TE~

beautiful, etc. Professor Cherniss argues that these are to be understood

as assertions of identity and not of attribution. He says: "Plato clearly
distinguishes two meanings of 'is x', namely (I) 'has the character x' and
(2) 'is identical with x', and states that alJ70 TO x and only aUTO TO X 'is
x' in the second sense". Thus, assertions like" Justice is just" or "Beauty
is beautiful" mean, as Cherniss puts it, "that' Justice' and 'just' or
'Beauty' and 'beautiful' are identicaf'. In general, he believes, " 'the
idea of x is x' means 'the idea of x and x are identical and therefore the
idea of x does not "have the character x" '''.16 Although Cherniss does
not say so explicitly, this analysis would presumably lead to the conclusion that the Third Man argument, as formulated in our passage, is deficient in assuming that "is large" can be predicated of largeness itself in
the same sense in which it can be predicated of any particular large thing.
Professor Vlastos joins Cherniss in thinking that "is" has two senses
- the so-called attributive and identity senses mentioned above - and
he agrees that Plato was aware of the difference between the two. But
he finds that Platonic sentences of the form 'A is B', where A is the name
of a form and B is an adjective, are ambiguous in still another way, which
was not evident to Plato. 17 In such cases, he says, 'A is B' can be read
either as meaning that the universal denoted by the subject term has the
attribute denoted by the predicate, or as meaning that whatever is an instance of the subject universal will eo ipso have the attribute denoted by
the predicate. E.g., "Justice is pious" could mean that the universal,
Justice, has the attribute of piety, or it could mean that whatever is just
is eo ipso pious. Interpreted in the first of these ways, it is what Vlastos
calls an "ordinary predication"; in the second, it is a case of "Pauline
predication" (after St. Paul's "Charity suffereth long and is kind"). In
terms of this distinction, one could say that if "Largeness is large", in
the Third Man argument, is taken as a Pauline predication it is plausible
but renders the argument unsound (because largeness will not be
predicated in the same sense of both itself and the other large things),
whereas if it is taken as an ordinary predication it becomes false or even
"sheer nonsense". 18
3. THE SENSES OF "IS"

All of these noted commentators, with anyone of whom it is risky to

differ, seem to me to base their analyses on two inter-related but highly
doubtful assumptions:
(a)

That there are, in English

Of

in Greek, at least (wo distinct

IDENTITY AND PREDICATION IN PLATO

35

of "is" (EurL) , viz., the "is" of identity and ti-Jat of

predication; and
That if sentences like "Beauty is beautiful" are to be meaningful, let alone true, the word' 'is", as it occurs in them, cannot have the same sense it has when it occurs in sentences like
"The Mona Lisa is beautiful". 19
sen~es

(b)

Of course Plato nowhere explicitly asserts any such principles as (a) or

(b). Consequently, some of his critics have felt that he has indeed "no
understanding of philosophical syntax" and that "he fails to realize how
great is the gap between universals and particulars"; while his defenders
strive to protect him by showing that he is after all aware of the point
of (a) and that he does not mean that beauty is beautiful in the same sense
of "is" in which a particular painting is beautiful. But critics and
defenders alike seem to agree that to fly in the face of (a) and (b) is to
risk making dreadful mistakes.
So let us first consider (a). Needless to say, the doctrine that there are
several senses of "is", including an identity sense and a predicative sense,
was not invented by Professors Cherniss and Vlastos; it has a long history
and by now is received in many quarters as philosophical gospel. 20 As
evidence for the distinction, one is likely to be given examples like "Scott
is the author of Waverley" and "Scott is human", and one is assured
that the first of these means that Scott is identical with the author of
Waverley (and not that he has the author of Waverley as an attribute since presumably the author of Waverley is a human being, not an attribute), while the second means that Scott has the attribute Humanity,
with which he is obviously not identical.
Do such considerations as these, together with corresponding ones for
Greek, suffice to show that "is" and EurL are not used in the same sense
in both kinds of case? They do not. The following analogy will help
establish this point. Suppose that I have a number of straight sticks,
which I am comparing directly with one another as to length. I report my
observations by using the phrase "is no longer than", making statements
that have either the basic form 'A is no longer than B' or are obtained
from elements of this form by (possibly repeated) application of negation, conjunction, and quantification. Further, I employ 'A is the same
length as B' or 'A matches B' as short for 'A is no longer than Band B
is no longer than A'; and 'A is shorter than B' abbreviates the statement
'A is no longer than B, and it is not the case that B is no longer than A'.
Now it is clear that whenever 'A is no longer than B' is true of a couple
of sticks, either 'A matches B' or 'A is shorter than B', but not both, will

36

BENSON MATES

be true of those sticks; and whenever either of the latter is true, the
former will be true. But obviously this does not suffice to show that "is
no longer than" is here ambiguous, having sometimes the sense of
"matches" and sometimes that' of "is shorter than". When I say 'A is
no longer than B' of a couple of sticks that happen to be equal, I am using
the phrase "is no longer than" in exactly the same sense as when I apply
it to a couple of which the first is shorter than the second. Note further
that to say 'A matches B' amounts to saying 'A is no longer than B' and
something more; and likewise for 'A is shorter than B'. If this 'something
more' were obvious from the context, I could communicate the fact that
two sticks match by simply stating the first component of the conjunction. Thus, if the context makes 'B is no longer than A' obviously true,
I can, as a practical matter, employ simply 'A is no longer than B' to convey the information that the two sticks match. But, again, this would not
show that "no longer than" sometimes means "matches" and the rest
of the time means "is shorter than".
To spell out the intended analogy between "is no longer than" and
"is" is probably unnecessary, but I hope that the reader will forgive my
doing it anyway. The point is that perhaps the "is" of identity and the
predicative "is", so-called, can both be defined in terms of a more
primitive "is", in a manner similar to that in which "matches" and
"shorter than" were defined above in terms of "no longer than". In fact
Leibniz 21 and, if I am not mistaken, certain Polish philosophers beginning with LeSniewski,22 have done just that. Leibniz defines 'A is the
same as B' as 'A is Band B is A', and 'A is (a) B' as 'A is B but B is not
A'. Analogously to the situation with the sticks, we have the result that
whenever 'A is B' is true either 'A is the same as B' or 'A is (a) B', but
not both, will be true, and each of the latter implies the former. Thus,
the fact that in "Scott is the author of Waverley" we can replace "is"
by "is the same as" and get a true sentence, while if we replace "is" by
"has as a property" we get a sentence that is false or nonsensical, in no
way shows that in this sentence "is" means "is the same as". We can also
carryover the point about what happens when the truth of one of the
conjuncts is part of the background infQrmation or is in some other way
too plain to need stating.
Leibniz was defining "same" in terms of "is" for a sort of regimented
Latin, where (because of the lack of a definite article and because of certain features of the regimentation) the indicated types of transformation
work better than they do in English. I do not wish to claim that in Plato's
Greek tad behaves in relation to TD:VTOIJ in exactly the way Leibniz sug-

IDENTITY AND PREDICATION IN PLATO

37

gests for est and idem. Nevertheless the relation may well be similar
enough to justify suspicion that the sort of evidence usually adduced in
support of the multiple sense hypothesis for f.aTL does not at all rule out
the possibility that that verb may be used in a single sense everywhere.
We shall return to this matter in connection with (i) - (vii) below.
In determining whether a word or other expression has more than one
sense, the unwary may be tempted to make still other fallacious inferences. In modern introductions to logic, for example, one often finds
it said that there are two senses of the connective "or"; the "exclusive"
and the "inclusive" senses. Sometimes, we are told, "or" is used in a
sense that excludes the possibility that both disjuncts are true, while in
other occurrences it has a sense that allows such a possibility. (Then one
is usually informed that for reasons of simplicity, etc., logicians have
decided to use the word, or a corresponding symbol, in the inclusive sense
only: a disjunction counts as true if and only if at least one of the disjuncts is true.)
Now it turns out that finding indisputable cases of the exclusive sense
of "or" in the natural language is not quite so easy as might be thought.
If I tell you that I shall either go to the concert or stay home and read
a good book, it is clear enough that I am not allowing the possibility that
I might both go to the concert and stay at home; but it is also clear that
we do not need to postulate an exclusive sense of "or" to account for
the exclusion, for the content of the disjuncts suffices to eliminate the
possibility that both might be true. (Note that even after the logician has
given his "inclusive" sense to the symbol "v", he uses it, without change
of sense, in disjunctions like' P v -P' , where it is impossible that both disjuncts be true). So, in order to have critical cases before us we must look
for disjunctions which are such that (a) the whole disjunction will be considered false if both disjuncts are true, and (b) it is at least possible that
both disjuncts be true. But even in these cases we must beware of such
contribution as the context or background information may make to the
inferences the hearer will draw from the disjunction. For example, if my
daughter has been expressing a wish to go to the concert and also to buy
a recording of the symphony that will be performed there, and I have
responded that it's certain we cannot afford both of these, and even
doubtful whether we can afford either, then, when I finally say, "All
right, you may go to the concert or buy the record", it will be obvious
to her that the possibility of both is excluded. But again the responsibility
for the exclusion need not be pinned on the "or"; rather, it seems more
properly attributable to the background information. In short, the fact

38

BENSON MATES

that a given disjunction is taken in such a way as to exclude the possibility

of both disjuncts being true may often (and perhaps always) be accounted for without supposing that the word "or" is being used iJ1 an
exclusive sense.
Analogously, it may be the case that whenever a certain type of
substantive, such as e.g. a proper name, occupies the predicate position
in a Leibnizian sentence 'A is B', then' B is A' follows from' A is B' . On
this basis, if 'A is Socrates' is true, then 'Socrates is A' is also true; hence
'A is Socrates' will be true if and only if 'A is identical with Socrates' is
true; and still there will be no ground in this for saying that "is" is used
in one sense in, e.g., "the teacher of Plato is Socrates" and in another
in "the teacher of Plato is wise".
Now, if the kinds of evidence indicated above are not sufficient to
show that Plato sometimes uses "is" in the sense of "is identical with"
and sometimes uses it in a "predicative" sense, what sort of evidence
would justify that conclusion? The following might seem to be what we
need. On the one hand, we note that Plato holds principle (I) mentioned
at the outset, together with its consequence, (3). On the other hand, at
Parmenides 158A5 - 6 we find
(5)

It is impossible for anything but the one itself to be one.23

In the context of Plato's philosophy, (3) and (5) look incompatible.

Thus, if he seriously means to assert both, and if the large itself and the
one itself are not identical for him, and if he is in full possession of his
logical powers when he writes each of these two sentences and has not
changed his mind between times, it would seem justifiable to conclude
that one or more components common to the two are ambiguous. From
here it is but a short step to the conclusion that "is one" is ambiguous,
and from there to the further conclusion that "is", or "one", or both
are the culprits.
But this whole argument collapses because it is not clear that Plato
seriously meant to assert (5). After all, (5) occurs in a dialogue, in the
mouth of a character representing a philosopher with whom Plato does
not agree on the very matters under discussion. 24 Moreover, this
philosopher, above all others in the history of philosophy, is notorious
for playing fast and loose with "is". If anyone is to be charged with taking "to be" in (5) as synonymous with "to be identical with", it had better be Parmenides, and not Plato. Consequently, although we have here
the right kind of evidence, in this particular case the proof fails because

IDENTITY AND PREDICATION IN PLATO

39

it is not certain that the inconsistent statements are reaJly asserted by a

single author .
Summing up, then, I find no conclusive evidence that there are any
such senses of "is" as the so-caJled "is" of identity, the "is" of predication, or the "is" of Pauline predication; afortiori I see no reason to suppose that Plato, knowingly or unknowingly, used the word equivocaJly
in these various purported senses.
4. "BEAUTY IS BEAUTIFUL". AND THE LIKE

Let us next consider assumption (b), that there is something wrong with
Platonic sentences like "Beauty is beautiful". Why are so many
philosophers and other scholars ready to teJl us that such sentences, if
taken literaJly, are "sheer nonsense"? It seems that the principal reason
- and this is surely paradoxical - is that the Platonic metaphysics has
been swaJlowed, hook, line, and sinker, and has then been interpreted in
such a way as to rule out part of itself. That is, one first agrees that beauty (or, let us say, Beauty) is an abstract entity, eternal, changeless, existing or subsisting in a world apart, while particular beautiful things alf
belong to the world of sights and sound, and, indeed, are beautiful
precisely because of how they look, sound, or in other ways affect the
senses . Then one infers that things so utterly different as these, belonging
even to different categories (whatever that means), cannot have attributes in common; e.g., that neither "is beautiful", "is good", nor any
other predicate can be true of the abstract entity Beauty if taken in the
same sense in which it is true of particular concrete objects.
These notions are occasionaJly reinforced by the mistaken idea that
unless we subscribe to some sort of theory of types, which would declare
it nonsensical to attribute a property to itself, we shaJl inevitably faJl into
Russell's Antinomy and related contradictions. But, as is well known,
type theory is not the only device, nor even the preferred one, for
avoiding the fundamental antinomies; so that if Plato wishes to make
statements like "Beauty is beautiful" he is thus far in no particular
danger from the side of logic.
Plato formulates his puzzling reflexive assertions in various ways. The
most common of these is of special interest. Instead of using a standard
abstract noun in the subject position, he employs the adjective with the
article, thus producing what appear to be literal counterparts of the
English sentences "the large is large", "the beautiful is beautiful", "the
just is just", "the holy is iJoly", etc. What do these statements mean?

40

BENSON MATES

We are told by the grammarians that such an expression as TO XCiAOV

("the beautiful") is ambiguous in Greek; inter alia it can refer to the
abstract entity, Beauty, or to a typically beautiful object, or to the
beautiful object that is under discussion in the given context. 25 According to this, "the beautiful is beautiful" is dreadfully ambiguous,
meaning perhaps "beauty is beautiful", or "whatever is beautiful is
(perhaps eo ipso) beautiful", or "the beautiful object (we've been talking about) is beautiful".
But the advice from our grammarians is less than satisfactory, for it
is formulated in such a way as to presuppose the Platonic distinction between abstract entities and particulars, whereas, presumably, we should
not have to accept Plato's metaphysics in order to understand the workings of the Greek language. One is also left with the uneasy feeling that
the only evidence these experts have for the ambiguity of the Greek expressions is the lack, in each case, of a single corresponding English or
German expression by whi.:h the Greek term may be translated at all of
its occurrences. But does this show that there is something wrong with
the Greek?
It is striking that in the dialogues no interlocutor ever hesitates a moment before agreeing to TO XCiAOV XCiAOV Ean ("the beautiful is
beautiful"), TO O[XCiLOV O[XCiLOV Ean ("the just is just"), and the like;
nobody ever says "Wait a minute; that doesn't make sense" or even "I
don't quite follow you, Socrates". The reason, I think, is that for any
Greek such a sentence would be a logical truth, in the Quinean sense that
(a) it is true, and (b) every result of substituting another adjective for its
only non-logical constant is equally true. In short, such a sentence would
be felt as obviously and trivially true. 26 The same holds, Mutatis mutandis, for .statements like 0 fan XA[VT] XA[VT] EaTl ("what (a) bed is, is (a)
bed"); they too satisfy the Quinean criterion for logical truth. Thus, the
various reflexive' assertions, when formulated in these fundamental
ways, seem not only true but even trivially true.
But Plato goes much further. In the relevant contexts he clearly uses
regular Greek abstract terms interchangeably with the corresponding
adjective-plus-article expressions just described. His readiness to do this,
I believe, is based on a logical or linguistic error of monumental import
for the subsequent history of philosophY. The sentences TO XCiAOV
XCiAOV Ean ("the beautiful is beautiful") and b nCiQ(hvwv X CiAO 5 Ean
("the Parthenon is beautiful") are apparently of similar structure, and,
as the latter informs us that the object denoted by b nCiQOfVWV is
beautiful, it is tempting to interpret the former similarly, I.e., to take TO

IDENTITY AND PREDICATION IN PLATO

41

xaM" as the name of something now asserted to be beautiful. But clearly

this is a mistake, analogous to the well-known error of treating words like
"nothing" and "something" as though they were names. 2? For what we
have here is in effect a device for universal quantification; the article TO
operates on the entire sentences and not just on the adjective immediately
following. This becomes even more obvious when Plato strengthens the
assertion to aUTO TO xaAO" xaM" fun ("the beautiful itself is beautiful"
or "the be<!utiful as such is beautiful"), which is plausible (and trivial)
if understood as "whatever is beautiful is eo ipso beautiful" but which
is now taken as though it predicated beauty of something named aUTO
TO xaM". In the plausible interpretation the word aUTO is treated properly as a modal operator governing the entire following sentence; read the
other way it appears to function only to specify further what is purportedly denoted by TO xaM".
Once it is assumed that expressions like TO xaM" ("the beautiful") or
aUTO TO xaM" ("the beautiful as such") name something, it is natural
to identify that something with beauty, i.e., to use these expressions interchangeably with TO xaAAos ("beauty"). This is exactly what Plato
does. In discussions where TO xaM" ("the beautiful") is used to express
generality he is willing to substitute TO xaAAOS (' 'beauty") for it; similarly
for TO OLxaw" ("the just") and ~ [nxawuv"l1 ("justice"), TO DUW" ("the
pious") and ~ DULOTlIS ("piety"), and so on. This by itself would cause
little difficulty, for if, e.g., the abstract term ~ oLxawuv"lI ("justice")
were merely in effect an abbreviation of TO OLxaw" ("the just"), we
could read a puzzler like I~ oLxawuv"lI DULa fUTL ("justice is pious")
as merely an alternative way of saying TO OLxaw" DULO" fun ("the just
is pious").
But unfortunately each abstract term also occurs in other contexts with
other kinds of predicates, and when in those contexts it is interchanged
with the corresponding adjective-plus-article phrase, the results may be
statements that can no longer be understood as modalized generalizations about particulars. Thus, e.g., aUTO TO OLxaw" lObs fun ("the just
itself is a form") cannot be taken as meaning "whatever is just is eo ipso
a form".28
We see, therefore, that Platonic statements purporting to be about a
given universal, e.g. justice, fall into two categories: (I) those (relatively
harmless ones) that can be "translated down" into modalized generalizations about the individuals falling under the universal, and (2) those
others that cannot. The latter, e.g. "Justice is eternal", constitute the
distinctive essentials of Plato's metaphysics.

42

BENSON MATES

The upshot of all this is as follows. There is no reason to doubt that

"is large" is used by Plato in the same sense in "The large is large" or
"The large as such is large" as it is in "The Parthenon is large". Furthermore, "The large is large" is, for Plato and any other Greek, selfevidently true. Things only begin to go awry when "the large" and "the large
as such" are taken as names, interchangeable with "largeness". A first
result of such interchange is that" Largeness is large" acquires the status
of an obvious truth; another is that "The large as such is an idea,
changeless, eternal, etc.", which is false if properly understood as a
mod ali zed generalization, now appears to be true. Many other problems
arise, including the crucial one that if any predicate I{J is denied of the
Ideas, we shall have both 'No Idea is I{J' and 'The I{J is I{J' as true. 29
Despite all of these confusing complexities, however, I believe it possible to maintain that Plato uses the verb "to be" in a single sense
throughout - a single sense in terms of which some of the other senses
that have been proposed can be defined. Of course, he is not writing in
a formalized language, and we know better than to look for exact defini~
tions and rules to cover even a philosopher's use of a natural language.
But, very roughly speaking, his usage seems to be in accord with some
such scheme as the following: For any terms A, B, C,
(i)
(ii)

(iii)

'A is the same as B' ('A = B') is true if and only if 'A is B'
and 'B is A' are true;30
'A isp B' ("Pauline predication") is true if and only if (a) for
all terms D, 'D is B' follows from 'D is A' and (b) for some
terms D, E, 'D = E' does not follow from 'D is A and E is
A',3l
'A is (a) B' ("ordinary predication") is true if and only if 'A
is B' is true but 'A = B' and 'A isp B' are not.

While we are at it, we may add:

(iv)
(v)
(vi)
(vii)

'A is' is true if and only if, for some term D, 'A is D' is true;32

'A is similar to B in respect to C' is true if and only if 'A is

C' and 'B is C' are true;
'A is similar to B' is true if and only if, for some term D, 'A
is similar to B in respect to D' is true;
'A is A' is necessarily true.

If Plato's usage is more or less along these lines, then we can expect that,
for him, whenever 'A is B' is asserted, then 'A is the same as B' or 'A
isp B' or 'A is (a) B' could also be asserted. Of these three, the last is inconsistent with each of the first two, though the first two are consistent

IDENTITY AND PREDICATION IN PLATO

43

with each other. As examples of sentences that come out true according
to the above scheme, we have "Socrates is identical with the teacher of
Plato", "Socrates is a man", "The just isp good", "The just is eternal"
(but not "Justice isp eternal"). In each case, the corresponding sentence
with the primitive "is" will be true; i.e." "Socrates is the teacher of
Plato", "Socrates is a man", "The just is good", and "The just is eternal" are all true in the same sense of "is". We shall also have such results
as that if "The statue is large" is true, then "The statue is similar to the
large" and "The large is similar to the statue" will also be true.
I hasten to acknowledge, however, that the matter is very much more
complex than these suggestions might indicate. A more satisfactory account would at least replace (i) - (vii) above by corresponding principles
for Plato's Greek, and difficult problems of word order and the placement of the article would have to be dealt with. Still further complications will result from Plato's use of the abstract noun and other expressions as apparently synonymous with the corresponding adjective-plusarticle. So the most that can be claimed for the above scheme is that it
shows one way in which the copula could be used univocally everywhere
and yet give rise to the kinds of texts that have made scholars consider
it ambiguous. 33
Returning in conclusion to the Third Man Argument, we may note that
it is fortunate for Plato that there is another way out besides that of
declaring that "is large" is ambiguous. For, as has often been noted, that
sort of ambiguity would render almost unintelligible his important doctrine that the particulars are likenesses of their corresponding ideas. The
text most clearly illustrating this is in the Symposium,34 where Socrates
describes a hierarchy of beautiful things; there are beautiful bodies, but
more beautiful than these are the beautiful souls, and the beauty of the
laws and of the various branches of knowledge ranks still higher. Most
beautiful of all, he says, is beauty itself. Then he goes on to explain in
detail exactly why beauty itself is more beautiful than anything else. 35
Unlike the other beautiful things, it is eternal, neither corning to be nor
passing away; unlike them, it is not beautiful in one respect and ugly in
another; nor beautiful from one point of view and ugly from another;
and so on. If the predicate "is beautifui" were not used in a single sense
throughout this comparison, the passage would be very dark indeed; for
to say that beauty itself is more beautiful than a beautiful soul, but in
a different sense of "is beautiful", would be like saying that light travels
faster than sound, but in a different sense of "fast".36
Thus Plato cannot very well join those who would save him from the

44

BENSON MATES

Third Man argument by finding an ambiguity in the "is" or the "large"

of "is large". His various assertions not only do not require us to
postulate such ambiguity, they actually forbid it. Therefore, the other
way out, namely, that of allowing the large to be large by virtue of itself,
would seem to be his only real alternative.
NOTES
1 A Swedish translation of an earlier version of this paper is included in a privately published memorial volume for the late Professor Anders Wed berg (Enfilosofibok, Stockholm,
Bonniers, 1978, pp. 66 - 84). I wish to acknowledge my indebtedness to Professor
Wedberg's chapter on Plato's Theory of Ideas, in Plato's Philosophy of Mathematics,
whi~h was one of the earliest systematic and lucid expositions of the matters here considered. Even after a flood of further literature by other authors it still must be ranked as
one of the best accounts available.
2 My treatment of the second formulation of the Third Man, at Parmenides 132D - 133A,
would be exactly analogous to what I have to say here about the first formulation. Cf. Note
13.
3 Parmenides 132B5 - 6. Why would it "no longer be subject to the consequences just now
mentioned"? Because, I suppose, the thought of the large, unlike the large itself, would
not necessarily be large.
4 Vlastos (1973), p. 344n8; Cornford (1939), ad lac.
S Thus, fl' must be construed predicatively at 13ICIO. 132B5. 132B7.
6 Cpo Wed berg (1955), p. 30 (3b).
7 Of course "character" might in this connection be used metalinguistically, as it were, to
refer to those Greek adjectives and nouns for which we notice that Plato postulates corresponding forms. Thus, we observe that corresponding to the adjective XOI}..(J" he
postulates one form. and similarly for various other nouns and adjectives. But when he says
(3) he cannot mean "corresponding to the adjective P.f"{OIS there is exactly one form", for,
whatever he may have in mind, he' is clearly not talking about words.
8 Phaedo 102D6 - 7.
Of course this assertion requires detailed argument, for which I do not have space here;
I include this paragraph only to indicate why I do not follow the common practice of formulating the issue in terms of "characters" and "corresponding forms".
10 I am talking about the argument as thus staled, and not about one or another possible
formalization of it by means of the notation of modem logic.
II E.g . Ross (1951), p. 87.
12 Here and in many other places throughout this paper I use single quotes as quasi-quotes.
Cf. Quine (1947), pp. 33 - 7.
13 Note how, in lines 132A6-1O. Parmenidcs eases into the claim that there is another
form of largeness: at 132A7 he says that "some one large will again appear", which strongI, suggests but does not strictly imply that it is a different form: and then this conclusion
is ostensibly re-stated at 132AlO as "anotherform will show up ... " (my italics). but only
if "some one large will again appear" is taken in the sense of "again, some one large will
appear", does it follow from the general principle that Socrates has admitted; "again"
here is metalogical, meaning "by another application of the principle".
In the second formulation of the Third Man I locate the one and only fallacy at the corresponding point, i.e., at 132E7, where it is concluded that "besides that form, another
form will always appear ... " (my italics).

IDENTITY AND PREDICATION IN PLATO

14

45

De Soph. EI. 189b36 - 9.

Russell (1945), p. 127.

Cherniss (1957), pp. 258 - 9.
17 Vlastos (1973), p. 307.
18 Vlastos (1973), p. 252ff. Cp o also Vlastos (1973), pp. 234 - 6, 270ff, and 318ff.
19 For reasons of style I have not always stated explicitly that my remarks are intended to
apply not only to the English expressions specifically quoted but also to the Greek expressions that correspond to them. Since part of the problem before us is in effect that of determining what corresponds to what, I recognize that this policy introduces a certain amount
of confusion, for which I apologize.
20 Cf. e.g . Russell (1903), p. 64n; Wittgenstein (1921), 3.323.
21 Couturat (1903), p. 382; Schmidt (1960), pp. 475, 479.
22 See Luschei (1962), p. 144ff.
23 vuv Of ivi I'fV dvctL 1rh~V aVT~ T~ lVI c,ouvaTov 1rOU.
24 It is immediately preceded and followed by a group of arguments so obviously fallacious
that we should insult Plato's intelligence if we supposed that he accepted them . E.g.,
(157C - D): "If something were a part of a multiplicity, among which it was itself included,
it would be a part of itself, which is impossible .. . "; and ultimately the absurd conclusion
(166C): "it seems that, whether there is or is not a one, both that one and the others alike
are and are not, and appear and do not appear to be, all manner of things in all manner
of ways with respect to themselves and to one another" - to which poor Socrates has to
reply. "Most true"!
25 Brugmann-Schwyzer (1961), Vol. II, p. 175; Kuhner (1898), Part II, vI. I, p. 266'ff;
Smyth (1920), pp. 272-4; Goodwin (1894), p. 204.
26 Vlastos' claim (Vlastos (1973), p. 249n77), that "justice is just" at Prot agoras 330C is
not taken as self-evident but is inferred, is puzzling . Note that c.ea in 330C7 need not be
read as strict logical "therefore" but only as "so then".
27 Cpo such fallacies as "Nothing is colder than ice; I have nothing up my sleeve; therefore,
what I have up my sleeve is colder than ice", or "Something just bit me; you gave me
something for Christmas; therefore, what you gave me for Christmas just bit me". Further, although any German is a German and any German can tell you where Goethe was
born , that does not mean that there is some German whose name is "any German" and
who can tell you where Goethe was born.
28 When a binary predicate, such as "equal to" or "similar to" is under consideration,
the mistake leads to even more painful consequences. On the same basis as before we now
have as obvious truths aUTa Ta ioa ioa faT! ("equals as sllch are equal") and OIUTa
Ta 01'0101 OI'Ola faT! ("similars as such are similar"); if these are interpreted as saying,
respectively, that what is denoted by "equals as such" is equal, and what is denoted by
"similars as such" is similar, and if the phrase "equals as such" is interchangeable with
"equality", and likewise for "similar as such" and "similarity", we arrive at a pseudoproblem as to whether equality and similarity are singular or plural. For if equality consists
of the equals as such, does it not consist, as Geach 1956), p. 76) suggests, of at least two
absolutely equal things? In short, the very same mistake that takes Plato from "the
beautiful is beautiful" to "beauty is beautiful" will also take him from "equals are equal"
to "equality is/ are? equal" .
to Thus, e.g., both "no idea is plural" and "plurality is plural" will be true, and there are
many other such examples. Vlastos (1973), pp. 259ff, tries to protect Plato from these contradictions by interpreting "plurality is plural" as a Pauline predication . On the other
hand, he recognizes (pp. 262 - 3) that "beauty is beautiful", in Diotima's speech in the
Symposium, has to be taken as an ordinary predication. Hence he is forced to hold that
15
16

46

BENSON MATES

Platonic statements of the form 'the", is '" are sometimes Pauline, sometimes ordinary. But
it seems to me that whenever 'the", is rp' is asserted in the dialogues it is put forward on
the same basis. To suppose with Vlastos (p. 265) that "justice is just" in the Prolagoras
is Pauline, while "beauty is beautiful" in the Symposium is not (pp. 262 - 3), should be
a last resort; far better to suppose that Plato uses "is" univocally but has not thought out
what to do about the difficult cases.
30 This condition for identity may be too weak. In the Prolagoras, in a discussion initiated
by the question "whether virtue is one, and justice, temperance, and piety are parts of it,
or whether these things that I have just now mentioned are all of them names of the same
one thing" (329C - D), it is concluded from "justice is pious" and "piety is just" that
"justice is either the same as piety or maximally similar to it" (33IB, cpo 333B); and from
considerations indirectly establishing that every temperate act is wise and every wise act is
temperate it is concluded that temperance and wisdom are one (333B). (Cp. Vlastos (1973),
pp. 243 - 6). Following this, Socrates begins what is plainly an attempt to show that
temperance and justice are one; and it looks as though his argument, never completed, was
going to involve establishing that every temperate act is just and every just act is temperate.
Thus, he seems to be trying to show, perhaps only to discomfit Protagoras, that
"wisdom", "temperance", "justice", and "piety" all name the same thi~. Whether he
or Plato actually believed this, is irrelevant; the crux of the matter is whether the course
of the argument sho"'s what he thinks would have to be the case if the various identity
statements were true. However, the references to similarity suggest that perhaps the truth
of 'A is the same as B' requires something more than that of 'A is B' and 'B is A', at least
when A and B are names of ideas.
31 Clause (b) is designed to eliminate the possibility that A is a name or description of a
particular. Otherwise, since e.g. "If anything is Socrates, then it is eo ipso a man" is true,
we should have "Socrates is a man" as a Pauline predication.
32 Thus two possibilities suggest themselves for 'A is not': (I) for no term B is 'A is B' true,
or (2) for some term B, 'A is not B' is true.
33 As emphasized in the text, there is in general no hope of finding simple, exact rules to
cover the usage of a given author writing in a natural language. The following may help
to indicate at least a significant subset of the Greek examples I seek to catch with clauses
(i) - (vii). In forming substitution-instances of a given clause:
(I) Any adjective, count noun, or proper name, prefixed by the definite article, may be
substituted for a variable in subject position.
(2) Any adjective or noun, with or without the article, may be substituted for a variable
in predicate position.
(3) Where, in the given clause, the same variable occurs both in subject and in predicate
positions, it is to be replaced in subject positions by an expression with the article if and
only if it is replaced in predicate positions by the same expression without the article.
(4) An abstract term (e.g., ~ Ol){atOUUP'I) is interchangeable with the corresponding adjectival phrase (TO M){atOP) or a fun phrase (a fun M){atOP).
Cases involving complex terms may, it is hoped, be treated by analogy with the foregoing
principles.
Some examples:
Of (i): TO OL){CnOp ){aL TO autOp mUToP fun is true iff TO M){atOP amop fun and TO OUtOP
M){atop fun are true. 0 Ew){ear"s ){aL 0 otOaU){aAOS IIAaTwPos mUToP fun is true iff 0
Ew){ear"s otOaU){aAOS IIAaTwPos fun and 0 otOaU){aAOS IIAaTwPos Ew){ear"s fUTL are
true.
Of (ii): TO M){atOP Cx-yalJol' fun (as a Pauline predication) is true iff (a) for all terms D,

IDENTITY AND PREDICATION IN PLATO

47

'D &."(CiIJOV Eun' follows from 'D ot)(CltOV Eun' and (b) for some terms, D, E, 'D = E'does
not follow from 'D ot)(CltOV Eun )(Cit E o[)(CltOV Eun'.
Of (iii): TO O[)(CltOV &')([V'1TOV Eun (or ~ O!)(CiWUUV'1 &')([V'1TOV Eun as an ordinary
predication, is true iff TO ot)(CltOV &')([V'1TOV Eun is true but not as a Pauline predication,
and TO ot)(CltOV )(Cit CiUTO TO &')([V'1TOV TCiUTOV Eun is not true.
34 Symposium, 210fr.
35 Op. cit., 21IA-B.
36 Perhaps I should state explicitly that, according to the interpretation I am advancing,
the sentences "Beauty is beautiful" and "The soul is beautiful" are both true fOf Plato
when the "is" is taken in what I am calling its "primitive" sense. When it is taken in the
identity sense, only the first sentence is true, and when it is taken in the sense of "ordinary"
(to us) predication, only the second sentence is true. As I have indicated in (v) and (vi),
"The soul resembles Beauty" will be true on the basis of the two sentences mentioned, with
"is" understood in the primitive sense in both cases.

BIBLIOGRAPHY
Brugmann, K. and E. Schwyzer: 1966, Griechische Grammatik, 3rd ed., Munich.
Cherniss, H.: 1957, 'The Relation of the Timaeus to Plato's Later Dialogues', American
Journal of Philology 78, 225 - 66.
Cornford, F. M.: 1939, Plato and Parmenides, Routledge & Kegan Paul, London.
Couturat, L.: 1903, Opuscules et fragments inedits de Leibniz, Presses Universitaires de
France, Paris.
Geach, P. T.: 1956, 'The Third Man Again', The Philosophical Review 65, 72 - 82.
Goodwin, W. W.: 1963, A Greek Grammar, London.
Kahn, c.: 1973, The Verb 'Be' in Ancient Greek, D. Reidel, Dordrecht.
Kiihner, R.: 1898-1904, Ausfiihrliche Grammatik der griechischen Sprache, Hannover
and Leipzig.
Luschei, E. c.: 1962, The Logical Systems of Lesniewski, North Holland, Amsterdam.
Quine, W. V.: 1947, Mathematical Logic, Harvard University Press, Cambridge.
Ross, W. D.: 1951, Plato's Theory of Ideas, Clarendon Press, Oxford.
Russell, B.: 1903, Principles of Mathematics, Cambridge University Press, London.
Russell, B.: 1945, A History of Western Philosophy, Simon & Shuster, New York.
Schmidt, F.: 1960, Leibniz: Fragmente zur Logik, Akademie-Verlag, Berlin.
Smyth, H. W.: 1956, Greek Grammar, Harvard University Press, Cambridge.
Vlastos, G.: 1973, Platonic Studies, Princeton University Press, Princeton.
Wed berg , A.: 1955, Plato's Philosophy of Mathematics, Almquist & Wiksell, Stockholm.
Wittgenstein, L.: 1922, Tractatus Logico-Philosophicus, Routledge & Kegan Paul,
London.

Dept. of Philosophy,
University of California, Berkeley,
Berkeley, CA 94720, U. S.A.

RUSSELL M. DANCY

ARISTOTLE AND EXISTENCE

Aristotle tells us more than once that 'to be' is said in many ways,
wratever that means. I had better say straight off that I can find very
little in the present paper that tells us what that means. But in the course
of considering what it might mean, Owen, a long time ago (1960), told
us that, while some people held that 'being' has "a single meaning" in
all its applications,
Aristotle was one of those who denied this. In his view, to be was to be something or other
I

In a footnote to this sentence, he added

This is not to deny the distinction between elval n and ebal lX1lAWS.

But it certainly sounds as if it is to deny just that: the Greek phrases

translate as 'to be something' and "just 'to be' ".
Owen says, of course, a great deal more than I have quoted, but not
enough to remove my puzzlement on at least the first few dozen readings.
Eventually, I came to think I saw what he might have had in mind, and
wrote my initial puzzlement off to stupidity. But recent work 2 has shown
that the message, if it had b~en Owen's, has not got around. So I am here
going to state what I arrived at.
But then, the message turned out not to be Owen's: I found myself in
disagreement with some (by no means all) of his subsequent pronouncements on Aristotle's ways with being. 3 So what follows is an interpretation born out of misunderstanding. No doubt some, including
Owen, would think that it is not only Owen I failed to understand. In any
event, it seems to me right about Aristotle.
Still, it must be admitted that the theory I am going to hand Aristotle
is not one he anywhere outright states. The idea is that it will help us to
understand some of the things I;te does outright state if we suppose that
he held this theory. So the claim that he held the theory is underdetermined by the evidence. I take it that this situation is not unfamiliar
to readers of Aristotle.
I shall begin by stating the theory, as I think it emerged from Aristot-

49
S. Knuuttila and J. Hintikka (eds.), The Logic oj Being, 49-80.
1986 by D. Reidel Publishing Company.

50

RUSSELL M. DANCY

Ie's reflections on Plato's mistakes, and fitting it into Aristotle's essentialism. This occupies Section 1 below.
The theory is, in the first instance, a theory about the Greek word
'lvo:t' whose bearing on the English translation, 'to be', of that word
is not obvious. But I think that, mutatis mutandis, it does have
something to tell us about Anglo-Saxon being, and, ultimately, perhaps,
something about Being (or even Sein: but I do not want to commit
sacrilege). Accordingly, after stating the theory, I try to mutate the things
that need mutation to bring it to bear on English. Here I hope to irritate
any Heideggerians that are still with us. I also complain a bit about the
theory's fringe benefits, and it is probably at about this point that my
disagreement with Owen begins to set in (but I shall not be pursuing this
disagreement beyondproviding relevant footnotes). At any rate, I make
as many enemies as I can.
Lastly, although it is a bit of a letdown after twitting the Heideggerians
and so on, I turn to some texts with which this theory is to help us. On
this occasion I consider only two: 4 one in De interpretatione 11 that
famously conflicts with a passage in Categories 10, and one in
Metaphysics A 7, the chapter on 'being' in Aristotle's 'dictionary'. Even
with these texts, I shall be more cursory than the subject ultimately will
allow. But brevity may at least yield clarity of outline, and in a
preliminary statement such as this, that is perhaps best.
1. WHAT IT IS TO BE IN GREEK

The Greek that Plato and Aristotle wrote and probably spoke does not
possess separate words, as English does, for 'to exist' and 'to be,.5 Firstyear Greek textbooks 6 tell the student that the single word EaTtV is to
be accented differently depending on whether it means 'exist' or just 'is',
but this orthographic convention, known as 'Hermann's rule', has no
foundation in the writing of Greek or in what we can tell of its pronunciation in ancient times, and may have no foundation before 1801, when
Hermann ruled it 7. The Greek word l:~{aT(Xat'}o:t, which, in the sense 'to
be separated from', 'to stand out from', or just 'to stand out', is destined
to travel through its Latin cognate into French and thence to appear in
English as 'to exist', is not used to mean 'exist', at least, not in the time
of Plato and Aristotle.
So, for Plato and Aristotle, the claim that Socrates exists would be
represented by a sentence shaped like this one:
(1)

Socrates is. 8

ARISTOTLE AND EXISTENCE

51

Now, in fact, the sentence 'Socrates exists' has a peculiar ring to it. It
helps here to change tenses and subjects: "So Prester John really existed,
after all". Here the extra words, the tense, and, most importantly, one's
background knowledge about the use of the quasi-proper name 'Prester
John' help to give an air of realism to the enterprise. But for starters, it
is easier to ignore friction and wind resistance; anyway, the theory itself
will involve us in restoring some of the realism lacking in (1), or in its
more normal but still rather peculiar English equivalent, 'Socrates exists'. In particular, we shall attend to the relevance of background
knowledge. So I shall stick with (1).
In this and subsequent formulations I am deliberately ignoring one
subtlety. It has frequently been said that the English 'existential prefix',
"there is (such a NP as) x", is not used to make existential claims, 'x exists', for a variety of reasons. 9 I do not want to deny this claim. On the
other hand, I take it that, on occasion, the sentences "So Prester John
really existed, after all" and "So there really was such a person as Prester
John, after all" are interchangeable. And that admittedly sloppy
equivalence is enough for present purposes: our focus is not, ultimately,
on the English verb 'exists' anyway ..
So a sentence like (I) translates the English 'Socrates exists', and the
Greek word in it that corresponds to the word 'is' is the same word that
appears in the Greek for these:
(2)
(3)

Socrates is pale.
Socrates is [a] man [i.e., a human being].

Consider first Plato's response to this situation; Aristotle' grows out

of it.
At the end of Republic V, Plato tells us that, whereas the form named
'the beautiful' entirely (lI'ap7fAWs, 477a3) and purely (ffALXQLPWS, 478d6,
479d5) is, ordinary beautiful things both are and are not. It is tempting
to lend a semblance of naturalness to this contrast by replacing 'is', etc.,
with 'exists', etc. The temptation should be resisted, at least at first, not
because, as some 10 think, no rational man (much less Plato) would ever
espouse the absurdity of degrees of existence - this seems to me to condemn too many philosophers too quickly to the asylums - but because
that replacement severs the contrast from the argument for it. What condemns the fair Helen II to being tumbled about (XVALPOf'i7CXL, 479d4)12 in
the region between what purely is and what purely is not is the fact that,
although she is beautiful by comparison with a pot (cf. Hi. Ma. 289a,
with Phd. 74a-c and R. V. 479ab), she is not beautiful by comparison

52

RUSSELL M_ DANCY

with the goddess Aphrodite. Throughout an important period of his life,

Plato had the unfortunate habit of concluding from this that Helen both
is beautiful and is not beautiful. 13 He managed to some extent to kick
this habit, but not before writing the Republic: indeed, here he indulges
in it with gusto and a further twist, for he goes on from the conclusion
that Helen both is and is not beautiful to the conclusion that Helen both
is and is not.
That, then, is the contrast between Helen and the beautiful: Helen
both is and is not, in as much as she both is beautiful and is not beautiful,
but the form, the beautiful, is purely, that is, it is unadulterated by any
admixture of nonbeing, in as much as it is simply beautiful, and in no
way is it not beautiful. 14
We could put the view underlying these moves - or rather, one of the
views - in terms that make it sound more philosophical by saying that,
where we have a true predication'S is P', Plato is regarding the predicate
'P' as giving a mode of-being of the subject S; and, where we have'S is
not P', P is a mode in which S lacks being. I should prefer to put it differently, and I shall; but even put this way it is possible to see, as Plato
himself came to see, that it is thoroughly unsatisfactory for his purposes.
For clearly, as he says in the Sophist, "for each of the forms, there is
much that it is, but an indefinite plurality [of things] that it is not" (Sph.
256e5 - 6). 15 So there are lots of modes in which the form, the beautiful,
lacks b~ing, and the contrast with Helen evaporates. And that, indeed,
is just what happens to the contrast in late dialogues such as the
Sophist: 16 the difficulty I am pointing to is one of which the Academy
became acutely aware. And it is this that leads to Aristotle's theory.
But this is better seen if we abandon the false depth of modes of being
for a more austere formulation, as follows. Plato is, in the Republic,
allowing the inference:
(E)

S is P

->

Sis,

along with its negative counterpart:

(NeE) S is not P

->

S is not.

And what makes (E) and its negative counterpart plausible to him is the
idea that the consequent in each case is merely a simplification of the
antecedent, just as in
(P)

Socrates is a pale man

->

Socrates is pale.17

For this sort of simplification to work, the residue after the simplifica-

ARISTOTLE AND EXISTENCE

53

tion must have the same significance it had before: its contribution to the
unsimplified original must be just what is left. Failure to meet this requirement is obvious where there is a gross change in the sense of one or
another word, as, perhaps, in
(L)

Socrates is a lousy provider

-+

Socrates is lousy,

but there are subtler failures. Consider

(G)

Socrates is a good cobbler

-+

Socrates is good.

Here we need not, and, I think, should not, say that 'good' has changed
sense from antecedent to consequent. 'Good' is what Geach once 18
called an 'attributive' adjective: a noun or noun phrase needs to be supplied that will answer the question 'a good what?'. And, although in a
specific conversational context it may be clear that the noun to be supplied is 'cobbler' ("The cobblers in Athens are all terrible." - "No,
Socrates is good.' '), we are talking about justified patterns of inference,
and that means we need a conclusion we can, so to speak, carry away
with us. But once we have carried away 'Socrates is good', the noun we
should supply to complete the sense is no longer 'cobbler', but,
presumably, 'man'.
All this applies to (E) (and to (NCE. At the very least, the 'is' in its
antecedent must be the same 'is' as that in its consequent. In particular,
there can be no shift from an alleged predicative sense of 'is' to an alleged
existential sense of 'is'. Since, as I shall be arguilig in the next section,
there are no such senses of 'is', this is not a problem. But also there must
not be any of the more subtle problems such as beset (G). And here, as
we shall see, trouble may arise. Aristotle is going to abandon (E), as aid
Plato, but in neither case does it have to do with shifting from a
predicative to an existential sense of 'is'. In Aristotle's case, it is not so
much a matter of abandoning (E) as of restricting its range of operation
while preserving the idea that the consequent in the allowable cases is a
simplification of the antecedent, and so contains the same 'is'.
But still, I must part company with those whose distaste for degrees
of existence leads them to read it out of Plato. As I have said, "Socrates
is" is, as far as I can tell, the way a Greek goes about asserting the existence of Socrates. Aristotle himself may be found using this format in
Cat. 10, 13 b 16-17,19, etc. For realism, one must imagine a conversation, say, between Xenophon and Cyrus. Xenophon has been talking
about Socrates, and Cyrus denies that there cuuld be anyone that
strange; Xenophon replies "Oh, Socrates exists, all right; I was at a party
with him just last month". Here a Greek might say (the Greek for) "Oh,

54

RUSSELL M. DANCY

Socrates is, all right; .... " So Plato is, as far as I can tell, committed
to degrees of existence, as we might phrase the matter; when he says that
the forms 'purely' and 'entirely' are, what he means can be put into
English by saying that their hold on existence is absolute because they are
what they are irrefragably, while Helen's hold is relative and weak,
because some of what she is, if not all of what she is, she also is not.
Then what is it that condemns (E)? I have already said something
about that, although in the 'deep' terminology; before putting it more
austerely, there are a couple of obvious problems with it that are not
directly covered by the theory I shall have Aristotle adopting, which
ought to be mentioned so as to avoid raising false hopes.
First, there is a problem recognized and dealt with by both Plato and
Aristotle: the notorious problem of what is not. Unicorns (a modern
favorite) and goat-stags (Aristotle's favorite) do not exist: they are not.
But goat-stags are believed in by some and used as examples by others.
If we apply (E) to that last sentence, we end up having them in the zoo
that is reality. The restriction on (E) needed here is not the one I am
primarily after, although it is close by, in Aristotle's mind. As I see it,
it would involve isolating the predicates, like 'believed in' and 'used as
examples', that give trouble, call them 'intensional predicates', and pronouncing (E) unfit for use with intensional predicates. Aristotle does not,
I think, see it quite that way. But I am not going to worry any more about
that here.
A second interesting problem not covered by the theory to come is provided by certain adjectives such as 'fake', 'mythical', and, particularly
for Aristotle, 'dead'. A fake diamond is no diamond. So we do not want
it to follow from 'that diamond is fake' that that diamond is (exists), and
Aristotle does not want it to follow from 'that man is dead' that that man
is. Here again we may simply label the problem and shelve it: call such
adjectives 'adjectives alienantia' or 'alienating adjectives'; 19 (E) does not
work for alienating adjectives as predicates.
Now what is the trouble with (E)?
As Plato uses it and (NeE) in the Republic, every predicate true of S
contributes equally to its existence, and every predicate false of its
detracts equally from its existence (ignoring intensional predicates and
adjectives alienantia). I shall refer to this as the 'democratic' attitude
toward S's predicates.
For Plato, the democratic attitude is a problem because there are many
things a form is not, yet its existence must be unadulterated. For Aristotle, the democratic attitude is a problem for a different reason.

ARISTOTLE AND EXISTENCE

55

The democratic attitude toward S's predicates suggests, as an account

of what S's existence consists in: the logical sum of all the predicates true
of S. But that account of S's existence ties S down entirely 'too firmly.
Take Socrates as our candidate S, and let him be pale. First, then: he can
lose his pallor at the local parlor and become dark; he is still Socrates.
Of course, we could simply clock his pallor, and then his being pale on
Thursday afternoon would be part of his existence, along with his being
dark Thursday night. But, again, he might perfectly well have been there
Thursday night, as much Socrat~s as ever, even if he had missed his appointment at the parlor and remained white as a sheet. The trouble is that
the loss of Socrates' pallor is perfectly compatible with the retention of
Socrates, while with the loss of other things - notably, the loss of his
life, and perhaps the loss of his humanity - we seem to lose something
more: namely, Socrates.
The view just registered is by now pretty familiar, and equally controversial. It rejects the democratic attitude toward a subject's predicates
in favor of a form of elitism that we can call 'essentialism;. This is a red
flag to certain bulls. These bulls can be found in quite different pens. In
one, there are the innocent democrats we have already met, for whom
any of Socrates' predicates is part of his being as much as any other:
these, in effect, believe that all relations are internal, and we might think
of them as Idealists. The other pen has in it a contemptuous lot who share
with the democratic Idealists an egalitarian attitude toward Socrates'
predicates but will have nothing to do with any of them: these, who insist
that none of Socrates' predicates is any part of his being, we may call
'Existentialists' .
I shall not here argue against any of these anti-essentialist bulls; I simply content myself with name-calling. It is not news to anyone that Aristotle is an essentialist. Let us just suppose he is right.
We shall need some of the terminology in which he expresses his
essentialism.
The predicate in
(3)

Socrates is a man

is one that holds of the subject 'by virtue of itself', 'in its own right', 'of
itself', or per (or propter) se. These are varying translations of the single
phrase' xed)' a,vro' .20 Let us adopt the first one: the predicate in (3) holds
of the subject by virtue of itself. By contrast, the predicate in
(2)

Socrates is pale

56

RUSSELL M. DANCY

holds of its subject accidentally, 'by virtue of an accident' (xaTa)

avp.{3e{3T/ x os).

This distinction pertains to the relationship between a predicate and its

subject, but Aristotle attaches it to the sentence at different points at different times. Initially, the reflexive pronoun (auTo, 'itself') refers back
to the subject: its gender, number, and person are frequently determined
by the corresponding features of the subject (as in Met. Z 4. 1029b
14-16, quoted below; for some other samples, see ilI8. 1022a 26-27,
28,31, [bis], 34, Z 3. 1029a 20,5. 1030b 19). But sometimes the phrase
goes along with the predicate, when the predicate is taken in isolation,
with only a more or less dim view of a possible subject: then the predicate
becomes a 'by-virtue-of itself' term (as in An. post. A 4. 73 a 34ff, 22. 84a
13 - 17, Met. Z 4. 1029b 9, etc.).
These different ways of employing the distinction are plainly connected, but it is not easy to say how. We might try, as a first approximation: a term T is a by-virtue-of itself term iff, whenever any subject S is
T, S is by-virtue-of itself (viz., S) T. This can only be an approximation,
since counterexamples or putative counter-examples abound: 'that pare
thing is a man', e.g., which has as predicate a by-virtue-of itself term,
but pale things are not by virtue of themselves (i.e., by virtue of being
pale things) men. Aristotle takes the line that 'that pale thing is a man'
is not of the form'S is T' (An. post. A 22. 83 a 1-17). The fact is, I think,
that the 'by-virtue-ofitself' terminology carries with it an intensional element, and when Aristotle uses that terminology in An. post. A 22 in denying that 'that pale thing is a man' is of the form'S is T', he :mports
an intensional element right into the form'S is T'. But that is part of
another story. 21
There is another point in the sentence at which Aristotle will occasionally attach his distinction, and this will be part of the present story:
he is prepared to distinguish the 'is' that occurs in (3) from the 'is' that
occurs in (2), so that the 'is' of (3) is a by-virtue-of itself 'is', and that
of (2) is an accidental 'is'. Here I suspect a blunder, of a type to be considered in the next section.
For the present, we can make do with the initial form of the distinction: Socrates is by virtue of himself a man, but accidentally pale. It is
in these terms that Aristotle defines 22 the phrase that is conventionally
translated as 'essence' (Met. Z 4. 1029b 13 - 14):
... the essence of each thing is what it is said [to bel by virtue of itself (luTL TO TL

~P

eben iXCxUTW 0 'Af-yeTcXI xcd), aUTO).

In fact, the phrase (TO TL 1}v elvcn + dative) conventionally translated

57

ARISTOTLE AND EXISTENCE

'essence' is a curious one,23 and what is of interest here is that its core
is an occurrence of the infinitive 'to be'. So, instead of the conventional
"the essence of each thing", as above, I shall adopt the more literal
"what it is24 for each thing to be". I take this to amount to things like
'what to be is for each thing', 'what each thing's being is, or consists in',
etc. Then the above passage will be retranslated, along with its immediate
sequel, as follows (1029 b 13 -16)
... what it is for each thing to be is what it is said [to be) by virtue of itself. For it is not
[so) that for you to be is for [you) to be educated,zs for you are not by virtue of yourself
educated. Therefore, [it's) what [you are) by virtue of yourself. (lcrTL TO TL ~V
elvw l}(ixcrT~ 0 >.i-YfTW }(CII')' aUTO. ob -yixe lcrTL TO crOL dva, TO JLovcr,)(~ dvw.
-yae }(aTa
cravTov d JLovcr,}(os. 0 aea }(aTa cravTov.)

ou

If we take this to give us different senses of 'to be' for Socrates on the
one hand (for whom 'to be' means 'to be a man') and Bucephalus on the
other (for whom 'to be' means 'to be a horse'), we are committing
another blunder of the type mentioned above and to be dealt with below.
We need not. We might just take it that filling in the predicate tells us
under what conditions 'Socrates is' (i.e., 'Socrates exists') is true, and
leave open the a)ternative possibility (which is, in fact, the right one) that
what determines which predicate to fill in is, not 'is', but 'Socrates'. For
the remainder of this section, I shall take it that this alternative possibility
has not been ruled out, and leave further discussion for the next.
Consider once more the two inferences

(E)

S is P

-+

S is

and
(NCE) S is not P

-+

S is not.

It is not that the essentialist view we have just handed Aristotle requires
him to reject the unrestricted use of (E) (again, leaving intensional and
alienating predicates out of the picture). In the first instance, what he
must reject is its converse, by restricting its application to cases in which
'P' is a predicate essentially true of S: cases in which S is by virtue of itself
p.26 And so he must reject (NeE): only where S is P by virtue of itself,
essentially, does it follow from'S is not P' that S is not.
But all this has, given the background, an indirect impact on (E) as
well. For the justification for the inference (E) was supposed to be simp)e
simplification, the sort of move that justifies
(P)

Socrates is a pale man

-+

Socrates is pale

58

RUSSELL M. DANCY

but breaks down for

(L)

Socrates is a lousy provider

-+

Socrates is lousy.

For (P), that meant that 'pale' had to have the same role in antecedent
and for
(G)

Socrates is a good cobbler

-+

Socrates is good.

(L) and (G) were ruled out on the ground that the simplified residue had
changed character in the course of the simplification. (P) was to be allowed because its residue was constant. And so, if (E) is to survive, it
must show a constant residue.
Consider, then, the instance of (E) that reads
(el)

Socrates is pale

-+

Socrates is.

This is to be an inference with a detachable consequent. So we write its

consequent on a slip of paper and put it to sea in a bottle. But when it
washes up on a distant shore, its readers, who know that "what it is for
each thing to be is what it is said t.o be by virtue of itself", and know
that what Socrates is by virtue of himself is a man, will supply 'a man'
as the completing noun. So (el) is a failure, and (E) is not in general true.
But no such failure is involved with
(e2)

Socrates is a man

-+

Socrates is.

(E), restricted to essential or by-virtue-of itself predicates, survives.

In fact, confining (E) to essential predications is overly restrictive.
Consider
(e3)

That pale thing is a man

-+

that pale thing is.

We should not count 'that pale thing is a man' as a straightforward essentialpredication, and, in fact, Aristotle would not so count it. To the extent that he is willing to count it as a predication at all, he thinks of it
as accidental: he thinks of it as grounded in, or having for its truth conditions, 'that man is pale' (see Section 4 below). And, as we have noted,
in the Posterior Analytics he would prefer to legislate it away: "either it
doesn't predicate at all, or it predicates not simply, but accidentally"
(~TOL p:"ocxp.ws xcxrrnoeeiv, 7} xcxrq'Yoeeiv p.Ev 1'1, a1rAWS, XCXTCx
CTvp.{3e{31/xos oE xcxrq'Yoeeiv, 83 a 15 - 17).
For simplicity of outline, let us allow him the legislation he tries to pass
in the Posterior Analytics. We shall expand our horizon slightly in Section 4 below.

ARISTOTLE AND EXISTENCE

59

We are now in a position to explain how Aristotle can suppose that to

be is to be something or other without giving up the distinction between
'to be something or other' and just 'to be'.
Take the latter, the distinction, first. According to this, dven n, to be
something, or 'to be' + Pred., is not the same as dven cnrAws, just 'to
be#'. The latter phrase is often translated 'to be without qualification',
or, worse, 'to be absolutely'. These translations tempt one to make too
much of the distinction. In fact, all Aristotle's phrase 'Elven a1rAWS' does
is point to the 'is' in sentences such as
(I)

Socrates is.

Here we have just 'is', not 'is' + Pred. Such occurrences mark existential
sentences: if there were an existential sense of 'is' in Aristotle, this would
be it. But there is plainly no special existential sense here; there is a merely
syntactic difference between 'to be' followed by a predicate and 'to
be #'. The latter results flom truncating a predication. So even 'to be #'
is, implicitly, to be something.
But, as we have seen, not just any predication may be so truncated:
the predicate must be restorable. Under the simplifying assumption that
such locations as 'that pale thing is a man' can be ruled out, the type of
predication that allows truncation is an essential predication. So while to
be# is to be something, it is not to be any old thing. In fact, (1) is elliptical for
(3)

Socrates is a man.

That is the theory. Let us translate it into English .

2. BEING AND EXISTING IN ENGLISH

The word 'is' does not behave the same way in normal English: sentence
(I) seems to be used in English primarily to translate philosophical
Greek. There is a more or less archaic use of the verb 'be' that survives
in recitations of Hamlet's soliloquy and Owen's example 'Arrowby is no
more',26 and our ability to understand the locution may give a
fingernail-hold for the theory in English, but it would be nice to have
more.
The first thing to notice is the oddity already mentioned about
(4)

Socrates exists.

I have in mind not the oddity that can be corrected by changing tenses

60

RUSSELL M. DANCY

and adding adverbs, but the requirement placed on understanding the

sentence of some background knowledge associated with the name
'Socrates': if you spring it on someone not in the know, he may not have
the faintest idea what is going on. So also with the sentence-form
(5)

There is (such a - as) Plato.

Without the parenthetical material, properly filled in, this does not even
scan.!7 And properly filling in the blank requires some fairly substantial
covering noun or noun phrase: just 'thing', for example, will only work
in quite special cases ("My good man, there is such a thing as decency,
you know", and so on). 'Plato' is the name of a computer language as
well as that of a philosopher, and, for that matter, I could use it to name
my pet cat, my pet theory, my favorite number, the national debt, the
coin Washington skimmed across the Delaware, or the value of that coin
in real terms on Christman Day 1776 given in terms of the Laspeyres index. In the absence of a covering phrase that has some bite to it, we do
not know what is being talked about.
It is not altogether clear, in that case, what pigeon-hole to place the
ignorance in. Some will, no doubt, feel the urge to argue that it is not
linguistically relevant. The pattern for such an argument might be this:
invoke the sense-reference distinction and couple it (although this would
not have made Frege very happy), with Mill's doctrine of proper names.
According to that doctrine (translated into Fregese), proper names do
not have sense. So when someone says 'I'm studying Plato' (or 'I'M
STUDYING PLATO'), and it is not clear to you whether she is trying
to grasp the theory of recollection or the theory of programmed learning,
that does not mean you have failed to get the sense of what she is saying:
you have got all the sense there is to get if you know the rest of her words
and, perhaps, the minimal fact that 'Plato' (or 'PLATO') is a proper
name.
This seems to me one of the many places where the sense-reference
distinction and its alliance with Mill's theory of proper names lead us
astray. It seems to me clear that there is something about the sense of
what she said that we are not getting, and somethin associated with the
proper name in it. 28 But I am not able to layout a proper account of the
matter, at least not here. So I shall simply assert that (4) and (5) require
for their understanding, for grasping their sense, some fairly substantial
covering noun or covering noun-phrase, and admit that I do not know
whether the 'explication' (I shudder at the word) of making sense belongs
to the science of semantics, that of pragmatics, or necromancy.

ARISTOTLE AND EXISTENCE

61

If this is correct, Aristotle is at least partly right: the demand for an

informative covering noun or phrase is his demand for an answer to the
question 'what is Socrates?' to complete the sense of 'Socrates is'. It is
no accident that this is close to some of the things that have recently been
said 29 in connection with identity by way of reviving Aristotle's
essentialism.
But as things are, an argument may suggest itself that would undermine any further steps in beating Aristotle's views into Anglo-Saxon
shape. For it may seem that, whatever the facts about Greek (which is,
after all, a dead language and perhaps should be allowed to rest in peace),
we do in English have a special sense for 'is' that is plainly existential.
Consider the 'is' in

(6)

There is such a person as Plato,

(again, embedded in a conversational context that gives it naturalness),

or, better, the 'was' in 'there really was such a man as Prester John'.
These verbs can be replaced by 'exists' and 'existed', respectively. But the
'is' in 'Plato is a person (who lives in Athens ... )' and the 'was' in
'Prester John was a man (who ... )' cannot be so replaced. So the 'is'
in (6) is existential.
I take it no one really falls for such replacement-arguments without
further support. But, in the absence of any alternative account of the 'is'
in (6), this one may cause consternation. So I shall say something about
how an alternative account might go, without pretending to completeness or even correctness. I think we can see enough to tell that these
are the right lines. 30
Consider the following sentences:
(7)
(8)
(9)
(10)
(11)
(12)
(13)

There arose a furious clamor.

There ensued a riot.
There followed the entire retinue.
There fought by his side the bravest of the Romans.
There resulted the worst disaster he remembered.
Outside her window there sang a popular singer.
There sat next to me the saddest man in the world.

We do not need a special existential sense of 'arise', 'ensue', 'follow',

'fight', 'result', 'sing', or 'sit' to explain how we can have such sentences
alongside their counterparts with the verb in the 'predicative' position.
The latter are sentences such as

62

RUSSELL M. DANCY

( 7')
(10')
(13' )

A furious clamor arose.

The bravest of the Romans fought by this side.
The saddest man in the world sat next to me.

And what we require is (to put it in a way that was fashionable in

linguistics last week or the week before) a transformation that will get us
from such sentences (or the structures that underly them) as (7') - (13')
to (7) - (13). The transformation would be labeled 'THERE-insertion'
and would look a little like this:
(T)

X - NP! - V (NP2) y X - there - V (such NP 2 as) Y NP!.

At least, this would generate (7) - (13) from (7') - 13'). And, just
among others, it would generate 'There is such a man as Plato' from
'Plato is a man'.
I said I would not pretend to correctness. And (T) is, accordingly, not
correct. It would generate some things we - or at least I - do not want.
E.g., starting from
(14')

These books weigh a lot

We could use (T) to derive

*(14)

There weigh a lot these books.

And it is not clear where the stuff represented by 'X' in (T) should go:
for the move from (13') to (13), (T) seems all right as it stands ('next to
me' is carried along with the verb), but with
(15')

Mice are in the bathtub

(T) would generate

*(15)

There are in the bathtub mice

where what we (at any rate, I) really want is

(15)

There are mice in the bathtub.

So the rule requires a lot of tinkering with.

The point is that, blemished as it is, this is the right sort of thing: we
do not want to rewrite the dictionary to take account of all these newly
discovered existential senses of all these verbs. But if this is the right 30rt
of thing, 'is' becomes just one verb among others subject to the transformation, and bears no special sense when it follows 'there' as in (5) or (6):

ARISTOTLE AND EXISTENCE

63

it is simply the same verb as in 'Plato is a man' with the surrounding

sentence reprocessed. 31
So we are left with 'simple sentences': subject-predicate sentences.
These frequently employ 'is'. But this word is, after all, quite meaningless: it is merely a syntactic device for connecting subjects and
predicates where the predicates are not already verb phrases, just as the
word 'thing' is a device for turning adjectival phrases and others (e.g. 'of
the mind') into noun phrases (e.g., 'things of the mind'), and just as 'to'
is merely a device for turning verbs into infinitives (e.g., 'to walk', 'to
be or not to be,).32
So Heidegger is right:
But being remains unfindable, almost like nothing, or in the end entirely like that: The
word 'being' is finally, then, only an empty word. It means nothing actual, tangible, real.
Its meaning is an unactual vapor. 33

But for Heidegger, this represents a lamentable loss: apparently, for the
Greeks before Plato, Erven was chockful to bursting with meariing. And
about that, Heidegger is wrong. But fifty percent is not bad.
So far all we have done is to show that the part of Aristotle's theory
that seems at first provincial to Greek may, in fact, have something corresponding to it in English. But, of course, that is not all there is to the
theory. For one thing, there is also its essentialism. This is not, on the
fact of it, a linguistic matter: the doctrine of essentialism is as plausible
or implausible in English, I take it, as it is in Greek. But it scmetimes
sounds as if the doctrine might not be plausible in Chinese,34 or
Nootka,35 or Rortyspeak .36 Here I am going to leave these languages,
along with others of which I am innocent, on one side.
But there is one thing about Aristotle's employment of his essentialism
that must be mentioned, that is closely connected with the presence or
absence of different senses of 'is' in English or in Greek.
To get at this, first consider another distinction that some have alleged
to be pertinent to English and Greek being: that between identity and
predication. 37 There is certainly a distinction here: it is between claims
like
(16)

Dr. Jekyll is Mr. Hyde.

on the one hand, and claims like

(17)

Dr. Jekyll is schizoid

64

RUSSELL M. DANCY

or
(18)

Dr. Jekyll is an addict

on the other.(16) states an identity; (17) and (18), I think,38 do not, and
I shall characterize both as predications.
Philosophers are, notoriously, myopic, and when they apply their
magnifying glasses to sentences that do different things, they usually can
only manage to bring into focus a single word. 39 Here the word is 'is'.
They profess to spot the 'is' of identity in (16) and the 'is' of predication
in (17), and then fall to arguing over the character of the 'is' in (18).
I shall call this habit of supposing that every difference in character
from one sentence to the next must be locatable in single ambiguous
words the 'fallacy of the magnifying glass'. We do not need to pull the
word 'is' out of those sentences and go into Angst over the meaning of
being. The situation is completely described by saying that 'is' is followed
in (16) by a singular definite noun phrase, in (17) by an adjective, and
in (18) by an indefinite noun phrase. Sentences that show the structure
of (16) state identities; the others are predications. But this has nothing
to do with the occurence of 'is' in different colors: it is what comes next
that counts. 40
Aristotle does not commit the fallacy of the magnifying glass in connection with 'dvOIL', identity, and predication. 41 But he is prone to the
fallacy; conceivably he is its inventor; probably42 he is its most influential
perpetrator. And if he falls into it anywhere, he falls into it with 'is'. He
thinks that the fact that 'man' and 'pale' relate differently to Socrates
makes for a different 'is' in
(3)

Socrates is a man

from the one in

(2)

Socrates is pale,

the former is an 'is' xed)' CXUTO, a by-virtue-of itself 'is'; the latter is an
'is' XCXTO: avp.(3f(31]XOS, an accidential 'is' (see, e.g., Met. d 7. 1017a 7 - 8,
and below, Section 4).
Unless there is more to it, this is just as much a case of the fallacy of
the magnifying glass as the idea that 'is' varies in sense from identities
to predications.
Perhaps it is a natural mistake. That Socrates is a man is essential to
him; .that he is pale is accidental to him. These are different relationships.
They are both covered by the word 'is'. So it is natural to talk as if there

ARISTOTLE AND EXISTENCE

65

were two varieties of 'is'. In the sequel, I shall occasionally go along with
Aristotle's talking that way.
But I do it under protest. If this way of talking is not to be explained
away merely as a way of talking, but is to be taken as marking a genuine
distinction in senses of 'is', or concepts of being, or whatever, it is a case
of the fallacy of the magnifying glass. It would be a great relief to get
Aristotle off this particular hook. I do not at present know how to do it.
Rephrased in these terms, and still leaving 'that pale thing is a man'
out of account, the theory of Section 1 becomes this. The 'is' of (3) is
a by-virtue-of itself 'is', and permits cancellation of the predicate. The
'is' of the residue, 'Socrates is', is then a by-virtue-of itself 'is', and also
an 'is' &'1I'~Ws, 'is#', since it no longer has an attached predicate. But
these are two different characterizations: roughly, the latter is syntactic
and the former semantic. To go one step farther along the road to perdition, we might say: the existential 'is' is, paradigmatically (see Section 4),
a by-virtue-of itself 'is', but the by-virtue-of itself 'is' is, in the first instance, a predicative 'is', and only becomes existential by cancellation of
the predicate.
3. DE INTERPRETA TIONE II. 21" 25-28.

In the final lines (21 a 18 - 34) of De into 11, Aristotle

is investigating the rule opposite to the one just dealt with: before he was investigating when
things predicated on their own could also be predicated together. but in these lines he is
investigating when things predicated together can also be predicated separately.

So Ammonius (in De into 210. 17 - 20)43 rightly says. Aristotle has just
been pointing out, for example, that the argumentative rule we might call
'Addition':
(A)

S is Adj. & S is NP

-+

S is Adj. NP

does not always work: from 'Socrates is good', where we have "simply
'good' " or 'good #' (b1l'~ws ayat'Jos, 21 a15) and 'Socrates is a cobbler',
you cannot use (A) to get 'Socrates is a good cobbler' (20 b 35 - 36).44
Now he notes that the rule we have been calling 'Simplification' also
needs restricting. He is very brief about this: he mentions all three of the
exceptions we listed in Section 1 above, namely things that are not (see
21 a 32 - 33), dead men (see 21 a 23), and inferences to existence based on
the wrong predicate, and it is not easy to see what he is trying to say about
them. I have excused myself from dealing with the first two here. What
he has to say about the third is this.
21 a 18 -21 tell us:

66

RUSSELL M. DANCY

It is true to speak of the particular [more literally, 'of the something'] simply as well, for
example, [to say that] the particular man [is a] man, or the particular pale man [is] pale,
but not always ... (ah'1t?~~ 0' fUTiv fl1f'fiv xCiTa TOU TtVO~ xCii (X1f'hW~, olov TOV Ttva avt?eW1f'OV avt?eW1f'OV ij TOV Ttva hfUXOV avt?eW1f'OV hfUXCIV' oux afi Of. ... ).

It is not clear what the first of these examples is telling us we are permitted to do,45 but the second is plainly allowing us the inference labeled
'(P)' in Section 1 above. 46
Aristotle then points out that simplification will not work with 'dead
man': in cases of this kind, where there is a contradiction between the
two terms antecedent to the simplification, the rule never works. (There
is a puzzle here: apparently Aristotlethinks 'this is a dead man' can be
true, although 'dead man' involves a sort of contradiction.) But in other
cases, sometimes it works and sometimes it does not;
for example, Homer is something, e.g., a poet; well, is he, therefore, or not? For 'is' is
predicated accidentally of Homer; for because he is a poet, not by-virtue-of itself, 'is' is
predicated of Homer. (21 a 25 - 28: wune OJL'1eO~ fUT! Tt, oiov 1f'OI'1rT,~' &e' oOV
XCi! EUTlV, ij ou; xCiTa UUJL(jf(j'1XO~ ,),ae XCiT7J')'OefiTCiI TO EUTIV TOU 'OJL~eOU' OTt ,),ae
1f'OI'1rT,~ iUTlV, aM' OU XCit?' CiUTO, XClT7J,),oefiTClI xCiTa TOU 'OJL~eOU TO EUTlV.)

The traditional way of taking these lines is as saying that it does not
follow from
(19)

Homer is a poet

that
(20

Homer is.

And along with this, often the denial that 'is' by-virtue-of itself applies
to Homer is understood to be the denial that the existential 'is' applies
to Homer (and then Aristotle must have had advanced views about the
'Homeric question').
In my view, neither of these things is correct. That is, to be as explicit
as possible: Aristotle is not denying that (20) follows from (19), and
"'is' by-virtue-of itself" is not the 'is' of existence. 47 I take up these
points in reverse order. Both bring in the theory of Section 1. The latter
is a simple application of it.
The theory of Section 1 tells us that the 'is' of 'Homer is a man' is a
}(cxf)' CXUTO 'is', a by-virtue-of itself 'is', yielding 'Homer is #', and that
this is the existential claim, 'Homer exists'. And the theory tells us that
the 'is' of (19) is a }(CXTa (JUP.{3f{3.,.,}(O~ 'is', an accidental 'is', and so
simplification may not be performed on (19). That is what Aristotle is
here telling us. Since he is giving as a reason for prohibiting simplifica-

ARISTOTLE AND EXISTENCE

67

tion the claim that 'is' is not here used xat'}' aUTO, he cannot mean by
this claim that 'is' is not here used existentially. That, after all, is the
question. He is simply saying: since (19) is not an essential predication,
and its 'is' is therefore not xat'}' aUTO, the predicate may not be canceled
by simplification to yield 'is' a7rhWS, the "existential 'is'''.
This way of putting it yields considerable ground to the fallacy of the
magnifying glass, but it certainly shortens the work.
So much for the second point: simplification cannot be applied to (19)
because its 'is' does not apply to Homer xat'}' aUTO. But then the first
point is clear: Aristotle is not saying that (20) does not follow from (19),
but that it does not follow by simplification. He puts forth the question:
"Homer is ... a poet; well, is he, therefore, or not?" In its context it is
not hard to read this, not as asking "does it follow, by some devious
means or other, that Homer is?", but "does the move we are talking
about apply here?"
This may, at first sight, seem unnecessary. But there are two good
reasons why we should not take Aristotle to be saying that it does not
follow from 'Homer is a poet' that 'Homer is'.
The first is that we know from elsewhere that, when it comes to the
question whether (20) follows from (19), he should say that it does. This
is a point he considers in Categories 10, and quite unambiguously
decides: when there is no Socrates, 'Socrates is healthy', 'Socrates is
sick', 'Socrates is blind', etc., are false, and their negations true. The apparent conflict of that passage with De into 11 is a notorious cruX. 48 In
my view, there is no conflict.
The second reason for avoiding making Aristotle deny the entailment
is that there is a very simple argument in favor of it, based in part on the
De into passage itself. Homer is a poet. But poets are, after all, human;
so they are men. So Homer is a man. But there the 'is' is a xat'}' aUTO
'is', and we can simplify. So Homer is. This argument is, admittedly,
phrased in terms of my interpretation of the passage. But there is very
little about it that is speci fie to that interpretation. On virtually any
understanding of the passage, the inference has to fail because the 'is' of
(19) is not xm'}' aUTO applied to Homer, where the 'is' in 'Homer is a
man' is applied to him xat'}' aUTO. So all we need is the concession that
poets are men, and we are away. It is extremely difficult to see how tha!
could be denied. (It may be worth noting that it would not help to say:
well, for all (19) has to say, Homer might have been a god. Then we
would have another alternative, all right, but it would get us to the unwanted conclusion just as easily.)

68

RUSSELL M. DANCY

4. METAPHYSICS 67. 1017" 7 - 30

Here, especially, more needs to be said than I shall say.

This chapter gives us four main headings 49 under which to rank 'that
which is', 'to be', or 'is' (TO DV 1017 a6, bl,TO flvw a24, 31, 35, TO i!anv
a31). About the latter two, the 'is' that signifies that something is true
e31 - 35) and 'to be' signifying actually or potentially being (a35 - b9),
our theory has nothing special to say. But the first two are accidental being and by-virtue-of itself being, and that is what our theory is about.
But even here it will not explain everything. In particular, it will not
explain, unaided, the most controversial of Aristotle's claims under these
headings: the claim that 'to be' in the second of these ways of conceiving
it shows variation from one category to the next. So, although I shall say
a little to indicate how that claim might be dealt with, I shall not provide
a full defence.
It may seem that the sights have been lowered so far that we are in
danger of shooting our toes off, but there are still difficulties to be met.
Aristotle begins by stating that 'is' comes in an accidental variety and
a by-virtue-of itself variety (a7 - 8), and promptly gives examples of the
former
10):

e8 -

e.g., we say the just is cultivated, the man [is] cultivated, and
the cultivated [is a] man.
We may write:
(21)
(22)
(23)

The just [one] is [a] cultivated [one].

The man is [a] cultivated [one].
The cultivated [one] is [a] man.

The English is pretty awful. The bracketed material has nothing to correspond to it in the Greek: it serves to remind us that, where in English
the adjective 'cultivated' occurring as predicate is easily identified as a
predicate because it is an unsupplemented adjective, in Greek the
predicate adjective '/lovau(os' needs no supplementation in order to be
treated as a noun phrase.
These examples are to be construed as making reference to particular
people in each case, the fellow holding up that lamppost over there, say:
Aristotle is not talking about a maxim he and his friends like to utter to
the effect that the just man is a cultivated man.

69

ARISTOTLE AND EXISTENCE

He adds another example which, he says, is similar

(24)

nO - 11):

The cultivated [one] builds houses.

(Here 'builds houses' is a single verb, 'oixooop.f:iv'.) Notice that it is

relevantly dissimilar to the preceding cases: it contains no 'is', and what
we are talking about is accidental 'is'. Aristotle knows this, and promptly
explains that (24) is so
because it-is-accidental to the housebuilder to be cultivated, or to the cultivated [one to be
al housebuilder (all - 12: OT! (Jup./3i/3rl'lt T~ oixooop..yJ P.OU(JLX~ dvOIL ij T~ P.OU(JLX~
oixooop.",,).

This imports the needed 'to be'. Aristotle is plainly not saying that (24)
shows accidental 'to be', but its parapnrase, which he seems to think is
more fundamental, does. So even (24) rests for its truth on accidental
being.
Aristotle's comment on (24) leads him to state the truth conditions for
his examples: the general form for aJl, including the paraphrase for (24)
(but not (24) itself) is "that this is this signifies that this is-accidental to
this" (al3 - 14: TO -rae TOOf: dvcn TOOf: aT/p.CtLVH TO aup.{3f:{3T/XEvcn Tct>O(;
TOOI':). More particularly, in cases like (21), both terms, 'just' and
'cultivated', are-accidental to the same thing e16) - namely, it appears,
to the existent thing designated by the subject term; and in cases like (23),
the subject term is-accidental to the predicate term (al7 - 18).
These are paraphrases of sentences employing 'is' in predicative position. They do not themselves employ 'is' in that role, but in one of them,
the paraphrase for (22), 'is' does occur. When Aristotle :.ummarizes in
a19 - 22, it becomes clear that this occurrence is important:
Well then, things said to be accidentally are so said either because both belong to the same
[thing that) is, or because that [i.e., the predicate-term)so belongs to [a thing that) is, or
because that, to which that of which it is predicated belongs, is.

These are occurrences of 'is # ': the claims are existential claims, and they
are taken to justify the occurrence of 'is' in the predications (21) - (23).
He is assuming stating (21) commits one to the existence of something
both just and cultivated. And presumably he thinks that this requires the
existence of, say a man who is both just and cultivated. 51
In a 18 - 19 he had made an incidental comment that pointed that same
way: just as (23) is aJlowed because the cultivated belongs to the man,
so also the not pale is said to be, because that to which it is-accidental is (OUTW Of 'Af-yf701L
xed TO p.~ 'Awxov dvOlL, OT!

c;,

(JuP./3i/31/XfV, iXE;VO E(JT!V).

70

RUSSELL M. DANCY

So
(25)

The not pale [thing] is #

is, marginally or parenthetically, acceptable.

There are two possible derivations for its 'is # '.
One may be suggested by the comment just quoted, viz.:
(3)
(I)

Socrates is a man.
Socrates is # .

(I) is to follow from (3) in the prescribed way. Assume that it is Thursday
night, after Socrates' appointment at the tanning parlor, and
(26)

Socrates is [a] not pale [thing].

The using this and something like Leibniz' Law, we might get (25) from
(I). If we did, the 'is # ' of (25) would be that of (I), and so, ultimately,
that of (3), and so, again, a )Cod)' cxilTo 'is'.
But the immediately preceding context suggests an alternative. The
parenthetical comment of al8 - 19 is attached to a statement of the truthconditions for (23), 'The cultivated one is a man': this is so because the
cultivated is accidental to the man, and so, Aristotle adds, even the not
pale is said to be.
That sounds as if Aristotle had the following derivation in mind. Start,
as before, with (3), and assume (26); this yields
(27)

The not pale thing is a man.

Then cancel the predicate, yielding (25). If we do it this way, the 'is #'
of (25) is that of (27), and we have a case of 'is #' which Aristotle would
describe as also an accidental 'is'.
That was ruled out only on the theory under the restriction provided
by An. post. A 22. 83 a l - 23, that made 'that pale thing is a man', and
presumably (27) along with it, not a case of predication. But there was
nothing intrinsic to the theory that brought this restriction on. We could
not allow simplification to operate on
(2)

Socrates is [a] pale [thing],

because, once context-free, its predicate would not be restorable: the

answer to the question 'what is Socrates?' would lead to the completion
'a man'. But if we now count (27), et al., as accidental predications, we
have cases in which the restoration would work as well as it ever does:
the question 'what is the pale thing?' would be answered by 'a man'. (Of

71

ARISTOTLE AND EXISTENCE

course, to answer the question, you have to know what is being referred
to, but that is just as much true of 'what is Socrates?' as of 'what is the
pale thing?'.)
The effect of this understanding of the passage, to which I am inclined,
is to pry apart the 'existential "is" , and the xm'}' aUTO 'is' even farther:
the 'is' of (25) is the former but not the latter. Still, it would remain so
that the 'is' of (25) is dependent for its presence on a xat'}' aUTO 'is': that
of (3). So the general message so far is that standing behind every accidental 'is' there is a by-virtue-of itself 'is'.
What, then, about this latter 'is'? Aristotle has this to say
(1017 a 22 - 27):
As many things are said to be by-virtue-of themselves as the figures of predication signify,
for in as many ways as [they) are said, in so many ways 'to be' signifies. So, since of things
predicated, some signify what [itl is, some what-[it)-is-like, some how-big, some relative-to
what, some to do or to undergo, some where, some when, 'to be' signifies the same things
as each of these. (xali' aUTa Of dVaL hf-YfTat ouang Uf/p.aLVft Ta C1X~p.aTa rij5
xarrnogLa5' OUC<XW5

-yag

hi-YfTat,

TouaUTaXW5

TO

dvat

Uf/P.C.LVft.

inL

ouv

TWV

xCtrnogoUp.fVWV Ta p.fv TL lUTt CTf/P.OtLVft, Ta Of 71'OtOV, Ta Of 71'OUOV, Ta Of 71'g05 Tt, Ta

71'Otfiv
?raUXftv, Ta Of ?rov, Ta Of 71'OTf, IxauTVJ TOVTWV TO dVaL TauTo CTf/P.c<LVft.)

Of

About the only thing that is agreed on heie 52 is that Aristotle thinks that
'is' varies somehow from one category to the next.
Our theory tells us that the 'is' he has in view is that occurring in essential predications, 53 the paradigmatic source for occurrences of 'is #'. So
whatever variation there is will show up in existence-claims as well.
But what essential predications are at stake? The passage recalls
another, Top. A 9. There Aristotle lists his categories, naming the first
one 'what [it] is', as here, and then says (103 b 27 - 39):
It is clear from them S4 that one who signifies what [it) is sometimes signifies [a) substance
sometimes [a) how-big, sometimes [a) what-like, sometimes one of the categories
[xaTfl-yogLwV). For when, with a man set out, one says that what is set out is a man or an
animal, he says what [it) is and signifies a substance; when, with a pale color set out, he
says that what is set out is pale or a color, he says what it is and signifies [a) whatlike. And,
similarly, if, with a magnitude of a cubit set out, he says that what is set out is a magnitude
of a cubit, he says what it is and signifies [a) how-big. And similarly in the other [cases):
each such [term), both if it is said about itself [lav Tf aUTO neL aUToii) and if its genus is
said of it, signifies what [it) is, but whenever [it is said) about another [thing). it signifies,
not what [it) is, but how-big or what-like or one of the other categories.

[ouuLav),

Here Aristotle unhesitatingly moves from the point that "all premisses
signify either what [it] is or how-big or what-like or one
of the other categories" (b25 - 27), where the sentences under consideration are ones like our old friends (3) and (2) and

(1I'QoTauHs) . ..

(28)

Socrates is a cubit tall

72

RUSSELL M. DANCY

to sentences like (3) and

(29)
(30

This [color] is off-white

This [viz., off-white] is a color

and
(31)

This [height] is a cubit.

The sentences on this latter list are answers to 'what is it?' questions;
those on the former list are answers to various question, 55 one of which
is 'what is it?' (asked of the entities that Aristotle will call 'substances,).56
In the 'what is it?' list, the question is cut free from the first category,
that of substances, and ranges over the entities that are signified in
sentences on the variable list. (3), 'Socrates is a man', occurs on both
lists. 57 It shows an 'is' by-virtue-of itself, as we have been saying. So do
the other sentences on the 'what is it?' list. Here Aristotle describes those
sentences by saying that in each something is either "said about itself"
or "its genus is said of it"; the former terminology is familiar from An.
post. A 22.83 a24ff,58 and is, for present purposes, merely a variation on
the 'by-virtue-of itself' terminology.
The transition in Top. A 9 from the variable list to the 'what is it?' list
is fairly abrupt, but not as abrupt as that in Met . .:1 7, for Aristotle does
not there even bother to give examples. 59 He merely sketches the variable
list ("of things predicated, some signify what [it] is, some what-like,
... " 1017a24 - 27) and says that 'to be' will have a single force (to use
as neutral a word as I can think of) for each entry on the list (a27). But
he is discussing 'to be' }lod}'
so it looks (at least to me)60 as if he
is making the same transition.
He follows it with a comment that has caused consternation 61
e27-30):

auro,

for there is no difference between 'a man is flourishing' and 'a man flourishes', or between
'a man is walking' or 'cutting' and 'a man walks' or 'cuts', and similarly in the other cases
(OVt'JfP "yare Ola",eeEl TO apt?eW1rO~ U"y1a[pwp fUTIP ij TO apt?eW1rO~ U"y1a[PEI, OV& TO apt?eW1rO~ l3aOllwP lUTIP ij Tep.pwp Toii apt?eW1rO~ l3aOlIEl ij Tep.PfI" bp.o[w~ Of xal f1rl
TWP aAAwp).

On any reading, this is going to be elliptical. The most natural one seems
to me this: just as, at 1017 a lO-12, he had introduced an example
(sentence (24), "the cultivated one builds houses") that did not employ
'is', and tried to show that it rested on an occurrence of 'is' anyway, so
here he notices that on his list of categories there are two at least that are
easily invoked in predications without using 'is': to do and to undergo;

ARISTOTLE AND EXISTENCE

73

and others are easily imaginable, even if the category-name does not, as
it does in these cases, make it obvious. 62 So he points out that here, too,
there is an 'is' in the offing: where you have a sentence containing a verb
other than 'to be', it may be replaced by a periphrastic verb phrase containing 'to be'. 63 The examples employ 'is' accidentally, but this would
not have bothered Aristotle, for the 'figures of predication' apply in the
first instance to the sentences on what we were just calling the 'variable'
list, where all the occurrences of 'is' but the first are accidental anyway.64
I have passed by the important and difficult question of why Aristotle
thinks that the 'is's of (3), (30), and (31) must all be different while those
of (29) and (30) (e.g.) are the same. I have, in fact, very little to say here.
Btlt there is, I think, a little more to it than one might think from what
has so far been said.
For, at this point, one might be tempted merely to re-invoke the fallacy
of the magnifying glass: Aristotle is, one might think, merely transferring a distinction that properly pertains to the predicate to the 'is' that
precedes (in English). 6S
But, in fact, Aristotle has an argument available, and, ~lthough a full
examination is beyond the scope of this paper, I should at least like to
state it, and then stop.
First, recall that our theory dictates that differences in the force of
essential 'is' carry with them differences in the force of existential 'is'.
This is to some extent usable as a two-way street: differences in the force
of 'is #' should show differences in the force of by-virtue-of itself 'is',
if the latter simply is the former with predicate uncanceled. 66
Notoriously, Aristotle denies that there is a genus of beings, that is,
a genus labeled 'that which is' (An. post. B 7. 92b I3-14). Less
notoriously, he provides an argument for that claim, in Met. B 3.
998 b 22-28 (see also Top. ~ 1. 121 a 14-19 for a related, but different,
argument): 67 a genus, he says, cannot be predicated of something unless
one of its species is predicated of that thing, and no species can be
predicated of its own differentia; so, if beings formed a genus, its differentiae could not be beings, that is, they would not exist, which is
absurd.
I do not at present see how to make anything convincing out of this.
But, as they say, tomorrow is another day.

74

RUSSELL M. DANCY

NOTES
I G. E. L. Owen, 'Logic and Metaphysics in Some Earlier Works of Aristotle', in
I. During and G. E. L. Owen (eds.), Aristotle and Plato in the Mid-Fourth Century (Studia
graeca ct latina gothoburgensia, XI, Giitenborg, 1960), pp. 163-90. The quotations are
from p. 165.
2 Particularly W. Jacobs, 'Aristotle on Nonreferring Subjects', Phronesis 24 (1979),
282 - 300. On pp. 297 - 98 (Note 6) Jacobs is critical of my note on De into II . 21 "25 - 27,
Appendix II to Sense and Contradiction: A Study in Aristotle (Reidel, Dordrecht, 1975),
pp. 153 - 55. Others have found that appendix obscure, but it was only Jacobs's note that
made me aware how grossly I might be misunderstood. The present effort (see Section 3
below) mayor may not help.
1 Especially in 'Aristotle on the Snares of Ontology', in R. Bambrough (ed.), New Essays
on Plato and Aristotle (Routledge & Kegan Paul, London, 1965), pp. 69 - 95.
4 The present paper is part of a larger project; I hope in carrying that out to remove some
of the more outrageous demands on the reader's credibility that the present paper presents.
S There are different verbs for 'to be' in the Greek of the period: for example, 'iJ1ro,QXfLv'
is sometimes used in Aristotle where we might expect 'dvcn' and translate 'to exist' (see
Bz. indo 788 b43ff). But it is also used occasionally as a copula (see ibid.). It could not be
used to 'disambiguate' 'dvO/l' if there were any ambiguity to disambiguate (see below).
6 E.g., F. H. Fobes, Philosophical Greek: An Introduction (University of Chicago Press,
1957), p. 51 Note I. Of course, this is not simply an artefact of introductory texts: it will
also be found in, for example, R. Kuhner and F. Blass, Ausfiihrliche Grammatik der
griechischen Sprache, Ier Teil (Hahn, Hannover, 1966) [reprint of ed. 3, 1890), Vol. I, p.
344, 90.2.
7 See here C. H. Kahn, The Verb 'Be' in Ancient Greek (Reidel, Dordrecht, 1973), Appendix II, 'On the Accent of lUTi and Its Position in the Sentence', pp. 420 - 34, esp. pp.
422-24.
8 For Greek examples (a great many) see Kahn, The Verb 'Be', Ch. VI.
9 E.g., N. Fleming and N. Wolterstorff, 'On "There is" " Philosophical Studies (U.S.)
11 (1960), 41 - 48; G. Vision, 'Existentials and Existents', Theoria47 (1981), 1- 30; Y. Ziv,
'Another Look at Definites in Existentials', Journal of Linguistics 18 (1982), 73 - 88.
10 G. Vlastos, 'Degrees of Reality in Plato', in R. Bambrough (ed.), New Essays on Plato
and Aristotle (Routledge & Kegan Paul, London, 1965), pp. 1 - 20, reprinted in Vlastos's
Platonic Studies (Princeton University Press, 1973, 1981). My disagreement with this article should not mask the fact that I am greatly indebted to it.
II Here I pass by an interpretative possibility given a fair run for its money by J. Gosling,'
Republic, Book V: Ta lI'o~a xcx}..o" etc.' , Phronesis 5 (1960), 116 -128. See also F. C.
White, 'J. Gosling on Ta 1/'o~a xcx}..o,', Phronesis 23 (1978), 127 - 32, 'The "Many" in
Republic 475a - 480', Canadian Journal of Philosophy 7 (1979), 291 - 306, and Gosling's
'Reply to White', idem 307 - 14. I am not convinced by Gosling, but I think the issue has
little bearing here.
12 The treatment of what might as well be a static situation (Helen's being simultaneously
beautiful and not beautiful) as if it involved change (Helen's vacillating between one and
the other) is characteristic: see Aristotle, Met. A 6. 987"32 - bIO, etc., and T . Irwin, 'Plato's
Heracleiteanism', Philosophical Quarterly 27 (1977), 1- 13 .
" For a review of the evidence with a rather different conclusion, see C. Kirwan, 'Plato
and Relativity', Phronesis 19 (1974),112 - 29. I seem to be adopting what Kirwan calls the

ARISTOTLE AND EXISTENCE

75

'relativist interpretation' of R. 479, and, worse, I am going to reject the one thing that he
says "is surely right in this interpretation, that it does not attribute to Plato any doctrine
about existence": see below. As for the 'relativist interpretation' in general, I cannot follow
Kirwan's emphasis on the formulation of the conclusion in 479b6 -7: there it is said that
big things are no more big than small, and Kirwan wants to say that that does not mean
they are both big and small. But they are picked out as big things, so they are at least that,
and if they are no more big than small, it follows from that that they are small. I am not
sure how much of an issue there is between us here. I certainly do not want to say that Plato
at the time of writing the Republic was aware that 'big', 'beautiful' and other relational
predicates were somehow special, and that is one of Kirwan's primary targets.
14 It is here especially that I am in agreement with Vlastos (see Note 10 above). The
disagreement comes when I say, below, that the concept of being that operates in this argument just is Plato's concept of existence. Here I am siding with Owen, 'Aristotle on the
Snares of Ontology', p. 71. Owen partially retracts this in 'Plato on Not-Being', in G.
Vlastos (ed.), Plato: A Collection oj Critical Essays, Vol. I (Doubleday Anchor Books,
Garden City, N.!Y., 1970), pp. 223 - 67 and 265 - 67, but his retraction concerns only the
Sophist, as far as I can tell, and I am saying as little as I can about that dialogue here (see
the next paragraph).
15 On the Sophist's treatment of 'is' see M. Frede, Priidikation und Existenzaussage:
Platons Gebrauch von " .. .ist . .. " und" . .. ist nicht . .. " im Sophistes (Vandenhoeck
& Ruprecht, G6ttingen, 1967). Particularly relevant to my theme above is Chapter II, pp.
37ff.
16 W. KamIah, Platons Selbstkritik im Sophistes (C. H. Beck, Munich, 1963 [= Zetemata
33)) has much to say that is relevant here (see esp. Chapters V & VI), but he does not seem
to see in Sph. 256e5 - 6 the flat rejection of R. v 476 - 80 that I do.
17 I ignore here complications that have come to light in recent discussions of adjectives
(see, e.g., M. Platts, Ways oj Meaning [Routledge & Kegan Paul, London, 1979) pp. 16lff
on 'Adjectival Constructions' and the literature there cited, esp. 1. A. W. Kamp, 'Two
Theories about Adjectives', in E. L. Keenan (ed.), Formal Semantics oj Natural Language
[Cambridge University Press, 1975) pp. 123 - 55). I am, for purposes of expositiOl" adopting the simple-minded view that 'pale man' is merely a concatenation of 'pale' and 'man':
that, in Montague's terminology, 'pale' denotes an intersection function (see 'English as
a Formal Language', in Formal Philosophy: Selected Papers oj Richard Montague, ed.
by R. H. Thomason [Yale University Press, New Haven & London, 1974), pp. 188 - 221,
at p. 211). Roughly: the pale men are the pale things that also happen to be men. This, of
course, is false: what counts as a pale man varies with race, location, time of year, and so
on. (So far, then, 'pale' might be an attributive adjective in the terminology of the next
paragraph. In one of Kamp's two theories about adjectives, all adjectives are: see his The
Verb 'Be'). It is this that I am ignoring, and it does not i!ffect the point at issue.
18 P. Geach, 'Good and Evil', Analysis 17 (1956),33-42 (reprinted in P. Foot (ed.),
Theories oj Ethics [Oxford University Press, 1967), pp. 64 - 73 ab init.
19 See Geach, 'Good and Evil'. Kamp (The Verb 'Be', p. 125) calls them 'privative
adjectives' .
20 One of the faults of Ross's generally reliable (Oxford) translation of the Metaphysics
is the astonishing variety of renderings it shows for the crucial phrase 'xed}' aUTo': 'in virtue of itself' (0 18), 'propter se' (Z 4), 'in itself' (Z 3. 1029"20), 'of itself' ("24), 'in virtue
of its nature' (Z 5. 1030b 19 - 20), 'self-subsistent' (Z 6. 1031"28fl), and, no doubt, other
things as well.
21 There is a great deal more to be said here. See A. Code, 'Aristotle's Response to Quine's
Objections to Modal Logic', Journal oj Philosophical Logic 5 (1976), 159-86; F. 1.

76

RUSSELL M. DANCY

Pelletier, 'Sameness and Referential Opacity in Aristotle', Nous 13 (1979), 283 - 311, and
my 'On Some of Aristotle's First Thoughts about Substances', Philosophical Review 84
(1975), 338 - 73 (esp. 340 - 42, 365 - 68) and 'On Some of Aristotle's Second Thoughts
about Substances: Matter', Philosophical Review 87 (1978), 372 - 413. Frank Lewis finds
all of us objectionable in 'Accidental Sameness in Aristotle', Philosophical Studies
42 (1982), I - 36.
22 'Defines' should be taken with a grain of salt here, for in A 18. 10228 27 - 28 Aristotle
'defines' the phrase 'by virtue of itself by means of 'essence'. The point is that the two
go together, and there is no understanding the one without the other.
23 For a review of the literature, and an interpretation of the Phrase with which (I think)
I do not agree, see J. Owens, The Doctrine of Being in the Aristotelian Metaphysics (Pontifical Institute of Mediaeval Studies, Toronto, 1951, edition 2, 1%3), pp. 180ff.
24The Greek have the imperfect, 'was'. The explanation for this is not of importance in the
present context, but the one 1 opt for is one rejected by Ross, Aristotle's Metaphysics
(Clarendon Press, Oxford, 1924) 1 127: the imperfect is the 'philosophicM imperfect'. This
occurs in English as well: 1 get you to agree to'S is P', we continue talking, and 1 later want
to remind you of that, so 1 say'S was P' wasn't it?'. Here it could be that'S was P' was
'Two and two made four': there is no real implication of past ness about the tense. Often,
in Plato, the 'philosophical imperfect' is employed in appealing to a previously stated
definition. Ross (ibid.) objects that it "is used only when there has been an actual previous
discussion of the subject in hand, which is the case in but few of the passages in which TO
Tt 1}v flvm is used". But that is Aristotle for you: his (and the Academy's in general)
technical terminology is arrived at by detaching terms from their dialectical contexts. The
original context would have been an explicit definition, appeal to which would be made by
the phrase 'what it was for virtue to be', and then that just becomes a label for whatever
it is that virtue is correctly defined to be.
2S Alternatively: "for [something) educated to be", 1 have, admittedly, picked the translation that most favors my overall interpretation. But it could be done either way (and, in
fact, 1 arrived at it thinking in terms of the alternative translation).
26 'Aristotle on the Snares of Ontology', p. 71. For further examples, see OED s. v. 'be',
B 1 I (Vol. I, p. 717, col. 3).
27 This consideration plays a part in the literature cited in Note 9 above.
28 1 do not want to reject the sense-reference distinction, and yet1 do not find myself quite
comfortable with saying that proper names have sense. Among those who are comfortable
with saying that are P. Geach (Mental Acts [Routledge & Kegan Paul, London, 1957), pp.
66fO; M. Dummett (Frege: Philosophy of Language [Duckworth, London, 1973, 1981),
index 2nd edition only) s. v. 'proper names, sense of); and D. Wiggins ('Frege's Problem
of the Morning Star and the Evening Star', in M. Schirn (ed.), Studien zu Frege II: Logik
und Sprachphilosophie [Frommann - Holzboog, Stuttgart, 1976), pp. 221 - 55). S. Kripke
hedges here: see 'Naming and Necessity', in D. Davidson and G. Harman (eds.), Semantics
of Natural Language (Reidel, Dordrecht, 1972), pp. 253 - 355, at p. 322 (in the reprint,
Naming and Necessity [Harvard University Press, Cambridge, Mass., 1980), p. 127). See
also M. Lockwood, 'On Predicating Proper Names', Philosophical Review 84 (1975),
471 - 98, and J. Cargile, Paradoxes: A Study in Form and Predication (Cambridge University Press, 1979), Ch. 2, 'MiII's Theory of Names'.
29 P. Geach, 'Identity', Review of Metaphysics 21 (1967/68), 3 -12 (in Logic Matters
[University of California Press, Berkeley, 1972), pp. 238 - 47) and 'Ontological Relativity
and Relative Identity', in M. K. Munitz (ed.), Logic and Ontology (New York University
Press, 1973), pp. 287 - 302; D. Wiggins, Sameness and Substance (Harvard University

ARISTOTLE AND EXISTENCE

77

Press, Cambridge, Mass., 1980; a revision of Identity and Spatio- Temporal Continuity
[Blackwell, Oxford, 1967)); N. Griffin, Relative Identity (Oxford University Press, 1977),
and B. A. Brody, Identity and Essence (Princeton University Press, 1980), esp. Ch. I.
30 The following sketch took its departure from S. Soames and D. M. Perlmutter, Syntactic Argumentation and the Structure of English (University of California Press, Berkeley,
etc., 1979); see esp. pp. 46 - 52, some of the claims of which I shall implicitly be
challenging.
31 See also C. J. Fiilmore, 'The Case for Case', in E. Bach and R. T. Harms (eds.), Universals in Linguistic Theory (Holt, Rinehart & Winston, New York, etc., 1968), pp. 1-88,
esp. pp. 44ff; J. Lyons, 'A Note on Possessive, Existential and Locative Sentences', Foundations of Language 3 (1967), 390 - 96; 'Existence, Location, Possession and Transitivity',
in B. van Rootselaar and J. F. Staal (eds.), Logic, Methodology and Philosophy of Science
III (North-Holland, Amsterdam, 1968), pp. 495 - 504; Introduction to Theoretical
Linguistics (Cambridge University Press, 1968), pp. 388 - 90; K. Allan, 'A Note on the
Source of THERE in Existential Sentences', Foundations of Language 7 (1971), 1-18.
Allan's rl":ction of the attempt (made by Fillmore and Lyons) to treat 'there' as a locative
is correct, I think, but the above is independent of this issue.
32 This is a Fregean view: see 'Begriff und Gegenstand', Vierteljahrsschrift fiir
wissenschaftliche Philosophie 16 (1892), 192-205, at p. 94, where the copula is said to
serve "als blosst'~ Formwort der Aussage" (in Translations from the Philosophical
Writings of GOlllob Frege, ed. by P. Geach and M. Black [Blackwell, Oxford, 1960], p.
43 "as a mere verbal sign of predication"). The view is heartily endorsed by Geach: see
Reference and Generality (Cornell University Press, Ithaca, N.Y., 1962, 1968), p. 34,
where Geach ascribes the view to Aristotle as well, on the strength of An.pr. A I. 24 b l7 - 18
(where, unfortunately, Ross would delete the words that make Geach's case: see Aristotle's
Prior and Posterior Analytics [Clarendon Press, Oxford, 1949, 1957], 290 ff: the parallel,
De into I. 16"16 - 18, which Ross cites, seems to me to make the case for the deletion quite
strong). For Frege, see also M. Dummett, Frege, p. 214. Both Frege and Geach want to
retain special senses of 'is': Frege, the 'is's of identity and existenc~ (Ioc. cit.), and Geach,
that of existence as opposed to predication (see 'Assertion', Philosophical Review 74
(1965),449 - 65 [= Logic Mailers 254 - 69): on p. 460 (265) he rails against those who "two
thousand years and more after Plato's Sophist, will wantonly confuse [the 'is' of predication) with the existential 'is' "). See here the references in Note 37 below. For Aristotle,
see also H. Bonitz, 'Uber die Kategorien des Aristoteles', Sitzungsberichte der Kaiserlichen
Akademie der Wissenschaften, Phil.-his!. Klasse, 10 (1853), 591 - 645, p. 601 (available as
a separate reprint with the original pagination, Wissenschaftliche Buchgesellschaft, Darmstadt, 1967).
J3 "Aber das Sein bleibt unauffindbar, fast so wie das Nichts oder am Ende ganz so. Das
Wort 'Sein' ist dann schliesslich nur ein leeres Wort. Es meint nichts Wirkliches, Greifbares, Reales. Seine Bedeutung is ein unwirklicher Dunst." Einfiihrung in die Metaphysi~
(Max Niemeyer Verlag, Tiibingen, 1957), p. 27. In R. Manheim's English translation
(which I have departed from in the above), An Introduction to Metaphysics (Yale, New
Haven, 1959), the passage occurs at the bottom of p. 35.
14 Cf. Tsu-Lin Mei, 'Subject and Predicate: A Grammatical Preliminary', Philosophical
Review 70 (1961), 153 -75.
35 Cf. B. L. Whorf, 'Languages and Logic', in Language, Thought, and Reality: Selected
Writings of Benjamin Lee Whorf, ed. by J. B. Carroll (MIT Press, Cambridge, Mass.,
1956), pp. 233-45.
36 Cf. R. Rorty, 'Genus as Matter: A Reading of Metaphysics Z-H', in E. N. Lee,

78

RUSSELL M. DANCY

A. P. D. Mourelatos, and R. Rorty (eds.), Exegesis and Argument: Studies in Greek

Philosophy Presented to Gregory Vlastos (Humanities Press, New York, 1973), pp.
393 - 420, p. 403; also, Philosophy and the Mirror oj Nature (Princeton University Press,
1979), p. 120.
37 On what follows, see also Frede, Priidikation und Existenzaussage (Note 15 above);
Lockwood, 'On Predicating Proper Names; (Note 28 above); B. Mates, 'Identity and
Predication in Plato', this volume, pp. 29 - 47; J. Hintikka, 'Semantical Games, The
Alleged Ambiguity of "is", and Aristotelian Categories', Synthese 54 (1983), 443 - 468,
and 'The Varieties of Being in Aristotle', this volume, pp. 81 - 114.
'" Others prefer to inflate (18) to 'there is some addict that Dr. Jekyll is the same as', call
this an identity, and, as an alleged consequence, call (18) and identity. But these moves will
turn any predication into an identity; there is nothing but confusion along these lines.
39 Cf. J. L. Austin, Sense and Sensibilia (Clarendon Press, Oxford, 1962), pp. 87 - 102.
40 See also C. H. Kahn, The Verb 'Be', p. 400, Note 33 : " . .. She is his Wife illustrates
the is of identity under conditions of monogamy, but not under polygamy. Surely the grammar of the sentence is the same in either case." This argument is picked up by C. J . F.
Williams, What is Existence? (Clarendon Press, Oxford, 1981), pp. 10- 12. But Williams
still wants a separate existential sense.
41 J . L. Ackrill, in Aristotle the Philosopher (Oxford University Press, 1981), p. 25, says
" . . . Aristotle points out that 'is' does not always assert identity. It also ... serves to
ascribe a characteristic to something." He is discussing Aristotle's refutation of Eleatic
monism in Physics A 2 - 3, but I cannot find Aristotle pointing this out anywhere in these
chapters.
42 There is a slight basis for hope: Aristotle does not, in fact, have any Greek that directly
translates the English "has many senses" or "is ambiguous". He says such things as "is
said in many ways", and perhaps he should be taken to be saying something weaker than
"has many senses". Hintikka once tried this out in a different connection (see Time and
Necessity: Studies in Aristotle's Theory oj Modality [Clarendon Press, Oxford, 19731, Ch .
I, 'Aristotle and the Ambiguity of Ambiguity'), but he appears not to want it in this connection (cf. ' "Is", Seman tical Games, and Seman tical Relativity', Journal oj Philosophical
Logic 8 (1979), 433-68, at p. 450 top). See also T. Irwin, 'Homonymy in Aristotle',
Review oj Metaphysics 34 (1980- 81), 523 - 44.
43 CAG IV 5. So also T . Waitz, Aristotelis Organon graece (Hahn, Leipzig, 1844) I 351,
Ackrill, Aristotle's Categories and De Interpretatione (Clarendon Press, Oxford, 1963), p.
148.
44 There is another question running through this passage, as to when two predicates of
a single subject combine to form a special sort of unity: see, e.g ., 21"10 - 14, and Ackrill's
comment, Aristotle's Cat. & De Int. 126f.
4S Perhaps: if this is a man , Socrates, then this is a man? See 21"2f, which makes this less
outrageous, bllt still not plausible.
46 I am unable to see what Jacobs (' Aristotle on Nonreferring Subjects', p. 287) makes of
these lines. He is at the very least denying the Interpretation Ammonius and I (and the
others cited in Note 43 above) accept. He apparently believes that we are still discussing
the rule of Addition . But he does not, as far as I can tell, explain these lines.
47 On both points I am in conflict with Owen: see 'Snares' 77, 82.
48 Cf. other attempts to resolve it: M. Thompson, 'Aristotle's Square of Opposition',
Philosophical Review 62 (1953), 251 - 65 (reprinted in J . M. E. Moravcsik (ed .), Aristotle:
A Collection oJCritical Essays [Doubleday, Garden City, N.Y ., 1967). pp. 51-72: see esp.
pp . 56 - 57 of the reprint); M. V. Wedin, 'Aristotle on the Existential Import of Singular

ARISTOTLE AND EXISTENCE

79

Sentences', Phronesis23 (1978), 179 - 96, Jacobs, 'Aristotle on Nonreferring Subjects'. My

resolution is closest to Thompson's.
49 This is deliberately vague (cf. Note 42 above). In E 2, 1026'33 - b2, Aristotle refers to
this four-fold scheme with the formula 'what is is said in many ways' (TO O~ .. Af-YfTOI<
7rOAAaxw~). In Z I. 1028'10 - 20 he uses the same formula in characterizing the subdivision
of Har')' aUTa DVTa by means of the categories. There is no indication, in Aristotle, or, as
far as I know, any of his commentators to the effect that these divisions differ in status,
and certainly none in any of his commentators to the effect that one of them divides the
uses of 'is' and the other senses of 'is' - until we get to J. W. Thorp, 'Aristotle's Use of
Categories', Phronesis 19 (1974), 238 - 56, who suddenly finds this distinction 'notorious'
(p. 238).
50 Cf. the employment of '{XftVO' as something like a predicate .. ariable or predicate-letter
in An. post. A 22. 83'24, 27.
51 Cf. F. Brentano, Von der mannigfachen Bedeutung des Seienden nach Aristoteles
(Herder, Freiburg, 1862; Olms, Hildesheim, 1960), p. 16. (In the English translation by R.
George, On the Several Senses oj Being in Aristotle [University of California Press,
Berkeley, etc., 1975], pp. 10-11.)
52 Since Thorp's article (op. cit., Note 49 above), there is not even agreement on this: cf.
his retranslation 011 p. 247. I cannot see how this translation is supposed to work.
5) So also Ross, Aristotle's Metaphysics I 306.
54 i~ aVTwv. Pickard-Cambridge (Oxford) translates "on the face of it"; Forster (Loeb)
takes it the same way, as does J. Brunschwig, Aristotle: Topiques I (Societe d'Edition 'Les
Belles Lettres', 1967), p. J3 ("de par la nature meme des choses"). But see also S. Mansion,
'Notes sur la doctrine des categories dans les Topiques', in G. E. L. Owen (ed.), Aristotle
on Dialectic: The Topics (Clarendon Press, Oxford, 1968), pp. 189 - 201 (p. 198: "a partir
de la").
55 Cf. C. H. Kahn, 'Questions and Categories', in H. Hiz (ed.), Questions (Reidel, Dordrecht, 1978), p. 227.
56 So the question 'what is it?' will not by itself mark out the category of substance.
Neither, for that matter, will the word 'substance': cf. 'the substance of everything relativeto something', Top. Z 8. 146b3, for example, and the official doctrine (applied primarily
to 'what it is') of Met. Z 4. 1030'17 - 27, 27 - bJ3.
57 It is misleading to say, as does Waitz, Aristotelis Organon II 447 (and endorsed by Mansion, loc. cit.) that 'Tt iUT!' is used in one sense (sensus) in 103 b - 22 and another in b27:
that would make 'Socrates is a man' ambiguous.
58 I have discussed this passage in Sense & Contradiction, pp. 100-102.
59 Except in '27 - 30, and these are not examples of Har')' aUTO dvOl< at all (see below). Supposing that they are intended as examples of Har')' aUTO dvOl< leads some to think that this
must include all predications, even 'Socrates is pale'. See, e.g., H. Maier, Die Syllogistik
des AristotelesII, 2 (H. Laupp, Tiibingen, 1900; Olms, Hildesheim, 1970), p. 32M (Note
I to p. 328), who thinks Aristotle is misspeaking himself here; H. Bonitz, Aristotelis
Metaphysica: Commentarius(Bonn, 1849; Olms,Hildesheim, 1960), p. 241; E. Buchanan,
Aristotle's Theory oj Being (Greek, Roman, and Byzantine Monographs, University,
Mississippi & Cambridge, Mass., No.2, 1962), pp. II - 13; M. T. Larkin, Language in the
Philosophy oj Aristotle (Mouton, The Hague, 1971), p. 88; K. von Fritz, 'Die Ursprung
der aristotelischen Kategorienlehre', Archiv jur Geschichte der Philosophie 40 (1931),
449-85, 488-96 (reprinted in F.-P. Hager (ed.), Logik und Erkenntnislehre des
Aristoteles [Wissenschaftliche Buchgesellschaft, Darmstadt, 1972], 22 - 79), p. 452 (p. 26
of the reprint); and C. Stead, Divine Substance (Clarendon Press, Oxford, 1977), p. 67f,
who finds the double classification of 'Socrates is pale' a "tiresome inconsistency".

80

RUSSELL M. DANCY

So also, perhaps, Larkin, Language in the Philosophy of Aristotle, p. 87 with Note 20,
but I find pp. 87 - 88 very confusing (see also last note).
61 Cf. Note 59 above, and Ross, Aristotle's Metaphysics I 307 - 308; Owen, 'Snares', p.
82, Note I; and Thorp, Aristotle's Use of Categories', p. 250.
62 In fact, Aristotle refers 'flourishing' (irYlcdvfLV) to the category of quality in Soph. el.
4. 166b 16 - 19; cf. also Cat. 8. 9" 14 -16.
63 So far, I agree withThorp. As he points out (Aristotle's Use of Categories, p. 250), this
interpretation fits Aristotle's use of the same examples in De into 12. 21 b 5 -10,10. 20"3ff.
64 Here I leave Thorp (Aristotle's Use of Categories, 252 - 254) for Ross (Aristotle's
Metaphysics I 307 - 308).
6S This is certainly the way many presentations make it sound, e.g. von Fritz, loc. cit. Contrast C. Kirwan, Aristotle's Metaphysics, Books r, .a., E (Clarendon Press, Oxford, 1971),
pp. 142f.
66 But it is still wrong to say, as does Owen, that J(a~' aUTO ov is "the (or an) existential
use of the verb" ("Snares" p. 82).
67 The argument of Met. B is reviewed by Thomas Aquinas in his comments on .a. 7; see
In duodecim fibros Metaphysicorum Aristotelis expositio, ed. by M.-R. Cathala and R. M.
Spiazzi (Marietti, Turin, 1950), p. 238 (889).
60

Dept. of Philosophy,
Florida State University,
Tallahassee, FL 32306-1054, U.S.A.

JAAKKO HINTIKKA

THE VARIETIES OF BEING IN ARISTOTLE

1. ARISTOTLE DOES NOT RECOGNIZE THE FREGE - RUSSELL AMBIGUITY

OF "IS"

In this paper, I shall try to enhance our understanding of Aristotle's

thought by relating it to certain contemporary problems and insights of
philosophical logicians. Now one of the most central current issues in
philosophical logic is a challenge to a hundred-year old dogma. Almost
all twentieth-century philosophers in English-speaking countries have
followed Frege and Russell and claimed that the words for being in
natural languages - "is", "ist", eun, etc. - are ambiguous between the
is of predication, the is of existence, the is of identity, and the generic
is. The significance of this ambiguity thesis has not been limited to
topical discussions but ha'S extended to historical studies, including
studies of ancient Greek philosophy. A generation or two of scholars
working in this area used the Frege - Russell ambiguity thesis as an important ingredient of their interpretational framework. Cases in point
are Cornford, Ross, Guthrie, Cherniss, Vlastos, Ryle, and (from the
German-language area) Heinrich Maier. Indeed, the Frege - Russell
distinction is still being invoked occasionally by Aristotelian scholars; see
e.g., Moravcsik(1967,p. 127),Kirwan(pp. 100-101, 141), Weidemann
(1980, p. 78) and Gomez-Lobo (1980- 81, p. 79).
However, many of us have by this time come to suspect that the
Frege - Russell ambiguity claim is completely anachronistic when applied to Aristotle. The sources of this dark professional secret are
various, ranging from G. E. L. Owen's brilliant studies of Aristotle on
being to Charles Kahn's patient examination of the Greek verb TO d/lcn.
Most of us good Aristotelians have nevertheless remained in the closet.
As was illustrated by the fate that befell the first major study in which
Plato's failure to draw the Frege - Russell distinction was noted, most of
the unliberated Aristotelians seem to have thought that to note Aristotle's failure to draw the distinction is to accuse him of an object logical
mistake. (See, e.g., Neal's introduction to Bluck.) Accordingly, we have
shied away from such impiety. It is time for some consciousness-raising,
however. It is not convincing enough merely to register the inapplicabili81

s. Knuuttila and .I. Hintikka (eds.),

(0) J 986

The Logic of Being, 81-114.

by D. Reidel Publishing Company.

82

JAAKKO HINTIKKA

ty of the modern distinction to Aristotle. (Cf. Benardete 1976 -77.) We

need a deeper understanding of the whole situation. In an earlier paper,
I have shown that there need not be anything logically or semantically
wrong with a theory which treats the verbs of being as not exhibiting the
Frege - Russell ambiguity. (See Hintikka, 1979.) More than that: not only can we now say that Aristotle's procedure is free from any taint of
fallacy; he may have been a better semanticist of natural languages than
Frege and Russell in this particular respect.
Hence I can without any impiety level at Aristotelian scholars the same
criticisms as Benson Mates recently (1979) directed at Platonic scholars,
viz. that they have been seduced by the modern myth that there is a
distinction between the is of identity, the is of predication, theJs of existence, and the is of generic implication, and to proceed to argue that
the distinction is not there in the Aristotelian Corpus. Relatively little
argument is in fact needed here. Not only is it the case that Aristotle, one
of whose main philosophical methods was to make conceptual distinctions, never labels the Frege - Russell contrast an ambiguity, or in his
own terminology, a homonymy. He does not even say that, in its different Frege - Russell uses, fan is "said in many ways" (7rOAAaxW~
Af-yfTm). (For the force of this terminological distinction in Aristotle, see
Hintikka (1959) and (1973), Chapter 1.)
What this means is in effect that Aristotle never officially
acknowledges the Frege - Russell distinction even as a difference between several uses, let alone as a difference between logically different
meanings of senses of esti.
To put these important points in somewhat different terms, in maintaining the unambiguity of TO dvm vis-a-vis the Frege - Russell distinction, Aristotle is not just a faithful Whorfian blindly following the
Weltanschauung implicit in the language of the tribe, as might be
suspected among other things on the basis of the absence of any separate
verb for existence in the ancient Greek. Aristotle was cognizant of the
controversies that had raged as to whether TO OV and TO V mean the same
or whether they have several different meanings. (See De Soph. EI. 33,
182 b 22ff.) Nor is Aristotle unaware of the dangers of uncritically assuming that what is, always is what it is, and not another thing, as is illustrated among other things by his criticisms of Parmenides in Phys. A
3. Nevertheless, his failure to acknowledge the Frege - Russell ambiguity
is deeper than a conscious choice between competing conceptual
schemes. Not only does he refuse to countenance the Frege - Russell
distinction as a homonymy between several different meanings. He does

THE VARIETIES OF BEING IN ARISTOTLE

83

ilOt always recognize the distinction as a separation between different

uses of the Greek words for being. More accurately speaking, he does
acknowledge some differences between the relevant uses, as we shall see,
but he does not co-ordinate them into a three-part of four-part
distinction.
Furthermore, Aristotle never resorts to the Frege - Russell distinction
in dealing with problems which we would deal with routinely in terms of
the distinction.
An example is offered by De Soph. EI. 5, 166 b 28 - 36, where Aristotle
is considering inter alia the fallacious inference form "Coriscus is different from Socrates" (i.e. "Coriscus is not Socrates") and "Socrates is
a man" to "Coriscus is different from a man" (i.e. "Coriscus is not a
man"). Here we would expect Aristotle to make a distinction between the
"is" of identity (used in the first premise) and the "is" of predication
(used in the second premise). His point has been so understood by Maier
(Vol. 2, p. 280), and there is sume primajacie evidence for doing so. For
instance, the terminological distinction Aristotle uses to expose the
fallacy, viz. between essential and accidental predication, will in Aristotle's later writings in fact assume (as we shall see) the force of a contrast
between those predications which have an element of identity and those
which do not. However, drawing the predication vs identity distinction
is not what Aristotle is doing here. What he actually does is draw a
distinction between transitive and nontransitive predication. The whole
force of calling a predicate accidental is to say that it does not exhibit the
appropriate transitivity: "It does not necessarily follow that all the same
attributes belong to all the predicates of a thing and to that of which they
are predicated". Applied to the example, this presumably means that "a
man" in the second premise is predicated of Socrates in a manner that
does not allow transitivity and that therefore the predicable "Coriscus
is different from x" , even though it is true of Socrates, does not have to
apply to "a man". This is not a distinction between two senses of "is",
identity vs predication, even though it is perhaps not too hard to see how
the latter distinction should have developed out of what Aristotle does
here. (I have tried to follow here Dancy 1975, Appendix II.)
Of course, it lies in the nature of things that we cannot expect to find
explicit evidence for Aristotle's refusal to hold the Frege - Russell ambiguity thesis. Aristotle is not consciously rejecting a distinction he is
aware of; he is completely oblivious to the very idea of the
Frege - Russell contrast. Hence it would have been virtually selfdefeating for him first to make the distinction and thereupon try to deny

84

JAAKKO HINTIKKA

it. Fortunately, even though Aristotle does not say in so many words that
the word est; does not exhibit the Frege - Russell ambiguity, he does say
as it were in the material mode of speech that the entities that would be
differentiated from each other in the Frege - Russell distinction (if
Aristotle had made it) in reality are one and the same thing. Thus in Met
r 2, lOO3 b 22 - 32, he writes (the text is Ross's and the translation is Kirwan's):
fi o~ TO OV )(al TO EV mUTOV )(al p.ia q,llCm,:r~ (noAoullftV aAA~AOtS wune I aex~ )(W Citnov aAA' oux ws IVI AO'YI" O"l/Aovp.fva (Otaq,feH Oi oullfv ouo' (xv bp.oiws iJ1rOAa/3wp.fP, aAAa
)(al 1reO fe'YOU p.aAAov) mUTo 'Yae ds av'oewr.os [)(al Orvllew1ros], )(al WV Ctvllew1ros
)(al Orvllew1f"os, )(al oux fTfeOV n O"l/AOt )(aTa T~V Af~tV f1raVQ[H1f"AOVP.fPOV TO fis OrvlleW1f"Os
)(QI ElS WV OrvlleW1f"Os (OijAOV 0' on ou xweil;TQt OUT' f1r1 'YfVfUfWS OUT' f1r1
q,lIoeas), bp.oiws Of )(QI f1f"t TOU ivas, wun q,QVfeOV on " 1reaull(1tS fV TOVTOtS mUTo
O"l/AOt, )(QI OUOEV fneOV TO EV 1rQea TO OV, ....

Suppose it true, then, that that which is and that which is one are the same thing - i.e.
one nature - in that each follows from the other as origin and cause do, not as being indicated by the same formula (though it makes no difference even if we believe them to be
like that - indeed it helps). For one man and a man that is and a man are the same thing;
and nothing different is indicated by the reduplication in wording 'he is one man' and 'he
is one man that is' (it is plain that there is no distinction in [the processes of) coming to
be or destruction); equa\1y in the case of that which is one. It fo\1ows obviously that the
addition indicates the same thing in those cases, and that which is one is nothing different
apart from that which is.

Although the text of the lines 1003 b 28 - 29 is especially messy, it is important to note that Aristotle is in the quoted passage employing est; in
what I shall later in this paper argue to be a purely existential use.
Nonetheless he is emphatically assuring us that this use is not different
from the identity sense of est; "is one and the same man". As to the
predicative sense, by the priority of the three cases announced a couple
of lines earlier, "he is a man" will be a further synonym for the two
phrases Aristotle mentions.
An even blander assertion to the same effect is found in De Soph. E/.
6, 169 a 8 - 10: "For the same definition (horos) applies to 'one single
thing' and to 'the thing' hap/os; the definition e.g. of 'man' and 'one
single man' is the same, and so, too, with other instances".
This confounds the first three members of the four-fold distinction of
Frege's and Russell's. As for the fourth it is clear that there is no
Frege - Russell type difference in meaning for Aristotle between the different occurrences of ;s in "Socrates is a man" and "a man is an
animal". If further evidence is needed for the total absence of the
Frege - Russell ambiguity thesis in Aristotle, it is easily forthcoming.

THE VARIETIES OF BEING IN ARISTOTLE

85

Even though these observations do not automatically solve any hard

interpretational problems concerning Aristotle, they help to clear away
misunderstandings. For instance, we can now see that Aristotle's formulas for what has later come to be known as essence, TO TL fun (what
[it] is) and TO TL 1}v dvcxt (what it is [for a thing] to be), exhibit for a true
Fregean an irredeemable ambiguity between predication and identity.
For Aristotle they express ipso facto both something's being such-andsuch and its being identical with some one entity. This is vividly shown
by the fact that Aristotle frequently used the very same formulas as a
name for his first category, substance, in spite of considering particularity ("separability and 'thisness''') as the main characteristic of
substances. No wonder Aristotle could thus raise the question, which in
our anachronistic ears may first sound paradoxical (Ross confesses that
it is for him "difficult to see the point of this question") whether or not
a substance is identical with its essence. (See Met. Z 6.) Similar remarks
apply to esti. In general, it is not possible to understand the questions
Aristotle is asking in such works as Met. Z without appreciating the
simultaneous presence of several Fregean meanings in esti in his
discussion.
2. THE NONAMBIGUITY OF ESTl DOES NOT PRECLUDE PURELY
EXISTENTIAL USES

It is especially important to realize exactly what is involved in Aristotle's

failure - or perhaps rather refusal - to make the Frege - Russell

distinction. What is denied in denying the Frege - Russell ambiguity
claim is not that the force of "is" or esti is different in different contexts.
Rather, what is ruled out is one particular explanation of these differences, viz. that they are occasioned by different meanings of the verb
"is". In other words, what is asserted is that such differences are always
traceable to the context and due to it. Indeed, it is an integral part of my
position that fun can have on different occasions in Aristotle different
Fregean uses. For instance, Aristotle can - and does - use esti with a
purely existential force. When one says "Homer is" ('O!L1)eO~ fUTO,
what is at issue is obviously the existence of a particular individual (Cf.
De Int. 11,21 a 25 - 27.) In general, when one asks Ei fun, one is asking
whether an entity or entities of a certain kind exist. (See Post. An. B
1 - 2.) Further examples of unmistakably existential uses of esti in
Aristotle are easily found; e.g. Cat. 10, 13 b 27 - 33, and not to speak
of Phys. VIII and Met. A (passim).
In this respect, my thesis differs sharply from what currently seems to

86

JAAKKO HINTIKKA

be the most popular reaction to the data that can be adduced against the
presence of the Frege - Russell ambiguity in Aristotle. According to this
competing view, esti is unambiguous because it has basically always the
predicative sense. Where it apparently does not, e.g. in the existential
uses listed above, we must understand the usage as being elliptical:
"Socrates is" on this view basically means "Socrates is something or
else". )There may be important restrictions as to what this "something
or else" can be, but they need not detain us here.)
This view seems to have been suggested by G. E. L. Owen, and it has
recently cropped up in slightly different variants. There is a sense in
which it probably comes close to being a true representation of what
things are like according to Aristotle's last and final conclusions.
Roughly, for any entity to exist is for it to be what it is, i.e., what it essentially is.
However, admitting this does not mean that in the force of the term
esti in Aristotle's actual argumentation is tacitly predicative. For one
thing, the identification just offered is probably only an approximate
one, anyway. It is not clear that for Socrates to exist is (apud Aristotle)
for him to be a man. Rather, on a closer look it seems (as Balme has
shown) very much as if for Socrates to exist is not so much for him to
exemplify (more generally, to develop towards exemplifying) the speciescharacteristic form of man, but rather to exemplify (more accurately,
develop towards exemplifying) the particular nature which consists in his
likeness to his parents. And it is not clear at all that Socrates' exemplifying this particular form is a predicative relation rather than an identity.
Be this as it may, even if the elliptical character of fun a1l'AWS is
perhaps a conclusion of Aristotle's arguments for his metaphysical
theory, it cannot for this very reason be a part of what he assumes in
them. When I reject the ellipsis theories, it is thus as a claim of what the
basic semantical force of esti and its con gates are for Aristotle, and not
as a possible feature of his ultimate metaphysical doctrine. However, in
the former sense I do reject it tout court, and hence also reject the
mistaken idea that it is somehow implied by the absence of the
Frege - Russell ambiguity assumption from Aristotle.
This puts on me the onus of commenting on the recent denials of any
purely existential uses of verbs for being in Aristotle. Suffice it here to
deal with one of the most recent putative arguments for the absence of
the existential uses in Aristotle or in certain parts of the Aristotelian
Corpus.
The ellipsis hypothesis has not been defended by its reputed originator

THE VARIETIES OF BEING IN ARISTOTLE

87

at any greater length. It has recently been discussed by A. Gomez-Lobo

(1980-1981). The part of the Corpus which Gomez-Lobo and his ilk
have to worry about most is clearly Post. An. B 1 - 2, where Aristotle
in so many words recognizes questions of simple being (El fun) besides
the three other kinds of questions which figure in an Aristotelian science,
viz. TO on, TO (iton, and TL fun.
Actually, strictly speaking Gomez-Lobo does not deny that a sentence
of the form [esti + noun phrase] can express mere existence in Aristotle.
He admits that e.g. Met. A 7, 1072 a 25 is a case in point. But he strives
to reduce greatly the scope of this way of reading Aristotle by removing
Post. An. B 1 - 2 (in fact, it seems, all of Post. An.) from its scope.
Hence a brief discussion of Gomez-Lobo's arguments are in order, for
if they were valid, much of the plausibility of my point would be lost.
The ei fun questions used to be taken without any further ado as questions of existence. Gomez-Lobo is entirely right in recognizing that the
situation has changed. The insight that Aristotle did not believe in the
Frege - Russell ambiguity and that the basic semantical meaning of esti
in Aristotle is hence neutral with respect to the different Fregean senses
of being certainly makes a fresh look at Post. An. B 1 - 2 necessary. Unfortunately, Gomez-Lobo fails to give the new look a run for its money,
for his arguments are inadequate in several respects. For one thing, most
of his discussion is predicated on a failure to understand in what sense
Aristotle thinks that El fun questions, like all four questions, amount to
looking for a middle term. "How can there be a middle term between a
single term and the predicate 'exists'?" he asks rhetorically. A
simple answer would be embarassingly obvious even if I had not pointed
it out thirteen years ago (in Hintikka 1972a). Aristotle is, as it were,
thinking of abbreviated syllogisms of the form
(*)

Every B is simpliciter
Every C is B
Hence: Every C is simpliciter

which result from a regular barbara syllogism by omitting the major

term, even though they are never found in Aristotle's writing in so many
words.
It is obvious that (*) requires a treatment of existence somewhat different from what contemporary philosophers have been used to.
,However, this is no argument against what I am saying. Even without
discussing any details here, it is patently clear on other grounds that we
have to shake our complacency concerning the adequacy of the received
Frege - Russell treatment of existence in logic.

88

JAAKKO HINTIKKA

It may be objected that quasi-syllogisms of the displayed form are

never actually put forward by Aristotle. The explanation is that he does
not need to do so. In the syllogistic structure of a science, the existence
of the 8's is always a consequence of the existence of a wider term, say
A. Hence Aristotle accomplishes the same effect by means of a regular
barbara syllogism as he accomplishes by means of (*), as long as a proviso is explicitly or tacitly added to the effect that it is only the widest
term that carries any existential force. This may perhaps be illustrated by
the following quasi-syllogism:
(**)

Every 8 is an A (and hence exists)

Every C is a 8 (no existential force)
Hence: Every C is an A (and hence exists)

I shall not discuss here what kind of treatment of existence is presupposed in (*) and (**).
My interpretation gains further credence from the fact that, according
to Aristotle, necessity is "carried downwards" in a syllogistic chain in the
same way as I have argued existence is. (Cf. Aristotle's theory of apodeictic syllogisms in Pro An. A 8 - 12, especially 9.) In the same way as in (**)
it is only the major premise that has to carry an existential force in order
for the conclusion to do so, in the same way we can obtain a necessary
conclusion from a barbara type syllogism if and only if the major premise
is a necessary one. In fact, it seems to me that Aristotle's treatment of
existence and necessity in the context of a syllogism are related to each
other very closely. 80th of them are based on the presence of an element
of identity in Aristotelian copula, whether or not it is in fact expressed
by him in terms of esti or not. For if the minor premise of a syllogism
like (**) expresses a numerical identity between each C and some 8 then
we must be able to say all the same things of each C as are said of each
8, for the former literally are among the latter. Hence the validity
(among other modal syllogisms) of the following form of barbara:
(***)

Every 8 is necessarily an A
Every C is (identical with) a 8
Hence: Every C is necessarily an A.

8e this as it may, there is plenty of collateral evidence that my construal of the role of existence in Aristotle's syllogistic theory is what
Aristotle in fact meant. Since the whole argumentative structure of
Gomez-Lobo's paper is thus mistaken, there is little that needs to be said
of the rest of his paper.

THE VARIETIES OF BEING IN ARISTOTLE

89

What I also find surprising is that there is conclusive evidence against

Gomez-Lobo in the very passages he is addressing himself to. Witness
this:
I mean if one is or is not simpliciter and not if [one isl white or not (Post. An. B 89 b 33).
fCTTLV ij 1'1, Q'JI"AWi Ai-yW, ahA' oux d AWXOS ij I'~.

TO 0' fi

How could Aristotle possibly have explained more clearly by the means
he had at his disposal that he was presupposing a purely existential use
of fi EUTL? It seems to me that we have to realize that Aristotle, like J. L.
Austin, ordinarily means what he says.
Ironically, Aristotle's very usage in Post. An. B I - 2 provides us with
further counter-examples to the ellipsis thesis. When Aristotle there asks
whether a middle term is (d fUTL piuo II , cf. 89 b 37 - 38,90 a 6), he cannot
but mean whether the middle exists, for he contrasts this question in so
many words with the question as to what it is.
There is elsewhere, too, excellent direct evidence against the ellipsishypothesis. In discussing in De Soph. EI. 5 the importance of
distinguishing the absolute and the relative uses of a term from each
other Aristotle writes (167 a 4 - 6):
OU -yae TCiUTO TO 1'1, flvcx! TL ){CX! Q'JI"AWi 1'1, dvCXI. q,CX!VfTCXt Of Ota TO 'JI"aef-Y-YVi TijS
XCX! I't){eOV Otcxq,ieHV TO flvcx; TL Toii flvCXI,.xcx! TO 1'1, dvcx! n Toii 1'1, flvCXI.

Ai~fWi

For it is not the same thing not to be something and not to be simpliciter, though owing
to the similarity of language to be something appears to differ only a little from to be, and
not to be something from not to be.

One can scarcely ask for more direct evidence. At the same time the
passage shows that, in spite of their differences, the predicative and the
absolute (existential) uses of esti are not unrelated, for they are the
relative and absolute uses of the same notion. The quoted passage hence
also offers evidence against ascribing the Frege - Russell ambiguity to
Aristotle.
What we have found deserves a few additional comments. First, my
defense of the presence of a purely existential use of esti in Aristotle is
squarely based on the absence of a purely existential meaning as
distinguished from its alleged predicative meaning and identity meaning.
For, in order for (**) to do the double duty of both establishing a
predicative link between C and A and at the same time carrying existence
assumptions downwards from higher wider terms to lower (narrower)
ones, the (**) must carry (for a Fregean) both a predicative sense and the
existential one. Morc generally, it is undoubtedly the illegitimate

90

JAAKKO HINTIKKA

assumption of the Frege - Russell ambiguity that has made it so difficult

for philosophers to appreciate Aristotle's treatment of existential
assumptions in the context of a syllogistically organized science.
My second point is a word of warning. Even though there are clear uses
of esti in Aristotle where the word has a purely existential force, because
Aristotle does not make the Frege - Russell distinction we cannot expect
to find a purely existential sense presupposed in Aristotle's general
remarks. Above, I described how according to Aristotle existential
assumptions filter downwards from the existence of the largest genus of
a science downwards to the other terms studied in it. This is amply confirmed by such passages as, e.g., Post. An. A 10, 76 b 3 -11. However,
as I have shown in Hintikka (1980), p. 141, these statements of Aristotle's cannot be understood to involve a purely existential sense of esti. Just
because Aristotle does not operate with the Frege - Russell distinction,
he is - mistakenly in this particular case, if you ask me - lending his
pronouncements also a predicative sense. This explains his statements to
the apparent effect that the widest premises of a science are the only truly
unprovable ones. In order to be able to save such statements, Aristotle
had to presuppose a non-syllogistic sense of proving or showing, a fact
which vividly illustrates the consequences of the absence of the distinction in Aristotle.
Third, our observations throw some fresh light on the ellipsis
hypothesis. In order to see it, let us assume, for the sake of argument,
that (**) is one of the ultimate premises of a syllogistic science in Aristotle. Then B typically expresses part of the essence of C. Now the C's are
shown to exist in (**) by means of the middle term B. This is true generally. As Aristotle puts it, even in questions of El fun what we are searching
is a middle term. (See Post. An. B. 2.) This is, from the vantage point
of Aristotle's philosophy of science, a large part of the cash value of saying that what it takes for C to exist is for it to be what it essentially is,
in our example (**), to be B.
However, this is only a part of the story. The formula "for an entity
to exist means for it to be what it essentially is" is seen to mean only that
the essence is the immediate cause (immediate explanatory term) of its
existence. In many contexts, it is more important to ask what the ultimate
cause of an entity's existence is. A glimpse at (**) shows that the answer
to this question is in terms of the megista gene into which each being must
fit. There are the Aristotelian categories whose existence thus seem to be
the logical basis of everything else's existence according to Aristotle. We
cannot go beyond categories, for "being is not a genus". (This is the

THE VARIETIES OF BEING IN ARISTOTLE

91

point Aristotle makes in Post. An. B 7 92 b 14.) Thus the ellipsis

hypothesis in any case is only a part of the story. For an entity C to exist
is perhaps for it to be what it essentially is in the sense that that is the
immediate ground of C's existence. In the total structure of beings,
however, the ultimate ground may be more important. And that is, not
C's being what it essentially is, but what it categorically is. Thus at the
very least the defenders of an ellipsis hypotheses should have two kinds
of ellipsis available to them, not just one.
This point can be illustrated by reference to Post. An. B 7,92 b 12 - 15:
Et'Tor "ort 0" lx7rOOft~fWI <!>or/lfV lxvor",("oriov dvm ofl"vvaOm a7rorv OTL laTtY, d It~ ouator
fi'1. TO 0' dvm ou" oua[or OUOfVt. ou "'(a(! ",(ivOl TO ov. lx7rOOfL~'1 a(!' faTm OTL faTLv.
Next, we say it is necessary for everything to be proved to be by a demonstration, unless
it is its ousia. But being is not the ousia of anything. For what is is not a genus. Therefore
there will be a demonstration that it is.

We can see what Aristotle is literally up to here. He comes very close to

a denial of the usual form of the ellipsis view. For what he is saying is
that the existence of anyone thing has to be demonstrated because it cannot be a part of its essence (ousia). In other words, in (*) the conclusion
(')

Every C is (i.e., exists)

cannot be thought of as being obtained by ellipsis from

(")

Every C is B

where B is a part of (or the whole) of the essence of C, for (hat would
presuppose an irreducible existential force of "is" in the latter and
thereby make being (existence) part of the essence of C. As an ultimate
(atomic) premise of an Aristotelian science (' ') thus does not have an existential import. Rather, Aristotle's point is that the existence of the C's
must be demonstrated, as in (*) or (**). It is not implied by C's essence
alone. Hence there is a sense in which the existl!nce of the C's is not due
to their being what they essentially are, but their falling under (as
established by a potentially quite long chain of appropriate scientific
syllogisms) the genus which characterizes the science in question. This
genus is what we do assume to exist in the science in question. As Aristotle continues,
And that is what the sciences as a mailer of fact do; for the geometer assumes what triangle
signifies but proves that is.

Even though Aristotle docs not press the point in the quoted passages,
if we proceed tov,'ard morc and more general sciences we never reach one

92

JAAKKO HINTIKKA

where existence is part of the essence of the genus, so that it need not be
assumed. That is what the quoted passage amounts to. Elsewhere
Aristotle indicates that the upward chain comes to an end with one of the
different categories. Hence the only case in which we are allowed to
assume (according to Aristotle) the existential import is "every C is D"
where D is a term for the category to which C belongs. In this sense, even
if we believe that such attributions of existence as (') are elliptical, the
omitted term is not the essence of C but the category of C.
This view of the role of existential presuppositions in Aristotle thinking seems to be confirmed by Met. H 6, 1045 a 34 - b 8.
A fourth observation may likewise be in order. Philosophers', difficulties in understanding all the different things that are going on in an
Aristotelian syllogism like (**) illustrate a more general methodological
moral. The Frege - Russell ambiguity assumption is built into all the better known formalisms of first-order logic (quantification theory, lower
predicate calculus). Since this assumption is completely foreign to
Aristotle, virtually all applications of modern logic to Aristotle are partly
anachronistic, and have to be viewed with considerable caution. This
flaw does not automatically invalidate them, however, nor does it make
such historical applications of formal techniques inferior to the work of
informal analysts of Aristotle's work, for most of the latter have likewise
been relying on the Frege - Rw;sell ambiguity thesis, as was pointed out
above.
A further general observation is the following: The Frege - Russell
distinction between different meanings of "is" and its cognates is correlated - at least roughly - with an ontological distinction between different kinds of entities. The is of identity equates in its clearest instances
particulars, modern philosophers' "individuals". In contrast, the is of
predication expresses the being of facts. Thus my thesis of the absence
of the Frege - Russell distinction in Aristotle is not without consequences
for the rest of his ontology. For instance, it is connected with the otherwise strange practice of Aristotle's in his syllogistic theory of scientific
explanation, where he treats the being (existence) of what looks like individuals and the being (occurrence) of facts or events on a par without
any apologies or explanations. (For examples, see, e.g., Met. Z 17, 1041
a 14 - 16, b 4 - 5.) Furthermore, another part of the same syndrome is
the important fact that, appearances notwithstanding, Aristotle does not
really have as sharp a notion of an individual (particular) as contemporary post-Fregean philosophers.

THE VARIETIES OF BEING IN ARISTOTLE

93

3. ESSENTIAL PREDICATION INVOL YES IDENTIFICATION

Thus realizing Aristotle's failure to make the Frege - Russell distinction

does not necessitate throwing overboard all earlier views concerning
Aristotle's treatment of being, even though it does necessitate a fresh
look at the evidence. (There seems to be a fashion in the current literature
to disparage earlier interpretations just because they did not take into account some important aspects of Aristotle's treatment of "is". Such a
failure may be regrettable, but it does not automatically invalidate all the
interpretations of a scholar.) Nevertheless, dispensing with the
Frege - Russell dogma opens the door to certain further lines of thought.
In the same way as the Aristotelian esti sometimes has existential force
and sometimes does not, in the same way it can sometimes have the force
of identity and sometimes does not. This possibility is connected with
Aristotle's central metaphysical doctrines. In Post. An. A 22, 83 a
24 - 30 we read:
"En Ta !!Ev ouoiOlv OT/!!OIivOVTOI one f'(flVO ii 01l"fe hflVO n oll!!OIivft '(018' ou
"OITlI-rOeflTOIL. 0001 oE 1'1, ouoiOlv oll!!OIivH, Crhha "OIT' CxAAOV V1I"O"ft!!fVOV Af-yfTOIL 0
1'1, fon y.~n (J1I"fe hflVO !!~n one hflVO n, OV!!(3f(3l1"oTOI, ot'ov "OITa TOU Crv8ew1I"ov TO
hfV"OV' ou -rae fonv b Cxv8ew1l"o~ oim o'ne hfV"OV oun 01l"fe hfU"OV n, Crh>.a
!~OV row~' one -rae !~OV fonv b Cxv8ew1l"o~.
Again, the things signifying an ousia signify of what they are predicated of precisely what
that thing is or precisely what is the particular sort of it; but the things which do not signify
an ousia but are said of some other underlying subject which is neither precisely what that
thing is nor precisely what is the particular sort of it, are incidental, e.g., pale of the man.
For the man is neither precisely what is pale nor precisely what is something pale; but
presumably an animal; for a man is precisely what is an animal.

It is not clear whether ousia here means essence or substance, but a comparison with such passages as Met. r 4, 1007 a 20 - 33 shows that the
former possibility cannot be excluded here. This is also shown by Aristotle's reference to incidental predication at a 27 - 28, complete with
Aristotle's stock example (AfV)(OV) of an accidental predicate. Hence
what we have found is that one important element in Aristotle's distinction between essential and accidental predication is that the former is an
assertion of identity whereas the latter one is not. In so far as predication
is expressed by means of a verb for being, this means that in essential
predication the verb is used to express identity whereas in an accidental
predication it is used to express predication. Once again, this difference
in use does not mean that Aristotle is thinking of esti as having different
senses or meanings.

94

JAAKKO HINTlKKA

This partial construal of the essential vs accidental distinction is further confirmed by such passages as Met. ~ 18, 1022 a 26 - 27 and Post.
An. A 4, 73 b 5 - 10. It is by any token an important element in Aristotle'~ ideas of essence and essential predication. Further arguments for the
same view have been presented by M. J. Woods, who argues that
"Aristotle held that a staement like 'Socrates is a man' was, despite appearances, to be construed as a statement of identity". In defense of this
view, Woods refers inter alia to Met. ~ 18, 1022 a 26-27; Z 4, 1029 b
28; Z 7, 1032 b 1 - 2; Z 8, 1034 a 8. As was noted above, the unmistakable
(albeit not exclusive) presence of this idea in Aristotle's mind is also
witnessed by the question he raises and discusses in Met. Z 6 as to
whether each thing is identical with its essence. Woods's thesis is essentially that Aristotle answers his own questibn affirmatively.
One intriguing feature of these observations is that they seem to turn
the contrast between essential and accidental predication into a prima
facie nonmodal distinction, whereas it has in recent discussions of
"Aristotelian essentialism" been treated as an almost paradigmatic example of a modal distinction. It is of course true that in spelling out the
distinction in systematic terms we soon get entangled with modal concepts, and it is equally true that in his discussion of the matter in Met.
Z-H Aristotle likewise resorts in the end to the notions of potentiality and
matter. Nevertheless it seems to me that the initial impression of a nonmodal distinction points to an interesting truth (and to an interesting
flaw in recent discussions). Even I cannot elaborate the point here, the
way in which the essential-accidental distinction involves modal concepts
is not through any direct appeal to them, but through a modal element
which there is in the very notion of an individual (for Aristotle, in the
notion of substance). Individuation, in short, is a process which inextricably involves a modal element.
My observation concerning the link in Aristotle between the essentialaccidental contrast and the difference between identity and copula - as
well as Woods's thesis - nevertheless represent only the first step in
understanding Aristotle's views on essence and substance. A telling indication of the problems one encounters in this direction is the fact that
Aristotle occasionally coulitenances also "accidental unities" like "pale
man" or "pale Socrates". In other words, in one of his moods he
somehow construes "Socrates is pale" and "man is pale" also as identity
statements , and not predications contrasted to " Socrates is man" or
"man is an animal " , which alone were supposed to be identities. Aristotle's reasons are nevertheless clear. As was noted earlier, Aristotle's key

THE VARIETIES OF BEING IN ARISTOTLE

95

terms for essence were double-barrelled, allowing a construal either in

terms of identity or in terms of predication. Small wonder, therefore,
that he was not consistent in his construal of essential attributions (in
contradistinction to accidental ones) as identities.
Even though it would take us far too far afield here to argue the point
fully, it nevertheless seems to me that Aristotle in the end subscribes to
the idea of essential predications as identities. In Met. Z 6 he concludes
his discussion by affirming the identity of each thing from its essence. As
he later formulates this point (see Met. Z 10, 1036 a 16 - 19), "the soul
just is animal, ... or the soul of each thing is each thing, and for a circle
to be [i.e., the "essence" of a circle] is a circle, and for a right angle to
be, i.e., the ousia of a right angle is a right angle ... " (translation largely
by Russell Dancy). When he resumes his main line of thought in Met Z
10 he nevertheless uncovers a mare's nest of further problems, which first
led him to hesitate about the identification. In Met. H 3, 1043 b 2 - 5,
he writes: "Sould and essence of soul are the same, but man and essence
of man are not, unless the soul is also called man; and although this is
so in one sense, it is not in another". These problems are due to the
presence of the material factor. In fact, the quoted passage is prefaced
by Aristotle by the explanation, "because the essence belongs to the form
and the actuality", i.e., does not include material aspects of a particular
substance which Aristotle identifies with its potentiality. As Aristotle
puts it in Met. H 6, 1045 a 22 - 23, "man is part matter and part form".
In short, the essence of a thing construed as a form cannot be identified
with the thing itself because the thing also contains matter.
However, when Aristotle comes around to offering his definitive solution to the individuation problem in Met. H 6, he ends up vindicating the
identity of a thing and its essence. This is accomplished by Aristotle by
considering matter and form the two sides of the same coin: matter is the
passive potentiality which is correlative to the active potentiality which
in turn is the form. Being correlative, the two are in a sense one;
their essence is the same. "There is no other cause of the potential
sphere's being an actual sphere; this was the essence of each." (Met. H
6, 1045 a 31 - 33). [OVOEV "fixe OTLV ai"TLov UeOV TOU T~V ovvixp.H
oc/>a'ieav VEe"fEip:
ElVa! oc/>a'ieav, a'A'ACx
TOUT' ~V
TO
Ti ~V
dvO't fxaTl'e~.] In the sense spelled out by Aristotle in Met. H 6, 1045
a 14 - 34, he thus ends up subscribing to the identity of a man and what
it is for a man to be, alias the essence of man. Hence in the last
Aristotelian analysis, essential predications are in a sense identities.
Be this as it may, what has been seen suffices to show that there are

96

JAAKKO HINTIKKA

extremely important connections between the different uses (not senses!)

of esti in Aristotle and his central metaphysical doctrines. Such uses
mean that on certain occasions the force of to esti is purely existental or
purely identificatory, even though these are not separate senses of the
verb. For this reason it is misleading to describe the meaning of esti (absolutely or on a certain occasion) in ancient Greek philosophy as one in
which the existential and copulative senses are "fused". (Cf. here Kahn,
1966, and Furth.) The converse image is more apt. Even though in the
basic meaning of esti we cannot tell the allegedly different Fregean senses
apart, contextual factors may on occasion have the effect of almost
separating from each other the different Frege - Russell forces and nearly eliminating all but one of them.
4. INSTANTIATION IN NATURAL LANGUAGES: A SYSTEMATIC VIEW

Suffice this as an indication of one line of thought opened by our observations. To return to the main theme of this paper, notice that from the
absence of the Frege - Russell ambiguity it does not follow that there
might not be other ambiguities about esti, over and above the nonambiguous differences in use which Frege and Russell mistakenly raised
to the level of ambiguities. Further light can be thrown on this question,
too, by means of recent topical insights. There is in fact another major
way in which recent logical and seman tical work on the concept of being
puts Aristotle in an interesting new perspective. In order to see what it
is, we have to stray temporarily away from Aristotle and discuss certain
topical problems in the logic of natural languages. I shall discuss them
in the case of English, even though similar things can be said of other
languages, including ancient Greel(.
These problems are as close to t.1e heart of all Sprachlogik as we can
hope to get. Any logician knows that the lifeblood of virtually all interesting logical techniques in that basic part of logic which is variously
known as first-order logic, quantification theory, or lower predicate
calculus are the rules of instantiation (i.e., rules for substituting names
or name-like terms for quantified variables). Now suppose w~ want to
deal with the logic of natural languages directly, without first attempting
the dubious and by this time largely discredited translation to formal
languages. Then our first task is to formulate. likewise directly for
natural languages, rules of instantiation far the quantifier phrases which
take over the role of quantified variables in natural languages. How can
we do that? How are we to deal with, say, a quantifier phrase like "every

THE VARIETIES OF BEING IN ARISTOTLE

97

white horse which Alexander rode" or "some small town where Socrates
lived", occurring in a context X - W? (We take here the general form
of these quanti fer phrases to be
every }
some

Y + wh-word + Z

where Z contains a "trace" to indicate where the wh-word was "moved

away from".) Now the obvious way of formulating instantiation rules
for such phrases is to legitimize a move from the sentence in which they
occur to sentences like
(I)

X - b - W if b is a white horse
and Alexander rode b

or, respectively,
(2)

X - d - Wand d is a small town and

Socrates lived in d,
where" b" and" d" are the respective instantiating terms. In general, the
output of an instantiation step is of the form
if
(3)

X - b - W

} b is a Y and Z'
and

where b is the instantiating term and Z' is like Z except that the trace has
been replaced by "b" with the appropriate preposition. (We have been
assuming that Y and Z are here singular.)
The details need not detain us here. What is of interest to us here is
an important difference between the situation in formal first-order
languages and natural languages. In the former, a single domain of
values for the substituting terms (e .g. my "b" and "d") is given. In the
latter, the entities referred to by the substitution-values have to be chosen
from different subdomains in different cases. For instance, in (I) b has
to be a living creature, whereas in (2) d has to be a location in space.
It lies close at hand for a logician to say that the only novelty here is
that natural languages employ many-sorted quantification theory (more
generally, many-sorted logic). And this need not by itself introduce any
complications (contrary to what is, e.g., implied in Moravcsik, 1976). In-

98

JAAKKO HINTIKKA

deed, many-sorted logics do not involve any serious new difficulties over
and above one-sorted ones.
Yet there is a new question present here. In many-sorted formal logics,
the sortal differences are indicated by notational conventions. How are
these differences marked in natural languages? How can one tell what
subdomain b or d must belong to?
Some clues are obvious, and the most obvious is the relative pronoun
which disappears in the process of instantiation. (These relative pronouns can be taken to be question words in a new role, except that
"what" is replaced by "that".) If the operative word is "who", the relevant subdomain consists of persons, if "where", of locations in space,
if "when", of moments (and/or periods) of time, etc. Further subdomains are introduced by prepositional phrases containing similar words,
for instance "like which" introducing a realm of qualities ("some color
like which you have never seen"). Clearly there is not a sharp one-to-one
correspondence between the ranges of natural-language quantifiers (my
"subdomains") and different relative pronouns (or other wh-words,
with or without prepositions or similar qualifiers), but a rough-andready correspondence certainly obtains.
The main discrepancy here is the fact that "what" covers several different subdomains. By asking, "what is X?", we can mean at least three
different things, to wit:
(i)
(ii)
(iii)

Which partic4lar entity is X?

What kind of entity is X?
What does X consist of?
(What is it made of?)

as illustrated by the following sample questions:

(i)
(ii)
(iii)

What is Sirius?
What is a gnu?
What is cordite?

Similar things can be said of Aristotle's Greek word esti: This ambiguity is made especially important in Aristotle's case by the absence of the
Frege- Russell distinction. For in terms of this distinction one could
distinguish the different what-questions (i) - (iii) from each other by the
different sense of "is" involved in them. But this distinction just is not
available to Aristotle.
In any case, the relative pronoun (or the corresponding question word)
cannot be the only clue to the choice of the subdomain. For one thing,

THE VARIETIES OF BEING IN ARISTOTLE

99

the whole relative clause can be missing from the quantifier phrase in
question, and hence be unavailable to supply any leads. Hence it is the
meaning of Yin (3) which must supply the main in formation as to which
subdomain (sort) we are dealing with. Presumably we must assume some
kind of semantical categorization of the terms (phrases) that can serve
as the Y in (3). In the case of simple terms these must be part of their lexical meaning. Since the Y's in (3) are basically predicate terms, we end
up in this way postulating a classification of all simple predicates of
English into certain equivalence classes. These classes wi\l be correlated
one-to-one with those subdomains of quantification, which we are dealing with, when using quantifiers in English, i.e., the largest classes of entities we can quantify over, and also correlated in a loose way with certain
wh-words and phrases.
The need of relying on Y for our choice of the subdomain is vividly
seen from the fact that if we try to eliminate Y (in the way in which we
could dispense with the relative clause), we would end up with an
ungrammatical expression. In order to preserve grammatically, we must
amplify the quantifier word itself so as to make it capable of conveying
the crucial information. For instance, some becomes someone,
something, somewhere, sometime, somehow, etc., where the added handle serves to betray the relevant sort (subdomain).
Furthermore, since each instantiation step (witness (3 introduces an
occurrence of "is", these correlated classifications are likewise correlated with a distinction between different uses of "is", viz., those that
could have originated from an application of the instantiation rules, plus
of course those that are logically on par with them.
Thus we are led to recognize four correlated multiple distinctions.
They distinguish from each other
(i)
(ii)
(iii)
(iv)

Certain wh-words (and phrases with wh-words).

Different kinds of simple predicates.
The largest classes of entities we have to recognize in the logic
of our language as domains of quantification.
Certain semantically different uses of "is". (In them we of
course cannot distinguish from each other the ises of identity,
existence, and predication.)
5. WHAT DO ARISTOTELIAN CATEGORIES CATEGORIZE?

At this point you are supposed to have a deja VII experience. For what
I have arrived at by means of purely systematic (logical and semantical)

100

JAAKKO HINTIKKA

considerations is to aJl practical purposes tantamount to Aristotle's

theory of categories. One of the most fundamental and most perplexing
questions concerning Aristotle's distinction between different categories
is: What is being distinguished from each other? What is Aristotle classifying in separating the different categories from each other? He uses different Greek question words or question phrases (7 fait, 7roaov, 7rOLOV,
7r(!OS it, 7rOV, 7rOT) as names for six of the categories, and the other labels
likewise go naturaJly with certain types of questions in Greek. This is
pretty much how he presents his categories in Top. 1,9. He envisages different kinds of entities "put before one" and classifies the different
things that can be said (and by implication asked) of it.
But when Aristotle introduces his categories in Cat. 4, they appear as
classes of simple predicates or "things that can.be said" of an entity.
Which are they?
The plot is thickened further by Aristotle's deeply ingrained habit of
considering categories as the widest genera of entities that can be logically considered together. This is seen for instance in Met. r 1, 1003 b 19ft"
or Post. An. A 22,83 b 10-17.
Furthermore, Aristotle repeatedly indicates that the distinction between the different categories goes together with a correlated distinction
between different uses of esti. What is more, occasionaJly he seems to run
the two distinctions together. For instance in Met. Z 1, 1028 a lOff "that
which is" is said to signify the different categories. See also Met. Ll 7,
1017 a 23 - 30.
Scholars have debated intensively which of these different things
Aristotle "reaJly" meant. For example, one persuasion maintains that
the categories represent the different kinds of questions one can (according to Aristotle) ask of a given entity. This view is in different variants
held by among others Ockham, Charles Kahn, Benveniste, and Ackrill.
Other scholars hold that Aristotelian categories are what he says they are,
predicables. Others, led by the formidable Hermann Bonitz, have have
held that categories were for Aristotle first and foremost the widest
genera of entities. "Sie bezeichnen die oberst en Geschlechter, deren
einem jedes Seiende sich muss unterordnen lassen", he proclaims (p. 623
of the original).
Still others have held that Aristotle's category distinction is primarily
a differentiation between several senses of esti, a reminder of the
"systematic ambiguity" of words for being in Aristotle. This view is
found, e.g., in Phys. A 2, 185 b 25 - 32. Among commentators, it has
been represented by Heinrich Maier, and in a sense it can be maintained

THE VARIETIES OF BEING IN ARISTOTLE

101

that G. E. L. Owen is another case in point. He has certainly been

followed by a host of younger scholars.
If we had not seen that Aristotle is completely free from the
Frege - Russell ambiguity assumption, we might also be puzzled by the
fact that the distinction between the different uses of esti in the different
categories sometimes appears prima facie as a distinction between different kinds of existential is (cf., e.g., the Topics discussion whether
'TO Oil is a genus), sometimes as a distinction between different kinds of
predicative is (cf., e.g., Pro An. A 37, 49a 6 - 9, read in conjunction with
the preceding chapter), and sometimes as one between different kinds of
identity. These different emphases in Aristotle have found their
fans. For instance, as Ross reports, "ApeIt regards the categories as
primarily a classification of the meaning of the copula 'is' " whereas
Bonitz stresses the existential and identity senses.
Some of the shrewder scholars have responded to this problem situation by suggesting that Aristotle was led to his distinction between the
different cate~ories by several convergent routes. For instance, AckriII
suggests that in Aristotle's classification there are two elements, first the
idea that different kinds of questions will have "categorically" different
answers, and second the idea of categories as the highest genera. This is
undoubtedly a step in the right direction. However, philosophers taking
such a line will nevertheless face the almost equally perplexing problem
as to why the different distinctions Aristotle had in mind should coincide
- or at least why Aristotle should have thought that they coincide. By
and large, they have not solved this problem. Ackrill says merely that,
"It is not surprising that these two ways of grouping things together
should produce the same results". This opinion simply will not stand up
to scrutiny when viewed in the cold light of contemporary analyses of
questions and answers. (Cf. Hintikka, 1976.) Contrary to what Ackrill
suggests, it is not at all clear that answers to different questions fall into
mutually exclusive classes which correspond to the widest classes of entities. For instance, it is perfectly legitimate to reply to the question
"Who is the head of the Academy?" not only by saying "Plato" or "a
man", but alternatively "a pale man", "the youngest brother of
Potone", or even "he is lying there", alI of which have to be pigeonholed
in different categories. Only by means of a further analysis can one
perhaps hope to eliminate some of these replies as amounting only to partial answers (or as supplying collateral information to back up the conclusiveness of an answer). Worse still, AckrilI's account is intrinsically inconsistent. For if the appropriate answers to different questions belong

102

JAAKKO HINTIKKA

to different categories, it is impossible to construe Aristotle's categories

as answers to one and the same question "What is it?", as Ackrill also
suggests. Even if what he says can somehow be salvaged in the last
analysis, it does not help us to understand what Aristotle's categories
really were in the least. Prima facie, it is far from obvious that the four
correlated distinctions we find in Aristotle should go together, and
Aristotelian scholars certainly have not supplied valid reasons why they
should do so.
Aristotle seems to be aware of the objection and tries to handle it, not
so much by explicit reference to questions as in terms of definition and
sameness. In Post. An. A 22 he argues that to define an entity means to
specify what substance it is and in Top. I 8 he argues that strict
(numerical) sameness means identity of substance. These explanations
presuppose Aristotelian distinction, however, and therefore cannot serve
as an independent theoretical motivation for it. A systematic theory of
questions and answers thus cannot in itself serve to explain the nature of
Aristotle's theory of categories, in the way some scholars seem to expect.
For the same reason, I shall not discuss the details of these two chapters
here.
6. ARISTOTLE RECONSTRUCTED

Now the brief analysis of the conditions of instantiation which I carried

out above, puts both Aristotle's theory and discussions thereof into a
new perspective. Led by purely topical (logical and semantical)
arguments, we have arrived at a remarkable reconstruction of Aristotle's
theory of categories. (My arguments have an even stronger theoretical
motivation than I have spelled out here, for they ensue from the basic
ideas of the highly successful approach to language analysis which I have
called game-theoretical semantics (see Hintikka, 1982, and Saarinen,
1979). We can now recognize all the apparently discrepant ingredients
of Aristotle's doctrine in the systematic situation revealed by my
analysis. Aristotle's use of question words and phrases as labels for
categories matches my use of wh-words as a guide to the subdomain involved in an instantiation. His view of categories as the different kinds
of simple things that can be said of an entity matches my classification
of the meanings of simple predicates as guides to the logical "sort" intended. His use of categories as the largest classes of logically comparable entities amounts to the focal point of my quest of the different
largest domains of quantification presupposed in a natural language,

THE VARIETIES OF BEING IN ARISTOTLE

103

and Aristotle's correlation of different uses of the word est; corresponds

to the automatic alignment in my treatment of the other distinctions with
certain differences in the use of the word "is".
What is more important, the correlation of these several distinctions
is seen not to be accidental or artificial. Its reasons lie deep in the logic
of the situation. Charles Kahn has suggested that the different
Aristotelian distinctions represent different strata in Aristotle's thinking.
That may very well be so, but we don't understand Aristotle unless we
also recognize the intrinsic logical connections between the different correlated classifications of his. No longer does it make any sense to ask
which of the several distinctions Aristotle "really" means, for they are
all inextricably intertwined. The extensive controversies that have been
prompted by this question are simply otiose. (This does not mean that
differences of emphasis are not called for here; cf. my comments below
on those who stress the ties between categories and question types.) The
interesting questions pertain instead (inter alia) to Aristotle's awareness
of the connecting links between the different dinstinctions. Indeed, it is
in spelling out the main interrelations between the distinctions which
converge in Aristotle's theory of categories that my "transcendental
deduction of the categories" goes essentially beyond those earlier
scholars who have emphasized the multi-faceted character of
Aristotelian categories.
Even though the reconstruction of Aristotelian categories which we
have just reached perhaps does not ipso facto solve any major interpretational problems, it yields valuable clues which help to understand
Aristotle and in mar:ty cases even promise further insights. For instance,
one problem we can now approach pertains to the relation of Aristotle's
theory to the facts of the Greek language.
Trendelenburg, Apelt, and Benveniste have claimed that the
Aristotelian distinction between different categories reflects certain
general features of the ancient Greek language. Ackrill's persuasive
arguments to the effect that what is distinguished from each other in the
category distinctions are not verbal expressions but entities may serve as
an antidote to such excesses. However, Ackrill's thesis does not imply
that Aristotle was not guided by logical structures which manifest
themselves in the grammar of the Greek language. I cannot try to write
either a transformational grammar or a game-theoretical semantics for
the ancient Greek language. Suffice it merely to point out that the grammatical facts which are highlighted by my treatment are less eye-catching
but subtler than those flaunted by Trendelenburg and Benveniste. They

104

JAAKKO HINTIKKA

pertain to such things as the identity (in form) of indefinite relatives with
indirect interrogatives in Greek, and the close relationship of both with
quantifier words. These features of the Greek grammar serve to link the
different correlated distinctions explained above to each other especially
closely, and thereby to motivate Aristotle's theory. If I had to find
linguistic evidence for my interpretation of Aristotle, that is the direction
in which I could (and would) go. Even on the present superficial level,
it is not hard to see that my treatment of instantiation works mutatis
mutandis even better with Greek thari with English.
Likewise, we are now in a position to draw an interesting conclusion
from our observations. The different classes of questions with which
Aristotle correlated his other distinctions were primarily indirect questions. The correlation depends crucially on an analogy between relative
pronouns and question words, and this analogy (or near identity) can obviously be best argued for by comparing with each other the logical
behavior of relative clauses and indirect questions. (An especially useful
Mittelglied here is the class of relative clauses without antecedents. Their
logic is remarkably similar to that of indirect questions.) Aristotle's
distinction between different categories is less a distinction between different question types as between question words, and it pertains to these
words in so far as they are doing duty of their relative clause twins.
This observation reflects somewhat unfavorably on those scholars
who have made much of the classification of questions as the alleged cornerstone of Aristotelian categories. It seems to me that their thesis remains unproven. Admittedly, the importance of the dialectical questioning games practiced in the Academy for Aristotle can scarcely be exaggerated. However, there is little evidence in the Topics or elsewhere that
the theory of categories was developed for (or from) such games.
7. CONCLUSIONS FROM THE RECONSTRUCTED ARISTOTELIAN THEORY.
CATEGORIES VS LOGICAL TYPES

One respect in which my reconstruction matches Aristotle's ideas is what

might first have seemed a blemish in it. It is the ambiguity of whatquestions (and of the corresponding mUltiplicity of that-clauses in quantifier phrases) registered above. In my reconstruction certain whatquestions correspond to the category of substance. Aristotle, too, relates
the category of substance to suitable what-questions. As Aristotle puts
in Met. Z I, 1028 b 2, the old question, "What is that w!1ich is?" really
amounts to "What is substance?" Furthermore, and importantly,

THE VARIETIES OF BEING IN ARISTOTLE

105

Aristotle's lists candidates for the status of substance in Met. Z, both in

Met. Z 3 and in Met. Z 13. These lists contain matter, form, and the individual consisting of both. These three correspond to the three senses
of what-questions listed in Section 4 above (in the reverse order).
Moreover, the fourth main candidate, the universal, can perhaps be considered (as Russell Dancy has suggested) a hangover from the earlier
academic practice of answering what-questions by reference to a genus
("the universal") and its differentiae. (This view is certainly supported
strongly by such passages as Met . .::l 28, 1024 b 4 - 6, Top. A 9, 103 b
36 - 37 and A 18, 108 b 22 - 23.)
We also have to raise one of the most crucial questions concerning
Aristotle's theory of categories. Is the theory correct as an analysis of the
"logic" of the Greek language (or of the English language)? Are there
differences between different languages vis-a.-vis Aristotelian categories?
The "transcendental deduction" of Aristotelian categories presented
above might seem to vindicate the main features of Aristotle's theory. In
spite of its persuasiveness, it nevertheless gives us only an approximation
to the true seman tical theory of natural language categories. It is based
upon assumptions which are only partly true, and hence it cannot be
taken as the final word on the subject. In the next section, I shall indicate
one specific limitation of my argument and consequently of our
reconstructed theory of Aristotelian categories. On a general theoretical
level, another major shortcoming of the theory is obvious, connected
with its relation to logical type distinctions. It is in a wider logical and
philosophical perspective clear that even Aristotelian category distinctions must in the last analysis be based on type distinction. Th(' latter
distinctions may not coincide with Russell's. Indeed, the types
(categories) of Frege and Russell seem to me too few and too far apart
to serve as a realistic basis of our Sprachlogik. But, whatever the requisite
types are, they must serve as the foundation of any viable distinction between different categories. In other words, some bridge has to be constructed from Aristotelian categories to logical ones to vindicate them.
How foreign modern type distinctions were for Aristotle is also illustrated by his deeply ingrained habit of bracketing together the obtaining of (what we would call) facts and the existence of individuals. (This
habit was noted above in Section 2.) This assimilation offers us in fact
an additional illustration of the absence in Aristotle of any real distinction between the is of predication, which expresses the obtaining of facts
and the is of existence.
This general problem is highlighted by the more specific observation

106

JAAKKO HINTIKKA

that Aristotelian categories turn out on my analysis to be quite different

from logical categories in the sense of logical types . (This point is relevant
here among other reasons because the contraty has been maintained by
Gilbert Ryle; cf. Ryle 1937 - 1938.) Not only is it the case that entities
of a different logical type (in what is roughly Russell's sense) belong to
the same category, as Socrates, man, and animal all belong to the socalled category of substance. There is a sense in which all categories come
close to containing entities of the same logical type. After all, they all
contain items which can be said of a substance like Socrates. For instance, the members of the so-called category of relation are not relations
for Aristotle, but relatives (relational predicates). This is amply shown
by his discussion of this category in Cat. 7, 8 a 35ff, especially his comments on correlatives and their epistemic and bntological interdependence. (Cf. Cat. 6,6 a 35ff; De Soph. EI. 31, 181 b 26 - 28; Top.
VI, 4, 142 a 28 - 31; and VI, 8, 146 b 3 - 4.) Likewise, quantities are not
for Aristotle what we would think of as quantities (e.g. a certain length),
but quantitative attributes (e.g. being of such-and-such length).
These observations are perhaps not very surprising. There is a sense ill
which the very "category" of relation (as distinguished from relational
predicates) came to its own only much later in the history of philosophy.
(Cf. Weinberg 1965.) However, the absence of relations proper from
Aristotle's categorical scheme highlights the problems it leads into. For
where else can he put relations? The only propositional form he seems
to recognize is the subject-predicate one. If some of those predicates are
relational, we need an account as to how some of them can be built of
relations. Alternatively, we need a reduction of relational propositions
to subject-predicate propositions. Neither task was attempted by
Aristotle, although the latter one was undertaken by Leibniz, whose
philosophy is in the last analysis much more Aristotelian than is usually
recognized. (Cf. here Hintikka 1972b.)
Another instance of the same unsatisfactory state of Aristotle's theory
is the discrepancy between Aristotle's treatment of time in the Categories
and in the Physics. Treating time as a category is simply inadequate for
a satisfactory categorial analysis of time. Merely including temporal
predicates as a class of predicables says next to nothing about the true
"categorial" structure of time. No wonder, therefore, that Aristotle offers an entirely different (and deeper) analysis of time in the Physics.
Similar remarks can be made about several other categories, especially
about the categories of quantity, place, and action.
These are illustrations of deeper and more widespread tensions in

THE VARIETIES OF BEING IN ARISTOTLE

107

Aristotle's thought. In treating (at least in its first stage) all categories on
a par Aristotle (as well as my rational reconstruction of his theory) fails
to give a deeper account of the rationale of category distinctions. It is for
this reason especially important to realize the differences between
Aristotelian categories and logical types.
It is here that Aristotle's relative neglect of the Frege - Russell distinction (even as a difference in use and not just as a difference in meaning)
becomes a handicap for him. Admittedly, it was claimed by Maier that
Aristotle's theory of categories was calculated to accommodate certain
distinctions between different senses of "is" (see Vol. 2, p. 291 ff).
Maier's distinctions include most of the Frege - Russell ones. Indeed,
Maier's first two distinctions are identificatory being vs accidentally
predicatory being, p. 280, and existential vs copulative being, p. 282. No
major insights are forthcoming from Maier, however, into the way
Aristotle managed to combine the Frege - Russell distinction with his.
doctrine of categories. For he firmly believes that. according to Aristotle,
"immediate reflection on the concept of being [Maier's emphasis) ...
forms the principle of division for the table of categories" (pp.
298 - 299). Maier's immediacy claim notwithstanding, Aristotle himself
does not trust immediate intuition here, but discusses the relation of
other categories to that of a substance. However, these arguments are
either calculated to show the dependence of other categories on that of
substance, or (which may come down to the same thing) to point out the
role of focal meaning in relating the being of the other categories to that
of substance. They do not rely on the kinds of distinctions which Maier
mentions or which are likely to be made by a twentieth-century logician.
Instead, they mark a slightly different point of partial contact between
Aristotle's theory of categories and modern type distinctions. For the
primacy of substances over members of categories came to mean for
Aristotle something very much like the claim that only substances are individuals in a modern philosopher's sense. This is shown by Aristotle's
frequent reference to a substance as a "this" (TC>O) or as a "this
something" (TOOf n). Another indication is given by Aristotle iq Met. Z,
3, 1029 a 27 - 29, where he says that "it is taken to be chiefly true of a
substance that it is separable and a certain this". [}(~ )'ae TO xWeWTOP
}(at TO TOOf n lJ7rCxexHP oO}(fi p,CxALaTa Til ova[~.) More generally,
Aristotle's use of separability (TO xwew rop) as a characteristic of
substance (cf. e.g., Met. Z 1, 1028 a 24 - 25) points in the same direction.
In Met. Z I, 1028 a 17 - 20 he says of the members of the other categories
that they "are called beings because they are, some of them, quantities

108

JAAKKO HINTIKKA

of that which in that way is [sc. of substances], some qualities, some affections, some something else" .
Similar statements are found elsewhere, e.g., in Met. r 2, 1003 b
5 - 10. This passage is an especially clear indication of the fact that the
Aristotelian primacy of substance is not due to any related recognition
of the Frege - Russell distinction, for it is in the very same chapter that
Aristotle denies (as we saw in Section 1 above) in effect most emphatically this distinction. However, Aristotle provides little by way of closer
analysis of this mode of dependence of the other categories on
substances. Nevertheless it seems fairly clear that Aristotle's celebrated
manoeuvre of considering the differences between the uses of esti in different categories as not being homonymous but instances of focal meaning (7I'eos gv) is squarely based on the idea of treating substances as
something very much like individuals in the sense of a Frege - Russell ontology. (Concerning Aristotle's attempted Aufhebung of category
distinctions along these lines, see Owen (1960) and (1965).)
8. CATEGORIES, MATTER, AND FORM

We have found plenty of indication that Aristotle did not consider the
doctrine of categories as being completely satisfactory in itself. On
systematic grounds, too, it can be argued that the doctrine so far expounded is only an approximation to the real analysis of the relevant
parts of our Sprachlogik. As a matter of historical record, in the
Organon Aristotle seems to be satisfied with this approximation.
However, in the Metaphysics, especially in rand Z - H, he realizes that
he has to go beyond it. In order to understand how he does this, it is once
again advisable to turn again to topical considerations.
My discussion in Section 4 above was based on a simple analysis of
quantifier phrases. If we look away from the relative clause, the structure
we have presupposed is simply something like the following:

--

QUANT

some
every
each

NP
...............NP

109

THE VARIETIES OF BEING IN ARISTOTLE

This is too simple to be realistic, however. Indeed, Joan Bresnan has proposed the following more refined analysis:
NP

/~

/'"
QP

Det

/~

some
every

NUM

PP

/"'"NP

of

e:~h r;~
Here Q marks among other things quantity classifiers, such as the
following:

one (s)
number
part
herd
gallon
ton

However, instead of this kind of Q we may have almost any correct

noun.
Examples of a quantifier phrases (sans the relative clause) to which
Bresnan's analysis applies are the following:
every two statues of bronze
many tables of rosewood
few cubes of ice.
Roughly, we can thus think of the rightmost NP (under the PP) of
Bresnan's analysis as specifying at least in a number of clear-cut cases,

110

JAAKKO HINTIKKA

the material of which the individuals over which the quantifier ranges as
being formed by imposing the individuation principle that goes together
with Bresnan's Q. The rightmost NP must in such cases involve a mass
term. The whole phrase thus expresses a kind of combination of matter
and a certain individuating principle. This principle is what Aristotle
calls the form, whereas the last NP specifies what Aristotle would call (he
matter. Thus the closer analysis of quantifier phrases just sketched is
closely related to the important Aristotelian contrast between matter and
form.
It is seen that this refinement is our analysis of quantifier phrases goes
beyond the conventional logical languages which can be traced back to
Frege and Russell. For in these languages, one starts from a given class
of basic individuals. The way they are constituted from more basic ingredients, such as matter and form, does not come up in them at all. Hence
the logic of Bresnan's refined analysis cannot be captured by means of
the usual logical languages, even when they are turned into many-sorted
languages capable of incorporating the reconstruction of Aristotle's
theory of categories outlined above.
From what has been said-it also follows that the systematic reconstruction of Aristotelian categories presented above can only be an approximation to the truth of the matter. It is based on an oversimplified
analysis of quanti fer phrases. This need not make my reconstruction of
Aristotelian categories any less interesting historically, however. On the
contrary, it seems that the pressures on the reconstruction due to its approximative character are very closely related to the reasons why the simple picture of categories so far adumbrated did not satisfy Aristotle,
either.
Thus we are beginning to see what light do these systematic observations throw on Aristotle's argumentation in the Metaphysics. First of all,
we can understand the role of one of the main concepts which Aristotle
did not use in the Categories but which he relies on heavily in
Metaphysics. That is the concept of matter. It is one of the main novelties
of the Metaphysics treatment of being and substance as compared with
the Categories.
Now the role of this concept is roughly the same in my systematic treatment as it is in Aristotle, for Aristotle, too, discusses how what in the
earlier approximation were unanalyzable values of quantifiers must now
be thought of as if they were combinations of matter and some individuality principle. This principle Aristotle labels form.
Several further similarities between Aristotle and my analysis can be
registered.

THE VARIETIES OF BEING IN ARISTOTLE

III

For one thing, Aristotle says that substances consist of matter and
form (cf. e.g., 1029 a 27 - 30). Moreover, apparently it is only substances
that do so, not members of other categories. For instance, in Met. H 4.
1044 b 8 - 9 Aristotle says that "things which exist by nature but are not
substances have no matter; their substrate is their substance". [ouo' Qaa
oh cj>vaH
J1.fV, J1.h
ouaLW Of, oux Ean TOVTOLS VA1], aAAO!
TO V7rOXfLJ1.fVOV ~ oua[a.] On the systematic side, too, it is at least questionable whether the full Bresnan analysis can be found among entities
other than individuals. Thus we can see how the introduction of the
matter-form contrast seriously upsets the symmetry between different
categories which they originally enjoyed in Aristotle's Categories.
Another relevant observation is that in many instantiations of the
Bresnan analysis we don't literally have to do with a clear matter-form
combination. In a large number of cases, we may have as the lexical instantiation of the rightmost NP, not a mass term but the plural of a count
noun. Then Q must be instantiated, not by a quantity word, but by a term
which indicates a structure the entities referred to by the count noun can
instantiate. Cases in point are the following:
some discrete set of moments of time
every large school of fish
many ordered groups of numbers.
We may describe these cases by saying that in them higher-order entities
are thought of as being from lower-order one. In contrast, earlier we
were dealing with the formation (construction) of individuals out of matter and form. It seems to me that Aristotle's notions of matter and form
are calculated to cover both formation processes. It is very dubious that
any unambiguous concept can bear such a burden. Hence we find here
some reasons to be suspicious of Aristotle's concept of matter. We have
to be very cautious, though. It might for instance seem that the more
refined analysis of quantifier phrases indicated above embodies a
mistaken admission to Aristotelian ways of thinking in that it in effect
disregards the modern contrast between count nouns and mass terms.
This distinction is linguistically much more dubious, however, than has
been realized in recent literature. Perhaps this is another direction in
which Aristotle is closer to the semantics of natural languages than
Fregean logic.
Be that as it may be, Aristotle's conception is far from unproblematic.
Aristotle's problems are increased by the fact that he assimilated the
form-matter contrast also to the traditional subject-predicate contrast

112

JAAKKO HINTIKKA

and to the actuality-potentiality contrast. It is doubtful that anyone concept can happily cover all these cases. Furthermore, the assimilation increases once again the distance between Aristotle's conceptual
framework and that of modern (Frege - Russell) logic. However, it
would require more space than I have here to follow Aristotle in these
ventures.
BIBLIOGRAPHY
Ackrill, l. L.: 1963, Aristotle's 'Categories' and 'De Interpretatione', Clarendon Press,
Oxford.
Albritton, Rogers: 1957, 'Forms of Particular Substances in Aristotle's Metaphysics',
Journal of Philosophy 54, 699 - 708.
Apelt, 0.: 1891, Beitriige zur Geschichte der griechischen Philosophie, Leipzig.
Balme, D. M.: 1980, 'Aristotle's Biology was not Essentialist', Archiv fur Geschichte der
Philosophie 62, 1 - 12.
Benardete, Seth: 1976 - 77, 'The Grammar of Being', Review of Metaphysics 30,
486-496.
Benveniste, Emile: 1966, 'Categories de pen see et categories de langue', in Problemes de
linguistique generale, Gallimard, Paris, pp. 63 - 74.
Bluck, Richard S.: 1975, Plato's 'Sophist': A Commentary, Gordon C. Neal (ed.), Manchester University Press, Manchester.
Bonitz, Hermann: 1853, 'Uber die Kategorien des Aristoteles', Sitzungsberichte der
Kaiserlichen Akademie der Wissenschaften, Phil.-Hist. Klasse, Vol. 10, pp. 591 - 645;
reprinted as a separate volume 1967 by Wissenschaftlicbe Buchgesellschaft, Darmstadt.
Bresnan, loan: unpublished seminar notes, 'On the Syntax of English Quantifiers', Stanford University.
Dancy, R. M.: 1975, Sense and Contradiction: A Study in Aristotle, D. Reidel, Dordrecht.
Dancy, R. M.: 1975, 'On Some of Aristotle's First Thoughts about Substances',
Philosophical Review 84, 338 - 373.
Dancy, R. M.: 'Aristotle and Existence', this volume, pp. 49-80.
Frede, Michael: 1967, Priidikation und Existenzaussage, Hypomnemata, Vol. 18,
Vandenhoeck & Ruprecht, Gottingen.
Furth, Montgomery: 1968, 'Elements of Eleatic Ontology', Journal of the History of
Philosophy 6, 111 - 132.
Gomez-Lobo, Alfonso: 1976 - 77, 'Aristotle's Hypotheses and Euclidean Postulates',
Review of Metaphysics 30,430-439.
Gomez-Lobo, Alfonso: 1980 - 81, 'The So-Called Question of Existence in Aristotle's An.
Post. 2. 1 - 2', Review of Metaphysics 34, 71 - 89.
Gomez-Lobo, Alfonso: 1981, 'Aristotle, Metaphysics H 2', Dialogos 38, 7 - 12.
Graeser, Andreas: 1978, 'Aristoteles und das Problem von Substanzialitat und Sein',
Freiburger Zeitschrift fur Philosophie und Theologie 25, 120 - 4 I.
Hintikka, laakko: 1959, 'Aristotle and the Ambiguity of Ambiguity', Inquiry 2,137 - 151:
reprinted with changes in Hintikka, Time and Necessity, pp. 1 - 26.
Hintikka, laakko: 1972a, 'On the Ingredients of an Aristotelian Science', Nous 6, 55 - 69.
Hintikka, laakko: 1972b, 'Leibniz on Plenitude, Relations, and the "Reign of Law" " in
Harry G. Frankfurt (ed.), Leibniz: A Collection of Critical Essays, Doubleday, Garden

THE VARIETIES OF BEING IN ARISTOTLE

113

City, pp. 152 - 190; reprinted (with an important change) in Simo Knuuttila (ed.), 1981,
ReJorging the Great Chain oj Being, D. Reidel, Dordrecht, pp. 259 - 286.
Hintikka, laakko: 1973, Time and Necessity, Clarendon Press, Oxford.
Hintikka, laakko: 1976, The Semantics oj Questions and the Questions oj Semantics,
North-Holland, Amsterdam.
Hintikka, laakko: 1979, '''Is'', Semaantical Games, and Semantical Relativity', Journal
oj Philosophical Logic 8, 433-468.
Hintikka, laakko: 1980, 'Aristotelian Axiomatics and Geometrical Axiomatics', in
laakko Hintikka et al. (eds.), Theory Change, Ancient Axiomatics, and Galileo's
Methodology, D. Reidel, Dordrecht, pp. 133 - 144.
Hintikka, laakko: 1982, 'Game-Theoretical Semantics: Insights and Prospect', Notre
Dame Journal oj Formal Logic 23, 219 - 241.
Hintikka, laakko: 1983, 'Semantical Games, the Alleged Ambiguity of "Is", and
Aristotelian Categories', Synthese 54, 443-468.
Kahn, Charles: 1966, 'The Greek Verb "To Be" and the Concept of Being', Foundations
oj Language 2, 245 - 265.
Kahn, Charles: 1973, The Verb 'Be' in Ancient Greek, D. Reidel, Dordrecht.
Kahn, Charles: 1973, 'On the Theory of the Verb "To Be" " in Milton K. Munitz (ed.),
Logic and Ontology, New York University Press, New York, pp. 1-20.
Kahn, Charles: 1976, 'Why Existence does not Emerge as a Distinct Concept in Greek
Philosophy', Archiv Jur Geschichte der Philosophie 58, 323 - 334.
Kirwan, Christopher: 1971, Aristotle's Metaphysics, Books r, ~, E, Clarendon Aristotle
Series, Clarendon Press, Oxford.
Loux, Michael 1.: 1974, translator and editor, Ockham's Theory oj Terms, Notre Dame
University Press.
Maier, Heinrich: 1896-1900, Die Syllogistik des Aristoteles, Vols. 1- 2, K. F. Koehler,
Stuttgart; reprinted 1969 by Georg Olms, Hildesheim.
Mates, Benson~ 1979, 'Identity and Predication in Plato', Phronesis 24, 211 - 229.
Moravcsik, 1. M. E. (ed.): 1967, Aristotle: A Collection oj Critical Essays, Doubleday,
Garden City.
Moravcsik, 1. M. E.: 1967, 'Aristotle's Theory of Categories', in Moravcsik (ed.), Aristotle: A Collection oj Critical Essays, pp. 125 -145.
Moravcsik, 1. M. E.: 1976, 'The Discernability of Identicals', Journal oj Philosophy 73,
587 - 598.
Owen, G. E. L.: 1%0, 'Logic and Metaphysics in Some Earlier Works of Aristotle', in 1.
During and G. E. L. Owen (eds.), Aristotle and Plato in Mid-Fourth Century (Studia
Graeca et Latina Gothoburgensia, Vol. ii), Goteborg, pp. 163 - 190.
Owen, G. E. L.: 1965, 'Aristotle in the Snares of Ontology', in R. Bambrough (ed.), New
Essays on Plato and Aristotle, Routledge and Kegan Paul, London, pp; 69 - 95.
Owen, G. E. L.: 1978-79, 'Particular and General', Proceedings oj the Aristotelian
Society 79, I - 21.
Ross, W. D.: 1924, Aristotle's Metaphysics, A Revised Text With Introduction and Commentary, Vols. 1-2, Clarendon Press, Oxford.
Ryle, Gilbert: 1937 - 38, 'Categories', Proceedings oJ the Aristotelian Society, Vol. 38, pp.
189-206; reprinted in 1971, Collected Papers, Vol. 2, Hutchinson, London, pp.
170-184.
Ryle, Gilbert: 1939, 'Plato's Parmenides', Mind 48, 129 - 151, 302 - 325.
Saarinen, Esa (ed.): 1979, Game-Theoretical Semantics, D. Reidel, Dordrecht.
Trendelenburg, Adolf: 1846, Geschichte der Kategorienlehre, Bethge, Berlin.

114

JAAKKO H!NT!KKA

Weidemann, Hermann: 1980, 'In Defense of Aristotle's Theory of Predication', Phronesis

25,76-87.
Weinberg, Julius R.: 1965, 'The Concept of Relation: Some Observation on Its History',
in Abstraction, Relation, and Induction: Three Essays in the History of Thought, The
University of Wisconsin Press, Madison, pp. 61 - 119.
Woods, M. J.: 1974-75, 'Substance and Essence in Aristotle', Proceedings 0/ the
Aristotelian Society, Vol. 75, pp. 167 - 180.

Dept. of Philosophy,
Florida State University,
Tallahassee, FL 32306- /054, U.S. 4.

THE CHIMERA'S DIARY

Edited by
STEN EBBESEN

Institute of Greek and Latin Medieval Philology, University of Copenhagen

[Editor's note: This paper reproduces the manuscript left by the chimera, but I have added
references to books and manuscripts, plus a few notes which appear in square brackets.
The reader will notice that the chimera has wisely disregarded accidental changes of
philosophers' choices of example when they need a composite animal. The chimera takes
remarks about, e.g., the goat-stag as remarks aimed at itself. As a matter of fact, Aristotle
and the Greek Aristotelian commentators prefer the goat-stag (TQOI-YMOIIPOS) and the centaur ({1r1rOXEJlTaIlQOS). In the Hellenistic period, the centaur, the scylla and the chimera are
the standard examples. In Latin medieval texts the chimera (inherited from Manlius
Boethius) is vastly more popular than any of the other composite animals.)

My feelings towards philosophers are mixed. For centuries they have

used me as an experimental animal, keeping me on a minimum of being.
In a way I may owe them my "life", but their experiments have weakened me so much that the end may be drawing near. If my weakness proves fatal, please inform the Centaur, Goat-Stag and Pegasus, who are my
next of kin. If the philosophers kill me, I expect them to keep at least one
of my relatives alive in order to continue the experiments. If we are all
doomed, I would like to secure us a place in man's memory. This is why
I have put together these eXtracts from my diary, recording the sufferings
to which I and my tribe have been subjected.
1. LONG B.C.

"In front a lion, in the rear a serpent, in the middle a goat" - that's
what I am. Zeus bless Homer for his excellent description [Iliad 6.181].
2.400 B.C.

Philosophers exhibit an un savoury interest in the being and the nonbeing. Gorgias claims that I am not. He seems to say, though, that my
claim to being is as good as anyone's because I may be thought about
115
S. Knuuttila and J. Hintikka (eds.), The Logic of Being, 115-143.
1986 by D. Reidel Publishing Company.

116

STEN EBBESEN

('PQoPiat?m). But he does not intend to save me that way. Only to show

that being thought about is no criterion of being, so that even if

something else is and I am not, men cannot single out that something
else. [Here the chimera may be wrong in detail, though hardly in
substance. Sextus Empiricus, Adversus Mathematicos 7.80, mentions
the chimera in a paraphrase of Gorgias' On the Non-Being (nQL TOU
p.r, OPTOS). He also mentions the Scylla. I doubt whether these examples
occurred in Gorgias' text. They do not occur in our other source for that
work, viz. Ps.-Aristotle's De Melisso, Xenophane, Gorgia (979a - 980b
in Bekker's edition of Aristotle; new edition ill Cassin (1980) 610-643,
at p. 64) = Bekker 980a).]
3.320 B.C.

This is an exciting century. Philosophers now seem to agree that I am

not, but thanks to Plato [Sophist] they no longer throw all us non-beings
into the waste-bin with the remark that "it shall never be established that
things which are not, are" [Parmenides F. 7]. One brilliant representative of this new flexibility is Aristotle whose lectures I used to attend
(invisibly, of course) until his recent decease. He once stated as a simple
matter of fact that people do say "the non-being is non-being" [Metaph.
4.2 1003blO]. But he would not accept an unqualified 'the non-being is'.
Thus he said that it does not follow that the non-being is because it is anobject-of-belief (oo~aaToP, I think I will henceforward render this
'opinable'), "for it is not the same to be something and to be simpliciter
though the similarity of expression makes it seem so" [Sophistici Elenchi
5 167al - 6]. He also explained the failure of the inference with the words
"for there is belief (or: opinion - oo~a) about it [i.e. the non-being] not
because it is but because it is not" - or rather (because his use of OTt
was ambiguous): "for there is belief concerning it, not [a belief] that it
is, but [a belief] that it is not" [De interpretatione II 21a32 - 33. Cf.
Analytica Priora 1.38 49a24 and the Greek commentators on the latter
text, viz. Alexander, APr., CAG 2.1 :368 - 369; Philoponus APr. J,
CAG 13.2:345; Leo Magentinus, APr. J, MS Vat. gr. 244: 229r, scholium
P1'/'.] Whichever interpretation is correct, I think he wanted to say that
the very point of saying 'the non-being is opinable' is to make it clear that
one will not subscribe to the unqualified 'the non-being is'. In the same
lecture he indirectly informed me that my friend Homer is no more,
because he treated the inferences 'Homer is a poet, therefore Homer is'
and 'the non-being is opinable, therefore the non-being is' as similar in
structure [De interpretatione II 2Ia25sqq.].

THE CHIMERA'S DIARY

117

Other memorable remarks of Aristotle's include, "Not-man is not a

name ... Let us call it an infinite name since it holds indifferently of
anything, whether being or non-being". [De interpretatione 2
I6a30- 33; 'since' etc. omitted by the best mss. and in Boethius' Latin
translation, but cf. Boethius, Int. 'C.'. 2a p. 62.7 Meiser: "et aequaliter
dicitur vel in eo quod est vel in eo quod non est".] He made a similar
remark about verbs with a prefixed 'not-' [De interpretatione 3
I6bI2-15]. Also, he said that "even the goat-stag signifies something,
but not, as yet, anything true or false, unless 'is' or 'is not' is added" [De
interpretatione 1 16al6 -18]; and "of that which is not, no one knows
what it is, only what the account or the name signifies when I say 'goatstag', but it is impossible to know what a goat-stag is" [Analytica
Posteriora 2.7 92b5 - 8; cf. Topica 4.1 121a21 - 25].
I must admit that I do not quite grasp all the implications of Aristotle's
statements about the non-being and fabulous animals. I hope that future
generations of men will help me understand him better.
4.50 B.C.

I have just read a beautiful poem by an Epicurean called Lucretius. He

gives us monsters a physical existence, thinking that the atomic pictures
thrown off by the animals of which we are composed may get mixed up
so as to produce monstrous pictures which men may perceive. But
he will not give us physical existence as sources of those pictures. The
freedom of atoms to combine is not so great, he holds, that beings consisting of parts belonging to different natural species can arise
[Lucretius, De rerum natura 2.700-717,4.722-748,5.878-924].
Others will only give us conceptual existence. It seems to be the common notion these days that there are two ways of forming concepts, (a)
directly on the basis of things met with in nature; (b) through creative
work on the materials gathered through direct confrontation with the
things. Creative work which consists in joining things which have not
been received together is called 'secondary combination' (7rtavpt'hUL~)
and is responsible for concepts of such monsters as me. [See Ebbesen
(1981), I: 191.]
5. A.D. 650

I once wished for help to understand Aristotle. During the last 500 years
or so there have been many Aristotelian commentators. By now they are

118

STEN EBBESEN

few. I therefore think it may be time to try to summarize what they have
said about us non-beings.
Some of the more recent writers have the rather crude notion that our
names are non-significative in the same way that 'blityri' and 'skindapsos' are; or that 'goat-stag', 'blityri' and 'skindapsos' share the property
of signifying non-existent things. [Stephanus, Int., CAG 18.3: 7.17 - 18;
Elias, Intr., CAG 18.1: 3.7 - 8; Ps.-Elias, Intr., 25.8, p. 52; David, Intr.,
CAG 18.2: 1.16-18; Anonymus Moraux, Intr., 111.107 -109 p. 80; cf.
Eustratius, APo., CAG 21.1: 95 - 96.] As if it made no difference that
you can describe a goat-stag or me, but not a blityri or a skindapsos since
no sort of meaning or notion is associated with those words - in fact,
they are "words" coined by philosophers precisely to show that it is
. possible to have an articulated string of sounds with n.9 meaning at all.
The better commentators - and even the naive ones, on occasion agree that 'chimera' and 'goat-stag' are significative and not to be confused with such nonsense "words" as 'blityri'. (After all, I have a
nominal definition [cf. Simplicius, Ph., CAG 9: 696.3sqq; Ps.Philoponus, APo. II, CAG 13.3: 359.26-360.9]). In their standard
theory, the semantic relation has three terms. The word signifies a concept which signifies a thing [cf. Ebbesen (l981a), 1: 141ff]. In my case,
they hold, the semantic relation is not satisfied as far as the third term
is concerned since there is no independently subsisting thing (or:
"nature") to be signified by the word via the concept; but at least there
is a concept of sorts. With 'blityri' there can be no semantic relation at
all, since both concept and thing are lacking. [Ammonius, Int., CAG 4.5:
29.8-9 (cf. 184-185); Elias, Cat., CAG 18.1: 129.15-17, Cf.
Boethius, Int. ed 2a p. 50; Ps.-Philoponus, APo. II, CAG 13.3:
362:32 - 363.2.]
I should like to know whether the commentators will allow me to be
in some Aristotelian category. The general attitude seems to be negative.
[See lamblichus apudPhiloponus, Cat., CAG 13.1: 9; Ammonius, Cat.,
CAG 4.4: 9 - 10; Dexippus, Cat., CAG 4.2: 7.20 - 24.] But perhaps Porphyry accepted me as a quasi-substance, in the sense that my name names
a concept of something which is not, as if it were a substance [Porphyrius
apud Simplicius, Cat., CAG 8: 11.6 - 12].
At any rate, if there is a concept corresponding to my name, it may
be argued that I have a sort of conceptual existence as long as people
think of me - though they may kill me by ceasing to do so - and that
it is possible to know and understand (E7rLamaiJw) what I am. This
knowledge and understanding (E7rtaT~/L7J, scfentia) will be as perishable

THE CHIMERA'S DIARY

119

as its object (E7rtaT7]TOII, scibile), being totally dependent on the human

mind which possesses the understanding. My being known and my being
become inextricably joined. Some have suggested that this may in fact
be the case [Porphyrius, Cat., CAG 4.1: 121.4-15. Cf. Boethius, Cat.,
PL 64: 229; Simplicius, Cat., CAG 8: 191; Ammonius, Intr., CAG 4.3:
39.14-40.6].
However, this line of though makes me similar to such abstract universals as the species "man". Too similar, most commentators think. Hence
thei introduce a distinction between "constructive thought" (E7riIlOteX)
and "mere constructive thought" (1/tt)..~ E7riIlOta). Abstract, post-rem,
universals are (concepts) derived from extra-mentally existing things,
and so they "are in constructive thought", but not "merely"; whereas
the concept of a chimera "is in mere constructive thought" for lack of
an extramental correlate. [See David, Intr., CAG 18.2: 116-117. Cf.
Elias, Intr., CAG 18.1: 49.19 - 20; Anonymus Moraux, Intr.,
111.102 - 109 p. 80. Cf. also David, Intr., CAG 18.2: 46.35 - 47.1, where
1/tt)..~ E7rillOta is distinguished from rpaIlTaaia].
By this move the commentators actually assign me a baser sort of concept than real things have, and so they can deprive me of my "scibility",
leaving only things associated with the finer sort of concept as possible
objects of genuine knowledge and understanding (scientia). They can
reaffirm the Aristotelian dictum "about that which is not, one cannot
know what it is" and they can deny that I have a being which consists
in being known (in the pregnant, Aristotelian sense of the word), though
they concede that man's thought may give me a precarious existence
which lasts as long as the thought about me. [See Ammonius, Intr., CAG
4.3: 39.14 sqq; David, Intr.,CAG 18.2: 108.25-109.5 & 114.1-6;
Elias, Intr., CAG 18.1: 47.3-9; cf. Boethius, Cat., PL 64: 229.]
Since the commentators allow me and my kin a being in the mind, they
can also allow us some predicates, such as 'not-just', 'nameable',
'opinable'. Ammonius in one place [Int., CAG 4.5: 184.25sqq.] imagines
somebody who [in imitation of Aristotle, Physics 4.1 208a30] asks,
"Where is that goat-stag of which we truly predicate 'not-just'?". The
answer given by Ammonius is, "In constructive thought". Predication
presupposes that the subject has been conceptualized - constructed in
thought - but not that it has existence before being constructed in
thought.
The predicates I truly have include infinite nouns and verbs, 'not-just',
'not-man', 'does-not-run', for instance [Ammonius, Int., CAG 4.5: 42,
52,184-185; Boethius, Int. ed. la, p. 60; ed. 2\ pp. 62, 69-70]. This

120

STEN EBBESEN

opens an interesting perspective. If these infinite terms can be truly

predicated of me, does it not follow that some finite predicates of the
same type can also be so? If I am not-just, not-man and do-not-run,
doesn't this mean that I am some other substance than man, equipped
with some other quality than justice, and exercising some other activity
than running? I have little hope, though, that the commentators would
admit as much. It looks as though the only finite predicates they will
allow me are such as indicate being in thought and having a name
('opinable', 'nameable' etc.), plus, I suppose, my nominal definition
('animal composed of goat, lion and snake'). My manuscript of
Boethius' second commentary on De interpretatione has a lacuna at the
vital point [Boethius, Int. ed. 2 a p. 70], but I think he wants to say that
in 'a/the centaur runs' running is predicated, but falsely. This indicates
that he thinks that all categorical propositions with finite predicates and
me for their subject are false when affirmative and true when negative.
But he does not say this in so many words.
Anyhow, the commentators agree that 'the chimera is not' is true. Ammonius [reported by Philoponus, APo. I, CAG 13.3: 323.12sqq.] once
added that since this cannot be otherwise, to know it is to have genuine,
"scientific" knowledge (7rL(JT~p,.,.,). Some others similarly state that I
neither am, nor will I possibly be [Boethius, Int. ed. 1a p. 60; Simplicius,
Ph., CAG 9: 517.9 - 10; also Michael Ephesius (12th c.), Metaph., CAG
1: 573.33 - 34]. They do not say whether the impossibility of my being
can be deduced from the principles of biology (essentially = Lucretius'
position) or whether I am ex hypothesi a creature that cannot be. In
either case, this bodes ill. If people become convinced that it is no contingent, but a necessary fact that I do not exist, my scarce ration of being
may be curtailed even further.
As I have mentioned, the commentators accept the truth of 'a/the
chimera is opinable', but they do their best to explain away its apparent
existential implication. One approach consists in saying that 'opinable'
means 'being in thought, but not extra-mentally" [Boethius, Int. ed.2 a ,
p. 376; cf. ed.
p. 166]. This interpretation of 'opinable' can be
bolstered up with the following interpretation of Aristotle's remark in De
interpretatione [11 21 a32 - 33]: "for there is opinion (and not scientific
knowledge) about the non-being not because it is (because then there
might be scientific knowledge about it) but precisely because it is not"
[Boethius, II. cc.]. Secondly, they sometimes follow Plato in
distinguishing between several kinds of being and not-being. The being
which the proposition 'the chimera is opinable' truly asserts I have is a

r,

THE CHIMERA'S DIARY

121

vile one - viler even than being-different-from

namely being-inopinion, from which it is not permissible to detach a simple "being" as
it stands in an indissoluble relationship to opinion [Thus Ammonius,
Int., CAG 4.5: 213; cf. Anonymus Tanin, Int. p. 98]. Thirdly, following
a suggestion of Aristotle's (De interpretatione 11 21 a26 - 28], they say
that in 'the chimera is opinable', 'is' is not primarily predicated of me
but of the opinable, so that the sense is "there is an opinion, and that
is an opinion of a chimera"; whereas 'is' is directly predicated of me in
the false proposition 'the chimera is' [Boethius, Int. ed. 2 a p. 376; cf.
Ammonius, Int., CAG 4.5: 184 & 211].
For good measure, rabid Platonists even deny me a concept (vorIlLCx).
The Peripatetic commentator Alexander of Aphrodisias has used me in
an explanation of one of Aristotle's arguments against the ideas [Alexander, Metaph., ad A 9 990b 14, CAG 1: 81 - 82; see also the text in
Harlfinger (1975),26-27]. If, the argument runs, thinking (vofiv) of
something requires a permanent, non-particular object - an idea -,
there must also be an idea of me, since people think of me. Unfortunately, Alexander's intention is not to equip me with an idea. He uses me in
much the same way as Gorgias did. Assuming that it would be silly to
posit an idea of me, he concludes that one should not posit ideas at all.
About a century ago the more Platonically oriented Asclepius delivered
a counter-attack against Alexander, in order to save the ideas - but not
mine. He pointed out that believing there to be an idea when something
is intellectually grasped, with the grasp called vorWL~, does not entail
believing in ideas of things which are not. For the sort of intellectual
grasp people have of such things is not a VOTJal~ but a cpCivTCiaiCi
[Asclepius Metaph., CAG 6.2: 74-75]. This seems to be the same sort
of trick as the one used to save the abstract universals. A distinction between two sorts of intellectual grasp is introduced and the inferior one
is associated with me. But I just wonder whether Asclepius' argument is
not circular. How does a man know that a thought of me is no vOTJaL~
except by knowing that I have no idea? The Aristotelian may claim that
he empirically knows that I do not exist in the particulars. and that
therefore the concept of me has no real foundation. But can the Platoni~t
have a similar knowledge of which ideas there are? Is the non-existence
of an idea of me vouched for by the empirical fact that he sees no mirrorimage of such an idea in the world of coming-to-be and passing-away?
Or did Asclepius have a reason for considering my non-existence a
necessary fact? I just wonder.

122

STEN EBBESEN

6. A.O. 1260

After several dull centuries, the last one and a half have been exciting.
I have often wondered whether I am an individual or a universal. [The
same doubt has haunted the editors of CAG. Some print XLJLmQex and
give the references in the index nominum; others use lower-case and put
the word in their index verborum.] It is now clear that I am a universal.
An endangered species, in fact. Thus one introduction to logic says,
Est quoddam universale quod praedicatur de nullo actualiter, sed de pluribus secundum
intellectum, ut chimaera [Anon., Logica Cum sit nostra, LM 11.2: 432. 11 - 12; cf. Anon.,
Logica Ut dicit, LM 11.2: 387).
[There is one sort of universal which is predicated of nothing actually but of several
things in thought; thus (the) chimera.)

One recent philosopher claims that the opinable is my genus, saying,

cum dicitur 'chimaera est opinabilis', haec est enuntiatio, et est de universali non universaliter sumpto. Unde si dicatur 'chimaera est opinabilis', praedicatur hic genus de specie.
Et sciendum quod istae differentiae 'universale', 'singulare' ut hic sumuntur non sunt differentiae entis sed rei, et non sumitur ibi res secundum quod dicitur res quae habet esse actuale, sed dicitur res omne illud quod potest apprehendi ab anima [Anon., Int., MS Oxford
Bod!. Canon. misc. 403: 43rB; minor scribal errors have been tacitly corrected).
[When we say 'a/the chimera is opinable', this is a statement and it is about a universal
not taken universally. In fact, when we say 'a/the chimera is opinable' a genus is predicated
ofa species. It should be understood, though, that the differences 'universal' and 'singular'
in the sense employed here are not differences of (a/the) being but of (a/the) thing; and
'thing' is not used in the sense in which a thing is so called because it has actual being, but
'thing' is used to cover everything which may be grasped by the mind.)

Like most people these days he holds that any proposition of the form
'A is (a) B' may be expanded ("expounded", they say) so as to yield "A
is (a) B being", in which 'B' modifies ("restricts") 'being' in the same
way that 'rational mortal' restricts 'animal' in 'rational mortal animal'
[Gp. cit. 48rA, quoted in lewry (1978),128. lewry mistakenly gives the
reference as 47vB]. The implication is that 'a/the chimera is opinable'
may be expounded as 'a/the chi mara is (an) opinable being'.
This philosopher is in express opposition to another group which tried
to get rid of the being that appears to be assigned me in that proposition
by distinguishing between two senses of 'is'. According to them, 'is}' =
"'is' secundum adiacens" is short for 'is 2 being'; whereas 'is 2 ' = "'is'
tertium adiacens" is a mere copula, i.e., solely a sign of composition of
subject and predicate. Their 'is 2 ' cannot be subjected to "exposition",
i.e. it cannot be considered shor.t-hand for 'is being' or any other phrase.
Consequently, the move from "the chimera is opinable' to "he chimera

THE CHIMERA'S DIARY

123

is' would be a move from 'is 2 ' (the "substantive 'is' ", as some say) to
'is l ' (which they call "adjective"); and so the move is illegitimate as it
hinges on the wrong assumption that 'is' has one sense only.
[An excellent collection of 13th-century texts bearing on this question
(as well as the others dealt with by the chimera in the present entry) is
found in Lewry (1978) and (1981a - b). The interpretation of est tertium
adiacens as a mere copula is at least as old as Abelard (see, e.g., his
Dialectica, pp. 135, 162, 164; cf. De Rijk (1981a- b)). It is earnestly
defended by Robert Bacon (early 13th c.) in his Syncategoremata; text
in Braakhuis (1979), I: 131 - 135). Robert's view is criticized in an
anonymous scholium on Sophistici Elenchi 166b37, in M. Osterr. Nat.bib!., lat. 166: 184v:
Item quaeri potest de paralogismis quos ponit Aristoteles. Sunt autem huiusmodi: 'quod
est opinabile, est; sed quod non est, est opinabile; ergo quod non est, est'; eodem modo:
'chimaera est opinabilis, ergo chimaera est'; alius paralogism us talis: 'hoc non est homo,
ergo non est'. Videtur quod in primo paralogismo sit aequivocatio, quoniam proceditur ab
hoc verbo 'est' secundum quod est substantivum ad idem secundum quod adiectivum.
Dicitur 'est' dupliciter sumi: quandoque scilicet est adiectivum, quando[que] simpliciter
praedicatur et non ponitur in numero, ut 'Socrates est'; quandoque est substantivum, ut
quando ponitur in numerum, ut 'Socrates est homo'. Fit ergo processus in praedicto
paralogismo secundum diversas acceptiones huius verbi 'est', scilicet in una est secundum
quod est substantivum, in altera secundum quod adiectivum; et ita erit ibi aequivocatio vel nulla est praedicta distinctio.
[Further, questions may be raised concerning the paralogisms in Aristotle's text. They
are as follows: 'what is opinable, is; but what is not, is opinable; therefore what is not, is';
in the same manner: 'althe chimera is opinable, therefore althe chimera is'; another
paralogism goes like this: 'this is not a man, therefore this is not'. It is arguable that in the
first paralogism we have to do with equivocation because there is a move from the verb
'is' in its substantival function to 'is' in its adjectival function. 'Is' is presumed to be
capable of two uses; sometimes it is adjectival (namely when it is precticatcd absolutely and
does not enter in a count, as in 'Socrates is'); sometimes it is substantival (namely when
it does enter in a count, as in 'Socrates is a man '). So, in the above paralogism the move
from premisses to conclusion is accompanied by a shift in the interprelation of the verb
'is', which in one proposition is laken in its substantival funclion, in another in the adjectival one; this, then, must be a case of equivocal ion - or else the said dislinclion is void.]

This question is solved as follows:

Ad hoc quod obicilur de primo paralogismo dicendum quod non eSI ibi aequi\ ocalio, sed
processus secundum quid el simpliciler, quoniam 'esl' secundum sC pracdical esse
simpliciler el secundum aClum, haec aulcm determinatio 'opinabik' Irahil ipsum ad
nominandum esse quodam modo, scilicet esse secundum opinionem; unde proccdilu ibi ab
esse quodam modo, scilicel secundum opinionem, ad esse simpliciler el in aLIU. Unde dici
POlcsl quod nihil esl dicere [ctislingucre cod.] hoc verbum 'esl' e"c aeqlliv()cum, nee quod
quandoquc CSI subslanlivum, quandoque adieClivum: semper cnim esl substanli\ulTl, sive
ponilur in numerum sive praedicelur simpliciICr.]

124

STEN EBBESEN

[As regards the criticism raised concerning the first paralogism, the answer must be that
there is no case of equivocation but a move from in-some-respect to absolutely, because
'is' in itself predicates being absolutely and actually, whereas the determination 'opinable'
makes it name being in some way - being in opinion, that is. Hence a move is performed
from being in some way - in opinion - to being absolutely and actually. Thus it may be
said that there is no foundation for claiming that the verb 'is' is equivocal, nor for claiming
that it is sometimes substantival, sometimes adjectival. For it is always substantival,
whether it enters into a count or is predicated absolutely. 1

The distinction between the predicative and the existential 'is' has met
with general disapproval, at least from the time of Robert Kilwardby (c.
1240). He flatly denies that 'is' is equivocal, and holds that the right way
to expound both 'is l ' and 'is 2 ' is "is being" [see text in Lewry (1978),
128]. This exposition brings out that a predicate has both matter and
form. The fo!'m is the means relating the matter to..the subject. When 'is'
is secundum adiacens, as in 'a/the man is', the matter of the predicate
is being simpliciter. If 'is' is tertium adiacens, the matter is a specified
sort of being - substantial in '(a/the) man is (an) animal', accidental in
'(a/the) man is just'. In his commentary on De interpretatione, Kilwardby says:
Dubitatur postea si praedicetur tertium, propter hoc quod in omni enuntiatione est medium
hoc verbum 'est' et subiectum et praedicatum extrema, cum praedicatur tertium adiacens;
et nihil unum et idem potest esse medium et extremum; ex quo sequitur quod non
praedicatur tertium. Sed intellege quod uno modo extremum, alio modo medium: ratione
compositionis medium, ratione substantiae sive rei verbi extremum. Est enim hoc ipsum
'est' praedicatum secundum materiam et formam; et dico formam praedicati medium per
quod comparatur praedicatum [medium) subiecto, et praedicatum secundum materiam
dir:o rem verbi; et ex hiis fieri unum praedicatum, ut cum dico 'homo est' id est 'homo est
ens', et sic praedicatur hoc ipsum 'est' secundum adiacens. Et quam vis sit copula aliquo
modo tertium (scilicet non ordine sed numero), quia iIIud cui adiacet (scilicet praedicatum
secundum materiam) non ponit in numero cum eo, non dicitur hoc ipsum 'est' esse tertium
adiacens, sed secundum, cum subiectum sit quod adiacet et praedicatum est quod adiacet;
et quia in hac 'homo est iustus' et 'homo est animal' praedicatur esse specificatum (scilicet
per substantiale et accidentale), quod quidem ponit in numerum cum esse simpliciter, ideo
dicitur hoc verbum 'est' in talibus praedicari tertium adiacens, ita scilicet quod sit tertium
numero, non ordine. Sic ergo aliquando praedicatur hoc verbum 'est' tertium adiacens.
[MS Cambridge, Peterhouse, 206: 7?rB.)
[Then doubt is raised whether it is predicated third. For in every statement the verb 'is'
is the mean and the subject and the predicate the extremes, when it is predicated as a third
supplement; and no one thing can be both mean and extreme. It follows that it is not
predicated third. But it should be realized that it is an extreme in one way, a mean in
another way: a mean as far as the composition is concerned, an extreme as far as the
substance or content of the verb is concerned. For this very 'is' is a predicate in respect of
matter and form. By a predicate's form I understand the mean through which the predicate
is related to the subject. By predicate in respect of matter I understand the verbal content.
Now, my claim is that these two together constitute one predicate, as when I say '(a/the)
man is' = '(a/the) man is being', and thus 'is' is predicated as a second supplem~nt. And

THE CHIMERA'S DIARY

125

although the copula is in a way a third - not in order, that is, but in number -, as that
to which it is a supplement - viz. the predicate in respect of matter - does not add to the
count together with it, 'is' is not said to be a third supplement but a second one, since the
subject is one supplement and the predicate is another supplement. And since in the propositions '(a/the) man is just' and '(a/the) man is an animal' the being which is predicated
is specified (as substantial or accidental), and this adds to the count together with being
absolutely, for this reason the verb 'is' is said to be predicated as a third supplement in such
propositions, i.e., in such a way that it is third in number, not in order. Thus, then, 'is'
is sometimes predicated as a third supplement.)

There can be no doubt that for Kilwardby the matter of the predicate
in 'a/the chimera is opinable' is opinable being. In fact, there has been
quite a debate about the status of opinable being vis-a-vis simple being.
Is the one somehow included in the other? If opinable being were a subjective part (species) of being, it would be permissible to argue, 'this is
opinable, ergo this is'. The argument would be as good as 'Socrates is
a man, ergo Socrates is an animal'. But since the substitution of 'the
chimera' for 'this' would render the antecedent true and the consequent
false, people deny that opinable being is a species of simple being. But
then it might seem that opinable being is a wider term than being since
both the (actually) being and the non-being is opinable. [So already Alexander Aphrodisiensis, Topica, CAG 2.2: 359.17 -18 ad Top. 4.6
127a34sqq.). This view would imply that 'this is, ergo this is opinable'
is a sound inference. But then the converse inference, 'this is opinable,
ergo this is' would commit the fallacy of consequent rather than the
fallacy secundum quid et simpliciter which Aristotle says it commits.
Kilwardby, and many others, solve this difficulty by distinguishing
between a proper and an improper sense of 'opinable'. In the proper
sense, they hold, it means "in opinion only, not really". When
'opinable' is taken in this sense, the actually being does not fall under the
opinable, and so there is a fallacy secundum quid et simpliciter, but not
a fallacy of the consequent in 'this is opinable, ergo this is'. However,
in an improper and general sense 'opinable (being)' does comprise actual
being as well as non-being or imaginary being. When 'opinable' is taken
in this sense, there is a fallacy of the consequent in 'this is opinable, ergo
this is' - but then, did not Aristotle himself say that one and the same
argument may be fallacious for more than one reason?
[For an early (12th century) treatment of the above que~tion, see
Anonymus Cantabrigiensis, Commentarium in Sophisticos Elenchos,
MS Cambridge St John's College D.12: 90rB - vA (on Arist. SE 166b37).
The following texts are all doctrinally close to Kilwardby's: Anonymus
Monacensis, Commenlarium in Sophislicos Elenchos, MSS Admont

126

STEN EBBESEN

241: 28rA - B & CLM 14246: 11 vA (presumably a little earlier than

Kilwardby); Albertus Magnus, Expos. SE 1.3.7, ed. Borgnet 2: 570
(directly influenced by Kilwardby, as shown in Ebbesen (1981 b; Anon.,
Scholium in Arist. SE 166b37, MS Cambridge Gonville and Caius
466/573: 137v; Anonymus e Musaeo, Commentarium in Sophisticos
Elenchos, MS Oxford Bodleian E Musaeo 133: 3rA. Kilwardby's own
text (Commentarium in Sophisticos Elenchos, MSS Cambridge
Peterhouse 205:291rB-vA & Paris BN lat. 16619: 18rB-vA) runs like
this (discrepancies between the mss. are few and not noted here):
Adhuc dubitatur de primo paralogismo quem ponit in littera (SE 167al), in quo dicit esse
fallaciam secundum quid et simpliciter. Videtur enim quod ibi sit magis fallacia consequentis quam secundum quid et simpliciter cum dicit 'est opinabile, ergo est', hac ratione: quandocumque ali quod argumentum sic se habet quod sequitur econverso et non sic, ibi est
fallada consequentis; sic autem est in praedicto paralogismo; ergo in ipso est fallacia consequentis. Maior huius ration is patel. Minor etiam manifesta est, cum omne quod est potest
cad ere in opinione et non econverso. Penitus eodem modo opponitur de secundo
paralogismo. Ibi enim est fallacia consequentis a destructione antecedentis cum dicit 'non
est homo, ergo non est'; econverso enim sequitur, sic: 'non est, ergo non est homo' .... Ad
primum dicendum quod, sicut dicit Aristoteles in secundo huius, nihil impedit plures occa
siones fallendi esse in una eademque oratione. Unde dico quod in prima oratione, si
sumatur 'opinabile' communiter, est fallacia consequentis. Uno modo opinabile enim communiter est omne illud quod cadit in opinione, et hoc modo non sequitur 'est opinabile,
ergo est', immo tenet econverso. Sumpto ergo opinabili communiter est in hoc paralogismo
tam consequens quam secundum quid et simpliciter. Ipso autem sumpto proprie non est
ibi fallacia consequentis: tunc enim non sequiter econverso, sic: 'est, ergo est solum in opinione'. Hoc enim est proprie opinabile quod est solum in opinione. Un de dico quod Iicet
in hoc paralogismo sit tam fallacia consequentis quam secundum quid et simpliciter, est
tamen hic proprius secundum quid et simpliciter, et dico "proprius" quia secundum quid
proprie sumitur opinabile. In alio autem paralogismo, cum dicitur 'asinus non est homo,
ergo non est', est fallacia consequentis et similiter fallacia secundum quid et simpliciter.
Inducit tamen ipsum propter peccatum secundum quid et simpliciter. Dicunt tam en
quidam quod non est ibi fallacia consequentis. Ponunt enim quod ista duo possunt simul
stare 'est homo' et 'non est', et dicunt quod Caesar est homo et cum non est. Et cum
arguitur contra eos sic: 'est homo, ergo est', dicunt quod hoc argumentum non valet, et
hac ratione: omne praedicatum essentiale et universale dicit esse habitudinis; quandocumque autem praedicatur hoc verbum 'est' secundum adiacens, dicit esse temporis sive esse
ut nunc. Unde dicunt quod cum arguitur sic 'Caesar est homo, ergo Caesar est' fit fallacia
aequivocationis, quia proceditur ab esse habitudinis ad esse temporis vel ad esse ut nunc.
Utrum autem dicant verum vel falsum non magnam habet dubitationem. "j
[Next, doubt is raised concerning the first paralogism which he gives in the text claiming
that it commits the fallacy of "in-some-respect-and-absolutely". For it is arguable that the
fallacy involved in that of consequent rather than that of in-so me-respect and-absolutely,
and for the following reason: Whenever an argument is thus constituted that the reverse
inference holds, but not the one stated in the argument, there is a case of the fallacy of con
sequent. But that is the situation with the paralogism in question. Therefore there is a
fallacy of consequent in it. The major of this reasoning is evident. The minor is also clear

THE CHIMERA'S DIARY

127

since everything that is can occur in opinion, but not vice versa. Exactly the same objection
is raised concerning the second paralogism. For there the fallacy of consequent from destruction of the antecedent is committed when he says 'it is not a man, therefore it is not,
as it follows vice versa, in this way: 'it is not, therefore it is not a man' .... As for the first
point, the response must be that, as Aristotle says in Book II of this work, nothing prevents
joint occurrence of several possible reasons of fallacy in the same utterance. Accordingly,
my position regarding the first utterance is that if 'opinable' is taken in a general sense,
there is the fallacy of consequent. For in one way the opinable is, in a general sense, all
that which can occur in opinion; and in that way 'it is opinable, therefore it is' is a non
sequitur, whereas it does hold vice versa. So when the opinable is taken in the general sense,
there is in this paralogism as well a fallacy of consequent as in-some-respect-andabsolutely. But when the opinable is taken in its proper sense, no fallacy of consequent occurs. For then it does not follow vice versa; thus 'it is, therefore it is in opinion only', as
that is properly speaking opinable which is in opinion only. Accordingly, I submit that
although there is in this paralogism as well a fallacy of consequent as in-some-respect-andabsolutely, more properly speaking there is here a fallacy in-some-respect-and-absolutely
- and I say "more properly" because the opinable is taken in the proper sense in somc
respect. But in the other paralogism, when it is said 'a/the donkey is not a man, therefore
it is not', there is a fallacy of consequent and likewise a fallacy in-some-respect-andabsolutely. But he introduces the paralogism for the sake of its vicious in-some-respectand-absolutely. Some, however, claim that no fallacy of consequent occurs there. For they
submit that the claims 'it is a man' and 'it is not' are compatible, and they say that Caesar
is a man even when he is not. And when they are confronted with the arg).lmcnt 'it is a man,
therefore it is', they say that this argument is not valid, and for the following reason: Every
essential and universal predicate expresses habitudinal being; but whenever the \crb 'is' is
predicated as a second supplement it expresses temporal being (or "being as of now"). Accordingly, they say that when it is argued as follows, 'Caesar is a man, therefore Caesar is'.
a fallacy of equivocation is committed because a move is performed from habitudinal being
to temporal being (or: "being as of now"). Whether they are right or wrong on this score
is scarcely a mattcr of doubt.]

Long ago, after reading Manlius Boethius, I asked whether my infinite

predicates confer some sort of existence on me. This question, whether
infinite terms "posit anything" is now being eagerly discussed. [The
question "utrum nomen infinitum ponat aliquid" is a standard one in
13th-century commentaries on De interpretatione. For Kilwardby and
his contemporaries' treatment of the matter, see Lewry (1978). 45fT For
later treatments, see (e.g.) Petrus de Alvernia, Quaestiol1c\ sUf7cr lihrulll
Peri hermeneias, MS Paris BN lat 16170: 102rB -103r13; Simon de
Faverisham, Quaestiones super libra Perihermeneias pp. 159 -- 160.]
Kilwardby holds that an infinite noun, such as 'not-man', docs signify
substance with quality like any other noun. But the quality is lack-of-aquality (privatio qualitatis) and the substance, which the term does posit
or at least "leave" (derelinquere) , is unspecified being, common to real
physical being and being in notion (ratio), expression (dictio), or (mere)
opinion. [Texts in Lewry (1978) 48ff. A characteristic passage from his

128

STEN EBBESEN

commentary on Anatytica Priora, MS Cambridge Peterhouse 205:

1l6vA:
Si autem tunc quaeratur expositio, dico quod sic exponitur 'non homo': "ens quod non
est homo", sed hoc quod dico "ens" in hac expositione non dicit aliquid ens secundum
naturam, sed accipitur communiter ad ens secundum naturam et ad ens secundum rationem vel intellectum, et sic ponendo ens non ponit aliquid simpliciter.]
[But if an exposition is demanded, I say that 'not-man' is to be expounded as "being
which is not (a) man". But when I say "being" in this exposition, this does not mean
something which is physically, but the exprrssion refers indifferently to what is physically
and what is notionally or conceptually; and when positing being in that way, it does not
posit anything in an absolute sense.]

It follows that, in Kilwardby's view, 'the chimera is (a) not-man' does

not claim that I have some substantial being. Nor does 'althe chimera
does-not-run' contain a claim that I exercise any activity; all that is
posited or left is an unspecified being - in fact just as is the case when
the predicate is 'is opined' (opinatur) [see Lewry (1978), 62J or 'is
opinable' in the broad, improper sense.
I am but the most extreme example of that which is not actually. Less
extreme examples are extinct or not yet actualized species, and individuals of the past. Philosophers often perform an experiment in
thought, saying "suppose there were no men". Their standard example
of an individual of the past is Caesar (who has taken over the role Homer
used to have). Now, if people can admit that there is a sort of being which
is predicable of both me and actually existing things, they should have
little difficulty in admitting that there can be true affirmative 'is'propositions about non-actual natural species or a defunct individual.
Some do indeed claim that there is a way of being which renders this
possible. They call it "habitual being" (esse habituate) and explain that
no stronger being is needed for purposes of verification, and that all
essential predications may be interpreted as not positing any stronger being. Hence 'Caesar is a man' is true even when Caesar is not (because he
is dead), and 'man is an animal' is true whether the species has actual
representatives or not. [For the 13th-century debate about habitual being, see Ebbesen and Pin borg (1970), Braakhuis (1977, 1981), Lewry
(1981a, b), De Libera (1981), Fredborg (1981).J
There would seem to be several interpretations of esse habituate, viz. (I)
= esse habitu (as dinstinct from esse actu) , i.e. an incomplete way of being, being as a tendency to be actualized in a certain way. Thus 'man is
(habitually) an animal' means "the nature of man is such that man can
be actualized and an actual man will be an animal"; (2) = esse
habitudinis or esse consequentiae (distinguished from esse temporis or

THE CHIMERA'S DIARY

129

esse ut nunc), i.e. a way of being which consists in entering into a relationship. Thus 'man is an animal' may mean "there is a relationship

(habitudo), viz. that of species to genus, between man and animal, such
that man entails animal and 'this is a man' entails 'this is an animal".
In other words, 'man is an animal' contains the true claim that there is
a relation of entailment (consequentia) between man and animal. [The
reader should be aware that the chimera follows general 13th-century
usage in talking about entailment as a relation between terms as well as
propositions. However, Roger Bacon (Compendium Studii Theologiae
p. 57) shows awareness that this usage is dangerous:
Adhuc cavillant de esse habitudinis, sed hoc in propositione (pro nomine ms & ed.) habet
locum, et ideo destruetur postea, cum de propositionibus fiet sermo.l
[They further use the trick of talking about habitudinal being. But this belongs in a proposition, and so it will be demolished below when I shall talk about propositions.l

(3) = esse habitum, i.e. to be present in human minds as a habitus, a

structure reflecting an intellectual grasp of something. In this sense a
habitus de homine is knowledge or opinion about man.
Though not invented to help me, habitual being is tailored to my
measure. On int!!rpretations (I) and (2), all I need in order to become as
respectable as the natural species is a quidditative definition, such as 'an
opinable being with such and such differentiae" or a recognition that my
nominal definition is as good as a quidditative one. On interpretation (3)
I just need people not to stop recognizing opinion as a habitus alongside
the true grasp, understanding - and then I defy them to find a waterproof criterion by which to distinguish opinion from understanding! As
a matter of fact, Nicholas of Cornwall seems willing to declare 'the
donkey is a not-man' and 'the chimera is a not-man' equally true when
interpreted as propositions about habitual being [Commentarium in De
interpretatione, MS Oxford Corpus Christi 119: 126vB:
De nullo (hoc MS) non ente praedicatur vere actus entis, ut currere etc. Sed de termino infinito vere praedicatur actus entis. Ergo etc. Minor patet per hoc quod sequitur 'asinus currit, ergo non homo currit'.
[About no not-being can an act of the being be truly predicated, such as running, etc.
But of an infinite term an act of the being can be truly predicated. Therefore etc. The minor
is evident from the fact that this follows: 'a donkey is running, therefore a not-man is
running' .1

The answer to this argument is:

Potest dici secundum sententiam Bocthii quod terminus infinitus nihil ponit, respondendo
huic rationi sic, quod nomen infinitum commune est ad ens et ad non ens secundum
Boethium: eo quod est in quolibet, quod est et non est, ideo neutrurn ponit; de co tamen
vere potest dici utrumque et ipsurn de utroque, surnpto esse habituali.]

130

STEN EBBESEN

[Answering the argument one may follow the view expressed by Boethius, viz . that an
infinite term does not posit anything, and say that an infinite noun extends indifferently
to the being and the not-being according to Boethius. Inasmuch as it is in just anything that
is or is not, it posits neither, but either may be truly said of it, and itself it may be truly
said of either, when the being in question is taken to be habitual.l

But habitual being has its enemies. Even Kilwardby dislikes it. For it
is clear that 'Caesar is a man' can have an esse actuate - interpretation,
not only an esse habituate - interpretation. But then, why should it be
different with 'Caesar is'? In short, 'is' becomes equivocal and 'Caesar
is a man, therefore Caesar is' will be a valid inference if 'is' is taken in
the same sense in the antecedent and the consequent, but invalid if it is
taken in the habitual sense in the antecedent and in the actual sense in
the consequent. Although he himself uses the habitus/actus distinction
in another, but not unrelated, context [De artu scientiarum
433 - 434, p. 150], Kilwardby will have none of such "an equivocal 'is'
[See the end of the extract from his Etenchi commentary, supra .]
Yet, doesn't he make being equivocal when he says that simple real being has a finite meaning and may be infinitized so as to leave unspecified
being, whereas unspecified being has no finite meaning and cannot be infinitized (or, if it can, the result will be "in-no-way-being")? [Kilwardby,
Cammentarium in Anatytica Priara, MS Cambridge Peterhouse 205:
116vB
Notandum autem quod non dicitur proprie infinitari ens nisi secundum (n.s. iecrio incerra)
quod dicit ens simpliciter secundum naturam et veri tat em et tunc privatur simpliciter ens
et derelinquit ens secundum opinionem sive secundum rem, illud tamen non ponitur per
ipsum. Si autem accipiatur ens commune secundum animam sive opinionem et additur
negatio, puto quod sola negatio erit, quia ilia privatio nihil derelinquit, cum privet tam ens
secundum rationem quam ens secundum rem. Sic enim omnino privatur ens, et sonat idem
quod nullo modo ens. Ens autem in tali communitate a<:<:eptum non l'st ali<:uius finitac
significationis, et ideo in tali acceptione non infinitatur . Et ho<: cst quod prius diximus talc
nomen non infinitari cuius non est aliqua finita significatio qualita!is.l
[But it should be noticed that being cannOl properly speaking be infinitized except in so
far as it means being absolutely, physically and truly. And in that case being. is eva<:uated
in the absolute sense, and it leaves being in opinion or in reality, but that is not posited by
it. If. however, we take it as general being in the mind or in opinion and add a negation ,
I think that the result will be a pure negation as this cvacuation Icaves nothing since it
evacuates as well being in notion as being in reality. For in this way being is totally
evacuated. and it means as much as "in no way being". Howevcr. being in this general
sense has no finite signifi<:ation, and so it cannot be infinitizcd when taken in this sense.
And that is what we expres sed earlier by saying that a name whIch is such as to have no
finite signification of quality cannot be infiniti/ed.l

Moreover, Kilwardby's unspecified and (in the broad sense) opinable

being shares with habitual being the property of being truly predicable

THE CHIMERA'S DIARY

131

as well as of things which are as of things which are not - the very property that makes habitual being an abomination to those who (like
Robert Bacon from Oxford) are equipped with a robust sense of reality.
By the way, Bacon is funnily old-fashioned in some respects. Thus it
reeks of the 12th century when he says that a proposition one of whose
terms has no referents is nonsensical (not: false) [Compendium Studii
Theologiae p. 62]. I think some 12th-century philosophers considered
sentences about me nonsensical. [On this aspect of 12th-century logic,
see Ebbesen (1981c).]
Something must happen. Bacon's semantics is bizarre and complex.
With others, senses of 'is' and modes of being tend to proliferate. This
may be to my advantage, but I do not think people will accept it in the
long run. Even Kilwardby's philosophy is not immune to the tendency
to multiply ways of being. And he lacks a good comprehensive theory to
back the claim that 'is' can mean either real or (in the strict sense)
opinable being, according to which environment it has in the
proposition.
7. A.D. 1300

For some thirty years a brand of philosophers whom I think one ought
to call "modists" have dominated the scene. They use the notion of
"analogy" to explain the semantics of 'man' in 'this is a dead man, ergo
this is a man' [cf. Ebbesen, 1979], and of 'is' in 'the chimera is opinable,
ergo it is'. They claim that 'is' is an analogical term. When it stands
alone, it means real being; when with an addition, the sort of being the
addition indicates. Thus the addition of 'opinable' makes 'is' stand for
being in the mind, which is a deficient sort of being ("esse deminutum").
Faced with the objection that an addition cannot change the significate
of 'is', they counter that if 'to be' had been a normal univocal term, this
would have been true. But as a matter of fact, it is "born" with the
liability to influence from additions, and so it is OK that it should change
its meaning according to the circumstances. They also say that my 'is'
and my 'opinable' mix in such a way as to form one predicate, so that
it is not possible to claim that one may take the 'is' out and predicate it
separately on the ground that 'opinable' may be detached and separately
predicated (however that was to happen). There is nothing revolutionary in most of this; but I think it is a step forward that they have
found a way to classify 'is' with other terms, so that they need fewer ad
hoc-rules for that verb. [For the above, see Incerti auctores, Quaestiones
super Sophis:icos Elenchos, CPhD vii: quu. 56, 89; Simon de

132

STEN EBBESEN

Faverisham, Quaestiones veteres super libro Elenchorum, quu. 22 - 23;

Anonymus Pragensis, Quaestiones super Sophisticos Elenchos, MS
Praha MK L.66: 89rA - B, quo 33; Radulphus Brito, Quaestiones super
Sophisticos Elenchos, quo 43, MSS Bruxelles BRoyale 3540-47:
526rA - B & Salamanca BUniv. 2350: 186vB -187rA.]
"Habitual being" is not a term the Parisian modists favour, but for
all that, many think they need something of the sort - essential being,
quidditative being or whatever - to guarantee the permanence of objects
of understanding, and thus the possibility of Aristotelian science.
Boethius of Dacia is an interesting exception (or: was - I believe he is
dead now, but I shall continue to talk about him in the present). He has
a robust sense of reality and even goes so far as to declare that at a time
when no man exists, the proposition 'man is man' is false, whereas 'man
is not man' is true [Text in Ebbesen and Pinborg (1970), 15 - 16]. I have
no doubt that he would also claim that 'the chimera is a chimera' is false
when no chimera exists, i.e., according to his lights, always. And surely
he would hold that 'the chimera is not a chimera' is true in the same circumstances. What horror!
On the other hand, he accepts that things which do not exist in the exterior world may exist in the mind and even be objects of scientific
knowledge. Thus it is possible to understand the phenomenon of an
eclipse or of thunder when there is no eclipse or thunder. For something
to be scibile it is only required that it have causes due to which its being
is necessary or possible. In a sophism called Omnis homo de necessitate
est animal he says:
Rebus corruptis non oportet scientiam de rebus corrumpi, si ita sit quod illae res habeant
causas et principia ex quibus sunt possibiles cognosci. Nam sicut causae effectus inducunt
effectum ilium, sic cognitio illarum causarum facta in anima de necessitate inducit cogni
tionem illius effectus in anima, unde illo modo quo res habet causas et principia cognita,
illo modo habet scientiam. Eclipsi enim non existente aut animali vel planta de eis non est
scientia qua scitur quod ipsa sint (hoc enim falsum est), sed est scientia de eis reman ens in
anima qua scitur ex qui bus causis et principiis et elementis habet fieri unumquodque illorum si fiat, ita quod ex aliis ullumquodque illorum est impossibile fieri; et est scientia de
eis qua scitur quis est modus generationis uniuscuiusque illorum ex suis principiis, ita quod
per alium modum ipsum fieri ex suis principiis est impossibile . [Preliminary text; critical
edition to appear in CPhD in preparation.]
[If things are destroyed, it is not necessary that understanding of the things be destroyed
if it is the case that these things have causes and principles due to which they have the
possibility of being known. For just as the causes of some effect produce that effect, so
the knowledge of those causes, once established in the mind, by necessity produces
knowledge of that effect in the mind, since the way in which understanding attaches to a
thing is the same as the way in which the thing has known causes and principles. If, for
example, an eclipse does not exist - or an animal or a plant - ,there is no understandin'l

THE CHIMERA'S DIARY

133

of them by which it is understood that they are (since this is false), but persisting in the mind
there is an understanding of them by which it is understood due to what causes, principles
and elements each of these things will be produced if it is produced, it being impossible for
any of them to be produced due to any other causes, principles and elements. And there
is an understanding of them by which it is understood which is the manner of generation
of each of these things from its principles, it being impossible for it to be produced from
its principles in any other manner.j

With Boethius the causal relationship has replaced being as the constant
factor that makes Aristotelian science possible. [Cf. Pinborg (1974,
1976).] - But of course some cannot resist the temptation to make this
a sort of being. Giles of Rome [Super Analytica Priara, quoted Pinborg
(1976), 247] speaks of an esse in causis which strongly resembles a
habitual being.
As far as I can see, what Boethius says amounts to saying that if you
can describe an object and state the precise conditions that will or would
actualize it, then that object is a legitimate citizen in the city of
understandable things. Now, if somebody could cogently argue that if a
she-goat were to conceive after an orgy during which she had had intercourse with both a snake and a lion, she would bear a chimera - 'would
the possibility of setting up such a causal explanation of my birth make
me a first-class potential entity that men could have scientific knowledge
of? On Boethius' theory it would. But he will not let me into the company
of understandable things, and so he stresses that I have no causes due to
which my being might be possible. I not only fail to be actual, I cannot
be actualized. In this I differ not only from things which must be actualized, but also from such generabilia as have causes due to which their
being is possible although they will not be actualized. In the sophism I
quoted above, he says:
Cum arguitur secundo "Quod non est, nemo potest scire quid est, ut tragelaphus et
sphinx" - verum est de illo quod non est penitus, ita quod non habet esse in re nec causas
ex qui bus esse suum est possibile; tale autem non est omne quod non est: eclipsis enim, cum
non sit, habet causas tamen in re ex quibus suum esse non solum est possibile sed etiam
necessarium, ut per easdem causas remaneat sua scientia apud animam .... Ad aliud dicendum, cum tu dicas quod omnis scientia est de ente: quia vel oportet iIIud esse ens actu vel
tale quod non habet esse, tamen causas et principia, ex qui bus suum esse est necessarium
vel possibile, habet; sicut apparet, cum eclipsis non sit, suum esse tamen necessario eveniet
ex suis causis (dico "possibile" propter generabilia: non enim omne genera bile generabitur,
habet tamen causas ex quibus ipsum esse est possibile).
[As for the second argument, that "Concerning that which is not, nobody can understand what it is (e.g., a goat-stag or a sphinx)", this is true of that which is not at all, so
that it does not have being in reality nor causes due to which its being is possible. But not
everything that is not is of this sort. An eclipse, for instance, even when it is not, does have
causes in reality due to which its being is not just possible but even necessary, so that in

134

STEN EBBESEN

virtue of these very causes an understanding of it may persist in the mind .... As for your
further claim that all understanding is about something that is, the response must be that
this will have to be either something actually being or such that without having being, it
does have causes and principles due to which its being is necessary or possible. A clear instance of the latter is an eclipse: even when it is not, its being will necessarily come about
due to its proper causes. (I say "possible" because of generable things, as not every
generable thing will be generated, yet does have causes due to which it is possible for it to
be).)

I think it requires some strong assumptions about how the machinery

of nature works to claim that one can distinguish between impossibilia
like me and possible, yet never actualized, generabilia. But perhaps
Boethius believes he knows that much about nature.
This faith in human insight into nature Boethius shares with lesser
philosophers, such as one Petrus (de Hibernia or de Alvernia - I really
cannot tell one from another in the host of Peters) in whose Quaestiones
super Analytica Posteriora I once read a discussion of the question
"whether it is possible to know what a goat-stag or a chimera is" [MS
Firenze BLaur. St. Croce Plut. 12 sin., 3: 36vB]. Echoing Plato's Sophist
(via Avicenna), Peter says, "About that which absolutely is not, it is not
possible to state anything" (de eo quod non est penitus non contingit aliquid enuntiare). But, he holds, it is possible to state something about me,
and so in some sense I must exist extramentally, and in some sense be
grasped by the human intellect. But thts happens in a tricky way. The
man who has a concept (intellectus) of me has, in reality, a concept of
the parts that I am supposed to consist of; and in so far as these parts
exist, there is something extramental to state something about. However,
the word 'chimera' does not primarily signify these parts but a compound of them, and that compound neither is nor can be; consequently
men cannot understand what it is but only what the word signifies
("nomen chimaerae significat coniunctum ex partibus quod non est nec
esse potest, et quia sic est, ideo non contingit intellegere quid est"). It is
different with donkeys, he says. Even if there were no donkeys, men
might understand what a donkey is since "the donkey is a nature which
by itself is not prohibited to be" ("Quamvis enim asinus non esset, tamen
asinus est aliqua natura quae de se non est prohibita esse, et ideo, si
adhuc non esset ... asinus, non esset simile de asino et chimaera").
Peter's "natures" have much in common with Asclepius' "ideas", and
I can only repeat myoid question. How does he know that I can have
no "nature"? Besides, I think Peter's semantics (typical of the University of Paris c. 1275) is unnecessarily clumsy. Why doesn't he use the notion of suppositio (reference)?

THE CHIMERA'S DIARY

135

8. AD 1340

It is simply impossible to keep abreast of the enormous output of learned

literature on me and other epistemological difficulties. But I think the
most important contributions to the discussion during the last three
decennia have been made by William of Ockham in England and John
Buridan in Paris. These two monstrously sharp intellects have combined
some of the most salient features of the views of Petrus de Alvernia and
Boethius de Dacia while kicking out some of the main notions of
modism, such as "natures", and reintroducing the notion of suppositio.
The result is terrifying.
At first sight, it may look promising that Ockham assimilates me to
regular universals and thinks that both they and I may be either actual
things in the sense of being qualities of the mind or fictitious entities
whose only being is being objects of thought (esse obiective) [Expositio
in librum Perihermeneias pp. 363 - 370]. But this is the only glimpse of
kindness I find in the writings of either man. These people distinguish
much clearer between word, concept and extramental thing, and between
signification and supposition than did their predecessors. And then the
onslaught sets in!
They agree that the word 'chimera' does signify something, and that
it is something which is also signifiable by means of normal general terms
- in short, it signifies my constituent parts. Incidentally, they are not
well versed in mythology. They believe I am a compound of goat and ox
or ox, virgin and dragon. But in signifying thus, the word is subordinated to a complex concept, similar in structure, Buridan says, to mancapable-of-neighing or man-horse . This may be expressed in a nominal
definition, such as 'a chimera is a compound of an ox and a goat', but
even this proposition is true only when interpreted in the sense' 'the word
'chimera' means the same as 'compound of ox and goat' .. or "the concept to which the word 'chimera' is subordinated is a complex concept
with the concept of goat and the concept of ox for its constituents" or
"if anything is a chimera, it is a compound of an ox and a goat" . If the
nominal definition is taken at face value, it is a false proposition, for no
affirmative categorical proposition can be true unless it is the case that
there is, was, will be, could have been or can be such a thing as the subject
signifies. In other words, any proposition of the form 'the/a chimera is
(was etc.) X' implies the assertion that 'this is a chimera' is or was etc.
true at some time. Buridan and Ockham agree that if 'chimera' is taken
in its natural sense, i.e. as purporting to refer to extramental entities,

136

STEN EBBESEN

there never is, was or will be a time when 'this is a chimera' is actually
or possibly true. It is simply impossible that I should ever be.
I think it is due to a slip of attention that Ockham in one place [Expositio super libros Elenchorum, p. 264] accepts the traditional assignment of a fallacy secundum quid et simpliciter in the case of 'the chimera
is opinable, therefore it is' on the ground that 'if something is opinable,
it is' is false. Buridan is adamant on this point: 'the chimera is opinable'
is simply false. Here follows an extract from his quaestio 'Utrum
chimaera sit intellegibilis' [Quaestiones super Sophisticos Elenchos, MS
Krakow Blag. 736: 70r - v, minor scribal errors tacitly corrected.]
Breviter ad istam quaestionem pono primo ill am conclusionem: haec propositio est falsa
'chimaera est opinabilis' vel 'chimaera est intellegibilis' vel 'chimaera est significabilis'. Ad
quam probandam supponatur primo quod ad veritatem propositionis affirmativae requiritur subiectum et praedicatum supponere pro eodem. Secundo supponatur quod omnis
terminus supponens supponit vel pro eo quod est vel pro eo quod potest esse vel pro eo
quod fuit vel pro eo quod erit. Tertio supponatur quod sicut impossibile est chimaeram
esse, similiter impossibile est chimaeram posse esse, similiter impossibile est chimaeram
posse fuisse, similiter impossibile est chimaeram posse fieri. Hiis suppositis probatur haec
conclusio. Si haec propositio est vera 'chimaera est opinabilis', cum sit affirmativa, per
prim am suppositionem eius subiectum et praedicatum supponerent pro eodem. Sed hoc est
falsum. Falsitas probatur: subiectus terminus pro nullo supponit, ergo falsum est quod terminus subiectus et terminus praedicatus pro eodem supponunt. Consequentia nota est de
se. Antecedens probatur: omnis terminus supponens supponit pro eo quod est vel potest
esse vel fuit vel erit, sed hic terminus 'chimaera' non supponit pro eo quod est nec pro eo
quod potest esse nec pro eo quod fuit nec pro eo quod erit. Igitur hic terminus 'chimaera'
non supponit pro aliquo. Maior est supposito. Sed minor patet ex eo quod per tertiam suppositionem impossibile est chimaeram esse, similiter impossibile est chimaeram posse esse,
sim!1iter impossibile est chimaeram fuisse, similiter impossibile est chimaram fieri. Consimilem conclusionem poneremus de vacuo: supposito quod impossibile sit vacuum fuisse
vel fieri, posse esse vel esse, haec est falsa 'vacuum est opinabile', similiter 'vacuum est intellegibile' et huiusmodi.
Secunda conclusio: Non obstante quod haec sit falsa 'chimaera est significabilis',
'vacuum est significabile', tamen hic terminus 'chimaera' significat aliquid, similiter hic
terminus 'vacuum' significat aliquid. Patet hoc ex eo quod hic terminus 'chimaera'
significat caudam draconis, similiter vent rem virginis et collum ... is et caput bovis;
similiter hic terminus 'vacuum' significat corpus et significat locum. Modo quodlibet
istorum est aliquid.
Circa quod notandum quod hic terminus 'vacuum' et hic terminus 'chimaera' vocalis vel
scriptus subordinantur termino mentali in significando non incomplexo sed complexo,
propter quod talis terminus mentalis veri us debet did oratio, licet extenso nomine posset
dici terminus, non tamen complexus complexione distante quae fit mediante hoc verbo
'est', sed complexione indistante, ut si fieret talis complexio 'homo asinus' vel 'homo
equus' ...... et quia tales orationes non supponunt pro aliquo, videtur quod huiusmodi
termini 'chimaera' 'vacuum' etiam pro nullo supponunt, postquam tali bus orationibus
subordinantur in significando; nec etiam oportet terminum supponere pro omni illo quod
significat, nam hic terminus 'album' ,licet significet albedinem. sicut patet per definitioncm

THE CHIMERA'S DIARY

137

eius exprimentem quid nomimis 'album est res habens albedinem', non supponit pro
albedine sed pro re subiecta albedini. Sic in proposito, licet hic terminus 'chimaera'
significet ventrem virginis, non tamen supponit pro eo. Similiter hic terminus 'vacuum'
licet significet corpus et similiter locum, pro nullo tamen eorum supponit.
[To deal briefly with this question, I first submit this conclusion: the proposition 'a/the
chimera is opinable' or 'a/the chimera is thinkable' or 'a/the chimera is signifiable' is false.
To prove this, let it be presupposed that for the truth of an affirmative proposition it is
required that the subject and the predicate suppone for the same. Secondly,let it be presupposed that every supponing term suppones either for that which is or for that which can
be or for that which has been or for that which will be. Thirdly, let it be presupposed that
just as it is impossible that a chimera is, so it is impossible that a chimera can be, and impossible that a chimera can have been, and impossible that a chimera can come to be. With
these presuppositions, this conclusion is provable as follows. If the proposition 'a/the
chimera is opinable' is true, then, since it is affirmative, according to the first presupposition its subject and its predicate would suppone for the same. But that is false. The falsity
of this is provable as follows. The subject term suppones for nothing, therefore it is false
that the subject term and the predicate term suppone for the same. The consequentia is selfevident. The antecedent is provable as follows. Every supponing term suppones for that
which is or can be or has been or will be; but the term 'chimera' does not suppone for that
which is, nor for that which can be, nor for that which has been, nor for that which will
be; therefore the term 'chimera' does not suppone for anything. The major is one of the
presuppositions, whereas the minor is obvious from the fact that according to the third
presupposition it is impossible that a chimera is, and similarly it is impossible that a
chimera can be, and impossible that a chimera has been, and i111possible that a chimera will
come to be. We could posit a similar conclusion with respect to the vacuum. On the presupposition that it is impossible that a vacuum has been or will come to be or can be or is, the
proposition 'a vacuum is opinable' is false, and likewise 'a vacuum is thinkable' and the
like.
Conclusion 2. Notwithstanding the fact that 'a/the chimera is signifiable' is false - and
likewise 'a vacuum is signifiable' - yet the term 'chimera' does signify something, and
similarly the term 'vacuum' does signify something. This may be seen from the fact that
the term 'chimera' signifies the tail of a dragon, and likewise the belly of a young woman
and the neck of a ... [illegible word) and the head of an ox. Similarly the term 'vacuum'
signifies body and signifies space. Now each of these things is something.
In this connection it should be noted that the term 'vacuum' and the term 'chimera' in
the sense of vocal or written terms are subordinated in signifying not to an incomplex but
to a complex mental term - for this reason such a mental term ought more truly to be
called a phrase, though through an extension of the name it might be called a term -, but
not one that is complex in virtue of the distant composition which is brought about by
means of the verb 'is', but in virtue of an indistant composition, such as woule! occur in
a combination of the type 'man ass' or 'man horse' .... And as such phrases do not suppone for anything, it appears that terms of the type represented by 'chimera' and 'vacuum'
also suppone for nothing, since in signifying they are subordinated to such phrases. Nor
is it required that a term suppone for all that it signifies. The term '(the) white', for instance, signifies whiteness, as is obvious from its nominal definition ("the white is a thing
having whiteness"), and yet it does not suppone for whiteness but for the thing that is the
subject of whiteness. So too in our case: although the term 'chimera' signifies the belly of
a young woman, it does not suppone for it. Similarly, although the term 'vacuum' signifies
body and space too, it does not suppone for either of these.)

138

STEN EBBESEN

It is no use, says Buridan, to defend 'the chimera is opinable' on the

ground that 'opinable' ampliates the subject term so as to make it stand
for present, past, future and possible entities alike, even such possible entities as never were nor will be. [The idea that 'opinable' ampliates is as
old as the 12th century. See Fallaciae Parvipontanae, LM 1: 564 - 566;
Tractatus Anagnini III, LM 2.2: 268; Tractatus de univocatione
Monacensis. LM 2.2: 338: Tractatus de DroDrietatihus sprmnnum. T.M
2.2: 729] Buridan holds that the ampliative force of 'opinabilis' does not
extend to impossible entities. Here are his own words [Summulae 7.4.2,
MS Krakow BJag. 662:83r]:
Secundus modus (sc. fallaciae secundum quid et simpliciter) est si additio sit pure
ampliativa, ut 'opinabile', 'intellegibile', 'possibile' etc. et multa alia. In hi is enim dictum
secundum quid praedicatur universaliter de dicto simpliciter et non econverso. Ideo est
bona consequentia affirmativa de dicto simpliciter ad dictum secundum quid et econverso
bona negativa, et non est bona affirmativa ex dicto secundum quid ad dictum simpliciter
nec econverso bona negativa, sed est fallacia secundum quid et simpliciter. Unde fiunt tales
paralogismi: 'Antichristus est possibilis, ergo Antichristus est' vel 'mons au reus est intellegibilis, ergo mons aureus est', vel 'vacuum est ens opinabile, ergo vacuum est' vel
econverso 'vacuum non est, ergo vacuum non est ens opinabile': omne enim ens est ens
opinabile et intellegibile, sed non omne ens intellegibile vel opinabile est ens. Hoc enim verbum 'est' vel hoc 'non ens' respectu huius verbi 'est' sine ulteriori additione non dicuntur
nisi de praesentibus, sed istae additiones 'opinabile', 'opinatum', 'intellegibile', 'intellectum', 'sci bile' , 'scitum' et huiusmodi faciunt ea stare non solum pro praesentibus sed etiam
pro praeteritis et futuris et hiis etiam quae numquam erunt vel fuerunt sed possunt esse.
Aliqui etiam ponunt in hoc modo ilium paralogismum 'chimaera est ens opinabile, ergo
chimaera est ens'vel 'chimaera est opinabilis, ergo chimaera est', sed non curo de hoc, quia
antecedens ita est impossibile sicut consequens, prout alias dictum est.
[The second mood of the fallacy of in-some-respect-and-absolutely occurs if an addition
is purely ampliative, like 'opinable', 'thinkable', 'possible' etc. and many others. For in
such cases that which is said in some respect is universally predicable of the same said absolutely, but not vice versa. Therefore there is a good affirmative consequentia from that
which is said absolutely to the same said in some respect, and a good negative one the other
way round; and there is not a good affirmative one from that which is said in some respect
to the same said absolutely, nor a good negative one the other way round, but there is a
fallacy of in-some-respect-and-absolutely. This is the basis of such paralogisms as 'Antichrist is possible, therefore Antichrist is' or 'a golden mountain is thinkable, therefore
there is a golden mountain' or 'a vacuum is (an) opinable being, therefore there is a
vacuum', or, the other way round, 'there is not a vacuum, therefore a vacuum is not (an)
opinable being'. For every being is (an) opinable and thinkable being, but not every
thinkable or opinable being is (a) being. For the verb 'is' or 'not being' with respect to the
verb 'is' are not said of any but present things when not accompanied by any further addition, but the additions 'opinable', 'opined', 'thinkable', 'thought', 'understandable',
'understood', and the like make them stand not only for present things but also for past
and future ones, and even for such as never will be or have been but can be.
Some also put under this mood the paralogism 'a/the chimera is (an) opinable being,
therefore a/the chimera is (a) being' or 'a chimera is opinable, therefore a chimera is'. But
I do not care about that since the antecedent is as impossible as the consequent, as has been
said elsewhere.]

THE CHIMERA'S DIARY

139

The "some people" (aliqUl), mildly censured toward the end of this
text, may be Ockham. But, in general, Ockham and Buridan agree, and
all the other affirmative categorical propositions which used to be accepted as true receive the same treatment as 'the chimera is opinable'
does at Buridan's hands. 'The chimera is capable of being signified', the
chimera is not-being', even 'the chimera is a chimera'. Ockham has the
cheek to remark [Summa Logicae, p. 288]: "No proposition in which
something is predicated of the name 'chimera' taken significatively can
be truer than the one in which the name 'chimera' is predicated of itself.
But it is consistent with this that neither one nor the other should be
true. "
In Buridan's and Ockham's philosophy I have ceased to be the menace
to the theory of science that I was as long as people thought the objects
of scientific understanding were things of which something may be
predicated, and as long as they did not clearly distinguish between expressions, concepts and extramental referents. With the new semantics
and theory of scientific knowledge, in which propositions are the object
of understanding, the old difficulties disappear. If it is possible to know
that 'this is an animal composed of a goat, a snake and a lion' can never
be true, and if it is clear that the fact that there is a word 'chimera'
equivalent to 'animal composed of. .. ' does not mean that the proposition 'the chimera is nameable' is true or that 'a chimera is an animal composed of ... ' or 'the chimera is definable' are true, then there is no
dangerous similarity between the word 'chimera' and words associated
with concepts to which actual or possible individuals correspond.
So this is what is left of me: a word and a corresponding complex concept. But not a trace of old-fashioned being. The price Buridan and
Ockham must pay for their achievement is to recognize that 'is' is
equivocal. Buridan, at least, is not afraid of that consequence. Indeed,
he distinguishes between 'is' as (1) an atemporal copula, (2) a temporal
copula, (3) an abbreviation of 'is being', and (4) a catachrestic way of
saying 'means the same as'.
The only comfort I can find in this situation is that Ockham and
Buridan may not be very interested in me for my own sake (they do not
even bother to give me the same constituents parts on the different occasions on which they state my nominal definition). They assume for the
sake of argument that there cannot be such a compound animal as the
chimera. But it does not matter to them as logicians whether this assumption is true, as long as people will grant them that some name may be
subordinated to a complex concept that can never be associated with extramental referents.

140

STEN EBBESEN

Buridan is still active teaching. I would like to discuss with him

whether he sees any difference between such impossibilities as me and a
man capable of neighing, the square circle, and a man who is not a man.
But after the proposition which states my self-identity has been declared
false I feel too tired to take up the discussion. They have nearly done me
in.
[Here ends the chimera's diary. Let me add a few references to Ockham
and Buridan. Ockham, Summa Logicae, pp. 88 - 89, 284 - 288, 507,
568; In librum Perihermeneias 434; In librum primum Sen tentiarum ,
Opera Theologica 3: 304, 4: 547. Buridan, Sophismata, pp. 21 - 47 (cf.
Roberts (1960; Quaestiones super Analytica Posteriora 1.9 (unpublished edition kindly put at my disposal by H. Hubien); Summulae 4.1.2,
4.1.3,4.1.4, pp. 181-185 Reina; on the senses of 'is', see Summulae
1.3.2 in Pinborg (1976b) 87 and the sources cited in Note 19 of Ebbesen
(1984). For the chimera's fate after the end of the diary, see Ashworth
(1977).]
ABBREVIATIONS
I In references to commentaries on Porphyry and Aristotle the following abbreviations
have been used:
intr. = Commentary on Porphyry's Introductio (Isagoge).
Cat. = Commentary on Aristotle's Categories.
Int. = Commentary on Aristotle's De interpretatione (Peri hermeneias).
APr. = Commentary on Aristotle's Prior Analytics.
A Po. = Commentary on Aristotle's Posterior Anaiytics.
Top. = Commentary on Aristotle's Topics.
SE = Commentary on Aristotle's Sophistici Elenchi.
Ph. = Commentary on Aristotle's Physics.
Metaph. = Commentary on Aristotle's Metaphysics.
2 For a number of works reference has been made to volume and page or column
(sometimes also line) in one of the following series. These works do not appear in the
bibliography below.
CAG = Commentaria in Aristotelem Graeca edita consilio et auctoritate Academiae
Regiae Borussicae, Vols. 1-23, Berlin, 1882-1907.
CPhD = Corpus Philosophorum Danicorum Medii Aevi, consilio et auspiciis Societatis
Linguae et Litterarum Danicarum editum, Copenhagen, 1955LM = L. M. de Rijk, Logica Modernorum, Vols. 1 - 2 (= Wijsgerige Teksten en
Studies 6, 16), Van Gorcum, Assen, 1962, 1967.
PL = Patrologiae cursus completus ... series latina ... accurante J. P. Migne, Paris,
1844ff.

141

THE CHIMERA'S DIARY

BIBLIOGRAPHY
Abaelardus, Petrus: 1970, Dialectica, ed. by L. M. de Rijk, 2nd ed. (Wijsgerige Teksten
en Studies I), Van Gorcum, Assen.
Albertus Magnus: 1890, 'Expositio Sophisticorum Elenchorum', in A. Borgnet (ed.),
Opera, Vol. 2, Paris.
Anonymus Moraux: 1979, Commentarium in Porphyrii Introductionem (my title), in Paul
Moraux, 'Ein unedicrter Kurzkommentar zu Porphyrios' Isagoge', Zeitschrijt fur
Papyrologie und Epigraphik 35, 55 - 98.
Anonymus Tanin: 1978, Int. = Anonymous Commentary on Aristotle's De Interpretatione (Codex Parisinus Graecus 2064), ed. by L. Tanin (Beitrage zur Klassischen
Philologie, Heft 95), Anton Hain, Meisenheim am Glan.
Aristoteles: 1831, Opera ex recensione Immanuelis Bekkeri, Vois. 1-2, Berlin.
Ashworth, E. 1.: 1977, 'Chimeras and Imaginary Objects: A Study in the Post-Medieval
Theory of Signification', Vivarium IS, 57 - 79.
Bacon, Roger: 191 I, Compendium Studii Theologiae, ed. by H. Rashdall, Aberdeen.
= Commentarii in librum
Boethius, Anicius Manlius Severinus: 1877, Int. ed.
.Aristotelis ITEPI EPMHNEIAE, pars prior, ed. by C. Meiser, Teubner, Leipzig.
Boethius, Anicius Manlius Severin us: 1880, Int. ed.2' = same edition, pars posterior,
Teubner, Leipzig.
Braakhuis, H. A. G.: 1977, 'The Views of William of Sherwood on Some Semantical
Topics and Their Relation to Those of Roger Bacon', Vivarium IS, 1 I 1- 142.
Braakhuis, H. A. G.: 1979, De J3de eeuwse tractaten over syncategorematische termen,
2 vols. (Diss., Leiden), Meppel.
Braakhuis, H. A. G.: 1981, 'English Tracts on Syncategorematic Terms from Robert
Bacon to Walter Burley', in Braakhuis et al. (1981), pp. 131 -165.
Braakhuis, H. A. G., C. H. Kneepkens, and L. M. De Rijk (eds.): 1981, English Logic
and Semantics from the End of the Twelfth Century to the Time of Ockham and
Burleigh, Artistarium, Supplementa I, Ingenium Publishers, Nijmegen.
Buridanus, Johannes: 1977, Sophismata, ed. by T. K. Scott (= Grdmmatica Speculativa
I), Frommann-Holzboog, Stuttgart - Bad-Cannstat.
Buridanus, Johannes: 1957, Summulae, tractatus 4, in M. E. Reina (ed.), 'Giovanni
Buridano, "Tractatus de Suppositionibus" " Rivista critica di storia della filosofla 12,
175 - 208 and 323 - 352.
Cassin, B.: 1980, Si Parmenide, Cahiers de philologie pub lies par Ie Centre de Recherche
Philologique de l'Universite de Lille III, Vol. 4, Editions de la Maison des Sciences de
I'Homme, Presses Universitaires de Lille.
De Libera. A.: 1981, 'Roger Bacon et Ie probleme de I'appellatio univoca', in Braakhuis
et al. (1981), pp. 193 - 234.
De Rijk, L. M.: 1981a, 'Die Wirkung der neuplatonischen Semantik auf das mittelalterliche Denken', Miscellanea Mediaevalia 13,19-35.
De Rijk, L. M.: 1981b, 'Abailard's Semantic Views in the Light of Later Developments'.
in Braakhuis et al. (1981), pp. I - 58.
Ebbesen, S.: 1979, 'The Dead Man is Alive', Synthese 40, 43 - 70.
Ebbesen, S.: 1981a, Commentators and Commentaries on Aristotle's Sophistici Elenchi,
3 vols. (Corpus Latinum Commentariorum in Aristotelem Graecorum VII, I - 3), E. J.
Brill, Leiden.
Ebbesen, S.: 1981b, 'Albert (the Great?)'s Companion to the Organon', Miscellanea
Mediaevalia 14, 89 - \03.

142

STEN EBBESEN

Ebbesen, S.: 1981c, 'The Present King of France Wears Hypothetical Shoes with
Categorical Laces. Twelfth-century Writers on Well-formedness', Medioevo 7,
91-113.
Ebbesen, S.: 1984, 'Proof a.nd its Limits According to Buridan, Summulae 8', in Z. Kaluza
and P. Vignaux (eds.), Preuve et raisons al'Universite de Paris. Logique, ontologie et
thiologie au XIV' siecle, Vrin, Paris, pp. 97 - 110.
Ebbesen, S. and J. Pinborg: 1970, 'Studies in the Logical Writings Attributed to Boethius
de Dacia', Cahiers de I'lnstitut du Moyen-Age grec et latin, Universite de Copenhague
3.
Ps.-Elias,lntr. = Pseudo-Elias (Pseudo-David), Lectures on Porphyry's lsagoge, ed. by
L. G. Westerink, North-Holland, Amsterdam, 1967.
Fredborg, K. M.: 1981, 'Roger Bacon on "Impositio vocis ad significandum" " in
Braakhuis et al. (1981), pp. 167 -191.
Harlfinger, D.: 1975, 'Edizione critica del testa del "De ia-eis" di Aristotele', in W. Leszl
(ed.), II 'De ideis' di Aristotele e la teoria Platonica delle idee (Accademia Toscana di
scienze e lettere 'La Colombaria', Studi 40, pp. 17fO, Olschki, Firenze.
Kilwardby, Robert: 1976, De Ortu Scientiarum, ed. by A. G. Judy (Auctores Britannici
Medii Aevi 4). The British Academy/The Pontifical Institute of Mediaeval Studies.
Lewry, 0.: 1978, Robert Kilwardby's Writings on the Logica Vetus (Diss., Oxford).
Lewry, 0.: 1981a, 'Robert Kilwardby on Meaning', Miscellanea Mediaevalia 13,
376- 384.
Lewry, 0.: 1981 b, 'The Oxford Condemnations of 1277 in Logic and Grammar', in
Braakhuis et al. (1981), pp. 235 - 278.
Lucretius: 1947, De rerum natura, ed. by C. Bailey, 3 vols., Oxford.
Ockham, Guillelmus de: 1979, Expositio super libros Elenchorum, ed. by F. del Punta
(Opera Philosophica et Theologica, Opera Philosophica 3), St. Bonaventure University,
St. Bonaventure, N. Y.
Ockham, Guillelmus de: 1978, Expositio in librum Perihermenias, ed. by A. Gambatese
and S. Brown (Opera Philosophica et Theologica, Opera Philosophica 2), St. Bonaventure University, St. Bonaventure, N.Y.
Ockham, Guillelmus de: 1977 -1979, Scriptum in librum primum Sen ten tiarum, Ordinatio, ed. by G. I. Etzkorn and F. E. Kelley (Opera Philosophica et Theologica,
Opera Theologica 3-4), St. Bonaventure University, St. Bonaventure, N.Y.
Ockham, Guillelmus de: 1974, Summa Logicae, ed. by Ph. Boehner, G. Gal, and S. Brown
(Opera Philosophica et Theoiogica, Opera Philosophica I), St. Bonaventure, N. Y.
Parmenides: 1966 12 , Fragmenta, in H. Diels and W. Kranz (eds.), Die Fragmente der Vorsokratiker, Vol. I, Weidmann, Dublin/Zurich.
Pinborg, Jan: 1974, 'Zur Philosophie des Boethius de Dacia. Ein Oberblick', Studia
Mediewistyczne IS, 165-185.
Pinborg, Jan: 1976a, 'Diskussionen urn die Wissenschaftstheorie an der Artistenfakultiit',
Miscellanea Mediaevalia 10, 240 - 268.
Pinborg, Jan: 1976b, 'The Summulae, Tractatus I De Introductionibus', in J. Pin borg
(ed.), The Logic of John Buridan (Opuscula Graecolatina 9), Museum Tusculanum,
Copenhagen pp. 71-90.
Rijk, L. M. de: see De Rijk.
Robert Kilwardby: see Kilwardby.
Roberts, L. N.: 1960, 'A Chimera is a Chimera: A Medieval Tautology', Journal of the
History of Ideas 21, 273 - 278.
Roger Bacon: see Bacon.

THE CHIMERA'S DIARY

143

Sextus Empiricus: 1935, Adversus Mathematicos Vll- Vlll, in Sextus Empiricus, with an
English translation by R. G. Bury, Vol. 2, Loeb Classical Library, London/Cambridge
Mass.
Simon de Faverisham: 1957, Quaestiones super libroPerihermeneias, in Opera omnia, Vol.
1: Opera logica, tomus prior, ed. by P. Mazzarella, Cedam, Padova.
Simon de Faverisham: 1984, Quaestiones super libro Elenchorum, ed. by S. Ebbesen et al.
(Studies and Texts 60), Pontifical Institute of Mediaeval Studies: Toronto.

Institute for Greek and Latin

Medieval Philology
University of Copenhagen
Njalsgade 92, trappe 2,2
DK-2300 Copenhagen S, Denmark

KLAUS JACOBI

PETER ABELARD'S INVESTIGATIONS INTO THE MEANING

AND FUNCTIONS OF THE SPEECH SIGN 'EST'

Although Peter Abelard was the most distinguished teacher of logic of

his time, a logic understood to be the science of argumentative discourse,
he was not destined to found a new philosophical tradition. The
historical situation offers at least a partial explanation - the pace of
philosophical and theological research was so brisk in the twelfth century
that many of the established schools enjoyed life spans of at most two
or three generations of teachers. The restlessness of the times is embodied
to a special degree in Abelard. 1 His writings include commentaries, in
many cases several to a work, on the logical works of Aristotle and Porphyry then available, handed down in the form of Boethius' translations,
and on Boethius' own logical works. Abelard has to take a number of
positions into consideration here: several commentaries on Aristotle by
ancient scholars, by Boethius, and by Abelard's own predecessors and
teachers, and furtHermore the grammatical theories of Priscian and
those deriving from Abelard's contemporaries. He discovers with
distinctive acumen that the tradition he is examining is disunited and full
of tensions on basic questions. It is in the analysis and discussion of these
tensions that he finds the field of his own philosophical research. He expects to reach solutions by intensifying the controversies, not by seeking
harmony. Thus he traces argument and counter-argument in great
thoroughness of detail and from a dizzying succession of points of view,
abandoning theses and offering countertheses. What his students could
learn from him was not so much a particular theory as his method of formulating and discussing problems.
The situation is much the same for us. If we turn to Abelard in our inquiry into the logic and semantics of the speech sign 'est', we must
discover anew the questions which concerned him. In the first part of this
paper, I will sketch some of the discussions conducted by Abelard in
order to make clear in what contexts he found himself confronting questions on the variations of meaning, function, or use of the expression
'est'. In the second part, I will group various theses which Abelard deal~
with appropriately. It is my intention to plot out the full range of the
theories discussed and to mark points of conflict. In the third and final
part, I will make some cautious comments on the deeper current of unity
145
S. Knuuttila and J. Hintikka (eds.), The Logic of Being, 145-180.
1986 by D. Reidel Publishing Compan v

146

KLAUS JACOBI

to be observed in Abelard's reflections, a current perhaps more easily

discernible to the modern eye then it was to Abelard himself.
1. DISCUSSIONS

1. The Distinction between 'per se significative' and 'co-significative'

Words and the Question to Which of These Two Classes the Copulative
'est' Should Be Assigned
In his commentary on the second chapter of Peri Hermeneias, Abelard
discusses the fact that Aristotle treated only nouns and verbs, disregarding prepositions and conjunctions. 2 Nouns and verbs are 'per se
significative', meaningful in and of themselves: each individual noun or
verb has a meaning which can be explicated. But prepositions and conjunctions are also elements of language which convey something - in
contradistinction to single letters or syllables. The difficulty is to arrive
at a precise determination of what they convey. 3
Some authors want to attribute meaning to prepositions and conjunctions when they are uttered alone, abstracted from all context. Yet compared with the way in which nouns and verbs have meaning, there is
something uncertain, even vague, about them. Only when brought
together with other words do they develop full, definitive meaning. 4
Abelard raises the objection that this view is incapable of providing an
adequate criterion for distinguishing nouns and verbs from prepositions
and conjunctions. For it can be maintained that the meaning of nouns
and verbs is also less definitive when they are uttered in the abstract, that
in discourse the understanding (intel/ectus) of each word is made more
precise by the other words in context. S Abelard tends toward the view of
other authors, according to whom prepositions and conjunctions uttered
in the abstract convey absolutely nothing. They function as speech signs
only when used together .with other words, words which are meaningful
in and of themselves. 6
What do we conceive of when we hear such 'co-significative' words
used together with significative words in a particular context? Abelard
finds that it is not difficult to paraphrase what most prepositions indicate. They deal with relations of direction, proximity, possession, or
origin, for example. In the case of some conjunctions, however, 'if' and
'and', for example,
it is not easy to specify what they signify or what image of which particular thing our
understanding of them is based upon. For if they have meaning, they can be understood,

PETER ABELARD ON THE SPEECH SIGN 'EST'

147

and to this end, there must be a thing or the image of a thing for our conception (intellectus)
to be based upon. But what sort of a thing or construction of the mind do we conceive of
when we use 'if or 'and'? Does it belong to the category of substance or to that of quality
or to some other category? It must belong to one of them if we are to be able to have a
proper conception of it. 7

It is likewise difficult to specify what is signified by the adverb of nega-

tion when it is placed before a connective. 8

Abelard inserts here a discussion of the copula, stating that a theory
of affirmative and negative statements expressing predication is an indispensable prerequisite for any theory of the joining together of such
statements. 9 Do 'est' and 'non est' have significative function when they
serve as copulas?
In the first part of the discussion Abelard seems to lean toward a
positive answer to this question. 'An animal endowed with reason' and
'an animal which is not endowed with reason' are complex expressions
which we understand as denoting something. If we join the two terms,
each of which signifies something existent, with the verb 'est' - 'animal
rationale est animal irrationale' - the result is a false statement. If we
add a negating word, the result is a new statement which is true. Now if
we can use the copula to join two denoting terms such that the result is
a false understanding (intellectus), and if we can make a true understanding out of a false one by employing a negating word, then it seems
that the copula and the negating word must have per se significative
function. 10
Nonetheless, Abelard makes it clear in the second part of the discussion that he does not share this view. With vigor he proposes a thesis of
his own: 'est' used as copula does not have significative function it has
affirmative function. 'Non est' has negating function. Although 'est'
and 'non est'
do not constitute conceptions, they cause a joining or separating in the mind of the things
of which the mind does have conceptions. Nonetheless, although they cause, they do not
signify joining or separating, because they do not themselves constitute conceptions; rather
they simply cause this joining or separating of conceptions. And so there are three acts involved in the understanding of a statement, namely the conceptions of the parts and the
joining or separating of them. It is not inconsistent for that act which is not in itself a conception to be a part of the understanding of the full statement. 11

According to the thesis thus proposed, the copula 'est' is merely a connecting sign in affirmative statements or a sign of separation when used
with a negating word. The hearer is directed to join the conception of the
first term with that of the second or to separate the.two conceptions from
each other. The validity of the thesis is not demonstrated here, neither

148

KLAUS JACOBI

by direct proof nor indirectly by refuting the counterthesis. No attack is

made on the view first discussed. Abelard merely shows its ineffectuality:
the joining and separating functions can be understood without deriving
them from the significative function. Yet Abelard fails to say why the
theory he has proposed is preferable to the traditional one. He takes up
the theme of conjunctions once again and comes to the conclusion that
just as 'est' and 'non est' are appropriate for joining and separating and not for signifying,
'st and 'non si' join and separate utterances which are appropriate for signifying
something, but do not signify anything themselves. For they do not contain conceptions
of anything, whether true or imaginary, but merely direct the mind to a particular manner
of conceiving of other things.12

The essays in this book investigate whether the word 'to be' is indeed
ambiguous, as is assumed in most modern discussions of logic. The first
discussion sketched from Abelard's writings teaches that caution is required at the outset, in the formulation of the question itself. If asked
whether the word 'to be' always conveys the same thing or whether it
conveys different things according to context and use, Abelard would
surely have been unable to favor either alternative. For this formulation
of the question assumes that 'est' has a significative function wherever
it is employed, an assumption which Abelard does not share. In at least
one use, namely when' est' stands between two terms as the copula, it has
connective function but no content, and is to be reckoned among the
logical functors,13 the so-called syncategoremata. 14.
This thesis was obviously not uncontroversial in Abela:-d's time.
Abelard shares the following assumptions with his discussion partners:
(1) A word has significative function if and only if it conveys
something.
(2) Every noun or verb has a categorial content which it conveys.
(They are thus categorematic words.)
(3) The copula does not have a definite categorial content to convey.
How the negative thesis (3) could be reformulated in the affirmative
is a matter of dispute.
(3a) The copula can convey every categorial content in an indefinite
manner. Which definite content is meant can be determined in context
from the terms between which the copula is placed.
(3b) The copula has no content to convey. It indicates that that which
the predicate term conveys is to be joined with that which the subject
term conveys.

PETER ABELARD ON THE SPEECH SIGN 'EST'

149

2. Theories on What Verbs Signify and the Question oj What the Word
'Esse' Signifies
A. Aristotle's Criterium jor Verbs: Consignificatio Temporis. 'Esse' as
a Linguistic Device jor the Verbalization ,oj Nouns
It is evident from the expositions of Abelard's Logica '/ngredientibus,ls

and Dialectica l6 that the theory of kinds of words was a central theme
in his time. In these works Abelard seeks to work out his own position
on a question on which many a debate had already taken place.
Following remarks by Priscian,17 grammarians q,ad attempted to
distinguish different kinds of words using the criterium of meaning.
They had manifestly taken the Aristotelian doctrine of categories as their
starting point. Each category was thought of as forming a distinct semantic range. In Abelard's writings the viability of this theory is debated only
as it applies to the verb. It can be inferred that the category of substance
encompassed only nouns, that of quantity numerals, that of quality adjectives, etc. and that there was contention as to whether the theory was
actually able to describe language adequately.
The theory teaches that the verb is a word which signifies an action or
a being-acted-upon. 18 In the discussion the objection is made that the
theory does not do justice to the full complexity of language. As counterevidence, verbs such as 'lies' (as in the sentence 'Cologne lies on the
Rhine'), 'live', 'have', and 'be' are cited. The reaction to this objection
is indicative of the kind of theory being sought. The validity of the thesis
is defended by drawing special distinctions. The object is obviously not
to make an empirical, descriptive statement as to the meaning of most
verbs but to-find a criterium which allows verbs to be distinguished from
other kinds of words. The defenders reply that the examples cited do indeed signify something other than an action or a being-acted-upon, they
signify something which must be ascribed to other categories. Since the
meaning of these words is heterogeneous, they contain a semantic ambiguity. But the different meanings, it is maintained, must be ascribed
to the different uses of the words. The words mentioned are used either
in the verbal or in the nominative function. In the verbal function they
signify either an action (active form) or a being-acted-upon (passive
form), and as such (called expressions of action) they have temporal
meaning (tense). In the nominative function, however, they signify
something belonging to another category (for example, that of being-in-

150

KLAUS JACOBI

a-position, of having, or of substance); thus as expressions of states they

have no temporal connotation. l9
Abelard considers this theory to be inadequate. With delight he
parades the distinctions which had been elaborated in its defense and unmasks them as mere ad hoc contrivances. This becomes evident in the
passage on 'esse' in particular. To be able to subsume this verb under the
general thesis that verbs denote an action or a being-acted-upon the
following auxiliary theses had been asserted: 20
(A) When used in the verbal function, 'esse' signifies something which
falls under the categories of action or being-acted-upon. In contradistinction to all other verbs, however, 'esse' signifies no definite action or being-acted-upon. It stands equivocally for any and every action
or being-acted-upon. Abelard reports that the grammarians had taught
that 'I am' "has the strength of say 'I love' or 'I read', or 'I run' or 'I
am loved' or 'I am read' etc.". It seems that, as a matter of fact, they
did teach that 'I am' signifies all acting and being-acted-upon confuse
and none praecise in the verbal function. 2l
(B) When used in the substantive function, 'esse' signifies existence.
Now one can maintain that not only substances have an existence, one
can also assert that accidents, say a quality such as a particular color are
to be found. Thus in this use too, 'esse' stands equivocally (or confuse)
for the existence of things falling under one of the various categories.
As far as Abelard is concerned, a theory which must make allowance
for three kinds of equivocation, namely between (A) and (B), within (A),
and within (B), does not really explain anything. At best, it makes one
aware of a problem. His critique is bitingly sarcastic, but will not be
reviewed here. 22 He rejects every attempt to distinguish verbs from nouns
according to meaning. It is possible to form both a noun and a verb for
every semantic content within each category. Aristotle did propose the
proper criterium to distinguish them: nouns and verbs differ only in that
verbs bear a definite time reference along with their meaning (consignijicatio temporis).23 Every noun can thus be transformed into a verb
by the addition of a tense indicator. But sometimes it can only be determined in the context whether a word is being used nominally or verbally,
that is, without or with temporal co-signification.
For 'amans' is either a noun or a participle. and both of these signify the same activity.
It is only the temporal connotation which allows one to distinguish when it is a noun and
when a participle. Likewise. 'ens' can be either a noun or a participle. both in the meaning
of existence, when it is used without or with temporal connotation. 24

In cases where no verb has been invented for a specific semantic content, one can form a substitute by combining a noun with a finite form

PETER ABELARD ON THE SPEECH SIGN 'EST'

151

of 'to be'. (It should be noted that Abelard's understanding of 'nouns'

included nouns, adjectives, and pronouns.) One proceeds in this way
with the conjugation of verbs in Latin (but in English and German, too):
some passive forms and other finite forms are not constructed by adding
a suffix but by using a participle together with a finite form of 'to be' .25
Abelard similarly regards verbs as a whole. There is no difference between 'sedere' and 'sedentem esse' or between 'sedisse' and 'sendentem
fuisse', neither in the principal meaning, nor in the temporal connotation. Language could function in this way in all kinds of matters: Verbs
could be formed such as 'substantiam esse, fuisse, fore', 'quantitatem . .. ', 'corpus . . .', 'hominem . . .', 'albedinem . . .', 'album esse,
fuisse, fore' .26 'Esse' bears no semantic content of its own here. It func- .
tions simply as a time word, as a substitute for a conjugational form.
In comprehensive and precise analyses of the logic of tenses Abelard
shows that an expression constructed from a noun or participle and a
helping verb must be conceived of as a single unit of speech. 27 Assuming,
for example, that we say of someone who is not walking at the moment
'erit ambulans', a contradiction would result if 'eri!' alone indicated the
future and the participle 'ambulans' retained present significance. Thus
'this person who is not walking now (as postulated), will be a person who
is now walking'. On the contrary, the view that the whole is a single unit
of speech is correct, so that the participle loses the meaning of being in
the present. 28 This analysis applies not only to the use of present participles but also to all sentences dealing (sensu diviso 2'1 with past or
future states of a present subject, as Abelard demonstrates in discussions
about the conversion of temporally definite statements and their use in
syllogisms. 3o
Abelard emphasizes that every significative word can be made into a
verb, that is, it can be given an additional temporal meaning by the
employment of 'to be'. "Just as some nouns signify things as concretely
and independently existing and others signify them by one of their properties, so is it with verbs also.,,31 Abelard's thesis can be restated
precisely with symbols: Let' to' symbolize the time of the statement (' to'
is 'now'), '<' the relation 'before', and' >' the relation 'after'. 'vI' is
a predicate. Then the following results:
(4.1)

(4.2)
(4.3)

'Xis 1/1' is equivalent to' X-/;-to';

'X was 1/1' is equivalent to 'X 1/1 - tm " tm < to';

'X will be 1/1' is equivalent to 'X - tn " tn > to'.

Just as '1/1 - t' the equivalent forms are units of speech.

The use of 'esse' to signify existential import (verbum substantivum)

152

KLAUS JACOBI

is to be kept strictly separate from its use as a time word. In the meaning
of existence, 'esse' is a full and independent verb. Thus in the symbols
chosen, it does not merely correspond to the conjugational form' - t',
but to the full unit 'I/; - t'. Abelard emphasizes that one may assert the
existence not only of substances ('homo est'), but of substantivated accidents as well ('albedo est'). Differing from the grammarians with
whom he is disputing, Abelard does not consider the verb of existence
to contain a semantic ambiguity. For him, 'est' does not have the force
of 'exists as a substance' or 'exists as a quality' or 'exists as a quantity'.
It "has in all cases the same meaning", that is "is something from among
the multitude of things which exist". 32
Now the thesis:
(5)
'Esse' as a full and independent verb signifies 'existence'
does not harmonize with thesis (2) formulated above. 'Existence' is not
a categorial content. It seems that thesis (2) will have to be reformulated
for 'esse' also to be counted among the significative words. But then isn't
thesis (3a) more natural than (3b)? Or if (3a) is too unclear, would it not
be better to come to a thesis (3a '), according to which' esse' has existential import even in copulative use? Why does Abelard consider it
necessary to maintain such a sharp distinction between the use of 'esse'
as a full verb from its use as a copula, to the extent that all significative
function of the copula is denied?

B. Temporal Co-signification and the Predicative Function

According to Aristotle in Peri Hermeneias the consignificatio temporis
allows the verb to be distinguished from the noun and thus serves to
define the verb. Aristotle adds to this definition a remark on the
predicative function of verbs: the verb "is always a sign for something
which is predicated of something else". 33 From Abelard's statements on
this passage it can be concluded that the sentence had caused earlier commentators considerable difficulty. Its location in the text was felt to be
inappropriate. If Aristotle indeed intends to devote Chapter 2 to the
noun and Chapter 3 to the verb in order to reserve for Chapter 4 the proposition as composed of noun and verb, then the location of the sentence
cited above in Chapter 3 is quite difficult to justify. This peculiarity
disturbs all the more when one discovers the sentence repeated shortly
thereafter.
Abelard reports no fewer than eight attempts to interpret the
passage. 34 He does not get involved in an exhaustive discussion, nor does
he betray which interpretation he considers best. Apparently, the

PETER ABELARD ON THE SPEECH SIGN 'EST'

153

systematic structure of Peri Hermeneias interests him less than the individual theses which Aristotle advocates. The correctness of the content
of Aristotle's remark is scarcely a subject of contention among the interpreters, however: "Verbs were above all invented to be predicated,
nouns that something be predicated of them". The "complete" proposition consists of at least a noun and a verb. 35
In Dialectica the function of .verbs to couple a predicate with a subject
is given even greater emphasis. In his commentary to Peri Hermeneias
Abelard said that this function is characteristic for verbs,36 but it was the
temporal co-signification which served as the criterium for the definition. Not so in Dialectica. Here the copulative function, on account of
which verbs were 'invented', is reckoned to be part of the 'concept' of
the verb,37 but temporal co-signification seems not even to be accorded
the recognition of being one of the verb's characteristic features. This
thoroughly new idea stands in sharp contrast to the long tradition
preceding it and must be investigated with great care.
In the section on nouns, Abelard first relates Aristotle's view on the
specific differences between the noun and the verb. 38 Then he starts to
argue against this view, the same he had defended in his commentary on
Peri Hermeneias. He asks if it is really so that only verbs, but not nouns,
bear a temporal connotation "so that they assign their main meaning to
the persons serving as subject in accordance with their tense". 39 Just as
one understands in 'curro' and 'currens' that running (cursus) is to be attributed to a person in the present, one understands in 'album' something
"which is determined by whiteness as an (accidental) form in the
present" ;40 and also when 'man' is used to denote something it is because
"a mortal animal substance endowed with reason is present", so that
"'man' amounts to 'is a mortal animal endowed with reason which exists
in the present' ".41 "Thus Aristotle seems to have been wrong in
distinguishing nouns from verbs by saying that nouns aie without temporal connotation. For nouns are shown to indicate tense, too, namely
that of the verb, that is, the present. ,,42
Does Abelard actually intend to contest all distinctions between nouns
and verbs? Does he take the extreme extensionalistic point of view that
nouns signify nothing more than their present denotata? The care which
Abelard devotes again and again to non-denotative words (that is, those
which have meaning, although they do not refer to anything which
exists)43 speaks against such a view. Then how are his remarks to be
interpreted?
'Album' does not mean 'white' here in the sense of a dictionary entry.

154

KLAUS JACOBI

It means 'something white', 'something that is white'. Abelard does not

treat' homo' as a word to be regarded in isolation, either, but as a word
which is attributed to someone, that is, one which is said of a particular
person: 'he is a man'. Accordingly nouns have temporal co-signification
if and only if they are predicated. 44 Of course the predication is only implicit when nouns are applied to things. A thing X is given the name 1{;,
but the 1{;-ness of Xis not asserted expressly. To this extent there remains
a difference between verbs and nouns' 'in the manner of signifying": the
noun serves to name X as something to which a definite substantial or
accidental form (for example, the quality 'whiteness') is to be ascribed.
But only when a verb is joined to the noun is a proposition created. It
is in the proposition that the inherence of something is asserted. 45 Nevertheless, a predication lies at the base of any naming. If the question is
asked whether a word '1{;' does in fact presently name something, that is,
whether the word is properly attributed, the answer consists of a proposition in which the description '1{;' moves to the predicate position. '1{;' is
then joined to the subject with a conjugational form of a verb in the present or with 'est'.
In Dialectica, the relative importance given to the copulative function
of the verb exactly reverses to that in the commentary to Peri
Hermeneias. It is no longer recognized as a mere feature of the verb, as
in the commentary, but as the criterium which defines the verb. Verbs
have copulative function and with it also temporal connotation. 46 Nouns
receive temporal connotation when they are used in such a manner as to
assert that a categorial content (their sense) subsists in a thing. 47 When
this use of nouns to name is made explicit, it is revealed to be predication.
The temporal significance does not belong to the signijicatio of a word
but to a specific function. When one refers to something by means of a
noun, one does predicate. The temporal co-signification is the result of
the implicit or explicit predication. If Abelard's line of thought were to
be followed to its ultimate consequences, one can imagine him giving up
the difference between nouns and verbs altogether and continuing to
observe only th~ functional difference between subject position and
predicate position. But this would have meant turning the whole structure of logic upside down. The Aristotelian tradition had sought to synthesize the proposition from the various units of speech. Now one would
have had to start from predication as the foundation of logic and then
work one's way back to that which the individual units of speech, such
as nouns and names, signify. It would hardly have been possible for
Abelard, who had developed his theses in the course of an exposition of

PETER ABELARD ON THE SPEECH SIGN 'EST'

155

traditional texts, to free himself from this tradition to such an extent as

to be able to recognise the far-reaching consequences of his line of
thought. It is doubtless easier for the interpreter of a later generation to
see them.

C. 'Est' as a Verb of Full Value and as a Copula

Abelard notes several times in the course of his discussions that propositions can be formed in two ways. First, a finite form of the verb serving
as predicate is joined to a noun, the subject. Second, a subject term and
a predicate term are linked together by a finite verb placed between
them. 48 When the verb serves as a predicate it has a double function: it
is both copula and that which is being predicated, that is to say, it joins
itself, its own meaning, to the subject. Only two verb~ are capable of
linking subject and predicate terms, that is of joining a meaning other
than their own to the subject. These are 'est' ('S is P') and 'nuncupatur'
('S is named p,).49
Now for Abelard, as shown above, propositions of the type'S I/; - t'
and those of the type'S copula-t 1/;' are identical in their semantic deep
structure. Why then does Abelard devote such great attention to the difference in surface structure? De Rijk has correctly recognized the target
of Abelard's interest: the difference allows Abelard to discuss the problem of "the basic ambiguity of the verb 'est' ".50
'Petrus est' is as much a complete proposition as 'Petrus currit'. 'Est'
has a double function: it is both that which is predicated of Petrus and
that which links the predicate with the subject. 51 When 'est' is thus
"predicated in the essential sense (proprie) " ,52 it links its own content to
the subject. Thus 'Petrus est' means 'Peter is something which exists'. 53
The question so heatedly disputed in Abelard's time (but before his
time, too, and afterwards again and again) was whether 'est' retains existential import when used as a copula. In his expositions on this matter
Abelard grants 'est' a special status, showing that no other verb (except
'nuncupatur') can be used as a connecting link between the subject term
and the predicate term. Contrary to first impressions, propositions such
as 'iste fit bonus' and 'iste videtur bonus' have a completely different
structure from 'iste est bonus' .54 In the first two examples, Abelard
regards 'good' as a means to define more precisely the predicates 'fit' and
'videtur'. 'Bonus' functions in just the same way as the object' Socratem'
in the sentence 'iste videt Socratem', where 'Socratem' limits the
predicate' videt'. The proof can be arrived at by making a conversion,
for here the complete verbal expression goes to the subject position. (For
example, 'Many an evil doer seems to be good' <---> 'Many a man

156

KLAUS JACOBI

who seems to be good does evil. '). The point is that 'to be' cannot be
regarded simply as a predicate which is defined more precisely by a
predicate noun. The meaning of the other example sentences can be
made clearer by seeing the limitation on the predicate as a namelyrider: 'This is becoming something, namely good'; 'This seems to be
something, namely good'; 'He sees something, namely Socrates'. One
may not, however, interpret linkages using the copula 'est' as meaning
'This is something, namely
Can we thus safely conclude that 'est" in copulative function does not
retain the existential meaning it has as a verb of full value? Abelard exhorts to caution in his further discussions. One thing has been clearly
established: that the copulative use of 'est' should not be confused with
its use as a predicate. The copula is not "predicated in the essential sense
(proprie)". Its function is to link not its own semantic content but that
of the predicate term to the semantic content of the subject term.
Nonetheless, even if the copula is not predicated proprie, it is indeed
predicated per accidens. 55 And although its principal function is to link
~ubject and predicate terms, it may be that the meaning 'existence' plays
a certain role when 'est' is used as copula.
In speaking of 'esse' Abelard points out that "there is always an existential import in its linkage" and "it allows us to determine that
another thing exists", even when it is used as a copula. 56 But he says this
with regret. 57 It would be ideal if the copula had absolutely no semantic
content and functioned as a purely syncategorematic connecting
symbol. 58 'Est' remains the word which comes closest to this ideal. Any
other verb would convey a specific categorematic content. This content
would make the verb unsuited to serve as a link between the subject and
yet another such content as found in the predicate term. 59 'Est' is
suitable for use as a copula because its own meaning preoccupies least
of all and can be most easily kept in the background or, one could even
say, suppressed by the subject and predicate terms.
One can find passages in which recognition is granted to the existential
import of the copula near other passages which emphasize that the existential meaning by no means belongs to the copulative function. My
understanding of this pecularity derives from the role Abelard accorded
to the science of argumentative discourse, this being to reflect on and explain language and yet not to shy away from correcting it and bringing
to it a greater precision. The discussions carried on in these passages need
not be summarized individually here. Nonetheless I consider it important
at least to show briefly which semantic types of propositions play a
special role in them.

vI'.

PETER ABELARD ON THE SPEECH SIGN 'EST'

157

In a first special case, the name of a person stands in the subject position of the proposition as an expression for a concrete substance. An expression for an accident stands in the predicate position. Take for exampIe 'Socrates est albus'. This predicate joins two items to the subject, first
the whiteness (albedo) as an accidental form (in adiacentia), and secondly something white (album) in existence (in essentia).60 Abelard is himself
unsure how the two should be weighted. In Logica '[ngredientibus' he
considers the speaker's real intention to be to predicate the attribution
or inherence of whiteness in the actual subject. The use of the concrete
term 'white' seems to be misleading here due to the existential import
which the copula bears. Because Socrates exists as something white, not
as whiteness, the abstract term cannot be used here, contrary to the
speaker's intention. In Dialectica. which was written later, Abelard
defended the form of the proposition which is naturally given in the
language, 'Socrates est albus' as being a sensible one. 61 The proposition
asserts that Socrates is one of the things which are white. In the first
chapter of Categories Aristotle makes a distinction between accidental
predicates, which "are in a substance" and substantial predicates, which
"are said of a substance". In reference to Socrates, 'white' belongs to
the former group. But the "inherence of attributed whiteness" is merely
"intimated" in this proposition. To make the meaning of this intimation
explicit is a task for a categorial analysis of predicates; it is not a task for
an analysis of the meaning and functions of predication.
A second special case, one which plays a considerable role ill
Abelard's investigations, is that of propositions whose subject terms
signify something which in principle or in fact does not exist. A few
classic examples: 'chimaera est opinabi/is', 'chimaera est non-existens'
and' Homerus est poeta' .!>l It is obvious that such propositions are not
to be understood as though the copula were asserting - even merely per
accidens - that such creatures as chimaeras ever existed or that Homer
were alive at the moment the proposition was uttered. Abelard and the
logicians whose writings he uses as groundwork tryout a long series of
transformations in search of a clear understanding of the meaning of the
sentences given as examples. Thus the sentence 'chimaera est opinabilis'
should be reworded as a sentence about someone who can imagine a
chimaera. 63 The true meaning of the sentence about Homer requires it
to be transformed into a sentence about a still existing poem written by
him or about a memory of him which has not died out. 64 The purpose
of the transformations is to return to sentences in which the verb 'to be',
in both in its full meaning and its copulative use, is dealing only with

158

KLAUS JACOBI

items which exist. Alternatively, the possibility is considered that such

sentences show best of all that 'est' in copulative use is fundamentally
non-literal and devoid of all inherent meaning. It depends on the semantic content of the predicate term, and not on the existential import of the
copula, whether one will or will not be in a position to infer that the subject term names something which exists. 6s For example, it can be inferred
from 'Petrus est homo' that Peter exists, but it cannot be inferred that
a chimaera exists from 'chimaera est opinabi/is' or 'chimaera est nonexistens,.66 Once "the heart of the semantic question of being has been
transferred from the meaning of the copula to the meaning of the
predicate noun used in the proposition" ,67 it is only a short step to the
theory that the copula is not an independent part of the proposition at
all, but that it, together with the predicate noun, should be regarded as
one single verbal expression. Abelard finds this step to be a daring one,
because it is so novel, but a solidly reasonable one. 68 The imposing
arguments found in the analysis of verb tenses which follows in Dialectica provide a strong recommendation for it.

3. The Theory of the Proposition and the Question of the Exact Meaning
of the Assertive Formula <ita est in re'
In Abelard's time, the following terminological distinctions were
regarded by logicians and grammarians as being fundamental: Every
grammatically well-formed combination of several per se significative
words (dictiones), with or without the addition of co-signific~tive words,
is called a word string (oratio).69 A word string is called 'incomplete
(oratio imperfecta)' when the person hearing the sequence of words cannot help but expect to hear an additional word or words. When such an
addition is not absolutely necessary, that is, when the combination of
words is such as to enable a (relatively) full understanding, the word
string is called 'complete (perfecta). ,70 Some kinds of minimal complete
word strings are questions, wishes, orders, and propositions. 71 Their
distinguishing sign is normally the presence of an inflected verb form.
When this form is absent, as in many orders and questions, the listener
takes it as implied. The proposition (oratio enuntiativa, propositio) is
defined as a "sentence signifying something which is either true or
false". ~2
With great pertinacity Abelard uncovers the problems hiding behind
these common distinctions and definitions. His treatment of the theory
of the proposition is particularly painstaking. He begins with questions
arising out of very simple observations. In understanding any word

PETER ABELARD ON THE SPEECH SIGN 'EST'

159

string, a person not only comprehends what is signified by each word,

he also, not less important, joins together the meanings which he has
comprehended. 73 Thus, for example, the difference between 'homo currens' and 'homo est currens' cannot consist of the copula's telling us to
join together the two already comprehended words 'homo' and 'currens', for this joining has already been attained in the course of coming
to understand 'homo currens'. Abelard goes even further: 'homo currens' conveys precisely the same joining of 'run' and 'man' as 'homo est
currens' and 'homo currit,.74 The copula or, as it were, the inflectional
ending of the verb brings no new understandable content to the terms
which are to be joined, when the consignificatio temporis is left out of
consideration. 75
The difference between a proposition and other "incomplete" or
"complete word strings,,76 does not lie in the content of what is comprehended. But then how is the traditional definition of the proposition
to be interpreted? Abelard answers 77 that the proposition signifies the
truth or falsity of something by stating (proponendo) what is the case or
what is not the case. An affirmative proposition differs from an
"understanding which connects various elements" and a negative proposition differs from an "understanding which separates various
elements" not in respect to content but in the way the content is "proposed" ("stated", asserted or contested). "For that reason declarative
sentences (enuntiationes) or propositions (propositiones) are expressly
named from the manner of proposing (proponendo). ,,78 Abelard supports his analysis by reformulating the propositions in such a way that
their assertive meaning is made explicit. In the sentence,
'it is true that Socrates is sitting (verum esl Socralem sedere)', the phrase in the subject position, 'that Socrates is sitting', produces the same understanding as 'Socrates is sitting', and
yet the phrase in the subject position cannot be called a 'proposition' because it is not
formed in a way which would enable it to be stated affirmatively or negative1y.79

The dictum proposition is is not a proposition; it merely comprises the

contents of that which can be proposed. A proposition is the assertion
of the truth or falsity of these contents. The meaning of 'is true' is explained as 'it is so in the thing'. 80 But as further analysis shows, this explanation can be misleading.
Simple words are normally not just signs for mere concepts but for
concepts of things existing in reality.
Nouns and verbs are ... namely signs in a twofold sense, for things on the one hand and
for concepts on the other. They are signs for things in that they constitute concepts be1()n):ing to the things, that is, concepts which heed a particular characteristic feature or

peculiaritv of the things. 81

160

KLAUS JACOBI

When Abelard later distinguishes between "valid" and "empty concepts", he does not deviate from his basic assumption: "valid" concepts
are those in which the basic intention to mean a thing is fulfilled. Abelard
explains the meaning of 'valid (san us)' in relation to 'concept (intellectus)' as follows. "Every concept, whether simple or complex, through
which we heed how the thing behaves, is valid. ,,82 This explanation is illustrated with an example: the word 'man' "yields a valid concept as long
as (at least) one man exists". 'Chimaera' and 'hircocervus' are examples
of words the concepts of which do not denote any existing thing; if there
were no such thing as man, the concept of 'man', too, would be empty.
The same holds for complex expressions and that of which they are
concepts. 83
Simple words or incomplete word strings such as 'homo currens' fulfil
their naming or descriptive function when there is at least one thing
which is (in a specific respect) as the word or word string conveys. Now
it is of great importance to Abelard that a distinction can be made
between
(6.1.) the truth criterium, 'the thing is in reality so as the proposition
says', and
(6.2.) the validity criterium, 'there is some real thing which falls
under this naming or description'.
His arguments are apagogic for the most part. He finds a theory
among contemporary logicians according to which the content of a proposition is to be identified with the content of its components. 84 He
shows in a penetrating analysis that this theory is unsatisfactory: it cC'nnot explain what is asserted in propositions of necessary consequences
such as 'si est rosa, est jlos,85 or in tautologies such as 'Socrates est
Socrates' .86 Abelard's critical deliberations cannot be presented here
individually.87 The existence of the proof that (6.1.) must be differentiated from (6.2.) will have to be taken for granted. But what exactly is
the difference between the two?
In the first chapter of Peri Hermeneias, anticipating the more
thorough discussion to come, Aristotle touches on the difference between (simple and complex) nouns and propositions. The noun alone
does not signify anything which can be true or false, he says, unless "being or not-being, whether absolutely or with a definite tense, is added. ,,88
Boethius interprets this passage as a reference to the deep structure of
every proposition. He reformulates such subject-predicate-propositions
as 'homo vivit' and 'Socrates ambulat' into 'homo vivens est' and
'Socrates ambulans est'. 89 The word order is surely not arbitrary; it is re-

PETER ABELARD ON THE SPEECH SIGN 'EST'

161

tained in such examples as 'Socrates philosophus est' and 'Socrates

philosophus non est'. The terms within each of the propositions 'A est
B' and 'A non est B' are joined together here into a single complex term
'AB'. In Boethius' view, being or not-being is to be predicated of this
complex term occupying the subject position.
Abelard starts to develop an opposing theory. 'It is so' and 'it is not
so' are impersonal expressions and are not what one normally understands as predicates. 'That A is 1/; (hominem vivere; Socratem ambulare),
is not a noun in the normal sense or function, even if it can be reformulated as 'the 1/;-ing of A'. It is at best a quasi-nomen. 90
That which is said in a proposition, '1/;(A)', is not a thing. 91 It can be
called the proposition's content,92 but then positive and negative contents, or states of affairs, would have to be admitted. 'Neg'ative things'
are an impossibility. There are only negative descriptions of things. This
consideration makes it clear that contents are not instantiations for propositions in the same way that things are instantiations for descriptions.
' ... is true ( ... is the case)' and' ... is false ( ... is not the case)'
as well as '. . . is possible', ' ... is necessary', and their negations are
impersonal expressions. 93 Grammatically they seem to be predicates for
the dicta which seem to be their subjects, but differently from other
predicates, they are not used denotatively.94 The denotative function of
predicates can be explained by substituting a 'something' in the subject
position as a realization (or a 'nothing' as a sign of the lack of realization): 'Something 1/; - t, namely .. .' (or, as the case may be, 'nothing
1/; - t'). If 'is the case' were to be predicated personally, the explanation
by 'something, namely ... ' would be appropriate here, too. The difference would merely be that not individual things but rather a particular
content would stand for 'something' here. The formulation 'something
is true, namely that. . .' may as a matter of fact appear in linguistic,
logical, or epistemological contexts,9S but Abelard contests resolutely
that this formulation, in which 'true' is used denotatively, functions
faithfully in the same way as ' ... is true ( ... is so in the thing)'. 96
'True' and 'false' should not be understood as general terms which could
be applied to particular cases. 97
But then how can the meaning of 'it is so in the thing' and 'it is not
so in the thing' be explained positively? Before attempting a solution,
one must take care not to be misled by the question. To a related question
on the words 'possible' and 'necessary', Abelard answers that these
co-signify (consignijicare) more than they bear a meaning in and of themselves. For we can
conceive of something in them only when they are applied to a word string in the subject

162

KLAUS JACOBI

position. And then they bring about that the things signified by the word string in the subject position are conceived of in a certain way, just as an interposed verb ('est') or a conjunction ('if. .. , then') whose coupling indicates necessity serve to do 50. 98

'. . . is the case' and '. . . is not the case' are syncategorematic expressions. Their meaning is nothing other than their function in a word
string. The content which Abelard mainly has in mind in his analyses are
events. The truth value of propositions about events is dependent upon
the point in time at which the proposition is asserted. It is the case in a
certain 'now' that Socrates is running, just as this is not the case in a different 'now'. The dictum as the content of a proposition does not receive
a specific truth value before it is actually asserted to be true or false.
Whoever understands the assertion comprehends the asserted content as
existing (or not existing, as the case may be) at the time the proposition
is asserted.
II. POSITIONS

The inquiry into the logic and semantics of 'est' was developed by
medieval authors as the inquiry into the nature of predication. It can be
seen from Abelard's discussions that the theory of propositions and
predication was a matter which provoked deep controversies. The discussion centers for the most part around the paradigm for all individual
analyses and does not occupy itself merely with puzzle solving within a
theory or the further development of an accepted one. The fact that all
parties appeal to Aristotle and Boethius only contradicts what I have said
at first glance. The deliberations of these authorities on semantics prove
on closer scrutiny to be less than firm, for they allow widely varying
interpretations.
(A) The common ground of all parties to the discussions consists of the
following assumptions:
(1) The search for a theory of the proposition is seen as a search for
a theory of predication. Following Aristotle, the basic structure of the
proposition is thus held to be that something is said about something else.
Relational propositions are put into predicative form. In 'A sees B', for
example, the object 'B' is considered to be a more precise determination
of the predicate. The first beginnings of a reorientation can be found
with William of Ockham: he investigates the supposition of both the
term preceding and following the relational word in relational propositions. The thought of constructing a theory of predicative propositions
on a necessarily more comprehensive theory of relations and thus of

PETER ABELARD ON THE SPEECH SIGN 'EST'

163

regarding concepts, that is functions with one empty position, as

borderline cases of relations, that is functions with several empty positions, was foreign to the Middles Ages.
(2) What is sought is one single structural description of the proposition which works for all forms of language without exception, that is, a
theory of the uniform logical deep structure of all kinds of predication.
Simple propositions can be observed in two different grammatical
forms. In the one, a noun in the nominative case is brought together with
an inflected verb ('Socrates currit'). In the other, a predicate noun is
coupled to the subject with .'est' ('Socrates est homo'; 'Socrates est
albus'). Now Aristotle had already accepted the equivalence of
'hominem ambulare' and' hominem ambulantem esse'. 99 The validity of
this view is taken for granted by all logicians who sought a unified theory
of predication. The manners of constructing a sentence are interchangeable in principle. 'Socrates currit' has the same meaning and the
same truth value as 'Socrates est currens' .
(B) Widely differing theories can be constructed on the basis of the two
assumptions listed above. The spectrum of discussion is exhaustive in
Abelard's time. The basic features, if not the finer details, of all kinds
of positions are developed and considered.
The equation' 'Socrates currit' = 'Socrates est currens" can be read
in both directions as an instruction on how to reformulate. This never
caught anyone's attention before Abelard. Nor did anyone after
Abelard, as far as I have been able to discover, consider the decision controversial whether the theory should be based on the two-part or the
three-part structural description. Even Abelard does not seem to
recognize the importance of this decision. He alternates between
statements which work from the three-part description and statements
which take the two-part description to be logically more transparent. I
consider his discovery that tensions between the two theories of predication begin to develop at this point to be of great significance.
(1) Normally, the three-part form is held to be an explanation of the
two-part form. The authority of Aristotle is not the only source of support for this view. For the structur,e.of language also seems to suggest it:
to resolve an inflected verb into a temporally definite copula and a participle seen as a predicate noun, just as language does to make certain
tenses and passive forms, is manifestly easier than to conceive of all combinations made with the copula as substitutes for verb forms which the
language has failed to "invent".

164

KLAUS JACOBI

Accordingly, the explicit logical standard form of an affirmative proposition has three parts: a subject term (about which something is
predicated) is coupled to a predicate term (which is predicated) with an
interposed 'est'. 100 The theory of predication then presents itself as the
theory of the coupling capacity of this interposed 'est'. Now linguistically, 'est' is a form of the verb 'esse', which signifies 'to exist'. 'Est' has
this meaning in all cases when it is used as secundum adiacens in such
propositions as 'Socrates est'. It signifies here the existence of that of
which it is predicated. The question arises whether 'est' also has existential import when it is used as tertium adiacens, that is, as the copula. In
the answer to this question two contradictory stand-points part company: The first position (1.1.) maintains that the two uses of 'est' are not
related; 'est' is equivocal. The second (1.2.) maintains that the two uses
are interrelated.
(1.1.) According to the equivocation theory, the copula has no semantic content, it simply functions as a coupling device. Thought out to the
last consequence this would mean including the copula among such conjunctions as "si (if. .. , then)'. 101 The justification for this thesis is to be
found approximately as follows: 102 'est' as secundum adiacens, that is,
as a complete predicate, joins its own semantic content, namely existence, to that of the subject term. The copula has, however, a different
function. Its function is to join the semantic content of the predicate term
to that of the subject term. The copula can exercise this function only
when it has no semantic content of its own, for it is not possible for a
word to join itself and something else to the subject simultaneously. This
last sentence, which is crucial to this line of argument, is supported by
a comparison with other per se significative verbs: Just as one cannot add
a further predicate to 'Socrates currit' or 'Socrates est currens', say
'homo', one can add no further predicate to 'Socrates est' or 'Socrates
est aliquid existens'. Consequently, 'est' does not mean 'est aliquid existens' in the sentence' Socrates est homo'. By varying it a bit, one can
use this argument to attack the opposing thesis. 103 If 'est' always meant
'est ens', then the predicate term in 'Socrates est ens' would be
superfluous, seeing that it would already be contained in 'est'. To say
'Socrates est ens' would amount to saying' Socrates est ens ens', etc.
Another argument with critical intent refers to propositions in which the
subject term signifies something which does not exist. 104 The proposition
'chimaera est opinabilis' would be false if the copula had existential import. Further, the proposition 'chimaera est non-existens' would provide
ground for the paradoxical inference that the chimaera exists (by virtue

PETER ABELARD ON THE SPEECH SIGN 'EST'

165

of the existential import of the copula) and does not exist (by virtue of
the semantic content of the predicate term). Consequently, 'Petrus est'
cannot be inferred from 'Petrus est homo', at least not by virtue of the
alleged existential import of the copula. \05
(1.2.) The point at which parties to the opposing position can assail the
equivocation theory is the following: the theory leaves one puzzle unsolved, namely why we use one and the same word, 'est', on the one hand
(as secundum adiacens), as both significative per se and copulative, and
on the other (as tertium adiacens), as merely copulative. This difficulty
weighs iheavily on the logicians of linguistic analysis of the time. If someone were to succeed in constructing a theory which explained the relationship of the two uses to each other convincingly instead of deriding
the double use of 'est' as a misleading accident of language, such a theory
would clearly be more comprehensive than the equivocation theory.
What must be demonstrated is that the word 'est' can join a term other
than itself to a subject and yet retain existential import. The opponents
of the equivocation theory want to do so by arguing for a stronger thesis:
the compatibility of existential meaning with the copulative function has
been established when one successfully derives the copulative function
from the existential meaning. And inasmuch as 'est' means everything in
being (essentia) indiscriminately, so it is said, it is capable of joining
together any and all content l06 and of being the sign of the "existence
of the thing" 107 in affirmative propositions. 'Socrates est' means
'Socrates is one of the things which exist'. When 'est' is used copulatively, so that a predicate term is added to it, what is thus determined is that
as which the subject exists, thus, for example, 'Socrates is (as) a man' or
'Socrates is (as) a person who is presently reading'. In the argument
brought to bear by the adherents of the equivocation theory that no word
can simultaneously join both itself and another term to the subject, they
overlook the special character of the semantic content of 'being': the
predicates added are not principally different from this semantic content
but are, rather, complementary predicative determinations of it. Thus,
in 'Socrates est ens' the speaker merely makes clear that he does not wish
to make a narrower determination of what Socrates is. Furthermore,
propositions whose subject terms stand for things which do not exist are
"non-literal" formulations. What they intend must be clarified by reformulating them. 'Homer' is the name of a person who no longer exists.
The proposition 'Homerus est poeta' is to be reformulated into 'the
renown of Homer is kept alive in human memory by his poetry'. 108 Propositions about the chimaera are to be interpreted as treating something

166

KLAUS JACOBI

which IS In the mind of the person thinking the proposition at that

moment. 109
(2) Instead of regarding the three-part structure 'subject term copula - predicate term' as the explanation for the two-part structure
'subject - inflected predicate', one can also regard constructions with
'est' and a predicate noun simply as substitutes for verbs which have yet
to be "invented". The equation' 'Socrates currit' = 'Socrates est currens' , is now to be read backwards, for the two-part construction of the
proposition is the logically more transparent one - the predicate is considered to be a complete unit.
In numerous passages in his works Abelard shows decided preference
for the two-part analysis of the proposition (with the one-part theory of
the predicate) over the three-part analysis of it (with the two-part theory
of the predicate). That here we have to do with more than a flash of inspiration to solve some special problems in the logic of tenses follows
from a passage of Dialectica: these discussions must be recognized as the
groundwork for a new theory of predication. In the book of Dialectica
entitled 'On the Parts of the Categorial Proposition', Abelard reviews
the three-part analysis but notes that he is writing here "following
general opinion", because he considers it important that the many people who adopt this thesis will be able to support it with arguments. He
does not fail to use this opportunity to refer the reader back to his own
two-part analysis as discussed in his book 'On Words with Definite
Meaning', which, he says, does not conflict with the material discussed
here but does offer a more elegant solution to the problem. 110
What distinguishes Abelard's theory from theories (1.1.) and (1.2.)
can best be described negatively. An analysis of the meaning and function of 'est' does not lie at the heart of the theory which regards the structure 'subject - inflected predicate' as the explicit logical standard form
of propositions. Propositions having 'est' as predicate ('Socrates est')
are not accepted as being in principle different from propositions with
other predicates (such as 'Socrates currit'). III It may be that the
understanding of the existential predicate poses special problems, but
this does not constitute the key problem. Constructions such as 'Socrates
est homo', 'Socrates est albus' and' Socrates est currens' are, as it were,
"non-literal" forms of predication. 112 In these constructions 'est' is not
the member by the exclusive means of which the terms may be joined to
one another, it is reduced more to functioning as a helping verb.
(C) It is necessary to distinguish between the broad question which

PETER ABELARD ON THE SPEECH SIGN 'EST'

167

general structural description of the proposition is most suitable and the

questions which are encountered in the course of the logical analysis of
particular propositions. Supposition theory as a theory of the interpretability of the terms of the proposition and modal theory as a theory
of the determinability of the linkage of subject and predicate are not
bound exclusively to anyone of the theories of predication distinguished
here and are compatible with all of them. The distinction both of the
kinds of supposition and of modalities is indeed suitable for a differentiating analysis, independent of the position one takes on the controversy
of predication theory, one which belongs to the philosophy of logic. One
may suppose that this contributed to the m0dest estimation among logicians of the time as to the importance of the discussion of underlying
principles. All interest was drawn by the ars sermocinalis then being
developed.
In scholarly writings on medieval predication theory, the inherence
theory and the identity theory are mentioned time and again as the two
competing models for predication theory. 113 According to the identity
theory, in an affirmative proposition one asserts that the subject term
and the predicate term stand for the same individual (or individuals);
thus both terms are interpreted extensionally. According to the inherence
theory, however, the predicate term must be understood intensionally:
it is asserted that the "universal nature" (form) signified by the predicate
term inheres in at least one (or in all) of the individuals denoted by the
subject term, 114 or, to say it another way, that the individuals denoted
by the subject term are of the universal form signified by the predicate
term. As has been shown,115 Abelard is familiar with the question
whether the predicate term indicates the form literally or the bearers of
the form. He discusses it in his analysis of propositions in which an accident is predicated of individuals, but the discussion is conducted in such
a manner that it can be applied only to this type of proposition or contents. So Abelard is by no means concerned to lay the groundwork for
a new theory of predication with this discussion. He is merely dealing
with a minor problem which was bound to come up in the course of
reflections on the deep structure of a particular type of proposition,
regardless of the position one started from.
III. REFLECTIONS

After Abelard, the two-part analysis of the proposition won no further

adherents. Subsequent logicians of the twelfth century did not feel com-

168

KLAUS JACOBI

pelled to refute it, either. Abelard's theory was simply not to be found
in the universe of discourse. It was considered obvious that it is the
copula which forms the proposition inasmuch as it relates the subject and
predicate terms to each other. Now as long as the equivocation theory
(II. B. 1.1.) and the theory maintaining the semantic and functional interrelation of the verb of existence and the copula (II. B. 1.2.) competed
with each other, the thought of ending the dispute by taking up the position counter to both (II. B. 2) would not have been far-fetched. But the
actual resolution came from an unexpected source.
In the thirteenth century the reception of Aristotle's 'Metaphysics'
took place in the European schools. This is the theory of being as such.
One of the methodological instructions given by Aristotle in order to
open up a new area of research is to pay attention to the various aspects
of the meaning of words and to investigate the interrelations between
them. 116 The semantics of the word 'est' inevitably became the central
theme of discussion now. Since the equivocation theory did not satisfy
the second part of Aristotle's instruction, the theory quickly came to be
regarded as obsolete. The opposing theory, on the other hand, fitted the
theoretical requirements of the time perfectly. 117
Is "the logic of being" a theme which ontology brings to logic? Is this
theme of central importance to logic only when logic is understood as being bound up with a metaphysical "theory of being"? The readers of this
volume will be in a better position to reflect on the question than an
author who has presently but a single voice in the concert.
When surveyed from the distance of history, Abelard's contributions,
as reported on here, show a common tendency. From the discussions of
his contemporaries, Abelard was familiar with the problems which confront someone trying to answer the question whether 'est' is semantically
ambiguous. His strategy was first to remove the word 'est' from the
limelight in which it had stood for his predecessors and contemporaries.
The logic and semantics of the word 'est did not necessarily constitute
the central problem at the basis of a theory of the proposition and the
sermocinalis scientia. Abelard's question was no longer why 'to be' was
the verb used to form the proposition. The question he had to answer was
why 'to be' could be substituted for inflected verb forms. Unlike the
adherents of the equivocation theory, he sought to account for this
linguistic phenomenon. Unlike the opponents of the theory, he attempted to do so not by offering a broadly encompassing theory,IIS but by
offering one which bore a minimal burden of argumentation. 'To be'
meant that something is simply given, that it exists, has existed, or will

PETER ABELARD ON THE SPEECH SIGN 'EST'

169

exist. And because it meant nothing more than this, it lent itself to functioning as an auxiliary verb to indicate the time for which the truth of
a proposition was to be asserted.
The theories of predication introduced in Part II of this essay are all
tenable and also flexible enough to come to terms with the material of
which such theories treat, namely with the proposition in normal
language. Abelard realized this. Although he is one who normally does
not let any opportunity for apagogic argumentation pass by, he makes
no attempt to demonstrate that the three-part analysis of the proposition
is inconsistent or not appropriate for the description of all types of propositions. Problems of the logic of tense and of propositions whose subject terms signify something which does not exist can also be resolved
within the framework of the copula theories. To this end it must be
shown that the meanings of the terms of a proposition are displaced or
amplified in a specific manner when the copula is put into the past or
future tense or when it is supplemented by a modal expression or an expression of knowing or being of an opinion.
Abelard is able to adopt the terminology of the copula theories without
reservation. When he refers to his theory as being "more elegant", he
is thinking of the problems just reviewed. But one must bear in mind here
that the logic of occasional sentences was accorded a much greater
relative weight in medieval systems of logic than in modern ones. For
Abelard, the occasional sentence is not a special case requiring the
development out of general logic of a special branch to be called the logic
of tenses. The occasional sentence is truly the normal case in his eyes, so
that logical theory must come to terms with it right from the beginning,
at its foundations. Thus a tense indicator is to be counted as a basic
feature of the predicate. It specifies the relationship between the point
in time at which the proposition is made and the content of the proposition. Is the two-part analysis indeed more elegant when seen in the context of logical theory as a whole and not merely of particular detail problems, assuming that the occasional sentence and a logic of tense corresponding to it playa central role?
What kind of a structure would a logic have to have which did not
from time to time merely bring in the two-part analysis to help clarify
specific problems but which took the two-part analysis as its foundation?
Up to now I have described the logical standard form of the proposition
rudimentarily as 'X Vt - t'. When, instead of a personal name, a general
term is placed in the subject position, distributors must be added specifying whether the predicate is to be asserted for all or only for some of the

170

KLAUS JACOBI

individuals which fall under the subject term. Now the denotative use of
a noun is based on a predication. 119 'A man' amounts to 'some thing that
is a man'. 'All men' amounts to 'every thing that is a man' .120 'Is' serves
here, as in all other cases, as a replacement for the tense indicator as contained in the verb ending. Abelard shows this to be the case in passages
on the proper conversion of temporally specific propositions. 121 He gives
the sentence, 'Every old man was once young (Omnis sen ex jui! puer)' ,
as an example. By interchanging subject and predicate, a universal affirmative proposition can be transformed into a particular affirmative proposition. Abelard emphasizes in his discussions that the tense indicator
of the predicate must also be transferred to the subject position: 'Some
of the .men who were once young ... '. Likewise, the original subject
must be given a tense indicator when it goes to the predicate position:
' ... are now old'. The original proposition can be reformulated in the
same fashion: 'All men who are now old were once young'. The form of
a proposition explained with the help of quantors and tense indicators
is then as follows:
'All/some which l/; - t, If' - t'.
It should be clear that a negating word can be added to the first predicate
as well as to the second. The tense indicator must also be given. Thus:
'All/some which [not] l/;-t(l/;-tovl/;-tm V l/;-t n ),
[not] If'-t(If'- (0 VIf'-tmVIf'-t n )'.

The transition from occasional propositions to propositions whose

truth value does not change with time, as, for example, in '(omnis) homo
est animaf, can be made without difficulty. Since one merely needs to
abstract from the time of the utterance of the proposition, the transition
consists of nothing more than an act of simplification. The tense no
longer needs to be indicated specifically, for this function will now be absorbed by a quantor which is to stand before the proposition. Thus we
have: 'Whenever some thing is a human being, it is an animal' .122 Expressed formally, this becomes:
'For every point in time t: All/some which l/; - t, If' - t'.
APPENDIX

After the completion of my paper, the editors have kindly given me the
opportunity to call the reader's attention to Kretzmann's article 'The
Culmination of the Old Logic in Peter Abelard' (in R. L. Benson and G.

PETER ABELARD ON THE SPEECH SIGN 'EST'

171

Constable (eds.), Renaissance and Renewal in the Twelfth Century, Harvard University Press, Cambridge, Mass., 1982, pp. 488 - 501). Kretzmann's admirably clearly written contribution shares to a considerable
extent the themes which I have dealt with here. At the centre of his discussion is Abelard's theory of predication, more specifically the logic of 'is'.
Kretzmann distinguishes between three-piece predication and two-piece
predication, and discusses the various theories in whose terms Abelard
examines the role of the copula in three-piece predication and that of the
substantive verb in two-piece predication.
I cannot accept Kretzmann's attempt to prove that a development took
place in Abelard's thinking on these matters, one which started from an
"lngredientibus theory", then a "Dialectica revision" and a "Dialectica
theory", ending with a "Dialectica suggestion" . As Kretzmann indicates, the "lngredientibus theory" was not, even at the time when it
was written, Abelard's last word on the problems connected with the
theory of predication. The" Dialectica theory" , like the" Dialectica suggestion", can be documented by reference to the Logica 'lngredientibus'
(see above, 1.1.; 1.2.b., Note 52 and Note 55; 1.2.c. resp. 1.2.a.). The array of different theories in Abelard's work is not to be explained in
chronological terms. As the title of his article shows, Kretzmann sees
Abelard as an autumnal figure, a man reaping the harvest which had
been grown on the ground of the 'old logic'. I see Abelard as more flexible and more restless. In his discussions, he neglects no question which
has once been asked nor any suggested answer; but it is just through this
placing in juxtaposition of a variety of suggestions that he allows himself
to be led along quite new paths. And a theory which challenges the whole
of previous tradition as emphatically as does Abelard's two-piece theory
of predication must be placed more decisively in the foreground of his
work than appears in Kretzmann's paper.
The fact that it is possible precisely to describe the points of disagreement between us shows at the same time to what a great extent we are
in accord. In fact it is not only the manner of our approaches which is
very similar, nor do we merely both discover different theories of
predication in Abelard's thought; we also name, describe, and evaluate
them in ways which are largely similar. Such broad agreement encourages me to believe that in our attempts to understand this author,
we find ourselves moving in the right direction .

172

KLAUS JACOBI

NOTES
Parts of the following essay have been published in German; see Jacobi (1980) and Jacobi
(1981). The invitation to contribute to the present volume has given me a gratifying opportunity to reexamine my earlier research, to incorporate supplementary material, and to
strive toward greater precision and clarity. I wish to thank C. Sam Farler for preparing the
English translation.
ICf. Jolivet (1969), Chapter IV; de Rijk (1980). Also compare Haring (1975), who explains
the meager transmission of Abelard's works as at least partially attributable to Abelard's
style of thinking and writing. His philosophical "works" were not written as books intended to be recopied and handed down but as records of his own thinking to be used in
teaching. A thesis which he adheres to with conviction at one point in his writings may reappear later or even in a reworking of the first source as being subject to doubt or in need
of revision.
2pt' cus Abailardus, Logica 'Ingredicntibus', Super Peri ermenias, 336,27 - 340,18. Compare the notes of textual criticism in Jacobi (1981) pertaining to the edition of Abelard's
commentary on Peri Hermeneias from which I quote.
3 Cf. Petrus Abailardus, Dialectica, 118,1 - 120,20.
4 Super Peri erm., 337,11-32.
5 Loc. cit. 337,41-338,7.
6 Loc. cit. 336,24 - 37; 337,33 - 40. Cf. Nuchelmans (1973), pp. 140 - 141.
7 Super Peri erm., 338,21 - 339,4; cited 338,39 - 339,4: 'Si'vero vel 'et' sive aliae multae,
quid significant, non est promptum assignare vel cuius rei imaginem eorum intel/ectus
habeat subiectam. Si enim significant, utique intel/ectum constituunt, ad quod necesse est
esse vel rem vel imaginem rei, in qua nilatur intel/ectus. Sed cuiusmodi rem vel cuiusmodi
figmentum imaginis per 'si' sive per 'et' concipimuS, numquid ad modum substantiae vel
qualitatis vel ad quem modum, ut sanus consistat intel/ectus?
8 Loc. cit., 339,5 -7.
9 Loc. cit., 339,7- 11.
10 Loc. cit., 339,11-19.
11 Loc. cit., 339,20-32; cited 339,24-32: Licet ('est' et 'non est') intel/ectus non conslituant, quandam tamen coniunctionem vel disiunctionem intel/ectarum rerum in anima
haberijaciunt, quam tamen coniunctionem vel disiunctionem non significant, licet haberi
jaciant, quia intel/ectum non dant in se, sed intel/ectorum coadiunctionem vel separationem habere nos jaciunt. Sunt itaque tres actiones in intellectu proposilionis, intel/ectus
scilicet partium, coniunclio vel disiunctio intel/ectarum rerum. Nec est incongruum, si ea
actio, quae intel/ectus non est, sit pars intel/ectus totiusproposilionis. Cf. 339,32 - 340,2.
Cf. Nuchelmans (1973), pp. 141-142; Tweedale (1976), pp. 231-234.
12 Super Peri erm., 340,2 - 6: Sicut autem 'est'vel 'non est' coniunctiva vel disiunctiva
sunt, non significativa, ita 'si' vel 'non si' significativQS voces copulant vel separant, ut ipsa
tamen non significent, cum nullius rei in se conceptiones teneant sive verae sive fictae, sed
animum inclinant ad quendam concipiendi modum.
13 cr. Tweedale (1976), p. 234.
14 Cf. Kretzmann (1982).
15 Super Peri erm., 346,1-351,22.
16 Dial., 130,6-8; 130,26-131,3; 132,21-133,28.
17 Priscianus, Inst. gramm., 1118, t. I, p. 55,8-9; VIII I, t. I, p. 369,2-3; XVII 14, t.
II, p. 116,26 - 27. Cited by Abelard, Super Peri erm., 346, I - 2; Dial., 132,38 - 133,2.

PETER ABELARD ON THE SPEECH SIGN 'EST'

Ui

173

Super Peri erm., 346,1 - 4; Dial., 130,6 - 7.

Super Peri erm., 346,4 - 12. Cf. de Rijk (198Ia), p. 30.

20 Super Peri erm., 346,13 - 24: Verbum etiam substantivum 'sum, es, est' dupliciteraccipi
dicunt, cum modo in vi verbi sumitur, modo in vi substantivi. In vi quidem verbi sumitur,
cum aequivoce pro quibuslibet verbis sumitur tam actionem quam passionem significantibus, ut scilicet tantumdem valeat quantum 'amo' vel 'Iego' vel 'curro' vel 'amor' vel
'Iegor' etc. Cum vero in ea significatione sumitur in qua est substantivum, tunc omnia in
essentia significat. Unde secundum aliam significationem est verbum, cum videlicet
quaslibet actiones vel passiones ut adiacentes significat, et secundum aliam institutionem
substantivum, cum scilicet quaelibet in essentia significat tam actiones vel passiones quam
res alias, et tunc quoque aequivoce sumitur, ut tantumdem valeat quantum vel 'substantia'
vel 'quantitas' vel 'qualitas' etc. quae in essentia significant res. Cf. the excerpts from
Glosule super Priscianum maiorem, which de Rijk (1967), p. \02, gives.
21 Cf. de Rijk (198Ia), p. 30.
<22 Super Peri erm., 346,36-347,22; 348,2-14; cf. Dial., 132,36-133,28.
23 Aristoteles, De into c.3, 16 b 6 - 9, Arist. lat. 7,1 - 5; Petrus Abailardus, Super Peri
erm., 346,25 - 36; 347,38 - 348,2; 348,15 - 28; 350,40 - 351,2.
24 Super Peri erm., 346,32 - 36: Nam 'amans' in designatione eiusdem actionis et nomen
est et participium, nec differt nisi in significatione tan tum temporis, quando nomen est et
participium. Similiter 'ens' et nomen et participium esse potest in eadem significatione
essentiae, si modo cum tempore modo sine tempore sumitur. Cf. 347,2 - 6; 348,15 - 28.
25 Loc. cit., 348,28 - 349,1.
26 Loc. cit., 348,21 - 28: ... singula nomina secundum proprias significationes verba requirent, si essent, qui imponerent, substantiva quidem, ut 'esse substantiam, fuisse, fore'
vel'quantitatem' vel'qualitatem' vel 'corpus' vel 'colorem' vel 'hominem' vel'albedinem',
sumpta vero verba ut 'esse quale' vel 'fuisse' vel 'fore' vel 'colora tum , vel 'album', sicut
'ambulare' vel 'ambulasse' vel 'sedere' vel 'sedisse' dicimus pro 'esse' vel 'fuisse ambulantem' vel 'sedentem'. Nil enim, inquit Aristotell?s, differre 'hominam ambulare' et
'ambulantem esse'.
27 Loc. cit., 349,1-350,39; 351,\0-22; cf. Dial., 138,11-141,3. Cf. Tweedale (1976),
pp. 285-292; 298-302; de Rijk (I981a), p. 28.
28 Super Peri erm., 349,6 - 19; cf. Dial., 138,26 - 139,3.
29 Super Peri erm., 350,19-39.
30 Dial., 139,12-140,14.
31 Super Peri erm., 346,25 - 28: Nobis autem placet omnia verba dici ab Aristoteles
quaecumque cum aliquo tempus habent significare, et sicut nomina quaedam res in essentia
significant, quaedam ex adiacenti proprietate, ita etiam verba. Cf. the passage cited in Note
26. As shown there, substantives signifying properties on the one hand ('qualitatem . . .',
'colorem . . .', 'albedinem esselfuisselfore') and properties as attributable on the other
hand ('quale . . .', 'coloratum . . .', 'album esselfuisselfore') can be made into verbs by using the helping verb 'esse'. Cf. further 347,38 - 348,2: Sicut ergo nomina substantialia inveniuntur, ita et verba, et sicut nomina quaedam sumpta sunt, ita etiam verba non solum
ab actionibus vel passionibus, verum etiam ab aliisformis, ut 'sedeo' a 'sessione' sumptum
est.
32 Loc. cit., 347,23 - 26: Volumus itaque verbum substantivum in eadem significatione
retentum, scilicet qua omnia significat in essentia, et substantivum esse et verbum et idem
semper notari in ipso. Veluti cum dicitur 'homo est' et 'albedo est', eundem ubique sensum
'est' verbum tenet, ac si diceret 'est aliquid de numero existentium'. The explanation of
'est' by 'est aliquid de numero existentium' is to be found at this point only in manuscript
19

174

KLAUS JACOBI

B, not In manuscript A, upon which B. Geyer's edition is based. Compare also 362, 20 - 23,
which is also to be found in Geyer's edition.
33 Aristoteles, De into c.3 16 b 7; 16 b 9 -10. Translatio Boethii, Arist. lat., 7,2 - 3; 7,5 - 6:
et est semper eorum quae de altero praedicantur nota . ... et semper eorum quae de altero
dicuntur nota est, ut eorum quae de subiecto vel in subiecto.
34 Super Peri erm., 351,23 - 354,5; cf. 357,16 - 363,24.
3S Loc. cit., 352,4 - 7: In quo innuit (Aristotelesj verba maxime propter praedicationem inventa, nomina vero propter subiectionem, ut ex his integram proposition is constitutionem
doceat.
36 Loc. cit., 351,23 - 31: "Etest semper'. Datadefinitione, qua omne verbum includit tam
copulativum praedicati quam non, tam rectum quam casuale, tamfinitum quam infinitum,
quandam proprietatem verbi supponit, ex qua vim maximam in propositione praedicativa,
de qua intendit, verbum habere monstrat. Unde in sequentibus dicet nul/am enuntiationem
absque verbo consistere. Haec est autem proprietas, quod verbum semper est nota, id est
copula, praedicatorum de altero, id est copulativum est praedicatorum, quae praedicata de
altero quam de ipsis verbis copulantibus necesse est praedicari.
37 Dial., 129, 21 - 26: Quod itaque dixit verbum semper esse notam eorum quae de altero
praedicantur, omne verbum monstravit habere officium copulandi praedicatum subiecto
nec iIIud "semper" ad temporum, immo ad verborum comprehensionem referendum est.
Potest enim verbum per se proferri nec aliquid copulare; semper tamen secundum inventionem suam copulativum est.
38 Loc. cit., 121,28-29; 122, 13-21.
39 Loc. cit., 122, 17 - 21: Nomina . .. non . .. sicut verba tempus consignijicant, ut
scilicet, quemadmodum dictum est, primam signijicationem subiectis person is secundum
tempus distribuere dicantur. Sed cur non?
40 Loc. cit., 122,22 - 29: Sicut enim 'curro' vel 'currens' cursum circa personam tanquam
ei praesentialiter inhaerentem demonstrat, ita 'album' circa substantiam albedinem
tam(quam) praesentialiter inhaerentem determinat; non enim album nisi ex praesenti
albedine dicitur. Unde et tan tum 'albi' nomen dicere videtur, quantum quidem praesentialiter albedine est informatum, sicut et 'currens' in quodam praesentialiter cursum parlicipat. Sicut enim substantivi verbi signijicatio, cui quoque tempus adiunctum est, verbis
adiungitur, sic et nominibus videtur. Cf. Twecdale (1982), p. 144.
41 Dial., 123,9-15: Quod itaque tempus verbis accidit, hoc etiam nominibus congruit,
praesens scilicet, sive ea sint substantiae sive adiacentiae vocabula. Sicut enim 'album' ex
praesenti albedine datum est, ita etiam 'homo' ex praesenti substantia animalis rationalis
mortalis, et quem hominem dicis, iam animal rationale mortale ipsum ostendis et tantumdem 'hominis' vocabulum sonat, quantum quidem praesentialiter 'est animal rationale
mortale'. Cf. de Rijk (1981 b), pp. 29 - 30.
42 Dial., 123,2 - 5: Male ergo per "sine tempore" nomina, quae etiam temporis
designativa monstrantur, Aristoteles verbis disiunxisse videtur; eiusdem, inquam, temporis
consignijicativa cuius et verba, idest praesentis.
43 Cf. Jacobi (1981), pp. 56 - 59; 67 - 68.
44 Tweedale's (1982) interpretation is different. He understands Abelard as associating a
tense with the isolated noun, generally the present (p. 146: cf. p. 144). In this case the noun
must change its meaning when it comes after a copula in past or future tense (p. 146). In
this interpretation the wording of the passage has not been attended to with sufficienti:are.
Nor is it clear how a noun in isolation is to convey a time connotation in addition to its
main meaning. Moreover, in Super Peri erm., 349,31- 33, Abelard speaks out against the
very theory which Tweedale attributes to him: Si qllis autem dicat 'homo' per adconiunc-

PETER ABELARD ON THE SPEECH SIGN 'EST'

175

tionem verbi 'fuit' transire in significationem praeteritorum, quia verbum adconiunctum

tempus praeteritum significat, non videtur recto. Compare the further discussions at
349,33 - 350,5.
4S Dial., 123,15 - 25: Non tam igitur in significatione temporis nomen a verbo recedere
videtur quam in modo significandi. Verbum enim, quod solum inhaerentiam facit, in eo
tempus quoque designat quod inhaerentiam rei suae ad subiectam personam in tempore
denotat. Nomen autem inhaerendi significationem non tenet nec aliquid quemadmodum
verbum inhaerere proponit, etsi rem aliquando ut inhaerentem determinet, ut 'album'
albedinem tamquam adiacentem atque inhaerentem significat, non tamen vel adiacere vel
inhaerere proponit, sicut verbum facit, quod etiam substantivi verbi copulationem adiunctam propriae significationi continet; tantumdem enim 'currit' verbum proponit quantum
'est currens' dicit. Cf. 148,26-30; 149,20-26. Cf. Tweedale (1982), pp. 144-145.
46 Dial., 129,18 - 21: In quo quidem ipse (Aristoteles) monstravit omne verbum cum officio copulandi vel ea quae tantum dicuntur de subiecto - nec scilicet sunt in subiecto ut 'homo' et 'rationale', vel ea etiam quae sunt in subiecto, temporis quoque significationem continere.
47 Loc. cit., 123,20-21, cited in Note 45.
4H Within the context of a different theme, Abelard gives great attention to impersonal propositions. Compare Tweedale (1976), pp. 244-272; the same (1982), pp. 147-148.
49 Super Peri erm., 359,9 - 363,24; Dial., 134,28 - 141,3; cf. Super Peri erm., 351,31- 34;
351,37-40; Dial., 159,11-160,13; 161,24-165,9; 167,6-168,10; 170,21-30. Cf.
Tweedale (1976), pp. 228; 242 - 244; 292 - 297; the same (1982), pp. 145 - 147; de Rijk
(1981a), pp. 21-28; the same (1981b), pp. 33-38.
so De Rijk (1981b), p. 33, section heading; cf. the same (1981a), p. 21, section heading.
SI Dial., 134,30-32; 161,28-32; Super Peri erm., 359,22-28; 362,7-10.
S2 Dial., 134,28-31; 135,6; Super Peri erm., 362,7-8.
S3 Dial., 135,6 - 8; Super Peri erm., 362, 20 - 23.
S4 Super Peri erm., 359,32 - 360,2.
ss Dial., 134,28 - 34; Super Peri erm., 362,4 -7. In Dial., 138,17 - 20, this thesis is ascribed to Aristotle; cf. De into c.l1, 21 a 26 - 28, Arist. lat. 25,15 - 17.
S6 Super Peri erm., 360,15 - 18: cum in essentia quaelibet significet, numquam ei copulatio
essentiae deest, quia ubique per ipsum proponitur aliquid aliud esse, etiam quando adiectivis adiungitur, veluti cum dicitur: 'iste est albus'. cf. 361,12-25; Dial., 131,4-6;
131,23-25; 138,7-10.
S7 Cf. de Rijk (1981a), pp. 21-24; the same (1981b), pp. 33 - 35. In both places de Rijk
is speaking about Super Peri erm., 360,9-361,3.
58 Cf. Super Peri erm., 362,23-34; Dial., 135,11 - 13; 136,37-137,6; 162,10-12.
S9 Super Peri erm., 362,13 - 20.
60 Loc. cit., 360,18 - 361 ,3; see de Rijk (l981a), pp. 21 - 24; the same (1981 b), pp. 33 - 35.
Th,: background for this thesis is formed by the distinction between esse in subiecto and
dici de subiecto, which Aristotle makes in Categories, Chapter 2. Cf. Petrus Abailardus,
Logica 'Ingredientibus', Super Praedicamenta, 126,27 - 133,31; 145,25 - 146,18; Super
Peri erm., 352, 13 - 22; 352,35 - 353,2; Super Top. (ed. dal Pra), 274,10 - 275,29; see de
Rijk (1967), pp. 204-205.
61 Dial., 131,33 -132,13.
62 Super Peri erm., 361,12 - 25; Dial., 135,9 - 138,26; cf. de Rijk (198Ia), pp. 24- 28; the
same (1981b), pp. 35-37; Tweedale (1976), pp. 292-297; Dial., 162,16-18;
167,6-168,2; 168,11 -169,28.
63 Super Peri erm., 361,22 - 25; Dial., 136,32 - 36; 168, 21 - 25.

176

KLAUS JACOBI

Dial., 135,24 - 25; 135,36 -136,36; 168, II - 17.

Super Peri erm., 362,21 - 25: 'Est' verbum . . . secundo praedicalum id ex propria inventione tenet, ut tam existentia quam non existentia copulet. Copulet inquam quod intelligitur tantum ex supposita voce. Loc. cit., 362,32 - 34: Unde interpositum tertium nil
signijicationis in se tenet, quod in intellectu copulet, sed tan tum rem praedicati suppositi.
66 Dial., 136,37 - 137,8: At vero mihi omnis ilia verbi praedicatio per accidens atque impropria videtur, quando ipsum, ut dictum est, tertium adiacens interponitur, cum non rem,
ut dictum est, praedicatam contineat, sed solius copulae officium habeat, ut in ea quoque
qua dicitur: 'Petrus est homo' vel '.. .albus'. Nec quidem quantum ad eius interpretationem pertinet, ex eo quod dicitur: 'Petrus est homo', inferri potest 'Petrus est', sed fortasse quantum ad praedicationem 'hominis', quod existentis rei tan tum nomen est. Si enim,
quia verbum copula interponitur, simpliciter praedica(re}tur, et ex eo quod dicitur:
'chimaera est opinabilis' vel 'non existens', chimaera esse concederetur.
67 De Rijk (198Ia), p. 27 .
68 Dial., 138,11- 17.
69 Cf. Aristote1es, De into 'c.4, 16 b 26-27, Translatio Boethii, Arist. lat ., '7,20-21:
Oratio autem est vox signijicativa, cuius partium aliquid signijicativum est separatum. Cf.
Petrus Abailardus, Dial., 146,34 - 148,16.
70 Cf. Boethius, In Peri herm., II, 8,30-9,5; Petrus Abaihirdus, Dial., 148,18-151,4.
71 Cf. Boethius, loc. cit., 9,6-15; Petrus Abailardus, loc. cit., 151,5-153,2.
72 Aristoteles, De into c.4, 16 b 33 - 17 a 4, Arist. lat., 8,6 - 10. The formulation cited stems
from Boethius, Dediff top. I, 1174 B7-8 . Cf. Nuchelmans(l973), pp . 132-135.
73 Super Peri erm., 330,11-18; cf. 325,24-326,15; see also Nuche1mans (1973), p. 143.
74 Super Peri erm., 327,1-4: Nam 'homo currens' vel 'homo non currens' nil intellectu
discrepare videtur ab 'homo currit' vel 'homo non currit', quia eadem est animi conceptio
hic et ibi, in coniunctione scilicet cursus ad hominem vel disiunctione.
7S Loc. cit., 327,7-13; cf. 329,29-330,36.
76 Cf. loc. cit., 327,27 - 40.
77 Loc. cit., 327,14-23.
78 Loc. cit., 327,40 - 41: Unde ex modo proponendi enuntiationes sive propositiones maxime dicuntur. Cf. Dial., 149,11 - 16: Sed dico hoc ad perfectionem orationis non sufficere,
ut quasi adiacentem homini albedinem vel cursum determinemus, nisi etiam adiacere
dicamus, quod sine verbo fieri non contingit. In hoc enim verbum a participio abundat
quod non solum personam per impositionem demonstrat aut ei cohaerentem actionem vel
passionem significat, verum etiam cohaerere dicit.
79 Super Peri erm., 327,23 - 26: Similiter et cum dicimus 'verum est Socratem sedere',
oratio subiecta, quae est 'Socratem sedere', eundem intellectum constituit, quem 'Socrates
sedet', nee tamen modum enuntiandi habet affirmando vel negando, ut propositio did
possit.
80 Cf. Dial., 154,4-29; 204,29-205,30; see also Tweedale (1976), pp. 217-219;
265-266.
81 Super Peri erm., 307,26 - 30: Nomina .. . et verba duplicem signijicationem habent,
unam quidem de rebus, alteram de intellectibus. Res enim signijicant constituendo intellectus ad eas pertinentes, hoc est naturam aliquam earum vel proprietatem attendentes. Cf.
Nuchelmans (1973), pp. 139-140.
82 Super Peri erm., 326,30-31: Sanus autem est omnis intellectus tam simplex quam compositus, per quem attendimus, ut res se habet.
83 Loc. cit., 326,31 - 37, cited 326,33 - 34: . . . ut 'homo' quamdiu homo subsistit, sanum
intellectum generat. Cf. 321,21- 36; also 332,30 - 334,17. Cf. Nuche1mans (1973), p. 144.
64

6S

PETER ABELARD ON THE SPEECH SIGN 'EST'

177

Super Peri erm., 365,21-30; Dial., 157,17-22.

Super Peri erm., 366,6-27; 367,2-9; 368,1- 8; Dial., 160,14-21.
86 Super Peri erm., 366,27 -40; Dial., 157,23 -158,3.
87 Cf. lolivet (1969), pp. 77-85; 201-203; de Rijk (1975), pp. 549-555; Nuchelmans
(1973), pp. 150-161; Tweedale(1976), pp. 216-278.
88 Aristoteles, De into C. I, 16 a 17 -18, Arist. lat., 6,1-2.
89 Boethius, In Peri herm. I, 44,9 - 10; 44,15 - 16; II, 48,32 -49,4; 49,18 - 21.
90 Dial., 150,13 - 20; cf. Super Peri erm., 366,37 -40; Dial., 160,26 - 29.
91 Super Peri erm., 365,31-38; 367,9-20; 368,1-30; Dial., 160,14-15; 160,23.
92 Dial., 160,23 - 26.
93 Super Peri erm., 361,30 - 36; 370,4- 22; 390,11- 393,27; Super Peri erm. B, 22- 31.
94 Cf. Nuchelmans (1973), pp. 149 - 150; 153 - 154; Tweedale (1976), pp. 244 - 272.
95 Super Peri erm., 370,11-15; cf. Nuchelmans (1973), p. 155; Tweedale (1976), pp.
264-265.
96 Super Peri erm., 370,15-22.
97 Super Peri erm. B, 30; see also Tweedale (1976), pp. 263 - 264.
98' Super Peri erm. B, 31, p. 20,19-25: Dicimus itaque 'necessarium' sive 'possibile'
in huiusmodi enuntiationibus magis consignijicare quam per se signijicationem habere; nil
quippe in eis est intelligendum nisi subiecte orationi applicentur, et tunc modum concipiendi faciunt circa res subiecte orationis sicut facit verbum interpositum vel coniunctio, si que
ad necessitatem copulat.
99 Aristoteles, De into C. 12, 21a 9- 10, Arist. lat., 27,3 -4.
100 Dial., 161,8 - 15: Est autem categoricarum natura secundum membra sive species
demonstranda. Sunt autem membra ex quibus coniunctae sunt, praedicatum ac subiectum
atque ipsorum copula, secundum hoc scilicet quod verbum a praedicato seorsum per se accipimus, veluti in ea propositione qua dicitur: 'homo est animal', 'animal' praedicatur,
'homo' vero subicitur, verbum vero interpositum praedicatum subiecto copulat; et in his
quidem tribus categoricae proposition is sensus perficitur. Most of the substantiating
quotations used here and below are taken from Abelard's writings. The writings of
Abelard's contemporaries have been preserved and edited only fragmentarily. It is in any
case interesting and characteristic of the disputative style of Abelard's thinking that he
takes on the whole gamut of possible positions in his writings.
101 Super Peri erm., 339,20 - 32; 340,2 - 6: see above, pp. 147 - 148.
102 Super Peri erm., 362,14-34; Dial., 159,11-18; cf. above 1.2.c.
103 Dial., 162,18-21.
104 Dial., 162,16-18; compare above, p. 157.
105 IJial., 137,3 - 8, cited above, Note 66.
106 See the interpolation in Glosule super Priscianum maiorem (ed. by R. W. Hunt,
'Studies on Priscian in the Eleventh and Twelfth Centuries, I: Petrus Helias and His
Predecessors', Mediaeval 'and Renaissance Studies I, (1943), 194 - 231; here:
226,28 - 227,14) which de Rijk (1967), p. 103, quotes: Cum . .. dico 'Socrates est', si consideremus 'est' in vi verbi, actiones quas signijicat, idest lectionem et alias, sub disiunct/one
copulat cum Socrate; si autem in vi substantivi, Socratem existentibus annumerare, idest
quod Socrates est unum de existentibus dicere, in tendo . ... Nam iIIud (verbum substantivum) positum in propositione diversas essentias convenienter ad se invicem iungere affirmamus. Nam ex hoc quod omnes res in essentia signijicat, aptum est ad hoc ut quaslibet
res sibi coherentes copulare possit, verbi gratia cum dico 'Socrates est animal', excluso
respectu verbalis significationis, non prorsus tamen verbi proprietate separata, copulat
enim ipsum 'est', quantum -ad officium quod exercet in oratione in, vi substantivi con-

85

178

KLAUS JACOBI

sideramus. Unde aperte rem animalis cum re hominis copulat. Non tamen negamus idem
'est' consideratum in ipsa oratione vim verbi obtinere; sed aliud est agere de vocibus per
se consideratis, aliud de eisdem ad vim et officium quod habent in oratione posite relatis.
Nom quantum ad vim huius oration is 'homo est animal', 'est' non per se tan tum sed cum
aliis hoc solum significat, quod ilia res que est homo sit ilia res que est animal. Hoc autem
ex vi verbi habere non potest, immo ex vi substantivi. For Abelard's view compare Super
Perierm., 347,23-27; 360,13-18.
107 Super Peri erm., 357,36-358,3.
108 Dial., 135,23 - 35. In Dial., 168,11 -169, 28, these interpretations are introduced as being the opinions of Abelard's teacher V. (Ulger).
109 Super Peri erm., 361,19-25; compare above, p. 157.
110 Dial., 165,3-8; compare the further reference to the earlier discussion 170,21-31;
compare also the formulation at 167,7 - 8: secundum eos qui 'est' tertium adiacens
praedicato non componunt, sed dictionem per se sumunt.
III Super Peri erm., 359,22 - 28; 362,7 - \0; 362,20 - 23; Dial., 134,28 - 32; 161,28 - 32;
compare above, p. 155.
112 Super Peri erm., 361,33 - 362,7; 362,23 - 36; Dial., 134,32 - 135,6; 136,37 - 137,6;
138,11 - 17; compare above 1.2.c.
III Cf. Moody (1953), pp. 32- 38; de Rijk (1967), pp. \02-106; 183 -186; 203 -205;
561 - 562; the same (1956), pp. XXXVIII - XL; Pinborg (1972), pp. 46; 53 - 55; Maieru
(1972), pp. 199 - 206.
114 Aristotle's explication of the proposition based on the word 'hyparchei' surely forms
the background here.
115 See above, p. 157.
116 Cf. Aristoteles, Top. Icc. 13 and 15.
117 Cf. William of Sherwood, Syncategoremata, pp. 70-71; Thomas Aquinas, In Peri
herm., L.I, l.V, 70-73; see also Zimmermann (1971).
118 The countertheory to the equivocation theory is easily joined with metaphysical
theorems about transcendentals and the actus essendi in the thirteenth century.
119 See above p. 154.
120 Cf. Tweedale (1982), p. 149.
121 Dial., 139,12 - 140,14; cf. Super Peri erm., 350,35 - 39.
122 Cf. Dial., 160,17 - 19. At Dial., 279,8 - 282,33, Abelard distinguishes between the case
that no more men exist and the case that none have as yet existed. In the latter case, 'animaf
would have a different meaning, namely no impositio for men.

BIBLIOGRAPHY
Aristotle, De Interpretatione, ed. L. Minio-Paluello, Oxford, 1949; Translatio Boethii, ed.
L. Minio-Pa1uello, Aristoteles Latinus II 1 - 2 (Corpus Philosophorum p1edii aevi
Academiarum consociatarum auspiciis et consilio editum), Desclee de Brouwer,
Bruges - Paris, 1965.
Aristotle, Topica, ed. W. D. Ross, Oxford University Press, Oxford, 1958.
Boethius, Commentarii in lib rum Aristotelis nEPI EPMHNEIAE, ed. C. Meiser. Prima
editio [I], Teubner, Leipzig, 1877; Secunda editio [II], Teubner, Leipzig, 1880.
Boethius, De differentiis topicis, Patrologia Latina, t. 64.
Peter Abelard, Dialectica, ed. L. M. de Rijk, Van Gorcum, Assen. 1956,21970.
Peter Abelard, Glossae super Peri ermeneias, ed. B. Geyer, in Peter Abaelards

PETER ABELARD ON THE SPEECH SIGN 'EST'

179

Philosophische Schriften, I. Die Logica 'Ingredientibus', (BGPhThMA 21,3), Aschendorff, Miinster, 1927.
Peter Abelard, Glossae super Periermenias B, cappo XII - XIV, ed. L. Minio-Paluello, in
Twelfth Century Logic. Texts and Studies II: Abaelardiana Inedita, edizioni di Storia
e Letteratura, Roma, 1958.
Peter Abelard, Glossae super Praedicamenta Aristotelis, ed. B. Geyer, in Peter Abaelards
Philosophische Schriften I. Die Logica 'Ingredientibus' (BGPhThMA 21,2), Aschendorff, Miinster, 1921.
Peter Abelard, Glossae super Topica, ed. M. dal Pra, in Pietro Abelardo, Scrittijilosojici,
Fratelli Bocca, Roma - Milano, 1954.
Priscian, Institutionum grammaticarum libri XVIII, ed. M. Hertz (Grammatici Latini, rec.
H. Keil, II, III), Teubner, Leipzig, 1855.
Thomas Aquinas, In libros Peri Hermeneias Expositio, ed. T. M. Spiazzi, Marietti, Rome,
1950.
William of Sherwood, Syncategoremata, ed. J. R. O'Donnell, Mediaeval Studies 3 (1941),
46-93.
Haring, N. M., 'Abelard Yesterday and Today', in Pierre Abilard - Pierre Ie Venerable,
Actes du Colloque international, Cluny, 2-9 juillet 1972, ed. R. Louis, J. Jolivet, J.
Chatillon, Editions du Centre National de la Recherche Scientifique, Paris, 1975, pp.
341-403.
Jacobi, K., 'Diskussionen iiber Priidikationstheorie in den logischen Schriften des Petrus
Abailardus. Versuch einer Ubersicht', in Petrus Abaelardus (1079-1142). Person,
Werk und Wirkung, ed. R. Thomas (Trierer Theol. Studien, Bd. 38), Paulin us, Trier,
1980, pp. 165 -179.
Jacobi, K., 'Die Semantik sprachlicher Ausdriicke, Ausdrucksfolgen und Aussagen in
Abailards Kommentar zu Peri hermeneias', Medioevo (1981), 41-89.
Jolivet, J., Arts du Langage et tMologie chez Abelard (Etudes de Philosophie Medievale
57), J. Vrin, Paris, 1969.
Kretzmann, N., 'Syncategoremata, exponibilia, sophismata', in The C-ambridge History of
Later Medieval Philosophy, ed. N. Kretzmann, A. Kenny, J. Pinborg, Cambridge
University Press, Cambridge, 1982, pp. 211-241.
Maieril, A., Terminologia logica della tarda scolastica, Edizioni dell' Ateneo, Roma, 1972.
Moody, E. A., Truth and Consequence in Mediaeval Logic (Studies in Logic and the Foundations of Mathematics), North-Holland Publ. Co., Amsterdam, 1953.
Nuchelmans, G., Theories of the Proposition. Ancient and medieval conceptions of the
bearers of truth and falsity, North-Holland Publ. Co., Amsterdam, 1973.
Pinborg, J., Logik und Semantik in Mittelalter. Ein Uberblick. Frommann-Holzboog,
Stuttgart - Bad Canstatt, 1972.
Rijk, L. M. de, Logica Modernorum. I1- J: The Origin and Early Development of the
Theory of Supposition, Van Gorcum, Assen, 1%7.
Rijk, L. M. de, 'La signification de la proposition (dictum propositionis) chez Abelard',
in Pierre Abilard - Pierre Ie venerable, Actes du Colloque International, Cluny, 2 - 9
Juillet 1972, ed. R. Louis, J. Jolivet, J. Chatillon, Editions du Centre National de la
Recherche Scientifique, Paris, 1975, pp. 547 - 555.
Rijk, L. M. de, 'Peter AbiiIard (1079-1142): Meister und Opfer des Scharfsinns', in
Petrus Abaelardus (1079- J142), Person, Werk und Wirkung, ed. R. Thomas (Trierer
Theol. Studien, Bd. 38), Paulinus, Trier, 1980, pp. 125 -138.
Rijk, L. M. de, 'Die Wirkung der neuplatonischen Semantik auf das mittelalterliche
Denken iiber das Sein', in Sprache und Erkenntnis im Mittelalter. Miscellanea
Mediaevaiia, Bd. 13/1, De Gruyter, Berlin-New York, 1981, pp. 19-35.

180

KLAUS JACOBI

Rijk, L. M. de, 'Abailard's Semantic Views in the Light of Later Developments', in

English Logic and Semantics. Acts oJ the 4th European Symposium on Mediaeval Logic
and Semantics, Leiden - Nijmegen, 23 - 27 April 1979, ed. by H. A. G. Braakhuis,
C. H. Kneepkens, and L. M. de Rijk (Artistarium Supplementa I), Nijmegen, 1981, pp.
I-58.
Tweedale, M. M., Abailard on Universals, North-Holland Pub!. Co., Amsterdam - New
York - Oxford, 1976.
Tweedale, M. M., 'Abelard and the Culmination of the Old Logic", in The Cambridge
History oj Later Medieval Philosophy, ed. N. Kretzmann, A. Kenny, and J. Pinborg,
Cambridge University Press, Cambridge, 1982, pp. 143 -157.
Zimmermann, A., "Ipsum enim ('est') nihil est' (Aristote1es Periherm. I, c. 3). Thomas
von Aquin tiber die Bedeutung der Kopula', in Miscellanea Mediaevalia 8, Walter de
Gruyter, Berlin - New York, 1971, pp. 282 - 295.

Philosophisches Seminar II
Albert-Ludwigs-Universitiit
Werthmannplatz
D-7800 Freiburg im Breisgau, B.R.D.

HERMANN WEIDEMANN

THE LOGIC OF BEING IN THOMAS AQUINAS

Being acquainted with the familiar distinction between the "is" of existence, the "is" of predication, and the "is" of identity, which Hintikka
has labeled "the Frege trichotomy" (1979, pp. 433f), a modern student
of Thomas Aquinas's doctrine of being cannot fail to realize that this
distinction, though it seems not to have been ignored by Aquinas, is overshadowed in his writings by another distinction between two semantically different uses of the verb "be", which he borrows from Aristotle. My
aim in this paper is, first, to examine how the two distinctions are related
to one another; secondly, to show that Aquinas, though drawing these
distinctions, does not commit himself to the assumption that the verb
"be" is genuinely ambiguous! and, finally, to elucidate how Aquinas
avoids such a commitment.
Since it is an ontological and to a certain extent even a theological
rather than a logical point of view from which Aquinas approaches the
problem of the semantically different uses of the verb "be", what he has
to say concerning the logic of being is split up into a lot of scattered
remarks, mainly the by-product of metaphysical reflections, from the
larger context of which. they have to be gathered and put together like the
pieces of a jigsaw puzzle. In view of this, not a few of the moves I shall
make in my following attempt to trace a coherent picture of Aquinas's
logical treatment of the verb "be" will be little more than conjectural.

The distinction which plays the predominant role in Aquinas's theory of

being is twofold rather than threefold and has indeed been called "a fundamental ontological dichotomy in Thomas Aquinas's thought" (Veres,
1970). In accordance with this dichotomy, for which he invokes the
authority of Aristotle (Metaph . .:l 7, 1017 a 22-35)2, we have to
distinguish between the use we make of the verb "be" to express the being of something which falls under one of the ten categories (i.e. the being of something which is either a substance, e.g. a man, or an accidental
property, e.g. the white colour of a man), on the one hand, and the use
181
S. Knuuttila and J. Hintikka (eds.), The Logic of Being, 181-200.
1986 by D. Reidel Publishing Company.

182

HERMANN WEIDEMANN

we make of the verb "be" to express the truth (i.e. the being the case)
of a true proposition,3 on the other hand:
Philosoph us, in V Metaphys. [... J, ostendit quod ens multipliciter dicitur. Uno enim
modo dicitur ens quod per decem genera dividitur: et sic ens significat aliquid in natura existens, sive sit substantia, ut homo, sive accidens, ut color. Alio modo dicitur ens, quod
significat veritatem propositionis ... (In II Sent. dis!. 34, q. I, a. I, C).4

Used in the first way, we are told by Aquinas, the verb "be" refers "to
the act of a being thing insofar as it is being, i.e. to that by which
something is called actually being in reality" (esse dicitur actus en tis in
quantum est ens, idest quo denominatur aliquid ens actu in rerum natura:
Quodl. IX, q. 2, a. 2 [3], c), e.g., to the act of living which is the being
of whatever is alive (In I Sent. dist. 33, q. 1, a. 1, ad 1). Understood in
this sense the term "being thing" (ens) applies to "something naturally
existing, be it a substance, like a man, or an accidental property, like a
colour" (In II Sent. dist. 34, q. 1, a. 1, c). Used in the second way,
however, the verb "be" serves to answer the question whether there is
(an est) such and such a thing (ibid., cf. S.th. I, q. 48,.a. 2, ad 2, De malo
q. 1, a. 1, ad 19). The distinction between the two uses of the verb "be"
so far considered thus amounts to distinguishing between two different
existential uses of this verb, which we may, following Peter Geach, call
its use in an actuality sense and its use in a there-is sense, respectively. 5
Whereas an entity A falling under one of the ten categories in virtue
of its having an essence, can be said to be in the sense that A actually exists as well as. in the sense that there is such a thing as A, an entity A
which lacks an essence, because it is rather a privation of some being than
a being in itself, can be said to be only in the latter sense. Sentences of
the form "A is" (i.e. "A exists"), where "A", to judge from the examples chosen by Aquinas, is a placeholder for concrete or abstract
general terms, like "(a) man" or "(a) colour", and for concrete or
abstract singular terms, like "Socrates" or "blindness", can thus be put
to two different uses. In analysing the second use Aquinas comes close
to the modern analysis of existence statements in terms of the existential
quantifier. What we convey by saying that blindness exists, to take one
of Aquinas's favourite examples, is, according to him, nothing but the
fact that it is true to say that something is blind (Dicitur enim, quod
caecitas est secundo modo, ex eo quod vera est propositio, qua dicitur
aliquid esse caecum: In V Metaph. lect. 9, no. 896; cf. De pot. q. 7, a.
2, ad 1). Being the privation of sight, blindness is not being in the sense
that it actually exists, however, because - unlike an animal that happens
to be blind - it has no essence which could be actualized in reality. Since

THE LOGIC OF BEING IN THOMAS AQUINAS

183

Socrates, e.g., has such an essence (or nature), to say that Socrates exists
(" Socrates est") is either to say, on the one hand, that he actually exists
(or is alive) as the human being that he essentially is, or to say, on the
other hand, that there is such a person as Socrates (cf. In V Metaph. lect.
9, no. 896).6
Of these two uses of the verb "be" the latter is regarded by Aquinas
as the more comprehensive; for whatever can be said to be In the sense
that it exists as a substantial thing or as an accidental property of such
a thing can also be said to be in the sense that its existence can be truly
affirmed, but not the other way round. The existence of blindness, for
example, can be truly affirmed by saying that there is blindness (caecitas
est); but none the less blindness is not to be found among the entities
which belong to the furniture of our actual world, because it is rather a
lack of actuality than something actually existing (cf. In II Sent. dist. 34,
q. 1, a. 1, c).
His distinction between these two existential uses of the verb "be"
aside, Aquinas is well aware of the difference between the "is" of
predication and the' 'is" of identity, as witness a couple of texts in which
he distinguishes between something's being predicated of something' 'in
the way of an identity" (per modum identitatis: S.th. I, q. 39, a. 5, ad
4) and something's being predicated of something "in the manner in
which a universal thing is predicated of a particular one" (sicut universale de particulari: ibid.). A predication of the latter kind, which in contradistinction to the so-called praedicatio per identitatem (S.th. I, q. 39,
a. 6, ad 2, In III Sent. dist. 7, q. 1, a. 1, c) he also calls (praedicatio) per
denominationem sive informationem (In III Sent. dist. 5, expos. textus),
is for Aquinas a predication "more properly" so called (magis propria
praedicatio: ibid.).7
Concerning predications properly so called, Aquinas draws a distinction between substantial and accidental predications, which embraces the
distinction between the two existential uses of the verb "be" already
mentioned, in that this verb, according to whether it is used in its actuality senseor in its there-is sense, either functions as a substantial predicate,
which corresponds to the question "What is ... ?" (quid est?), or as an
accidental one, corresponding to the question "Is there such a thing
as ... ?" (an est?) (cf. In II Sent. dist. 34, q. I, a. 1, c; In V Metaph. lect.
9, no. 896; De malo q. 1, a. 1, ad 19).
Aquinas's view that to use the verb "be" in its actuality sense and,
hence, existentially is to use it as a predicate which .corresponds to the
question of what a given thing is might seem rather odd. It is explained

184

HERMANN WEIDEMANN

by the fact that, to Aquinas's mind, it is only in accordance with a thing's

essence or nature that actual being (or existence) belongs to a thing; for
the actual being of a thing is nothing but the actuality of that thing's
essence (actus essentiae: In I Sent. dist. 33, q. 1, a. 1, ad 1). Despite this
intimate connection between a thing's having an essence and its being actually existent, which accounts for Aquinas's tendency to assimilate the
use of the verb "be" in its actuality sense with its use as a substantial
predicate, there is a difference between the essence of a thing and its actual existence, which is the actuality of its essence. This difference is
taken into account by Aquinas when, instead of drawing his standard
twofold distinction between the use we make of the verb "be" to refer
to "the essence of a thing or (to) its act of being", on the one hand, and
the use we make of it to refer to "the truth of a proposition", on the
other hand (De pot. q. 7, a. 2, ad 1), he occasionally contrasts being as
being true with being as being what a thing is (i.e. with the nature or
essence of a thing), on the one hand, and with being as being actually
existent (i.e. with the actuality of a thing's essence), on the other hand:
... esse dicitur tripliciter. Uno modo dicitur esse ipsa quidditas vel natura rei [... 1. Ali"
modo dicitur esse ipse actus essentiae [... 1. Tertio modo dicitur esse quod signific.!'
veri tat em compositionis in propositionibus, secundum quod 'est' dicitur copula (In I Sen'
dis!. 33, q. I, a. I, ad I).

As regards Aquinas's doctrine that in answering the question whether

there is such and such a thing we are using the verb "be" as an accidental
predicate which signifies the truth of a proposition, it is not selfexplanatory either, especially since it is this use of the verb "be" that is
taken by Aquinas to be its copulative one (ibid.; cf. In III Sent. dist. 6,
q. 2, a. 2, c, In V Metaph. lect. 9, no. 896, S.th. I, q. 48, a. 2, ad 2). For
a modern reader who has become accustomed to distinguishing the
copulative use of "is" and its use in the there-is sense as its respective
predicative and existential uses from its so-called veridicaluse 8 such a
doctrine looks highly implausible. Some comments on the view taken by
Aquinas are therefore in order.
From Aquinas's point of view it is by "signifying a propositional combination, additionally invented by the (human) soul in the act of linking
a predicate to a subject" (signijicat composition em proposition is, quam
anima adinvenit coniungens praedicatum subiecto: S. tho I, q. 3, a. 4, ad
2), that the copula signifies the truth of the sentences in which it occurs
(signijicat veritatem proposition is: S.th. I, q. 48, a. 2, ad 2). That is to
say that in its copulative use the verb "be" signifies that a predicate is
said to be true of a subject or, more accurately, that a predicate-term

THE LOGIC OF BEING IN THOMAS AQUINAS

185

is said to be true of what a subject-term stands for. 9 What Aquinas is obviously trying to say in this account of the copula is that, used in a simple
sentence of the form" Sis P" to make a statement, the copulative "is"
performs not only the predicative function of establishing a "propositional combination" 10 between the predicate-expression "P" and the
subject-expression" S" (where" S" stands proxy for general terms, like
"[a] man", II as well as for singular ones, like "Socrates"), but also the
assertive function of indicating the speaker's commitment to the truth
of what is expressed by the propositional combination of "S" and" P".
In other words, when used to make the statement that Sis P, the copula
serves not only as a propositional link between" S" and" P", but also
as an expression for the truth-claim which, whether it is warranted or
not, i.e., whether the sentence "s is P" is true or not, we are making to
the effect that" P" is true of what" S" refers to whenever we assert that
Sis P.
As for Aquinas's doctrine that it is by means of the copulative "is"
that the question whether there is such and such a thing is answered (cf.
S.th. I, q. 48, a. 2, ad 2), which seems to result from a confusion between
the predicative and the existential uses of the verb "be", it is accounted
for by the fact that existence statements of the form "There is F-ness"
or "F-ness exists" (e.g. "Blindness exists") are analysed by Aquinas in
terms of the corresponding predicative statements of the form
"Something is F', in which the "is" functions as copula. When we assert
the existence of a privation (e.g. the existence of blindness, which is the
privation of sight), "our intellect links that privation", we are told by
Aquinas, "to a subject as being something like a form (thereof)" (in-

tellectus componit privationem cum subiecto, sicut formam quamdam:

In II Sent. dist. 37, q. 1, a. 2, ad 3). This analysis can be expanded in
a straightforward way to the effect that any predicative statement of the
form" Sis P" can be viewed as the copulative counterpart of an existence
statement of the form "There is such a thing as S (that is) P", to which
it is equivalent.
That Aquinas indeed assumes such an equivalence to obtain is suggested, e.g., by his account of what it is for something coincidentally to
be. What is called "coincidentally being" (ens secundum [or per] accidens) is so called, Aquinas says in his Commentary on Aristotle's
Metaphysics (In V Metaph. lect. 9, no. 885), "on account of a comparison of an accident (i.e. an accidental property) to a substance; and
this comparison is signified by the verb 'is', when (a sentence: like) 'a man
is white' is uttered; whence this whole thing (called) a white man is

186

HERMANN WEIDEMANN

(something which is) coincidentally being":

ens secundum accidens prout hic sumitur, oportet accipi per comparationem accidentis ad
substantiam. Quae Quid em comparatio significatur hoc verbo 'est' cum dicitur 'homo est
a1bus'. Unde hoc totum, homo lest) 12 albus, est ens per accidens (ibid.).

This account seems to rest on the assumption that the predicative statement that a man is white is equivalent to the existence statement that
there is such a thing as a white man (or, in other words, such a thing as
a man who is white).
The relation between the so-called "Frege trichotomy" and what has
been called Aquinas's "fundamental ontological dichotomy" may be
depicted by the following diagram:

~"iS"Of

existence

predication
(properly so
called)

(predication of)
identity

in act

actuality sense

substantial

substantial (?)

true

there-is sense

accidental

accidental (?)

to be

II

In the preceding section it was shown that Aquinas's dichotomy between

being as actually being (or being in act) and being as being true is related
to the familiar Frege trichotomy between the "is" of existence, the "is"
of predication, and the "is" of identity in such a way that each member
of the former division may be subdivided in accordance with the latter
one, to the effect that the verb "be" can be put to two different existential as well as to two different predicative uses (presumably including
what Aquinas calls predications "per modum identitatis" [S.th. I, q. 39,
a. 5, ad 4]).
Does Aquinas, when distinguishing between different uses of the verb
"be", ever imply that this verb is genuinely ambiguous? The passages of
his writings so far considered favour a negative answer to this question,
which, as a matter of fact, is expressly given by Aquinas in his Commentary on Aristotle's Peri hermeneias (De interpretatione). Commenting on
the passage in which Aristotle says that "by itself" the word "being" is
"nothing" (Ch. 3, l6b 23f), Aquinas rejects Alexander of Aphrodisias'
explanation according to which the word "being" is said to be
"nothing" because it is "said equivocally of the ten categories": ...

THE LOGIC OF BEING IN THOMAS AQUINAS

187

scilicet ipsum ens, de quo dicit quod nihil est (ut Alexander exponit), quia
ens aequivoce dicitur de decem praedicamentis (In I Peri herm. lect. 5,
no. 70 [19]). "This explanation", Aquinas says, "does not seem to be
appropriate, for in the first place, 'being' is not, strictly speaking, said
equivocally, but according to the prior and posterior; whence, said absolutely, it is understood of that of which it is said primarily": t3
haec expositio non videtur conveniens, tum quia ens non dicitur proprie aequivoce, sed
secundum prius et posterius; unde simpliciter dictum intelligitur de eo, quod per prius
dicitur (ibid.) .

At first sight the passage just quoted might suggest that Aquinas'~ rejection of the view that the word "being" is used, strictly speaking, in an
equivocal or ambiguous way is meant by him to be confined to the use
we make of the verb "be" to express the actual existence of what falls
under one or the other of the ten categories. That this is by no means the
case, however, is shown by the fact, already pointed out, that on both
sides of his dichotomy Aquinas tends to assimilate the existential and the
predicative uses of the verb "be" by attributing to it, on the one hand,
the role of a substantial predicate when it is used in its actuality sense
and, on the other h.md, the role of an accidental predicate when it is used
in its there-is sense. In addition, Aquinas assimilates the predicative use
which, according to him, is made of the verb "be" even in a statement
of identity with the predicative use of this verb properly so called in the
following way:
In every true affirmative statement the predicate( -term) and the subject(-term) must in
some way signify what is really the same but conceptually different. That this is so is clear
in the case of statements whose llTedicate is an accidental one as well as in the case of those
whose predicate is a substantial one. For it is obvious that (the terms) 'man' and 'white'
(e.g.) are the same as regards the subject-thing (they signify), but different with regard to
the (respective) concept (under which they signify it); for the concept of (a) man is a different concept from that of (a) white (object). Similar considerations apply, when I utter
(the sentence) '(A) man is an animal'; for the very same thing which is a man is truly an
animal, because in one and the same subject-thing there is to be found both a sensitive
nature, on account of which it is called an animal, and a rational nature, on account of
which it is called a man. Hence in this case, too, predicate and subject are the same as
regards the subject-thing (they signify), but different with regard to the (respective) concept
(under which they signify it).
Even with statements in which the same thing is predicated of itself, this is in some way
the case. insofar as our intellect treats what it assigns to the subject(-position) as being on
the side of a subject-thing, whereas what it assigns to the predicate(-positilln) is treated bv
it as belonging to the nature of a form existing in a subject-thing, in 3(cordance "ith th~
saying that predicates are taken formally and subjects materially.
While it is the plurality of predicate and subject which answers to the conceptual difference, it is the (propositional) comhination (of subject and predicate) hy means of which
our intellect signifies the real identity (S.th. I. q. 13, a. 12. C).14

188

HERMANN WEIDEMANN

The Latin text runs as follows:

in qualibet propositione affirmativa vera oportet quod praedicatum et subiectum significent idem secundum rem aliquo modo, et diversum secundum rationem. Et hoc patet tam
in propositionibus quae sunt de praedicato accidentali quam in illis quae sunt de praedicato
substantiali. Manifestum est enim quod 'homo' et 'albus' sunt idem subiecto et differunt
ratione; alia enim est ratio hominis et alia ratio albi. Et similiter cum dico 'homo est
animal'; illud enim ipsum quod est homo vere animal est; in eodem enim supposito est et
natura sensibilis, a qua dicitur animal, et rationalis, a qua dicitur homo; unde hic etiam
praedicatum et subiectum sunt idem supposito, sed diversa ratione.
Sed et in propositionibus in qui bus idem praedicatur de se ipso, hoc aliquo modo invenitur, inquantum intellectus id quod ponit ex parte subiecti trahit ad partem suppositi,
quod vero ponit ex parte praedicati trahit ad naturam forrnae in supposito existentis, secundum quod dicitur quod praedicata tenentur forrnaliter, et subiecta materialiter.
Huic vero diversitati quae est secundum rationem, respondet pluralitas praedicati et
subiecti; identitatem vero rei significat intellectus per ipsam compositionem. IS

According to this text, there is a close affinity between predicative

statements properly so called and statements of identity to the effect that,
on the one hand, a predicative statement involves what has been called
an "identity factor in predication" 16 insofar as it states the identity of
what its subject-term stands for with what its predicate-term is true of,
whereas, on the other hand, a statement of identity exhibits a predicative
feature in that the identity it states is expressed as the inherence of a form
(or property) in a thing being subjected to it. 17 That is to say - to make
explicit what Aquinas seems to be implying - that, just as the predicative
statement that Socrates is white states the inherence in Socrates of the
property of whiteness as the identity of Socrates with something white, 18
the identity statement that Socrates is Socrates states the identity of
Socrates with himself as the inherence in Socrates of the property of being Socrates. 19
If Aquinas does not regard the verb "be" to be genuinely ambiguous
as far as the so-called Frege trichotomy of the different uses of "is" into
the "is" of existence, the "is" of predication, and the "is" of identity
is concerned, does he not assume a genuine ambiguity to obtain at least
with respect to his own dichotomy of the different uses of "is" into the
"is" as applied to what falls under one of the ten categories, on the one
hand, and the "is" as applied to a true proposition, on the other hand?
That he in fact does so, is suggested by his saying that Avicenna (= Ibn
Sina) "has been deceived by the equivocation of being (i.e. of the verb
'be')" (deceptus est ex aequivocatione entis: In X Metaph. lect. 3, no.
1982), which is equivocal, according to Aquinas, in that "the 'is' which
signifies a propositional combination is an accidental predicate",
whereas "the 'is' which is divided by the ten categories (i.e., which has

THE LOGIC OF BEING IN THOMAS AQUINAS

189

its meaning specified by the kind of entity whose being it affirms)

signifies the very natures of the ten genera" (ibid.).
That the phrase "aequivocatio entis", as it is used here by Aquinas,
is not to be understood in the sense of a genuine ambiguity of the verb
"be", however, but rather in the sense of a "multiplicity of
applications,,20 of this verb, is attested by the comparison which
Aquinas makes between the two members of his "fundamental ontological dichotomy,,21 in several passages of his writings.
Since this comparison throws some light on the reason why in sticking
to the dichotomy at issue Aquinas does not commit himself to the view
that with respect to it the verb "be" is genuinely ambiguous, it is worth
entering into its details. This will be the task for the next section.
III

Of the two modes of being that the verb "be" can be used to signify,
namely being as being actually existent under one of the ten categories
and being as being true, "the second is comparable to the first", we are
told by Aquinas, "as an effect is to its cause; for it is from a thing's being
in reality that the propositional truth and falsity follows which our intellect signifies by means of the word 'is' insofar as it is (used as) a verbal
copula":
... iste secundus modus comparatur ad prim urn sicut effectus ad causam. Ex hoc enim
quod aliquid in rerum natura est, sequitur veritas et falsitas in propositione, quam intellectus significat per hoc verbum 'est' prout est verbal is copula (In V Metaph. lec!. 9, no. 896).

The same point is made in Aquinas's Commentary on the Sentences

of Peter Lombard, where we read that "as signifying the truth of a propositional combination, in which respect (the word) 'is' is called the
copula, being is to be found in the combining and dividing intellect (i.e.
in the intellect forming affirmative and negative propositions) as the
completion of truth, but founded upon the being of the thing (thought
of), which is the act of (that thing's) essence":
Tertio modo dicitur esse quod significat veritatem compositionis in propositionibus, secundum quod 'est' dicitur copula; et secundum hoc est in intellectu componente et dividente
quantum ad veri 22 complementum, sed fundatur in esse rei, quod est actus essentiae (In I
Sent. dis!. 33, q. I, a. I. ad I).

What prevents Aquinas's ontological dichotomy from reflecting a genuine ambiguity of the verb "be" is, according to the passage just
quoted, the fact that being as actually existing (or falling under one of
the ten categories), on the one hand, and being as'being true (or being

190

HERMANN WEIDEMANN

the case), on the other hand, are two modes of being of which the latter
depends on the former in such a way that the truth of what we say is effected by and founded upon the actual existence of what we talk about.
As an attempt to account for the non-ambiguity of the verb "be" the
passage to be considered next deserves especial attention. Commenting
on Aristotle's statement, made in Peri herm. (De int.) 3, 16 b 24, that the
word "being" (that is to say, the verb "is") "co-signifies some (propositional) combination" (or "composition"), Aquinas remarks that, according to Aristotle, it does so because
it does not signify such a composition principally but consequently. It primarily signifies
that which is perceived by our intellect in the mode of actuality absolutely; for 'is', said
simply, signifies to be in act, and therefore signifies in the mode of a verb. However, the
actuality which the verb 'is' principally signifies is the actuality of every form or act commonly, whether substantial or accidental. Hence, when we wish to signify that any form
or act actually is in some subject, we signify it by means of the verb 'is' [... 1; and for this
reason the verb 'is' consequently signifies a (propositional) composition (In I Peri herm.
lect. 5, no. 73(22)).23

The Latin text runs as follows: 24

Ideo autem dicit [Aristotelesl quod hoc verbum 'est' consignificat compositionem, quia
non earn principaliter signlficat, sed ex consequenti; significat enim primo illud quod cadit
in intellectu per modum actualitatis absolute; nam 'est', simpliciter dictum, significat in actu esse; et ideo significat per modum verbi. Quia vero actualitas, quam principaliter
significat hoc verbum 'est', est communiter actualitas omnis formae 25 vel actus substantialis vel accidentalis, inde est quod cum volumus significare quamcumque formam vel actum actualiter inesse alicui subiecto, significamus illud per hoc verbum 'est' [... 1. Et ideo
ex consequenti hoc verbum 'est' significat compositionem.

Far from being an interpretation which could be said to reveal [he proper
sense in which Aristotle himself intended his saying that the word
"be(ing)" co-signifies (i.e. additionally signifies) some (propositional)
combination to be understood 26 the text just quoted is nevertheless
peculiarly enlightening with regard to Aquinas's own view of the relation
between being in the absolute sense of actually being (or being actually
existent) and being in the copulative sense of something's being.. true of
something else. In addition to what we have already been informed of,
namely the fact that being in the latter sense depends on being in the
former sense as an effect depends on its cause, the passage under consideration gives us the reason why it is by means of one and the same
word (namely the verb "be") that these two distinct senses can be
expressed.
For Aquinas, the actuality sense of the verb "be" seems to be, as it
were, its "focal meaning" (to borrow G. E. L. Owen's happy term),27

THE LOGIC OF BEING IN THOMAS AQUINAS

191

i.e. its primary sense, by reference to which it is secondarily and

derivatively used in whatever other sense it may have. If a given use or
sense of a word can be shown to be derivative "if it can be explained how
the word in question comes to have the sense that it has by reference to
another sense which is for that reason primary" ,28 Aquinas's account of
the relation between the use we make of the verb "be" in its actuality
sense and the use we make of it in its copulative sense can surely be said
to be an account of the derivability of the latter sense from the former
and, hence, an account of the fact that the verb "be" is not genuinely
ambiguous.
Since to use the verb "be" as copula (i.e. as a sign of propositional
combination) is for him "to signify that a (certain) form or act actual1y
is in some subject" (significare . .. formam vel acturrtactualiter in esse
alicui subiecto) , Aquinas is indeed able to explain the fact that the verb
"be" can be put to this copulative use by reference to the absolute use
we make of it to signify that something is actual1y existent - be it a
substance, the actual existence of which is the actuality of the substantial
form it has (or, for that matter, the actuality of the substantial act it performs, e.g., the actuality of the act of living performed by an animal; cf.
In I Sent. dist. 33, q. I, a. I, ad I), or an accidental property, the actual
existence of which is the actuality of the accidental form (or act) it identical1y is (e.g., the actuality of the form of whiteness or of the act of
running).
IV

A final comment on Aquinas's treatment of the verb "be", which wil1

serve to round out the picture, may conveniently be added to what has
been said so far. It must be noticed, to begin with, that, according to
Aquinas, the mode of being signified by the copulative "is" docs not
belong to the reality of our world, but only to the proposition-forming
activity of our intellectual soul: As opposed to the mode of being which
is "the act of a being thing insofar as it is being" (actus entis ill quantulII
est ellS: Quodl. IX, q. 2, a. 2 [3]. c), the copulative being (esse . .. , secundum qllO(} est copula verhalis: ibid.) is not to be found ill rerum narura,
but only ill actu anil71ae cOl71ponentis et dil'identis (ibid.; cf. In III Sellt.
dist. 6, q. 2, a. 2, c: "hoc esse non est in re, sed in mente quae coniungit
subicctunl cum pracdicato").
Since to be in the copulati\e sense of something's being something i"
for Aquina), to be trill' in the sense of its being the case (or a fal't) that

192

HERMANN WEIDEMANN

something is something, his view that the copulative being is, as it were,
only an intellectual being, "additionally invented by the (human) soul in
the act of linking a predicate to a subject" (S. tho I, q. 3, a. 4, ad 2), commits him to excluding what we would call facts from his ontology, reserving to them, in a sense which reminds one of Frege's dictum that "a fact
is a thought that is true",29 the peculiar status of true thoughts; or so it
seems.
That it is only substances and their substantial forms, on the one hand,
and accidental properties of substances on the other that Aquinas is willing to admit into his ontology is obvious from his assumption that only
things having an essence or a form, which makes them fall under one of
the ten categories, can be said actually to be (cf. Quodl. IX, q. 2, a. 2[3),
c; In II Sent. dist. 37, q. 1, a. 2, ad 3), because the actual existence of
anything is nothing but the actuality of its essence (cf. In I Sent. dist. 33,
q. 1, a. I, ad I) or its form (cf. In I Peri herm. lect. 5, no. 73[22]).
As for Aquinas's view that facts, as opposed to particular things and
their properties, are not real entities but mental ones to which something
in reality coaesponds, witness the account he gives of St Augustine's
definition of truth in terms of his ontological dichotomy in De ver., q.
I, a. I. The definition in question, according towhich "the true is that
which is" (verum est id quod est), we are told by Aquinas, can be accounted for either by saying that "it defines truth (only) insofar as it has
a foundation in reality, and not insofar as it belongs to the complete notion of truth that reality corresponds to (the way it is conceived of by our)
intellect", or by saying that, when the true is defined as that which is,
"the word 'is' is not (to be) taken in the sense in which it signifies the
act of being, but in the sense in which it indicates that our intellect is
establishing a (propositional) combination, i.e. in the sense in which it
signifies that a proposition(-al content) is being affirmed; the meaning
of (the words) 'the true is that which is' would then be that (truth obtains)
when something which is (the case) is said to be (the case)":
... dicendum quod diffinitio ilia Augustini datur de veri tate secundum quod habet fundamentum in re et non secundum id quod ratio veri completur in adaequatione rei ad intellectum. - Vel dicendum quod cum dicitur verum est id quod est, Ii est non accipitur ibi
secundum quod significat actum essendi sed secundum quod est nota intellect us componentis, prout scilicet affirmationem proposition is significat, ut sit sensus: verum est id quod
est, id est cum dicitur esse de aliquo quod est (ibid. ad I; cf. De ver. q. I, a. 10, ad I, In
I Sent. dis!. 19, q. 5, a. I, ad I).

What Aquinas seems to be contrasting here is, on the one hand, the
sense in which the expression "something which is" means "something

THE LOGIC OF BEING IN THOMAS AQUINAS

193

which actually exists" and, on the other hand, the sense in which it
means "something which is the case". This is confirmed by his reply to
the objection that from St Augustine's definition of truth it seems to
follow that nothing is false, because this definition implies that the false
is that which is not (De ver. q. 1, a. 10, obi. 1). After having repeated
that in defining the true as that which is, the definition at issue "does not
perfectly express the notion of truth but, as it were, materially only, save
insofar as (the verb) 'be' signifies that a proposition(-al content) is being
affirmed, so that we might say that true is what is said or thought to be
such as it is in reality":
ista diffinitio 'verum est id quod est' non perfecte exprimit rationem veri tat is sed quasi
material iter tan tum, nisi secundum quod Ii esse significat affirmationem proposition is, ut
scilicet dicatur id esse verum quod sic esse dicitur vel intelligitur ut in rebus est (ibid. ad I),

he points out that "in this way we might also say that false is what is not,
i.e., what is not such as it is said or thought (to be); and this is to be found
in reality":
et sic etiam falsum dicatur quod non est, id est quod non est ut dicitur vel intelligitur: et
hoc in rebus inveniri pot est (ibid.).

According to this account, the false is not to be identified with

anything that does not actually exist, but with something said or
thought to be the case without really being the case or, in short, with
something that is not the case.
Aquinas's concluding remark that "this is to be found in reality" does
not refer, of course, to that which is false in the sense that it is not such
as it is said or thought to be, i.e. to that which is not the case; for what
is not the case is, by the very fact of its not being the case, excluded from
reality . What the remark in question must be taken to refer to is rather
the fact that something is not such as it is said or thought to be, i.e., the
fact that something which is said or thought to be the case is not really
the case, which "is to be found in reality" (in rebus inveniri pOlesl) insofar (and only insofar) as reality does not correspond to what is said or
thought to be the case.
That does not mean, however, that what in fact is the case when reality
does correspond to what is said or thought to be the case is itself
something to be found in reality; for it is nothing but a propositional content (or a thought), nowhere to be found but in our intellect truly affirming it, whereas in reality there are only substances and accidental properties thereof, whose actual existence is that which corresponds or fails to
correspond to what is said or thought to be the case. This at least is the

194

HERMANN WEIDEMANN

view which Aquinas seems to have adopted, as witness his statement,

already quoted above, that "as signifying the truth of a propositional
combination, in which respect (the word) 'is' is called the copula, being
is to be found in the combining and dividing intellect as the completion
of truth, 30 but founded upon the being of the thing (thought of), which
is the act of (that thing's) essence" (In I Sent. dist. 33:q. 1, a. 1, ad 1).31
1f we have succeeded in putting together the scattered pieces of our
jigsaw puzzle to form a correct picture of the position held by Aquinas
with regard to the logic of being, we may note two interesting features
of this position. For one thing, it enables Aquinas to subscribe to a correspondence theory of truth which is far from being trivial; for the way
in which he tries to establish a connection between the two fundamental
modes of being and the two corresponding uses of the verb "be" which
we have seen him distinguishing throughout his work seems to have
forced upon him the view that what makes a predicative statement of the
form" Sis P" true is not simply the fact (or its being the case) that S is
P, but rather the actual existence of the property signified by "P" as a
(substantial or accidental) property of the subject-thing referred to by
"S".
That this is indeed a position which Aquinas would be prepared to defend is suggested by his saying, in a text already quoted above, that it is
"a comparison of an accident(-al property) to a substance which the verb
'is' signifies when (a sentence like) 'a man is white' is uttered": "Quae
quidem comparatio (accidentis ad substantiam sci!.) significatur hoc verbo 'est' cum dicitur 'homo est albus' " (In V Metaph. lect. 9, no. 885).
Another piece of evidence is the above-cited passage from Aquinas's
Commentary on Aristotle's Peri hermeneias in which the copulative "is"
is described as "signifying that a (certain) form or act actually is in some
subject" (lib. I, lect. 5, no. 73[22]).
In view of Aquinas's thesis that the copulative "is" signifies the truth
of the sentence in which it occurs (cf. S.th. I, q. 48, a. 2, ad 2, In III Sent.
dist. 6, q. 2, a. 2, c) the two passages just quoted strongly support the
assumption that, according to Aquinas, the sentence "a man is white"
signifies - as the condition of its truth and, thus, as the condition of its
being the case (or its being a fact) that a man is white - that whiteness
is actually existent as a property inherent in a man.
This brings us to a second point, which is closely related to the first.
When Aquinas declares, on the one hand, that it is a "propositional COlTlbination, additionally invented by our (intellectual) soul" (compositio
proposifionis. quam anima adillvenit: S.th. I, q . 3, a. 4, ad 2; cf. 111 II

THE LOGIC OF BEING IN THOMAS AQUINAS

195

Sent. dist. 34, q. 1, a. 1, c), and, on the other hand, that it is "the actual
being in some subject(-thing) of a (certain) form or act" (formam vel actum actualiter inesse alicui subiecto: In I Peri herm. lect. 5, no. 73[22])
which the copulative "is" signifies, he seems to have in mind something
like a distinction between a word's signifying something to the effect that
it expresses a sense and its signifying something to the effect that it has
a reference. In the light of such a distinction, which is crucial in cases in
which, instead of a form actually existing in reality, a privation of some
actual being (e.g. blindness) is combined with a subject-thing by means
of the word "is" (cf. In II Sent. dist. 37, q. 1, a. 2, ad 3), it must be
noticed, as regards the passage quoted from Aquinas's Commentary on
Peri hermeneias, that, though it is the copulative sense or meaning of the
verb "be" which he intends to explain there by reference to the sense in
which "to be" means "to be in act", what he does in fact explain by
reference to this actuality sense of the verb' 'be" seems dot to be, strictly
speaking, what the copulative "is" means, but rather what it refers to,
namely a certain form's actually being in something subjected to it.
What may have helped Aquinas in this way to blur the distinction between meaning (or sense) and reference, which he is elsewhere, following
Aristotle, careful enough to observe (cf. In I Peri herm. lect. 2, no. 15[5]:
" ... necesse fuit Aristoteli dicere quod voces significant intellectus conceptiones immediate et eis mediantibus res"), is the fact that in drawing
his distinctions concerning the different uses of the verb "esse" ("to
be") he is making prominent use of the verb "significare" ("to
signify"), which is, as it were, neutral with respect to the sense/reference
distinction. 32
NOTES
1 By saying that a word is ambiguous, I mean that it has several distinct senses and can
thus be put to semantically different uses; by saying that a word is ambiguous but not genuinely so, I mean that its different senses are in some way or other systematically con
nected, for instance in such a way that "a number of secondary senses depend upon a single
primary one" (Hamlyn, 1977173, p. 6). Cf. B. Miller's distinction between "casually ambiguous" and "systematically ambiguous" expressions (1975, p. 346).
2 The question to what extent Aquinas's account of this dichotomy is faithful to the position held by Aristotle himself is not easy to answer. In the present paper it will simply be
disregarded. For Aquinas's Commentary on Aristotle's Metaphysics cf. Doig (1972),
together with the critical review by Georg Wieland, Archiv fur Geschichte der Philosophie
57 (1975), 73 - 78.
3 It must be noticed that it is, in general, a declarative sentence, and not the propositional
content thereof, that Aquinas uses the term "propositio" to refer to.

196

HERMANN WEIDEMANN

Cf. De ente et essentia, Ch. 1, De pot. q. 7, a. 2, ad 1, and the other passages listed by
Veres (1970, pp. 92 - 97).
5 Cf. Geach (1969): "Existence in the sense of actuality (Wirk/ichkeit) is several times over
emphatically distinguished in Frege's works from the existence expressed by 'there is a soand-so' (es gibt ein - ). Indeed, he says that neglect of this distinction is about the grossest
fallacy possible - a confusion between concepts of different level" (p. 65). Cf. ibid., Ch.
4 ('Form and Existence', pp. 42 - 64), where Geach is discussing "what Aquinas meant by
his term esse, or actus essendi, 'act of existing' " (p. 42). For an aHempt to defend Geach's
view on the different senses of "exists", which has been criticized by Dummett (1973, pp.
386f) , see Miller (1975).
6 Cf. Weidemann (1979).
7 Cf. McCabe (1970, p. 77).
8 Cf. Kahn (1973), Ch. VII: 'The Veridical Use' (pp. 331- 370).
9 Like Aristotle before him, Aquinas seems to have muddled together what Geach calls
"two sorts of truth", namely "the truth of propositions, and the truth of predications"
(1972, p. 15).
10 For this term cf. Strawson (1974, pp. 20 - 22). "A truth-or-falsity-yielding combination
we call a propositional combination" (p. 21).
11 If in a sentence of the form
is P" a general term functions as subject-term, the
sentence in question is traditionally called an "indefinite proposition" (cf. In I Peri herm.
lec!. 11, no. 150[8)). Since its general subject-term is not explicitly quantified, such a
sentence, e.g., "(A) man is white" (homo est a/bus: ibid.) can be treated as logically
equivalent either to a particular proposition ("Some Sis P') or to a universal one ("Every
S is P").
12 The "est" of the Marietti edition which I have enclosed in square brackets should be
deleted; cf. no. 894: " ... hoc totum, quod est homo albus, est ens secundum accidens, ut
dictum est".
13 Oesterle's translation (1962, p. 51), slightly modified.
14 My translation (for different ones cf. McCabe (1964, p. 95), Malcolm (1979, p. 394).
" Cf. S.th. I, q. 85, a. 5, ad 3: " ... compositio autem intellectu> est signum identitatis
eorum quae componuntur. Non enim intellect us sic componit, ut dicat quod homo est
albedo; sed dicit quod homo est albus, idest habens albedinem; idem autem est subiecto
quod est homo, et quod est habens albedinem". For the context of this quotation cf. Note
17 below.
16 Veatch (1974, p. 419). Similarly, Malcolm speaks of an "identity aspect of predication"
(1979, p. 394 and passim). - Referring to S.th. I, q. 13, a. 12, c, and I, q. 85, a. 5, ad
3, Veatch launches a heavy attack on Geach to the effect that "Geach in his reference to
both of these passages never gives his readers any intimation that they both contain unequivocal assertions as to the presence of an identity factor in affirmative predication"
(Veatch, 1974, p. 418). In the face of such criticism it must be acknowledged that in the
original version of his paper 'Subject and Predicate' (1950), apparently unknown to
Veatch, Geach had commented on the latter of the two passages in question as follows
(with an additional reference to S.th. I, q. 39, a. 5, ad 5): "As regards the truth-conditions
of an affirmative predication (compositio), he [Aquinas] rejects the view that subject and
predicate stand for two different objects, which we assert to be somehow combined; on the
contrary, the truth of the predication requires a certain identity of reference. Thus, if the
predicate 'white' is to be truly attached to the subject 'man' or 'Socrates', there must be
an identity of reference holding between 'man', or 'Socrate,', and 'thing that has
whiteness' ('quod est habens a/bedinem'); the two names must be idem slihiecto. Notice

"s

THE LOGIC OF BEING IN THOMAS AQUINAS

197

that what is here in question is the reference of a descriptive name, not of a predicate;
Aquinas does not hold. indeed he expressly denies, that predicates like 'white' stand for
objects (supponunt). His theory is that if the predicate 'white' is truly attached to a subject,
then the corresponding descriptive name 'thing that has whiteness' must somehow agree
in reference with the subject" (p. 478; cf. Malcolm, 1979, p. 3951).
Unfortunately this comment, which would have rendered Veatch's criticism almost
pointless, is absent from the rather different version, referred to by Veatch, in which 'Sub
ject and Predicate' has been incorporated into Geach's book Reference and Generality
(1968, pp. 22-46; 1980, pp. 49-72).
11 Leaving aside the controversy between Geach and Veatch, which it is not my present
concern to settle, I should like to point out the following problem, which seems to have
gone unnoticed by both authors: The text of S.th. I, q. 13, a. 12, c, suggests that to account
for the truth-conditions of a sentence of the form" Sis P" in terms of the identity of what
the subject-expression" S" stands for with what the predicate-expression" P" is true of
is to answer the question of what the sentence refers to in reality, and that to account for
the conditions of its truth in terms of the inherence of the form of P-ness in what" S"
stands for is to describe the sense expressed by the sentence, i.e. the mode of conceiving
the real identity it refers to; the text of S.th. I, q. 85, a. 5, ad 3, suggests, however, that
it is just the other way round. For the passage quoted from this text in Note 15 above is
embedded in the following context: "Invenitur autem duplex compositio in re materiali.
Prima quidem formae ad materiam; et huic respondet compositio intellectus qua totum
universale de sua parte praedicatur [... J. Secunda vero compositio est accidentis ad
subiectum; et huic reali compositioni respondet compositio intellect us secundum quam
praedicatur accidens de subiecto, ut cum dicitur 'homo est albus'. - Tamen differt compositio intellect us a compositione rei; nam ea quae componuntur in re sunt diversa; compositio autem intellectus est signum identitatis eorum quae componuntur" (ibid.). Accor
ding to this account, it is the combination of the accidental property of whiteness with a
man that corresponds to the (true) sentence "(A) man is white" in reality, whereas the iden
tity of a man with something that is white (or has whiteness) seems to be what the sentence
in question expresses as the sense in which it is to be understood: " ... et secundum hanc
identitatis rationem intellectus noster unum componit alteri praedicando" (ibid.). Other
texts relevant to the problem of reconciling the apparently different accounts given in S.th.
I, q. 13, a. 12, c, and in S.th. I, q. 85, a. 5, ad 3, are S.th. I, q. 16, a. 2, c, S.th. III, q.
16, a. 7, ad 4, In VI Metaph. lee!. 4, no. 1241, In IX Metaph. lec!. II, no. 1898.
18 If a predicate-term like "white" is combined not with a singular subject-term like
"Socrates" but with a general one like "man", a word which, like "some" or "every",
specifies "the kind of identity of reference" (Geach, 1950, p. 478; cf. p. 479) required for
the truth of the resulting sentence must be added or understood from the context (cf. Note
11 above).
19 Cf. Geach (1950, pp. 4761).
20 Cf. Hintikka (1973, p. 6), together with the critical review by Dorothea Frede,
Philosophische Rundschau 22 (1976), 237 - 242.
21 Cf. Veres (1970).
22 I have conjectured "veri" instead of the "'sui" of Mandonnet's edition, which does not
make good sense. My conjecture is based on Aquinas's remark, immediately following the
quoted passage, "sicut supra de veri tate dictum est", which refers back to dis!. 19, q. 5,
a. I, ad 1: "Vel potest dici, quod definitiones istae dantur de vero non secundum com
pletam sui rationem, sed secundum illud quod fundatur in re". Other pieces of evidence
are In II Sent. dis!. 37, q. I, a. 2, ad I ("verum dupliciter potest considcrari. Vel secundum

198

HERMANN WEIDEMANN

quod fundatur in re [... J. Vel secundum quod completur operatione animae compositionem formantis") and De ver. q. I, a. I, ad I (" ... secundum id quod ratio veri completur in adaequatione rei ad intellectum").
23 Oesterle's translation (1962, p. 53), slightly modified.
24 For a detailed analysis of this text cf. Zimmermann (1971).
25 I have deleted the comma after "formae" in the Leonine and Marietti editions, because
it is misleading. As the clause "cum volumus significare quamcumque formam vel actum
actualiter inesse alicui subiecto" shows, the words "actualitas omnis formae[,J vel actus
substantial is vel accidentalis" are not to be taken in the sense of "actuality of every form,
be it a substantial or an accidental act" ("Wirklichkeit jeder Form, sowohl eines substanzialen wie auch eines akzidentellen Aktes": Zimmermann, 1971, p. 292), but in the sense
of "actuality of every substantial or accidental form or act" (in Oesterle's translation "or
act" is missing).
26 Cf. Weidemann (1982).
21 For a critical account of the idea of "focal meaning" cf. Hamlyn (1977178).
28 Hamlyn (1977178, p. 5). Although Hamlyn states this condition as a necessary one ("It
will be possible to show the use or sense to be derivative only if ... ": ibid., my italics),
he obviously assumes that it is also sufficient; cf. p. 6: "Thus the example satisfies the condition that a sense is derivative from another when an explanation of its derivability is in
principle forthcoming. Without that it would have been a case of straight ambiguity."
29 "Eine Tatsache ist ein Gedanke, der wahr ist": Gottlob Frege, 'Der Gedanke: Eine
logische Untersuchung', Beitr. zur Phi/os. des deutschen Idealism us 1 (1918/19),58 -77;
reprinted in Gottlob Frege: Logische Untersuchungen, herausgegeben und eingeleitet von
G. Pat zig (Vandenhoeck & Ruprecht, G6ttingen, 1966), pp. 30 - 53 (quotation, p. 50). For
an English translation cf. Gottlob Frege: Logical Investigations, ed. with a preface by P. T.
Geach, transl. by P. T. Geach and R. H. Stoothoff (Basil Blackwell, Oxford, 1977);
quotation, p. 25.
30 Cf. Note 22 above.
31 Cf. Aquinas's statement that "this man committing a sin (i.e., the fact that this man
commits a sin) is a kind of mental entity insofar as it is called true": "hoc quod est istum
peccare est quoddam ens ration is prout verum dicitur" (In II Sent. dist. 37, q. I, a. 2, ad
I). Cf. also In VI Metaph.lect. 4, no. 1241:" ... compositio et divisio, in quibus est verum
et falsum, est in mente, et non in rebus [... J. Et ideo iIIud, quod est ita ens sicut verum
in tali compositione consist ens, est alterum ab his quae proprie sunt entia, quae sunt res
extra animam, quarum unaquaeque est aut quod quid est, idest substantia, aut quale, aut
quantum, aut aliquod incomplexum, quod mens copulat vel dividit."
32 For helpful comments I am grateful to Gregg Beasley.

REFERENCES
Aquinas, SI. Thomas, De ente et essentia (Leonine edition, Vol. 43, Rome, 1976).
Aquinas, SI. Thomas, Scriptum super libros Sententiarum/super Sententiis Magistri Petri
Lombardi, Vols. I -II, ed. P. Mandonnet (Lethielleux, Paris. 1929); Vol. Ill, cd. M. F.
Moos (Lethielleux. Paris, 1933); quoted: In I (II. III) Senl.
Aquinas, St. Thomas, In ArislOlelis libros Peri hermeneias exposilio (Leonine edition, Vol.
I, Rome, 1882; Marietti edition [Torino/Rome], 1964); quoted: In I (II) Peri herm.
Aquinas, SI. Thomas, In duodecim libros Melaphysicorum Aristolelis exposilio (Marietti
edition [Torino/Rome], 1964); quoted: In I (II, ... ) Melaph.

THE LOGIC OF BEING IN THOMAS AQUINAS

199

Aquinas, St. Thomas, Summa theologiae (Leonine edition, Vols. 4 - 12, Rome,
1888-1906); quoted: S.th.
Aquinas, SI. Thomas, Quaestiones disputatae de veritate (Leonine edition, Vol. 22, Rome,
1970 - 76); quoted: De ver.
Aquinas, SI. Thomas, Quaestiones disputatae de potentia, de malo (Marietti edition
[Torino/Rome], 1965); quoted: De pot., De malo.
Aquinas, SI. Thomas, Quaestiones quodlibetales (Marietti edition [Torino/Rome], 1956);
quoted: Quodl.

(If not otherwise indicated, translations from the Latin are my own.)
Doig, James C., Aquinas on Metaphysics: A Historico-Doctrinal Study of the Commentary on the Metaphysics (Martinus Nijhoff, The Hague, 1972).
Dummett, Michael, Frege: Philosophy of Language (Duckworth, London, 1973).
Geach, Peter Thomas, 'Subject and Predicate', Mind 59 (1950), 461 - 482.
Geach, Peter Thomas, Reference and Generality: An Examination of Some Medieval and
Modern Theories (Cornell University Press, Ithaca and London [1962], emended edition 1968; third edition, 1980).
Geach, Peter Thomas, God and the Soul (Routledge and Kegan Paul, London, 1969).
Geach, Peter Thomas, Logic Mailers (Basil Blackwell, Oxford, 1972).
Hamlyn, D. W., 'Focal Meaning', Proceedings of the Aristotelian Society 78 (1977178),
1-18.
Hintikka, Jaakko, 'Aristotle and the Ambiguity of Ambiguity', in J. Hintikka, Time and
Necessity (Oxford University Press, Oxford, 1973), pp. I - 26.
Hintikka, Jaakko, ' "Is", Semantical Games, and Semantical Relativity', Journal of
Philosophical Logic 8 (1979),433 - 468.
Kahn, Charles H., The Verb 'Be' in Ancient Greek (The Verb 'Be' and Its Synonyms:
Philosophical and Grammatical Studies, ed. by John W. M. Verhaar, Part 6; Reidel,
Dordrecht, 1973).
Malcolm. John, 'A Reconsideration of the Identity and Inherence Theories of the Copula',
Journal of the History of Philosophy 17 (1979), 383 - 400.
McCabe, Herbert, St Thomas Aquinas: Summa Theologiae, Vol. 3: Knowing and Naming
God (la. 12 - /3). Latin text. English translation, Introduction, Notes, Appendices &
Glossary (Blackfriars [in conjunction with Eyre & Spottiswoode, London, and
McGraw-Hili Book Company, New York], 1964).
McCabe, Herbert, 'Categories', in A. Kenny (ed.), Aquinas: A Collection of Critical
Essays (Macmillan, London, 1970), pp. 54 - 92.
Miller, Barry, 'In Defence of the Predicate "Exists" " Mind 84 (197),338 - 354.
Oesterle. Jean T., Aristotle: On Interpretation. Commentary by St. Thomas and Cajetan
(Peri Hermeneias) , transl. from the Latin with an Introduction (Marquette University
Press, Milwaukee, 1962).
Strawson, P. F., Subject and Predicate in Logic and Grammar (Methuen & Co., London,
1974).
Veatch, Henry, 'SI. Thomas' Do,trine of Subject and Predicate: A Possible Starting Point
for Logical Reform and Renewal', St. Thomas Aquinas 1274 -1974: Commemorative
Studies (Pontifical Institute of Mediaeval Studies, Toronto, 1974), Vol. II. pp.
401 -422.
Veres, Torno. 'Eine fundamentale ontologische Oichotomic im Den ken dcs Thoma' \on
Aquin', Philosophisches Jahrbuch 77 (1970), 81 - 98.

200

HERMANN WEIDEMANN

Weidemann, Hermann, '''Socrates est" /"There is no such thing as Pegasus": Zur Logik
singularer Existenzaussagen nach Thomas von Aquin und W. Van Orman Quine',
Philosophisches Jahrbuch 86 (1979), 42 - 59 (translated into English by David J. Marshall Jr., in Contemporary German Philosophy 3 {I 983), 159-178).
Weidemann, Hermann, 'Aristoteles tiber das isolierte Aussagewort: De into 3, 16 b
19 - 25', Archiv fur Geschichte der Philosophie 64 {I 982), 239 - 256.
Zimmermann, Albert, '''Ipsum enim ('est') nihil est" (Aristoteles, Periherm. I, C. 3).
Thomas von Aquin tiber die Bedeutung der Kopula', Miscellanea Mediaevalia 8 (Walter
de Gruyter, Berlin/New York, 1971), pp. 282-295.

Philosophisches Seminar,
Universitiit Munster,
Dompiatz 23,
D-4400 Munster, B.R.D.

SIMO KNUUTfILA

BEING QUA BEING IN THOMAS AQUINAS AND

JOHN DUNS SCOTUS

I. THE METAPHYSICS OF ESSENCE IN THOMAS AQUINAS

Thomas Aquinas thought, like Aristotle, that individuals of the

sublunary world could be classified into natural classes or species on the
basis of their qualities. In this approach the invariant structure of the empirical world is seen as consisting of unchangeable bundles of qualities
which are always exemplified and which belong to the objects they
qualify as long as those objects exist. Some of these combinations of
qualities distinguish natural species. They specify classes whose members
are not distinguishable from one another by any qualifying differences
belonging to the invariant system of qualities mentioned above. I
In Aristotelianism, as Aquinas understood and accepted it, this empirical world order is taken to be based on a metaphysical deep structure
of reality. It is thought that there must be something which explains the
manifest fact of the invariant sets of qualities exemplified in objects,
which are divided into natural kinds. The metaphysical answer to this
question is found in the idea that the members of a species having the
same fixed bundle of qualities are actualizations of one of the forms of
substantial being which together constitute the ultimate structure of the
unchangeably actual order of being. These essences or common natures
do not exist as such, but they can have existence through their singular
inferiors. The invariant sets of properties are concomitant with these actualizations of essences. The point of this view is that the members of
species should not be considered simply as aggregates of the properties
mentioned in the empirical definitions of species. It is thought that in
some way the whole of the invariant properties in a singular being is more
than a sum of the parts, and this idea is formulated by speaking about
things as representatives of metaphysical forms of being. 2
In addition to essences, which make things to be what they are, there
is in Aquinas's Aristotelian metaphysics another non-empirical constituent of the empirical world. In Met. Z, 3, 1029a20 - 28 Aristotle argues
that characteristics of things other than substantial forms can be
understood as properties of the substance. The ultimate possessor of the
property of being such and such substance cannot be the individual
201
S. Knuuttila and J. Hintikka (eds.), The Logic of Being, 201-222.
!C) 1986 by D. Reidel Publishing Company.

202

SIMa KNUUTTILA

substance, but rather the prime matter as the bare substratum. Aquinas
understood this to mean that the prime matter, which does not exist as
such, is a metaphysical principle having existence as the substrate of the
singular instantiations of essences. 3
Thomas Aquinas says in De ente et essentia that the nature or essence
of things, absolutely considered, is neither one nor many. Were it many,
the humanity could not be actual in Socrates. And were it one, Socrates
and Plato would be one and the same as men. Because essences as such
cannot be one or many, they do not exist as such. Although an essence,
absolutely considered, is abstracted from any being, it is said to have being in singular things and as a concept in the mind. 4 The essences exemplified in the sublunary world have being as many, because they are
instantiated in matter, which functions as a principle of individuation in
such a manner that it generates a plurality of instances of species. 5
These metaphysical ingredients of the sublunary world have being in
the extramental reality as constituents of singular things. 6 The objects
which they constitute belong to the domain of existent objects, although
they themselves do not exist. From the metaphysical point of view we can
characterize subjects of existence we are immediately acquainted with as
actualities of essences individuated by matter. Aquinas believed that on
higher levels of the great chain of being, essences can be actualized
without being individuated by the prime matter, 7 and so he refers in his
more general accounts of the quidditive being simply to the individual instances of actuality of essences. 8
The essences, which as such are neither one nor many, are intelligible.
An essential part of Plato's and Aristotle's worldview is the belief that
there is an intelligible framework of the cosmos consisting in a system of
invariant forms of being. For them the most attractive part of the human
perfectibility is the possibility of ascending into an immediate knowledge
of this intelligible structure of the sensible reality. 9 Thomas Aquinas includes this kind of jelicitas theoretica in his conception of the final end
of man. lo However, there are some remarkable reformulations in his
handling of the subject.
Prime matter has no properties of its own and it is unintelligible.
Therefore, when essences are individuated by prime matter they become
parts of compositions which, because of the unintelligible factor present
in them, are not intelligible as such. II However, the human intellect can
u'nderstand that there is an intelligible structure in the extra mental reality, because it can abstract the intelligible essence or form from the
matter. 12 Through sensations we apprehend particular objects. From

BEING QUA BEING IN AQUINAS AND SCOTUS

203

empirical data called phantasms the agent intellect abstracts likeness of

the quiddity of sensible things. This intelligible species actualizes the
potential intellect,I3 and it forms the corresponding concept or, as
Aquinas also says, the mental word. 14 An important group of mental
concepts consists of the kind-universals. The names of species signify
directly these concepts and through them singular members of natural
kinds. 15
Through the actuality of essences as concepts in their intellects, human
beings are acquainted with the form-giving principles of reality and the
extramental world is intelligible to them. However, on the scale of intellectual beings men occupy the lowest position. Their intellect is in
many ways limited. 16 For example, it is only the essences of those objects
exemplified in the external world which are actual in the human intellect
in statu viae. 17 In some places Aquinas says that we can know the
essences through the properties of their actualizations. 18 But there are
many other passages in which it is stated that all or almost all essences
remain unknown to us. 19 The discrepancy between these statements may
be explained by the fact that according to Aquinas the human knowledge
of essences is always imperfect and partial. The human intellect does not
fully comprehend the intelligibility of the essences which are conceptualized in the intellect. Aquinas believes that in heaven the human intellect will be improved in this respect to the effect that it will fully
understand the intelligible nature of the essences it already knows partially. And in addition, it will come to know, in so far as it is possible
for the lowest intelligence to know, those actualized essences it did not
know in statu viae and essences which could have been instantiated but
which have not been. 2o
I have argued in other places that until the late thirteenth century the
use of modal concepts in Western thought was often based on paradigms
of modalities in which the meaning of the modal notions was not made
explicit by considering simultaneous alternative states of affairs. Instead
of this intensional interpretation, there was a tendency to understand
modal notions in extensional ways. Therefore it was natural to think, for
example, that no genuine possibility can remain forever unrealized.
Before Aquinas this "principle of plenitude" was explicitly accepted by
many medieval thinkers for natural modalities, but for theological
reasons it was usually denied for Divine possibilities. It seems that before
Aquinas this partial denial of the principle of plenitude was often considered as only one case of the many anomalies which must be taken into
account when supranatural things are discussed. 21 In Thomas Aquinas

204

SIMO KNUUTTILA

we find a more sophisticated theory of the unrealized Divine possibilities.

According to him, the Divine Omnipotence must be defined by referring
to absolute possibilities, which are expressed by statements whose terms
are not contradictory. The number of these Divine possibilities is much
greater than that of the actualized possibilities. Thomas Aquinas's absolute possibilities come close to what (since Duns Scotus) have been
called logical possibilities. Their role in Aquinas's thought is very
limited, however. 22 Certain general presuppositions made the extensional interpretation of modality attractive for Aquinas, and his doctrine
of the imperfect knowledge of essences seems to be one of the factors
which pushed him into this direction. Whenever we consider whether a
member of a natural kind could have a quality not as yet exemplified by
any member of that kind, Aquinas's theory of the partial knowledge of
essences suggests that the answer is always the same: we don't know the
essence as such and therefore we cannot know whether it is contradictory
or not that this kind of property is predicated of this substance. We know
that an actual instance of an essence can have those properties which
manifestly belong to the members of the species. This means that we
know that something is possible only if there is an example of it in the
extramental reality, although we believe that the actual world is not an
exhaustive manifestation of what is possible. As for unrealized essences,
we cannot know about any imaginable example whether or not it is
possible.23
2. AQUINAS ON THE ANALOGY OF BEING

In his Summa the%giae I, q. 13, a. 10 Thomas Aquinas writes as

follows:
Univocal names have exactly the same ratio and equivocal names have an entirely different
ratio. As for analogical names, a name with a certain signification is used in the definitions
of the same name considered as having other significations. So "ens" used in connection
with substances occurs in the definition of "ens" used in connection with accidents, and
"healthy" predicated of animals occurs in the definitions of "heal!hy" predicated of urine
or medicine, for urine is a sign of the health of an animal and medicine restores it.

Aquinas thought that what is immediately signified by a name is a

mental conception or a definition.24 This view is part of a philosophical
theory of meaning which was well accepted in Paris in the second half
of the thirteenth century. It was founded on a reinterpretation of some
important technical terms of the early medieval terminist logic in the light
of Aristotle's psychology and metaphysics of knowledge. 25 It was

BEING QUA BEING IN AQUINAS AND SCOTUS

205

thought that the area of being manifests itself to the human intellect in
different modes of being (modi essendl) and that there are fixed modes
of intellection (modi intelligendl) , through which a conceptual picture of
the extramental reality is formed in the intellect. Words as significant
units signify directly something which exists in the intellect. They refer
to extramental objects indirectly, through the intellect (sub modo
intelligendl).26 Bearing this triadic view of signification in mind, we can
say that in the first sentence of the above Guotation Aquinas claims that
a univocal name signifies in all its applications one and the same element of the conceptual correlate of the extramental reality. Correspondingly, an equivocal name signifies at least two different ingredients of
the intelligible order having no reference to each other in their
definitions.
In his characterization of an analogous name Aquinas says that in its
applications there is one basic use connected with a certain meaning. This
meaning is implied in the prima Jacie different meanings an analogous
word seems to have when predicated of various types of things. 27 This
corresponds roughly to what Aristotle says about the pros hen
equivocity.28 It is not quite clear how the new terminology was introduced in Scholasticism; at all events, it seems to have been commonly applied in the first half of the thirteenth century. 29 For example, in the
Summe Metenses written before 1220, the class of analogical terms is
delineated as follows:
Equivocity is understood in two ways, properly and commonly. The proper equivocity
belongs to signifying units (dictio) which are actually multiplex. The common equivocity
belongs to signifying units which are related to many, to one primarily and to others ex consequenti. And so analogical words are said to be equivocal, e.g., ens. unum, and aliquid.
They are primarily said of substance, and secondarily (per posterius) of quantity, quality
etc.; and similarly "healthy" is said primarily of animals and ex consequenti of urine and
drinks and food. 30

When Aquinas speaks about "healthy" in the above quotation, he

says that it is primarily applied to healthy organisms and secondarily to
things which restore or preserve or are signs of healthy organisms. This
means that in its various uses an analogous name signifies directly such
different elements of the mental conceptual system the definitions of
which are connected in certain ways with one basic definition. The
ordered group of meanings corresponding to the many uses of an
analogous name is, as a whole, a kind of microstructure belonging to the
intelligible order realized in the world and repeated in the mind. This is
implied in the distinction Aquinas sometimes makes between ratio propria and ratio communis of an analogous name. The ratio propria is the

206

SIMO KNUUTTILA

meaning or res signijicata signified by the primary mode of signifying.

It is included in the definitions of what is signified by the secondary
modes of signifying. Ratio communis is not a definite single element of
the intelligible order. It is not an additional analogical meaning, and its
formal structure is based on something like "to be related in some way
to ratio propria". 31 It could be understood as a disjunctive totality of the
analogous meanings, signified by the analogous word in those cases
where it is indifferent which of the analogates the word refers to.32
According to Thomas Aquinas, being (ens) is an analogous term.
When a member of the category of substance is spoken of as a singular
instantiation of an essence, it is handled by means of concepts which
make up the primary meaning of ens. Ens signifies directly, and has as
its focal meaning a mental conception of essence having existence and being individuated. 33 The ratio propria of ens is thus not simple. It is a complex idea whose elements include the concepts of essence and actual existence which, as Thomas Aquinas says, are really distinct. The real
distinction between essence and existence seems to mean that existence
is the act by which the essence is or has being. The act of existence actualizes forms of being which are metaphysically prior to existent
entities. 34 The interpretations of the details of this real distinction are
controversial. To reiterate, the focal meaning of ens is the idea of an individuated actuality of essence through an act of existence, and Thomas
Aquinas claims that this is what is primarily meant when ens is applied
to things in the world known to us.
The secondary uses of ens are those in which it is applied to the various
qualifications of substances classified in the Aristotelian doctrine of
categories. When these modifications of the actualizations of essences in
singular instances are considered as beings, they are necessarily
understood vis-a-vis the focal meaning. 3s
In addition to this horizontal analogy of being, there is a vertical one,
too. It is connected with the Neoplatonic construction of the hierarchy
of being. Actualizations of essences taking place at different levels are
not referred to by ens univocally. Quidditative beings at different levels
of the great chain of being can be understood as more or less perfect imitations of the full being of God, however, and for this reason there is a
basis for analogous uses of ens in the vertical order, toO. 36
In a well-known passage of De potentia Thomas Aquinas says that esse
is the actuality of all acts and the perfection of all perfections. 37 The
meaning of ens corresponding to this abstract notion of esse is the ratio
communis of ens. Even if we know of something nothing else but that

BEING QUA BEING IN AQUINAS AND SCOTUS

207

it is a being, we understand that ens is said of it in one of the many

analogous meanings included in the ratio communis of ens, i.e. in the
ordered system of analogous meanings. This does not mean that ens
could be used univocally of everything. 38 If ens is applied to something
so that it signifies directly the ratio communis, that something is thought
to be a being in one of the analogous meanings included disjunctively in
the ratio communis. 39
3. SCOTUS'S NEW THEORY OF BEING QUA BEING

John Duns Scotus's doctrine of the univocity of being is easily

understood against this Aristotelian-Thomistic background. He does not
directly discuss Aquinas' theories in his metaphysics. However, some
features of the ontological theory just delineated were commonly accepted in Scotus's time, and they were brought under attack in Duns
Scotus's thought.
In his Scriptum super primum Sententiarum of 1316 Peter Aureoli
states that in his time the common metaphysical view was that
the concept of being ... expresses in a prior way the ratio of substance and expresses the
other rationes by attribution; and it expresses each under its own proper ratio, not under
some common ratio in which they agree . It is nonetheless a disjunctive concept. not a
copulative concept. Wherefore. when we speak of something as a being. we immediately
conceive it as a substance or a quality or a quantity. we do not conceive some common
ratio. 4o

There are some recent studies on this disjunctive approach to the concept
of being which is the direct object of Scotus's criticism. 41
In one of his arguments for the univocal metaphysical concept of being
Scot us says that it is possible to know of something that it is without
knowing what kind of being it is.42 In the disjunctive approach this kind
of knowledge could be interpreted so that we know that one or another
of the analogous meanings of being is applicable to something although
we don't know which one. Scotus's point is now that when something is
known to be in this way, that something is at the same time conceived
recognizable, although we cannot identify it. 43 Things considered identifiable can be univocally referred to by ens, whatever they are.
Scotus's position could be explained to the defenders of the disjunctive
approach as follows: Imagine that an omniscient identifier has made a
list in which a proper name is given to anything tqat is in the actual world
and can be called a being in some of the many analogous meanings of
the word. Whenever we hear a name taken from that list we know of that

208

SIMO KNUUTTILA

baptized being that it is to be found in the world althought we don't know

it. It is because it can be identified and so we have an idea of its being
before we know, for example, which of the analogical meanings of being
can be applied to it. So there is, according to Scotus, a common and simple ratio of the notion of being and, contrary to the views of Aquinas and
the disjunctivists, it is taken to precede conceptually all other meanings
attached to the word "being".
In the arguments just mentioned Scotus seems to think that the
possibility of being identified as something is the most primary constituent of the positive nature of whatever is and, as such, is the real basis
for the univocal metaphysical concept of being. 44 Although the concept
of being delineated in this way is derived from our acquaintance with existent beings, real or mental existence is not a part of its meaning. Nonactual possible things are in principle identifiable, toO. 45 Therefore the
content of the notion of transcendental being is expressed by the formulation "that to which to be is not repugnant" (cui non repugnat
esse).46 In the metaphysical sense ens is a univocal word which can be truly applied to anything which can be thought of as identified and which
therefore is something positive and distinct from the absolute
nothingness of impossible things. 47
On the basis of his insight into the unity of the concept of being Duns
Scotus could offer a remarkable new view on metaphysics as a science.
It has one single object referred to by the univocal notion of being. This
notion has two basic uses. It applies properly to the possible subjects of
existence, and in addition to this quidditative use, it is applied
denominatively to qualifications and modifications which inhere in a
subject. Being (ens) is the first of the transcendentals. Other transcendentals studied in the metaphysics are (l) the attributes coextensive with being as such (passiones entis simpliciter convertibiles), (2) the disjunctive
attributes (passiones entis disjunctae) , i.e. the primary differences of being which in disjunction are coextensive with being (like "infinite-orfinite", "necessary-or-contingent", "actual-or-potential"), and (3) the
pure perfections (perfectiones simpliciter), which are the remaining attributes predicable of God and hence transcending the finite categories. 48
In addition to the argument just mentioned, there are many lines of
thought Scotus takes in arguing for the univocal concept of being. 49
However, as for the later influence of his writings, the most important
of the new metaphysical patterns of thought is the one in which the concept of being is connected with the idea of identifiability of things which
can be existent. 50 This view plays an important role in Scotus's modal

BEING QUA BEING IN AQUINAS AND SCOTUS

209

theory, too. In fact it is an essential part of the new understanding of

modality introduced by Duns Scotus.
As mentioned above, the modal paradigms commonly used in Western
thought until the late 13th century contained presuppositions which easily led one to thing that modal potions are in the last analysis reducible
to extensional terms. Although the medieval doctrine of Divine
possibilities motivated some new considerations, they did not lead to a
totally new philosophical understanding of the nature of modality before
Duns Scotus. He seems to be the first to formulate a general theory of
modality in which modal notions are interpreted intensionally and in
which no role is left to the view that actuality in the real historical world
should be taken as the criterion of the genuineness of an alleged
possibility. 5 1
In Scotus's modal theory the meaning of modal notions is connected
with the idea of considering several alternative states of affairs at the
same time. The domain of possibility is introduced as an area of conceptual consistency. It rna)' be considered as consisting of all possible individuals, their possible properties and their mutual relations. This domain of all possible states of affairs is structured so that it is divided
into equivalence classes on the basis of the relations of compossibility.
One of the classes into which logical possibilities (possibile /ogicum) are
partitioned is the actual world. Some logical possibilities are real alternatives of the actual world (possibile rea/e).52
Scotus uses this model in what could be called a sketch of the Divine
psychology. Some formal distinctions are drawn and formulated by
means of the medieval instantia naturae terminology. In the first instance
of nature all thinkable entities are produced in esse intelligibile by the
Divine Intellect. In the second instance of nature the content of the intelligible esse is the correlate of the Divine Omnipotence and the logical
possibilities can be considered as objectively actualizable. As such they
present themselves to the Divine Will in the next instance of nature, and
then one particular compossible set of them is chosen to be actual by the
Divine Will. 53
Scotus says in some places that the metaphysical notion of being is a
real concept. This means that it refers to something which transcends the
mental reality. 54 It follows from this that when the transcendental notion
of being is applied to possible beings having some kind of existence in
the Divine Intellect, it is not applied to them because they have this esse
intelligibile, but because there is something in themselves which causes
them to be in this esse as they are. Scotus seems to think that the Omnis-

210

SIMO KNUUTIILA

cience identifies every possible being because they can be identified.

Although they themselves have no kind of existence, they get esse intelligibile in the Divine Intellect, when it identifies everything which can
be. A possible being becomes identified as the possible being it is because
it is what it is. 55 As a possible being in the Divine Intellect it is thought
of as actual in those states of affairs in which it can occur, and it is implied that any intellect which could identify everything would find this
same being. This is the real aspect of the notion of being applicable to
anything which can be.
Individuals occur in the domain of possibility in all forms and configurations which are intelligibly possible. It is typical of Scotus's modal
thinking that according to it, individuals can have properties which are
mutually exclusive. Therefore they must be considered simultaneously as
members of several possible combinations. Although a possible being
can occur in several possible states of affairs at the same time, only one
of the possible alternatives is actual and correspondingly the possible being is existent only if it belongs to the actual world .56
According to this view, something can be considered as a metaphysical
individual, if it can be identified in a world by virtue of its individual
nature. 57 Because the same possible being can be thought of as actual in
several possible states of affairs, its identifiability is not bound to any
single world. It follows from this that there is in an actually existing being
a not purely mental distinction between its individual identifiable nature
and its actual existence, although they are not really distinct in the sense
that one of the two could exist apart from the other. 58
The pair "potential-actual" of the coextensive disjunctive attributes
of being implies that "existence" is a kind of predicate which belongs to
those metaphysical beings which are instantiated in the actual world. A
being which is existent in the actual world can be identified in an alternative possible world, too. Because of its existence it is not a merely
potential being.
We have seen that according to Thomas Aquinas quidditative beings
are actualities of essence in singular instances. It seems that in this approach the individual identity of a substantial being is virtually equated
with the real history of an actualized essence. It follows from this that
there are certain problems if it is said that an individual at a certain moment of time could at that time have properties different from those it
actually has. If the possibility of being otherwise is assumed to be realized, it is also assumed that the history of the individual would be different from what it actually is. However, if the identity of the thing under

BEING QUA BEING IN AQUINAS AND SCOTUS

211

discussion is bound to one concrete history, it cannot be thought of as

the subject of simultaneous alternative possible qualifications. So if we
look at one individual being at a certain moment of time, it seems to have
all its qualifications necessarily at that time. The alleged de re
possibilities of this subject, which remain unrealized, are in fact not genuine possibilities of this being. Its actual history is what it is, and alternative ways of being would make it another singular actualization of
essence. 59 This is, of course, in accordance with the above mentioned
tendency to treat genuine possibilities in such a way that it is demanded
that there must be place for their real actualization in one and the same
historical world.
Duns Scotus's intensional modal theory was developed as a conscious
alternative to traditional models in which modal notions in the last
analysis are merely means of speaking in a certain way of what happens
in the one and only actual world at different moments of time. The basic
new idea of considering different alternative states of affairs at the same
time demanded a new theory of individuals and existence. If there are
alternatives to an actual state of affairs at the same time, it seems that
the individuals occurring in the actual world must be treated
simultaneously as members of alternative states of affairs, too. And then
it is required that the identifiability of an individual is not bound to its
actual existence. It must be possible to speak about the same individual
in alternative states of affairs at the same time. It is clear that Scotus's
new doctrine of being qua being was developed in close connection with
his modal theory which emerged as a result of a critical discussion of the
traditional modal paradigms.
The essentials of this interpretation of modal notions played an important role in the fourteenth century modal logic developed on the same
lines by William Ockham, Pseudo-Scotus, John Buridan, and others. It
was thought that the logical properties of modal statements can be obtained by considering the set of equivalence classes of compossible
possibilities, which was taken to be an a priori domain of mod ali zed
terms and to determine the logical behaviour of modal statements. Quantification over possibilia belonged to the standard methods of the new
modal logic, and it was connected with a view similar to Scotus's theory,
according to which individuals are identifiable entities in the domain of
possibility (of which a part is actual) and they can be thought of as
members of alternative possible states of affairs at the same time. The
fourteenth century logicians were not very interested in the question of
how the same individual could be picked up in alternative states of af-

212

SIMO KNUUTTILA

fairs. This kind of identifiability of the entities in the domain of

possibilia was simply accepted as a structural presupposition of the
model of modality which was found philosophically most satisfying. 6o In
Duns Scot us it is explained in terms of the Divine Omniscience and the
individual essence (haecceitas).
One important figure in the history of the later influence of Scotus's
modal metaphysics is Francesco Suarez. In his Disputationes
Metaphysicae Suarez seems to reject the Scotistic univocatio en tis. But
as argued by Walter Hoeres, the rejection is formal only and in fact
Suarez broadly accepted the Scotistic teaching. 61 His concept of being in
vi nominis is applicable to actual and potential individuals and it is used
in a modal theory which is clearly influenced by Scotus's model described
above. 62 This is historically interesting, because Suarez's metaphysics
was particularly prominent and influential in the early modern period.
It is one of the historical links between the modal theories of Duns Scotus
and Leibniz.
Although the idea of compossible sets of possibilities, the distinction
between logical and real possibilities, the model of actual and possible
worlds, and some other general elements of Leibniz's modal theory are
more or less directly derived from the late medieval philosophy, there are
some striking differences between the modal metaphysics of Scot us and
Leibniz. One of them is that instead of speaking about the same individual existing in several alternative worlds Leibniz thinks that one individual substance can occur in one possible world only. What is prima
facie considered as an unactualized possibility de re belonging to an actual being is in fact an attribute belonging to the counterpart of the actual
being in a possible world. 63
Leibniz says that his doctrine of individual substances is influenced by
Aquinas's view according to which each angel is an infima species, i.e.,
the only possible instance of a specific substantial form. 64 He takes this
to mean that the factual history of an angel shows exactly what it is to
be this angel. We have suggested that this conclusion could be deduced
from Aquinas's metaphysics vis-a.-vis any individual, because it seems
that the identity of an individual substance is virtually bound to its role
of being the subject of one historical sequence. One corollary of this idea
is that at any moment an individual has necessarily every attribute which
belongs to it at that time. According to Leibniz something like this
follows from his fundamental in esse principle. Leibniz thinks that every
substance must have a complete notion which comprises every predicate
which refers to any moment in its history. The only difference between

BEING QUA BEING IN AQUINAS AND SCOTUS

213

Leibniz's view and what seems to be implied in Aquinas's metaphysics

is that Leibniz applies this idea both to the individuals of the actual world
and to the individuals of possible worlds. 65
If the general principles of Scotus's modal theory and Aquinas' theory
of what it is to be a quidditative being are amalgamated, the natural
result is something like Leibniz's counterpart theory. It is interesting that
this syncretistic view is consciously accepted by Leibniz.
NOTES
1 See, for example, An. post. B, 13, 96a20 - b 14; Thomas Aquinas, In II Post. an., \c. 13,
Met. Z, 12, 1037b28 - 1038a27; Thomas Aquinas, In VII Metaphys., \c. 12. For the fixity
of species, see PA 646a35-b2, GA 731b31-732al, De an. 415a26-bl and Thomas
Aquinas, In II de anima, \c. 7, n. 314, 317.
2 Met. Z, 17, 104la6-104lb33 and Aquinas, In VII Metaphys., Ic. 17; see also De ente
et essentia, c. 2, S.th. I, q. 3, a. 3 - 5, q. 39, a. 4, ad 3, q. 76, a. 4c, and 1. Owens, 'Common
Nature. A Point of Comparison between Thomistic and Scotistic Metaphysics', in 1. F.
Ross (ed.), Inquiries into Medieval Philosophy. A Collection in Honor of Francis P.
Clarke, Greenwood Pub!. Co., Westport, Conn. 1971, pp. 191-194.
3 S.th. I, q. 7, a, 2, ad 3, q. 66, a. 1-2. Aristotle's doctrine of matter and some modern

interpretations of it are discussed in R. Dancy, 'On Some of Aristotle's Second Thoughts

about Substances: Matter', The Philosophical Review 87 (1978),372-413. See also H.
Happ, Hyle. Studien zum aristotelischen Materie-Begriff, De Gruyter, Berlin, New York,
1971. For the doctrine of the bare substratum in Aristotle and some modern thinkers, see
M. 1. Loux, Substance and Attribute. A Study in Ontology, D. Reidel Pub!. Co., Dordrecht, 1977, pp. 107 - \12.
4 De ente et essentia, c. 2, In II de anima, \c. 12, n. 378 - 380.
l S.th. I, q. 84, 2c, In de trin., q. 4, a. 2c.
6 "esse substantiae compositae non est tan tum esse formae nec tantum esse materiae sed
ipsius compositi; essentia autem est secundum quam res esse dicitur. Vnde oportet ut essentia qua res denominatur ens, non tan tum sit forma nec tantum materia, sed utrumque,
quamvis huius esse suo modo forma sit causa." De ente et essentia c. 1. "Finitur autem
quodammodo et materia per formam et forma per materiam. Materia quidem per formam
inquantum materia antequam recipiat formam est in potentia ad multas formas, sed cum
recipit unam, terminatur per ilIam. Forma vero finitur per materiam inquantum forma in
se considerata communis est ad multa, sed per hoc quod recipitur in materia fit forma
determinate huius rei." S.th. I, q. 7, a. Ic; see also S.th. I, q. 3, a. 4c.
7 S.th. I, q. 3, a. 2, ad 3, a. 3c, q. 12, a. 4c, q. 50, a. 2c.
8 S.th. I, q. 3, a. 4c, q. 50, a. 2, ad 3, q. 54, a. Ic, q. 75, a. 5, ad 4. For further examples,
see 1. C. Doig, Aquinas on Metaphysics. A Historico-doctrinal Study of the Commentary
on the Metaphysics, M. Nijhoff, The Hague, 1972, pp. 255 - 275, 358 - 367.-5ee also 1.
Owens, 'The Accidental and Essential Character of Being in the Doctrine of St. Thomas
Aquinas', Mediaeval Studies 20 (1958), I - 40; reprinted in 1. Owens, St. Thomas Aquinas
on the Existence of God. Collected Papers of Joseph Owens, C. Ss. R, ed. by 1. Catan,
State University of New York Pre,s, Albany, 1980, pp. 52-96.
9 Plato, Rep. 521C-54IB, Aristotle, EN VI, 6, 1140b31-\141a20.
10 "Naturale desiderium rationis creaturae est ad sciendum omnia ilia quae pertinent ad

214

SIMO KNUUTTILA

perfectionem intellectus, et haec sunt species et genera rerum, et rationes eorum." S.th. I,
q. 12, a. 8, ad 4.
11 S.th. I, q. 79, a. 3c, q. 86, a. I, ad 3.
12 S.th. I, q. 54, a. 4c, q. 79, a. 4, q. 85, a. I.
13 Abstrahit autem intellectus agens species intelligibiles a phantasmatibus, inquantum per
virtutem intellectus agentis accipere possumus in nostra consideratione naturas specierum
sine individualibus conditionibus, secundum quarum similitudines intellectus possibilis informatur." S.th. I, q. 85, a. I, ad 4. "Species intelligibilis est similitudo ipsius essentiae
rei, et est quodammodo ipsa quidditas et natura rei secundum esse intelligibile, non secundum esse naturale, prout est in rebus." Quodl. VIII, q. 2, a. 2c. See also S.th. I, q. 14,
a. 12c.
14 "Unde necesse est quod species intelligibilis, quae est principium operation is intellectualis, differat a verbo cordis, quod est per operationem intellectus formatum; quamvis ipsum verbum possit dici forma vel species intelligibilis, sicut per intellectum constituta, prout forma artis quam intellect us adinvenit, dicitur quaedam species intelligibilis." Quodl.
V, q. 5, a. 2c.
IS "Nam primo quidem consideratur passio intellect us possibilis secundum quod informatur specie intelligibili. Qua quidem formatus, format secundo vel definitionem vel divisionem vel compositionem, quae per vocem significatur. Unde ratio quam significat
nomen, est definitio; et enuntiatio significat compositionem et division em intellectus. Non
ergo voces significant ipsas species intelligibiles, sed ea quae intellect us sibi format ad
iudicandum de rebus exterioribus." S.th. I, q. 85, a. 2, ad 3; cf. ad 2. For further texts and
literature pertaining to these questions see R. W. Schmidt, The Domain of Logic according
to Saint Thomas Aquinas, M. Nijhoff, The Hague, 1966, pp. 98-117, E. P. Mahoney,
'Sense, Intellect; and Imagination in Albert, Thomas, and Siger', in N. Kretzmann, A.
Kenny, J. Pinborg (eds.), The Cambridge History of Later Medieval Philosophy, Cambridge University Press, Cambridge, 1982, pp. 605 - 611.
16 S.th. I, q. 12, a. 4c, q. 55, a. 2c, q. 58, a. 3 - 4, q. 79, a. 2c, q. 84, a. 7c, q. 85, a. 5c.
17 S.th. I, q. 84, a. 7c and 8c, q. 85, a. Ic.
18 S.th. I, q. 13, a. 8, ad 2, III, q. 25, a. 5, ad 4.
19 S.th. I, q. 29, a. I, ad 3, q. 77, a. I, ad 7, q. 87, a. Ic, In II Post. an, Ie. 13, n. 533,
De spirit. creat., a. II, ad 3, ScG IV, c. I, De ver. q. 10, a. I. For further examples, see
W. H. Kane, 'The Extent of Natural Philosophy', New Scholasticism 31 (1957),90-92;
cf. A. Kenny, The Five Ways. St. Thomas Aquinas' Proofs of God's Existence (Studies
in Ethics and the Philosophy of Religion), Schock en Books, New York, 1969, p. 90.
20 S.th. I, q. 12, a. 8c and ad 4.
21 'Time and Modality in Scholasticism', in S. Knuuttila (ed.), Reforging the Great Chain
of Being (Synthese Historical Library 20), D. Reidel Publ. Co., Dordrecht, 1981, pp.
163 -207.
22 Ibid., pp. 208-217.
23 For some relevant texts, see Note 20, ScG III, 56, De un. verbi inc. a.l, In II Post. an.,
1.6.
24 See, for example, S.th. I, q. 13, a. I, a. 4, a. 9, ad 2, q. 85, a. 2, ad 3, De pot. q. 8, a. I.
2S See J. Pinborg, 'Speculative Grammar', in The Cambridge History of Later Medieval
Philosophy (Note 15 above), pp. 262 - 265, J. Pinborg, 'Die Logik der Modistae', Studia
Mediewistyczne 16 (1975), 39 - 97.
26 See, for example, S.th. I, q. 13, a. 1-4, q. 45, a. 2, ad 2, In VII Metaphys., Ie. I, n.
9; R. M. Mcinerny, The Logic of Analogy. An Interpretation of St. Thomas Aquinas, M.
Nijhoff, The Hague, 1961, pp. 49-66.

BEING QUA BEING IN AQUINAS AND SCOTUS

215

27 "Dicendum quod in omnibus nominibus quae de pluribus analogice dicuntur, necesse

est quod omnia dicantur per respectum ad unum; et ideo illud unum oportet quod ponatur
in definitione omnium ... sicut sanum quod dicitur de animali, cadit in definitione sani
quod dicitur de medicina, quae dicitur sana inquantum causat sanitatem in animali; et in
definitione sani quod dicitur de urina, quae dicitur sana inquantum est signum sanitatis
animalis." S.th. I, q. 13, a. 6c. See also R. M. McInerny, op.cit., pp. 67 -79; J. F. Ross,
'Analogy as a Rule of Meaning for Religious Language', International Philosophical
Quarterly 1 (1961), 468 - 502; reprinted in J. F. Ross, ed. (Note 2 above), pp. 35 -74. In
the same volilme, pp. 75 - 96, there is R. M. McInerny's paper 'Metaphor and Analogy',
reprinted from Sciences Ecctesiastiques 16 (1964), 273 - 289.
28 G. E. L. Owen, 'Logic and Metaphysics in Some Earlier Works of Aristotle', in I. During and G. E. L. Owen (eds.), Aristotle and Plato in the Mid-Fourth Century (Studia
Graeca et Latina Gothoburgensia XI), Almqvist & Wiksell, Gothenburg, 1960, pp.
163 - 190; J. Hintikka, Time and Necessity. Studies in Aristotle's Theory oj Modality, Oxford University Press, Oxford, 1973, pp. I - 26.
29 See H. Lyttkens, The Analogy between God and the World. An Investigation oj Its
Background and Interpretation oj Its Use by Thomas oj Aquino, Almqvist & Wiksell,
Uppsala, 1952, pp. 125, 159-162.
30 For the text, see L. M. De Rijk, Logica Modernorum II, I. The Origin and Early
Development oj the theory oj Supposition (Wijsgerige teksten en studies), Van Gorcum,
Assen, 1968, p. 475.
31 See R. M. McInerny, 'Metaphor and Analogy' (Note 27 above), pp. 81 - 90 and 'The
ratio communis of the Analogous Name', Laval Thtiologique et Philosophique 18 (1962),
9-34.
32 "Sed dicendum est quod unum dividentium aliquid commune potest esse prius altero
dupliciter: uno modo, secundum proprias rationes, aut naturas. dividentium; aliomodo,
secundum participationem ration is illius communis quod in ea dividitur. Primum autem
non tollit univocationem generis .... Sed secundum impedit univocationem generis. Et
propter hoc ens non potest esse genus substantiae et accidentis: quia in ipsa relatione entis,
substantia, quae est ens per se, prioritatem habet respectu accidentis, quod est ens per aliud
et in alio." In I Periherm., Ie. 8, n. 6. Quoted in McInerny, op.cit. (Note 26 above), pp.
96-97. See also S.th. I, q. 16, a. 6c.
33 S.th. I. q. 13, a. 10, De pot., q. 7, a. 7, In IV Metaphys., Ie. I, n. 539, In XI Metaphys.,
Ie. 3, n. 2197. See also Doig, op. cit., pp. 255-275, 358-367.
34 S.th. I, q. 3, a. 4c. See also the texts mentioned in Owens, Op. cit. (Note 8 above), pp.
248-9: In I Sent., d. 19, q. 2, a. 2, De ver. q. 27, a. I, ad 8, In Boethii de hebd., Ie. 2.
3S See the references in Note 33 above.
36 See Lyttkens, op. cit., pp. 266 - 283.
37 De pot., q. 7, a. 2, ad 9; cf. S.th. I, q. 3, a. 4c, q. 4, a. I, ad 3.
38 See Note 32 above.
39 Cf. In I Periherm., Ie. 5, n. 19-22.
I have neglected to discuss the relations between the predicamental and other kinds of
uses of ens in Aquinas, because they are treated by Hermann Weidemann in this volume,
p.181 - 200. There has been a vivid discussion of the interpretation of Aquinas' doctrine
of being among the contemporary thomists. The theories put forward are not purely
historical and an analysis of them lies outside the purview of this paper. See, for example,
J. Maritain, Existence and the Existent, trans!. by L. Galantiere & G. Phelan, Pantheon,
New York, 1948, E. Gilson, Being and Some Philosophers, 2. ed., Pontifical Institute of
Mediaeval Studies, Toronto, 1952; C. Fabro, La nozione meta]lSica di partecipazione

216

SIMO KNUUTTlLA

secondo S. Tommaso d'Aquino, 3 ed., Societa Editrice Internazionale, Torino, 1963; L. B.

Geiger, La Participation dans la philosophie de S. Thomas d'Aquin, 2 ed., Vrin, Paris,
1953; L. De Raeymaeker, The Philosophy of Being, trans\. E. Ziegelmeyer, Herder, SI.
Louis, 1957; 1. De Finance, Etre et Agir dans la Philosophie de Saint Thomas, 2 ed.,
Librairie Editrice de l'Universite Gregorienne, Rome, 1960; M. A. Kr'lpiec, 'The Theory
of Analogy of Being', in S. Kaminski, M. Kurdzialec, and Z. 1. Zdybickaleds.), Theory
of Being, Towarzystwo Naukowe Katolickiego Uniwersytetu Lubelskiego, Lublin, 1980,
pp.31-106.
40 Quoted in S. F. Brown, 'Scotus' Univocity in the Early Fourteenth Century', in De doctrina loannis Duns Scoti. Acta Congressus Scotistici lnternationalis 1966, Vo\. IV (Studia
Scholastica 4), Romae, 1968, p. 36.
41 S. F. Brown, op.cit .. , 'Richard Conington and the Analogy of the Concept of Being',
Franziskanische Studien 48 (1966),297 - 307; 'Avicenna and the Unity of the Concept of
Being', Franciscan Studies 25 (1965), 117 - 150; M. Schmaus, Zur Diskussion uber das
Problem der Univozitiit im Umkreis des Johannes Duns Skotus (Sitzungsberichte der
Bayerischen Akademie der Wissenschaften, Philos.-hist. K\' 1957, 4), Miinchen 1957.
42 Ord. I, d. 3, pars I, q. 1- 2, n. 27 - 29 (ed. Vat Ill), Ord. I, d. 8, pars I, q. 3, n. 68 - 69
(ed. Vat. IV).
43 Ord. I, d. 3, pars I, q. 3, n. 131-151. I don't discuss here the details of Scotus' ontology. However, it may be in order to mention shortly that according to him transcende;1tal being can be predicated in quid and in quale. It is predicated in quid of God and of
everything which is or can be a finite concrete subject of existence. And it is predicated in
quale of what is or can be a modification or qualification to the primary subject of existence. See A. Wolter, The Transcendentals and Their Function in the Metaphysics of
Duns Scotus, The Franciscan Institute, St. Bonaventure, N.Y., 1946, pp. 77-99. For
Scotus' doctrine of the univocity of being in general, see also T. A. Barth, 'Being, Univocity, and Analogy According to Duns Scotus', in 1. K. Ryan and B. M. Bonansea (eds.),
John Duns Scotus 1265 - 1965 (Studies in Philosoph'y and the History of Philosophy 3),
The Catholic University of America Press, Washington, D.C., 1965, pp. 210 - 262; C. L.
Shircel, The Univocity of the Concept of Being in the Philosophy of John Duns Scotus,
The Catholic University of America Press, Washington, D.C., 1942; L. Honnefelder, Ens
inquantum ens. Der Begriff des Seienden als solchen als Gegenstand der Metaphysik nach
der Lehre des Johannes Duns Scotus (Beitriige zur Geschichte der Philosophie und
Theologie des Mittelalters, NF 16), Aschendorff, Miinster, 1979.
44 Ord. I, d. 3, pars I, q. I - 2, no. 29: "Quilibet philosophus fuit certus, ilIud quod posuit
primum principium, esse ens, ... non autem fuit certus quod esset ens creatum vel increatum, prim urn vel non primum... . Confirmatur etiam, nam aliquis videns
philosophos discordare potuit esse certus de quocumque quod aliquis posuit primum principium, esse ens, et tamen propter contrarietatem opinionum eorum potu it dubitare utrum
sit hoc ens vel ilIud. Et tali dubitandi si fieret demonstratio concludens vel destruens aliquem conceptum inferiprem, puta quod ignis non erit ens primum sed aliquod ens posterius
primo ente, non destrueretur ilIe conceptus primus sibi certus, quem habuit de ente, sed
salvaretur in ilIo conceptu particulari probato de igne."
45 See, for example, Ord. I, d. 36, q. un., n. 60 (ed. Vat. VI): "Ita in proposito: homini
in aeternitate inest "non esse aliquid" et chimaerae "non esse aliquid", sed homini non
repugnat affirmatio quae est "esse aliquid" sed tantum inest negatio propter negationem
causae non ponentis, - chimaerae autem repugnat, quia nulla causa posset in ea causare
"esse aliquid". Et quare homini non repugnat et chimaerae repugnat, est, quia hoc est hoc
et iIlud ilIud, et hoc quocumque intellectu concipiente." Cf. also Note 50.

BEING QUA BEING IN AQUINAS AND SCOTUS

217

Ord. IV, d. 8, q. I, n. 2 (ed. Vives XVII), cf. Quod/. q. 3, n. 2 (ed. Vives XXV).
Ord. I, d. 43, q. un., n. 14 (ed. Vat VI): ... per ipsam potentiam "sub ratione qua est
omnipotentia" non habet obiectum quod sit primo possibile, sed per intellectum divinum,
producentem illud primo in esse intelligibili, et intellectus non est formaliter potentia activa
qua Deus dicitur omnipotens; et tunc res producta in tali esse ab intellectu divino - scilicet
intelligibili - in primo instanti naturae, habet se ipsa esse possibili in secundo instanti
naturae, quia formaliter non repugnat sibi esse et se ipso formal iter repugnat sibi habere
esse necessarium ex se." Whatever can be thought of is produced in intelligible esse by the
Divine Intellect. Possible objects are identifiable as members of possible worlds, of which
the Divine Will chooses one to be the actual world (cf. Lect. I, d. 39, q. 1 - 5, n. 62 - 63).
Possible individuals have a positive nature as identifiable candidates to existence, although
they as such have no kind of existence. Cf. Ord. I, d. 36, n. 61. For the connection between
Scotus' ideas of being and possibility, see also L. Honnefelder, op.cit. and 'Die Lehre von
der doppelten ratitudo entis und ihre Bedeutung fiir die Metaphysik des Johannes Duns
Scotus', Deus et Homo ad mentem I. Duns Scoli, Acta Tertii Congressus Scotistici Internationalis 1970 (Studia Scholastico-scotistica 5), Societas Internationalis Scotistica, Romae
1972, pp. 661-671.
48 See Wolter, op.cit.
49 See the discussions in the works mentioned in Note 43 above.
50 For the un~vocity of being in Ockham, see M. Matthew, The Concept of Univocily
Regarding the Predication of God and Creature According to William Ockham, The Franciscan Institute, St. Bonaventure, 1952; P. Boehner, 'Scotus' Teachings According to
Ockham I. On the Univocity of Being', Franciscan Studies VI (1946), 100 - 107; D. C.
Langston, 'Scotus and Ockham on the Univocal Concept of Being', Franciscan Studies 39
(1979), 105 -129. The influence of Scotus' doctrine of being on Suarez is discussed in W.
Hoeres, 'Francis Suarez and the Teaching of John Duns Scotus on Univocatio entis', John
Duns Scotus 1265 -1965 (see Note 43 above), pp. 263 - 290. Short remarks on later influence are to be found in L. Honnefelder, 'Duns Scotus/Scotismus II', The%gische
Rea/enzyk/opiidie 9, De Gruyter, Berlin, New York 1982, pp. 232 - 240.
51 Knuuttila, op.cit., pp. 217 - 234.
52 Basic texts are Lect. I, d. 39, q. 1 - 5 and Ord. I, d. 43, q. un. For a discussion of these
and some other texts, see Knuuttila, op.cit., pp. 217 - 234, 'Modal Logic', The Cambridge
History of Later Medieva/ Philosophy (see Note 15 above), pp. 353 - 355; 'Duns Scotus'
Criticism of the Statistical Interpretation of Modality', in Sprache und Erkenntnis im MiIte/alter (Miscellanea Mediaevalia 13/1), De Gruyter, Berlin - New York, pp. 441 - 450.
53 In addition to references mentioned in Note 47 above, see Ord. I, d. 35, q. un., n. 61;
d. 38, q. un., n. 10 (ed. Vat. VI).
54 For the notion of real concept in Scotus, see Wolter, op.cit. pp. 15 - 17, 65 - 66.
55 After the passage quoted in Note 45 above Duns Scotus states: " ... quia homini non
repugnat (sc. to be something), ideo est possibile potentia logica, ... et illam possibilitatem
consequitur possibilitas obiectiva, et hoc supposita omnipotentia Dei quae respicit omne
possibile (dummodo iIIud sit aliud a se), tamen ilia possibilitas logica, absolute - ratione
sui - posset stare, licet per impossibile nulla omnipotentia earn respiceret." Ord. 1, d. 36,
q. un., n. 61.
56 In Scotus's metaphysics actuality and potentiality (actus ct potentia) belong to the disjunctive transcendentals which are in disjunction proper to quidditative being. Actuality
as a transcendental attribute means existence. See Wolter, op.cit., pp. 145 -148. Wolter
quotes Ord. I. d. 7, q. 1, n. 72 where it is stated: " ... ens in communi non tantum dividitur
per actum et potentiam, sed etiam quodcumque genus entis, et quaecumque species et in46
47

218

SIMO KNUUlTILA

dividuum, quia sic albedo eadem primo est in potentia et postea in actu". Cf. Ord. II, d.
16, q. un., n. 5: "Illud enim individuum, quod nunc est in actu, iIlud idem fuit in potentia".
When an individual has become actual, that same individual was earlier a (merely) potential
individual, i.e. it was a not yet actualized member of that possible world which is the actual
one. It is a member of alternative possible worlds, too. See, for example, Ord. I, d. 41,
q. un., n. 7, d. 44, q. un., n. II (ed. Vat. VI).
S7 See, for example, Ord. II, d. 3, pars I, q. 5-6, n. 191.
S8 Cf. A. Wolter, 'The Formal Distinction', in John Duns Seotus 1265 -1965 (Note 43
above), pp. 54 - 60; 'Is Existence for Scotus a Perfection Predicate, or What?", De doetrina loannis Duns Seoti II (Note 40 above), pp. 175 -182.
S9 In S.th. I, q. 3, a. 5c Aquinas states: "omnia quae sunt in genere uno, communicant
in quidditative vel essentia generis, quod praedicatur de eis in eo quod quid est. Differunt
aut em secundum esse: non enim idem est esse hominis et esse equi, nec huius hominis et
iIlius hominis." For temporal necessity in Aquinas, see, e.g., In I Periherm., Ie. 15, n. 2,
Sent. I, d. 38, q. I, a. 5, ad 3, De ver. q. 2, a. 12, ad 4.
60 John Buridan, Traetatus de eonsequentiis, ed. by H. Hubien (Philosophes medievaux
16), Publications Universitaires, Louvain, 1976, pp. 27, 31 - 40; 58,4 - 60,56; 75, 196 -76,
204; William Ockham, Summa /ogieae, ed. by P. Boehner, G. Gal, and S. Brown
(Guillelmi de Ockham, Opera philosophica et theologica: Opera philosophica I), The Franciscan Institute, St. Bonaventure, N.Y. 1974, pars I, c. 72, pp. 215, 37-218,112. See also
S. Knuuttila, 'Modal Logic' (Note 52 above), pp. 355 - 357 and 'Topics in Late Medieval
Intensional Logic', in I. Niiniluoto and E. Saarinen (eds.), Intensional Logic: Theory and
Applications (Acta Philosophica Fennica 35), Societas Philosophica Fennica, Helsinki,
1982, pp. 32 - 38; Ockham's Theory of Terms. Part 1 oj the Summa logieae, transl. and
introd. by M. J. Loux, University of Notre Dame Press, Notre Dame, London, 1974, pp.
40-44.
61 W. Hoeres, 'Francis Suarez and the Teaching of John Duns Scotus on univocatio entis',
John Duns Seotus 1265 -1965 (Note 43 above), pp. 263 - 290.
62 Disputationes metaphysicae (reprinted from Opera omnia, Vives 1856 - 78), Olms,
Hildesheim, 1965, Vo!. II, pp. 176 - 177, 190- 203,207 - 223. There is a short discussion
on Suarez's views on possibility and reality in J. A. Trentman, 'Scholasticism in the Seventeenth Century', The Cambridge History of Later Medieval Philosophy (Note 15 above),
pp.826-827.
63 For Leibniz's modal theory, see H. Poser, Zur Theorie der ModalbegrifJe bei G. W.
Leibniz (Studia Leibnitiana Supplementa VI), Franz Steiner Verlag, Wiesbaden, 1969; H.
Schepers, 'Zum Problem der Kontingenz bei Leibniz. Die beste der moglichen Welten', in
Collegium Philosophieum. Studien Joachim Ritter zum 60. Geburtstag, Basel, Stuttgart,
1965, pp. 326-350; J. Hintikka, 'Leibniz on Plenitude, Relations, and the "Reign of
Law''', in S. Knuuttila (ed.) (Note 21 above), pp. 259- 286; B. Mates, 'Leibniz on Possible
Worlds' , in Logic, Methodology, and Philosophy of Science lIl, Proceedings of the Third
International Congress, ed. by B. van Rootselaar and J. F. Staal (Studies in Logic and the
Foundations of Mathematics), North-Holland Pub!. Co., Amsterdam, 1968, pp.
507-529.
64 See, for example, Discours de Metaphysique, Sections 9 - II, transl. in Gottfried
Wilhelm Leibniz, Philosophical Papers and Letters. A selection translated and edited with
an introduction by L. E. Loemker (Synthese Historical Library 2), 2 ed., D. Reidel Pub!.
Co., 1976, pp. 308 - 309.
65 Diseoursde Metaphysique, Sections 8 - 9, 13; the letter to Arnauld, 14 July 1686, transl.
in Loemker, op.cil., pp. 331 - 338.

BEING QUA BEING IN AQUINAS AND SCOTUS

219

BIBLIOGRAPHY
Aristotle, De anima, ed. with introd. and comm. by W. D. Ross, Oxford University Press,
Oxford, 1961.
Aristotle, Ethica Nicomachea, ed. by I. Bywater (Scriptorum Classicorum Bibliotheca Oxoniensis), Oxford University Press, Oxford, 1970 (first published 1894).
Aristotle, Generation of Animals, with an English Translation by A. L. Peck (Loeb
Classical Library), Heinemann & Harvard University Press, Cambridge, Mass. and
London, 1953.
Aristotle, Metaphysics. A Revised Text with Introduction and Commentary by W. D.
Ross, 2 Vols., Oxford University Press, Oxford, 1924.
Aristotle, Parts of Animals, with an English Translation by A. L. Peck (Loeb Classical
Library), Heinemann & Harvard University Press, Cambridge, Mass. and London,
1945.
Aristotle, Prior and Posterior Analytics. A Revised Text with Introduction and Commentary by W. D. Ross, Oxford University Press, Oxford, 1949.
Barth, T. A., 'Being, Univocity, and Analogy According to Duns Scotus', in J. K. Ryan
and B. M. Bonansea (eds.), 1965, pp. 210-262.
Boehner, Ph., 'Scotus Teachings According to Ockham I. On the Univocity of Being',
Franciscan Studies VI (1946), 100-107.
Brown, St. F., 'Avicenna and the Unity of the Concept of Being', Franciscan Studies 2S
(1965), 117 - 150.
Brown, St. F., 'Richard of Conington and the Analogy of the Concept of Being', Franziskanische Studien 48 (1966), 297 - 307.
Brown, St. F., 'Scotus's Univocity in the Early Fourteenth Century', in De doctrina loannis Duns Scoti. Acta Congressus Scotistici lnternationalis 1966, Vol. IV (Studia
Scholastico-Scotistica 4), Rome, 1968, pp. 35 - 41.
Dancy, R., 'On Some of Aristotle's Second Thoughts About Substances: Matter', The
Philosophical Review 87 (1978),372-413.
Doig, J. C., A quinas on Metaphysics. A Historico-doctrinal Study of the Commentary on
the Metaphysics, Nijhoff, The Hague, 1972.
Fabro, c., La nozione metafisica di partecipazione secondo S. Tommaso d'Aquino, 3 ed.,
Societa Editrice Internazionale, Turin, 1963.
Finance, J. de, Etre et Agir dans 10 Philosophie de Saint Thomas, 2 ed., Librairie Editrice
de l'Universite Gregorienne, Rome, 1960.
Geiger, L.-B., La Participation dans 10 philosophie de S. Thomas d'Aquin, 2 ed., Vrin,
Paris, 1953.
Gilson, E., Being and Some Philosophers, 2 ed., Pontifical Institute of Mediaeval Studies,
Toronto, 1952.
Happ, H., Hyle. Studien zum arislOtelischen Materie-Begriff, De Gruyter, Berlin, New
York, 1971.
Hintikka, J., 'Leibniz on Plentitude, Relations, and the "Reign of Law" , in S. Knuuttila
(ed.), 1981. pp. 259- 286.
Hintikka. J .. Time and Necessity. Studies in Aristotle's Theory of Modality, Oxford
University Press, Oxford. 1973.
Hoeres. W., 'Francis Suarez and the Teaching of John Duns Scotus on univocatio entis',
in J. K. Ryan and B. M. Bonansea (eds.). 1965, pp. 263 - 290.
Honnefelder, L.. Ens inquantum ens. Der Begriff des Seienden als solchen als Gegenstand

220

SIMO KNUUTTILA

der Metaphysik nach der Lehre des Johannes Duns Scotus (Beitrage zur Geschichtt! der
Philosophie und Theologie des Mittelalters, NF 16), Aschendorff, Miinster, 1979.
Honnefelder, L., 'Die Lehre von der doppelten Ratitudo entis und ihre Bedeutung fUr die
Metaphysik des Johannes Duns Scotus', in Deus et Homo ad mentem I. Duns Scoti. Acta Tertii Congressus Scotistici Internationalis 1970 (Studia Scholastico-scotistica 5),
Rome, 1972, pp. 661 - 671.
Honnefelder, L., 'Duns Scotus/Scotismus II', Theologische Realen~klopiidie 9, De
Gruyter, Berlin, New York, 1982, pp. 232-240.
John Buridan, Tractatus de consequentiis, ed. by H. Hubien (Philosophes medievaux 16),
Publications Universitaires, Louvain, 1976.
John Duns Scotus, Opera omnia studio et cura Commissionis Scotisticae, Civitas Vaticana,
1950 - .
John Duns Scotus, Opera Omnia, Vives, Paris, 1891 - 5.
Kane, W. H., 'The Extent of Natural Philosophy', New Scholasticism 31 (1957), 85 - 97.
Kenny, A., The Five Ways. St. Thomas Aquinas' ProoJs oj God's Existence (Studies in
. Ethics and the Philosophy of Religion), Schocken Books, New York, 1969.
Knuuttila, S., 'Duns Scotus's Criticism of the "Statistical" Interpretation of Modality',
in Sprache und Erkenntnis im Mittelalter (Miscellanea Mediaevalia 13/1), De Gruyter,
Berlin, New York, 1981, pp. 441-450.
Knuuttila, S., 'Modal Logic', in N. Kretzmann et al. (eds.), 1982, pp. 342 - 357.
Knuuttila, S., 'Time and Modality in Scholasticism', in S. Knuuttila (ed.), 1981, pp.
163 - 257.
Knuuttila, S., 'Topics in Late Medieval Intensional Logic', in I. Niiniluoto and E. Saarinen
(eds.), Intensional Logic: Theory and Applications (Acta Philosophica Fennica 35),
Societas Philosophica Feunica, Helsinki, 1982, pp. 26-41.,
Knuuttila, S. (ed.), ReJorging the Great Chain oj Being. Studies oj the History oj Modal
Theories (Synthese Historical Library 20), D. Reidel, Dordrecht, 1981.
Krllpiec, M. A., 'The Theory of Analogy of Being', in S. Kaminski, M. Kurdzialec, and
Z. J. Zdybicka (eds.), Theory oj Being, Towarzystwo Naukowe Katolickiego Uniwersytetu Lubelskiego, Lublin, 1980, pp. 31 - 106.
Kretzmann, N., A. Kenny, and J. Pinborg (eds.), The Cambridge History oj Later
Medieval Philosophy, Cambridge University Press, Cambridge, 1982.
Langston, D. c., 'Scotus and Ockham on the Univocal Concept of Being', Franciscan
Studies 39 (1979), 105 -129.
Leibniz, G. W., Philosophical Papers and Letters, trans!. and ed. with introduction and
notes by L. E. Loemker, 2 ed. (Synthese HistoricalLibrary 2), D. Reidel, Dordrecht,
1976.
Loux, M. J. (trans!. & introd.), Ockham's Theory oJ Terms. Part I oJthe Summa Logicae,
University of Notre Dame Press, Notre Dame, London, 1974.
Loux, M. J., Substance and Attribute. A Study in Ontology, D. Reidel Pub!. Co., Dordrecht, 1977.
Lyttkens, H., The Analogy between God and the World. An Investigation oj Its
Background and Interpretation oj Its Use by Thomas oj Aquino, Almqvist & Wiksell,
Uppsala, 1952.
Mahoney, E. P., 'Sense, Intellect and Imagination in Albert, Thomas, and Siger', in N.
Kretzmann et al. (eds.), 1982, pp. 602 - 622.
Maritain, J., Existence and the Existent (trans!. by L. Galantiere & G. Phelan), Pantheon,
New York, 1948.
Mates, B., 'Leibniz on Possible Worlds', in Logic, Methodology, and Philosophy oj

BEING QUA BEING IN AQUINAS AND SCOTUS

221

Science Ill; Proceedings oj the Third International Congress, ed. by B. van Rootselaar
and J. F. Staal (Studies in Logic and the Foundations of Mathematics), North-Holland
Publ. Co., Amsterdam, 1968, pp. 507 - 529.
Matthew, M., The Concept oj Univocity Regarding the Predication oj God and Creature
According to William Ockham, The Franciscan Institute, St. Bonaventure. 1952.
Mcinerny, R. M., The Logic oj Analogy. An Interpretation oj St. Thomas Aquinas, M.
Nijhoff. The Hague, 1961.
Mcinerny, R. M., 'Metaphor and Analogy'. Sciences Ecctesiastiques 16 (1964),273 - 289;
reprinted in J. F. Ross (cd.), 1971, pp. 75 - 96.
Mcinerny, R. M., The ratio communis of the Analogous Name', Laval Theologique et
Philosophique 19 (1962), 9 - 34.
Owen, G. E. L., 'Logic and Metaphysics in Some Earlier Works of Aristotle', in I. During
and G. E. L. Owen (eds.), Aristotle and Plato iti the Mid-Fourth Century (Studia
Graeca et Latina Gothoburgensia XI), Almqvist & Wiksell, Gothenburg, 1960, pp.
163 - 190.
Owens, J., 'The Accidental and Essential Character of Being in the Doctrine of St. Thomas
Aquinas', Mediaeval Studies 20 (1958), I - 40; reprinted in J. Owens, St. Thomas
Aquinas on the Existence oj God, pp. 52 - 96.
Owens, J., 'Common Nature: A Point of Comparison between Thomistic and Scotistic
Metaphysics', in J. F. Ross (ed.), 1971, pp. 185 -209.
Owens, J., St. Thomas Aquinas on the Existence oj God. Collected Papers oj Joseph
Owens, C. Ss. R. (ed. by J. Catan), State University of New York Press, Albany, 1980.
Pinborg, J., 'Die Logik der Modistae', Studia Mediewistyczne 16 (1975),39-97.
Pinborg, J., 'Speculative Grammar', in N. Kretzmann et al. (eds.), 1982, pp. 254-269.
Plato, The Republic with an English translation by P. Shorey, 2 vols. (Loeb Classical
Library), Heinemann & Harvard University Press, Cambridge, Mass., and London,
1930 - 35.
Poser, H., Zur Theorie der Modalbegrifje bei G. W. Leibniz (Studia Leibnitiana Supplementa VI), Franz Steiner Verlag, Wiesbaden, 1969.
Raeymaeker, L. de, The Philosophy oj Being (transl. by E. Ziegelmeyer), Herder, St.
Louis, 1957.
Rijk, L. M. de, Logica Modernorum II, I. The Origin and Early Development oj the
Theory oj Supposition (Wijsgerige teksten en studies 16), Van Gorcum, Assen, 1967.
Ross, J. F., 'Analogy as a Rule of Meaning for Religious Language', International
Philosophical Quarterly 1 (1961), 468 - 502; reprinted in J. F. Ross (ed.), 1971, pp.
35 - 74.
Ross, J. F. (ed.), Inquiries into Medieval Philosophy. A Collection in Honor oj Francis
P. Clarke (Contributions in Philosophy 4), Greenwood Publ. Co., Westport, 1971.
Ryan, J. K. and B. M. Bonansea (eds.), John Duns Scotus 1265 - 1965 (Studies in
Philosophy and the History of Philosophy 3), The Catholic University of America
Press, Washington, D.C., 1965.
Schepers, H., 'Zum Problem der Kontingenz bei Leibniz. Die beste der moglichen Welten',
in Collegium Philosophicum. Studien Joachim Riller zum 60. Gebur/stag, Schwabe &
Co., Basel, Stuttgart, 1965, pp. 326- 350.
Schmaus, M .. Zur Diskussion iiber das Problem der Univozitiit im Umkreis des Johannes
Duns Scotus (Sitzungsberichte der Bayerischen Akademie der Wissenschaften, Philos.hist. KI., 1957,4), Munchen, 1957.
Schmidt, R. W., The Domain oj Logic According to Saint Thomas Aquinas, M. Nijhoff,
The Hague. 1966.

222

SIMO KNUUTIILA

Shircel, C. L., The Univocity of the Concept of Being in the Philosophy of John Duns
Scotus, The Catholic University of America Press, Washington, D.C., 1942.
Suarez, Francisco, Disputationes metaphysicae (reprinted from Opera omnia, Vives, Paris
1856-78), G. Olms, Hildesheim, 1965.
Thomas Aquinas, In Aristotelis libros Peri Hermeneias et Posteriorum. Analyticorum expositio, ed. by R. Spiazzi, Marietti, Turin, 1964.
Thomas Aquinas, In duodecim libros Metaphysicorum Aristotelis expositio, ed. by R.
Spiazzi, Marietti, Turin, 1950.
Thomas Aquinas, Expositio super librum Boethii De Trinitate, ed. by B. Decker (Studien
und Texte zur Geistesgeschichte des Mittelalters 4), Brill, Leiden, 1955.
Thomas Aquinas, In Aristotelis libros Peri Hermeneias et Posteriorum Analyticorum expositio, ed. by R. Spiazzi, Marietti, Turin, 1964.
Thomas Aquinas, In Aristotelis librum De anima commentarium, ed. by A. Pirotta,
Marietti, Turin, 1959.
Thomas Aquinas, Opuscula philosophica, ed. by R. M. Spiazzi, Marietti, Turin, 1954.
Thomas Aquinas, Opuscula theologica, ed. by R. A. Verardo et al., Marietti, Turin, 1954.
Thomas Aquinas, Quaestiones disputatae I - II, ed. by R. M. Spiazziet al., Marietti,
Turin, 1953.
Thomas Al1uinas, Quaestiones quodlibetales, ed. by R. M. Spiazzi, Marietti, Turin, 1956.
Thomas Aquinas, Scriptum super libros sententiarum, Vols. I - II, ed. by P. Mandonnet,
Lethielleux, Paris, 1929.
Thomas Aquinas, Summa theologiae, in Opera omnia iussu Leonis XIII, P. M., edita,
Vols. 4-12, Vatican Polyglot Press, Rome, 1896-1906.
Trentman, 1. A., 'Scholasticism in the Seventeenth Century', in N. Kretzmann et al. (eds.),
1982, pp. 818-837.
Weidemann, H., 'The Logic of Being in Thomas Aquinas', this volume, pp. 181-200.
William Ockham, Summa logicqe, ed. by Ph. Boehner, G. Gal, and St. Brown (Guillelmi
de Ockham, Opera philosophica et theologica: Opera philosophica I), The Franciscan
Institue, St. Bonaventure, 1974.
Wolter, A. B., 'The Formal Distinction', in 1. K. Ryan and B. M. Bonansea (eds.), 1965,
pp. 45 -60.
Wolter, A. B., 'Is Existence for Scotus a Perfection, Predicate, or What?' in De doctrina
loannis Duns Scoti. Acta Congressus Scotistici lnternationalis 1966, Vol. II (Studia
Scholastico-Scotistica 2), Rome, 1968, pp. 175 - 182.
Wolter, A. B., The Transcendentals and Their Function in the Metaphysics of Duns
Scotus, The Franciscan Institute, St. Bonaventure, 1947.

Dept. of Systematic Theology,

University of Helsinki,
Neitsytpolku 1b,
SF-00140 Helsinki 14, Finland.

LILLI ALANEN

ON DESCARTES'S ARGUMENT FOR DUALISM AND THE

DISTINCTION BETWEEN DIFFERENT KINDS OF BEINGS

1. INTRODUCTORY REMARKS

It is a well-documented and widely-known fact that Descartes, in defending his new philosophy, relied heavily on Scholastic theories and notions, and that he also adapted them skillfully for his own purposes. l For
instance, in arguing that the mind and the body are radically different
and mutually independent substances, Descartes used a theory of distinctions which he seems largely to have taken over from Suarez. 2 However,
although Descartes's argument for the mind-body distinction has often
been discussed and analyzed in recent literature, 3 the theory of distinctions on which it is based and to which Descartes also refers in developing
his argument has not been extensively studied. This is somewhat
surprising. 4 For the interpretation of Descartes's dualism, as well as the
assessment of the argument supporting it is, it seems to me, dependent
on the sense given to the notion of a real distinction as used by Descartes
and his predecessors. It is my contention that Descartes's argument,
when interpreted in the light of traditional uses of this notion, is both
more cogent and less unproblematic than is usually thought. As I will try
to show, Descartes can at least be said to have answered, in a satisfactory
manner, the criticism of Arnauld, who was the most perspicacious and
serious of Descartes's contemporary objectors. 5 I do not want to claim
that Descartes's dualism is an unproblematic doctrine. My aim, rather,
is to show that many of the difficulties of this doctrine are due to ambiguities in the terminology which Descartes took over from
Scholasticism, as well as to the application he made of them. 6
Descartes's own theory of distinction which is exposed In the Principles, seems to be relatively simple and clear. (Cf. below, Section 5.) It
is important, however, to bear in mind that Descartes's use of traditional
Scholastic notions and concepts is not always very consistent. For
although Descartes developed and exposed his views and arguments in
terms of common, traditional concepts used by the School, he was an innovator in many respects. Whenever he thought it arpropriate, he also
liked to stress the originality of his views. The problem is that it is not
always very clear when Descartes, in using the terminology of the
223
S. Knuuttila and J. Hinlikka (eds.), The Logic oj Being, 223 - 248.
1986 by D. Reidel Publishing Company.

224

LILLI ALANEN

Scholastics, aligns himself with the views that inform this terminology,
and when he is on the contrary innovating, giving new senses to traditional terms. 7 The interpretation of the key-concepts of Descartes's
theory of distinctions, such as substance, essence, attribute, and mode,
for instance, is therefore delicate and problematical. This also contributes to the difficulty of evaluating the originality of Descartes's
arguments and views. Descartes claimed to be the first to assert that the
mind consists in "thought alone" or the faculty of thinking, and he
seems also to have considered his proof of a real distinction between the
mind and the body as unprecedented in force and cogency. 8 However,
the mere assertion of a real distinction between the mind and the body,
or between form and matter, seems to have been something of a commonplace among Scholastic philosophers and could therefore hardly, in
itself, justify any claims to originality. (Cf. below, Section 4.)
In order to compare Descartes's theory of distinctions with different
accounts of distinctions given by some of his predecessors, I will begin
by outlining briefly the general problem-context in which the medieval
theories of distinctions were developed and discussed. I will also present,
in a general manner, the main distinctions and concepts discussed by the
Scholastics which are relevant to the understanding of Descartes's theory
and his application of it in the proof for mind-body dualism.
2. SUBSTANCES, REAL BEINGS AND CONCEPTUAL BEINGS

Originally, the medieval discussions of different kinds of ontological

distinctions were centered on two different but related problems, the one
philosophical, and the other theological. The philosophical problem of
distinctions was connected with the discussion of the nature of universals, and the common source for this discussion was mainly Aristotle's
remarks on numerical, specific and generic sameness and difference. The
theological problem was that of the nature of the Trinity, and the very
term distinction seems first to have been introduced in the discussion of
this problem. 9 Difficulties related to these problems were dealt with by
nearly aU medieval philosophers, and different accounts or theories of
distinctions can be found in the works of all the major Scholastics, e.g.,
Bonaventura, Aquinas, Duns Scotus, Ockham and Suarez.
The solution proposed by these philosophers to the particular problems mentioned above is usually of a wider interest, however. Since
most of the medieval philosophers followed Aristotle in treating being
and unity as convertible notions,10 there is a close connection between

ON DESCARTES'S ARGUMENT FOR DUALISM

225

the general doctrine of being adopted by a particular philosopher and the

theory of distinctions defended by him: the latter constitutes, as one
author notes, a key to the understanding of the former. II The account
of these distinctions offers, notably, an answer to the questions of what
kind of entities or beings there are in the world, how they are related to
each other, and where or how the distinction between so-called real (i.e.,
extra-mental) beings or things and merely conceptual (i.e., mental) entitie~ should be drawn. This holds good, ajortiori, of Descartes, who was
concerned neither with the problem of universals, nor with that of the
Trinity, at least not directly. But Descartes's reduction of real things to
purely mental things or substances, on the one hand, and merely extended
substances, on the other hand, could be accomplished only at the price
of a radical change in the application of traditional terms such as
substance, essence, accident and mode. Descartes was consequently
forced to give his own account of the traditional ontological distinctions
discussed by the Scholastics. 12
While all the Scholastics, and Descartes with them, seem to agree on
a fundamental division of beings or things into real, extra-mental entities
(res, entia realis), and mere conceptual beings (entia rationis), there is
considerable disagreement on how this division should be drawn and also
on the question of whether there are, in addition to these two classes,
some kind of intermediate entities, and how these should be classified or
accounted for.
The origin of this difficulty can be traced to Aristotle's comments on
the various uses (or meanings) of the terms to be, being, substance,
essence and accident in the Metaphysics, and to his well-known and
notoriously problematic distinction between primary and secondary
substances in the Categories. 13 The question whether the categories introduced by Aristotle are to be understood as merely linguistic or logical
categories, or instead as metaphysical or ontological categories, seems to
be left open by Aristotle. Thus, while Aristotle seems to reserve the term
substance primarily for things existing per se, i.e., as a subject of predication, and as opposed to what is said to exist only in alio or as predicated
of something else, he also uses the term in a second sense, of that which
is predicated of a subject(e.g., the "shape" or the "form" of each thing;
Met.!:. 8, IOl7b23 - 25). Although these secondary substances are said to
be "separable" from the subjects of which they are predicated they are
not considered by Aristotle to exist independently of these. Without being capable of separate existence, they nevertheless are according to
Aristotle real beings, i.e., extra-mental or mind-independent entities or

226

LILLI ALANEN

things. The problem of accounting for these secondary substances posed

by Aristotle constitutes, one could say, the core of the problem of universals discussed by the Medievals.
3. MEDIEVAL THEORIES OF DISTINCTIONS

As regards the medieval discussions of the distinctions to be considered

here, a main difficulty, related to the above problem, was to determine
the criteria for particular distinctions between things and concepts. Most
of the medieval philosophers seem to have admitted the following three
basic distinctions:
(I) A real distinction (distinctio realis), i.e., a distinction between real
things or individuals in the extra-mental world. This distinction was
usuaIly defined as a distinction between thing and thing (inter rem et
rem), existing before the operation of the intellect.
(2) A purely mental distinction (distinctio rationis), i.e., a distinction
created by the mind (per opus intellectus).
(3) An intermediate distinction which was generally defined with
reference to the mind, but which, differently from the purely mental
distinction, was usually conceived as a distinction having a basis in the
nature of things and therefore corresponding to some kind of diversity
or non-identity outside the intellect. 14 There was, however, no general
agreement on how this intermediate distinction should be accounted for
or characterized. The controversial notion of a formal distinction introduced by Duns Scotus, for instance, can be seen as one attempt to
clarify the nature of this foundation in reality. 15
The answer to the question concerning the criteria of these different
distinctions depends of course on the sense given to the notion of a
"thing" and the conditions required for something to be qualified as a
"real thing". Following Aristotle on this point the Scholastic
philosophers usually considered as real things not only individual things
(substances in Aristotle's first sense of the term), but also their forms and
qualities. In so far as the qualities were understood as "real", i.e., mindindependent entities or things, they were considered either as universal,
common natures, or as individual qualities or "things' (res). The former
view was represented, roughly, by thinkers considered to be "Platonic"
realists and "Aristotelian" conceptualists, the latter by the nominalists.
According to the teaching common to most of these thinkers, the distinctions between entities such as matter and form, the body and the soul,
the mind and its particular thoughts, were considered as instances of a

ON DESCARTES'S ARGUMENT FOR DUALISM

227

real distinction, i.e., of a distinction between thing and thing. This means
that although these entities are actually united (e.g., matter and form
which are always found together and the composition of which was supposed to constitute the individual substances or substances in Aristotle's
primary sense of the word) they are separable in the sense that at least
one of them can be conceived as existing independently of the other:
more precisely, their existence is not necessarily bound to that combination in which they are actually found. Thus, most of the Scholastics seem
to agree on the possibility of conceiving, for instance, this form without
this specificity or this particular matter, or the body without the soul, or
the soul without these particular thoughts, ot, more generally, a
substance without its accidental qualities. This may be contrasted with
the distinction between merely conceptual or mental entities (ens rationis) , for instance, between 'man' and 'rational animal', or with the
more controversial distinctions defined as intermediates between the real
and the mental distinction, e.g., the distinction between rationality and
animality in man. Another instance of this intermediate or "formal"
distinction is that between the Persons of the Trinity, or God's attributes,
as his goodness, mercy and justice. The entities belonging to this intermediate class were often defined as conceptual (or formal) but they
were not regarded as created by the intellect. Although they were supposed, contrary to the purely mental entities, to have some kind of basis in
the nature of (extra-mental) thin~s, they were regarded as inseparable
from these things. Not being capable of separate existence these entities
were often characterized as different aspects of real things, existing in
these things before and independently of the operation of the intellect. 16
A case discussed in this connection, besides those mentioned above, was
that of the soul and its faculties: the faculties of the soul, according to
many authors, can be conceived as distinct entities in spite of the fact that
the soul is one and undivided and its faculties, consequently, are inseparable from the soul and from each other.
As noted before, there was, however, much controversy on how these
various distinctions should be classified. The Thomists, for instance,
used the term real distinction in a somewhat different sense and extended
it to items which, according to other authors, are separable only in
thought, as the soul and its faculties, essence and existence. But later
Scholastics, such as Duns Scot us, Ockham and their followers, seem to
have restricted the real distinction to things considered as separable in the
extra-mental reality. Hence, as opposed to the other distinctions assumed by these authors, a real distinction, for the later Scholastics,

228

LILLI ALANEN

always presupposes either a mutual or at least a non-mutual separability

with regard to the existence of the items or things considered. 17 Instances
of a mutual separability, in the case of actually composed or united entities, are a house and the elements of which it is built, or the body and
the hand, and so on. In the case of non-mutual separability, only one of
the united things can exist without the other. This criterion of separability with regard to existence is, as will be seen, important for Descartes too.
But contrary to some of his Scholastic predecessors Descartes insists that
a real distinction always requires a mutual separability. Thus, for
Descartes the contrast between a substance and its qualities, which
Descartes considers as "modes", cannot be an instance of a real distinction, since the modes cannot (according to the common teaching) subsist
in themselves independently of the substance in which they inhere. Such
distinctions, which do not satisfy the criterion of a mutual separability,
are not, according to Descartes, real but modal distinctions (cf. below,
Section 5).
4. POSSIBLE SEPARABILITY AND REAL SEPARABILITY

It is important to note that things which are considered as separable need

not, according to the later Scholastics, be actually separated in the extramental world; they need not even be naturally separable at all. It is sufficient that they can be conceived apart or independently of each other
without contradiction. More precisely, it suffices that one of two things,
which are perceived as united to each other, can be conceived independently of the other without contradiction, in order to conclude that
it is really distinct from the other. According to a principle that the
Scholastics generally agri'.:d upon, and that Descartes also frequently
relies on, the omnipotence of God is limited only by what is selfcontradictory.18 It follows from this principle that whatever can be conceived without contradiction is possible, i.e., can be created or posited
by the power of God. 19 Thus, although the mind cannot exist naturally,
according to the Scholastics, apart from the body thai it is supposed to
inform, nor the qualities without the substance in which they are perceived and independently of which they have no subsistence, the mind
and the body as well as the substance and its qualities can be said to be
really distinct by the above principle. For it is possible, at least, to conceive, without contradiction, the body as existing without the mind, and
the substance without its accidental qualities . Conversely, although the
soul or the form in general cannot be conceived as existing independently

ON DESCARTES'S ARGUMENT FOR DUALISM

229

of some individual body or particular piece of matter, nor the qualities

without some individual thing or substance of which they are predicated,
it is possible to conceive, without contradiction, the soul or the form, and
the qualities respectively, without this particular body or this particular
substance in which they are actually perceived. They are therefore, according to the later Scholastics, real things which are really distinct from
each other, in the sense that they can always be separated, i.e., created
apart, by the omnipotence of God. 20
Interpreted in the light of this principle (i .e. the principle of God's omnipotence) and the application made of it by the Scholastics, Descartes's
proof that the mind and the body are really distinct things does not seem
very controversial; on the contrary, it seems rather trivial. That it was
problematic in the eyes of Descartes's contemporaries is, however, not
difficult to understand. For, as mentioned above, Descartes restricts the
real distinction to mutually separable things, to things each one of which
can exist separately (i.e., be conceived without contradiction independently of each other). Such things, as Descartes and his contemporaries understood it, are individual things or substances in Aristotle's
primary sense. In order to prove that the mind and the body are really
distinct in this restricted sense Descartes had therefore to prove that the
mind is a substance in the primary and original sense of the word, i.e.,
a thing the existence of which can be known independently of, and as
Descartes also purports to show, before the existence of the body is
known. Although this is opposed to what most of the Scholastics considered as common sense, and challenges the Aristotelian definition of
the mind as the form of the body, the doctrine, interestingly, is not new
or unheard of in Scholasticism. For instance, the question of whether or
not individual substantial forms or qualities are numerically distinct entities and hence mutually separable from the individual things in which
they inhere was discussed before, and notably by Ockham, who, for his
part, did not hesitate to answer it affirmatively (cf. Note 20 above).
Ockham's position was however radical indeed, and it is clear from the
objections of the theologians Caterus and Arnauld, that the thesis advanced by Descartes, according to which the mind, considered in itself,
is a substance, i.e., a self-subsisting thing, or as Descartes also calls it,
a "complete" thing which can be known independently of the body,
could not be accepted without argument.

230

LILLI ALANEN

5. DISTINCTIONS ADOPTED BY DESCARTES

Before discussing this thesis and Descartes's argument for it, I will present, briefly, the theory of distinctions upon which Descartes's reasoning
is implicitly based, as it is developed in the Principles, I, 60 - 62. As
noted above it corresponds largely to the theory of Suarez. 2t Like
Suarez, Descartes admits three different kinds of distinctions, but he interprets them, characteristically, in his own way. They are: the real
distinction, the modal distinction and the distinction of reason.
The real distinction according to Descartes "is properly speaking
found between two or more substances" (AT VIII, 28; HR II, 243). By
a substance, according to a definition given earlier in the same text,
Descartes understands "nothing else than a thing which so exists that it
needs no other thing in order to exist". Absolutely speaking, there is only
one substance which can be said to fulfill this requirement, namely God:
other things can exist only by the help or concourse of God. Therefore,
Descartes adds, the Scholastics are right in saying that "the word
substance does not pertain univoce to God and other things" because
"no common signification for this appellation which will apply equally
to God and to them can be distinctly understood" (AT VIII, 24; HR
I, 239 - 249). However, the concept can be attributed to created
substances, in so far as these' 'need only the concurrence of God in order
to exist". And in this sense it can be attributed univocally, in Descartes's
view, to the soul and the body (AT VIII, 24 - 25; HR I, 240). As regards
the criteria or sign of a real distinction Descartes says that
we can conch;de that two substances are really distinct one from the other from the sole
fact that we can conceive the one ciearly and distinctly without the other. For in accordance
with the knowledge we have of God. we are certain that he can carry into effect all that
of which we have a distinct idea. (AT VIIl. 28; HR 1.243. Cf. also AT VII. 162 and
170-171; HR II. 53 and 59.)22

The modal distinction, as Descartes understands it, is of two sorts: (a)

"between the mode properly speaking, and the substance of which it is
the mode", and (b) "between two modes of the same substance" (AT
VIII, 29; HR I, 244). What distinguishes the modal distinction from the
real distinction is the fact that a substance can be clearly and distinctly
conceived without the mode which therefore differs from it, whereas a
mode cannot, reciprocally, be perceived without the substance (A T VII I,
29; HR I, 244). By a mode Descartes means "what elsewhere is termed
attribute or quality", when a substance is considered "as modified or

ON DESCARTES'S ARGUMENT FOR DUALISM

231

diversified by them" (i.e., by the qualities or attributes) (AT VIII, 26;

HR I, 241). More precisely, by a mode Descartes seems to understand accidental qualities, and he reserves, generally, the term attribute to the
permanent and essential properties of substances. Instances of a modal
distinction in the first sense are: that between a corporeal substance (e.g.,
a stone) and its figure or movement, or between a mind and its act of affirming or recollecting; and in the second sense: the distinction between
the form and the figure of the stone, or between the act of recollecting
and the act of affirmating in the mind. It is characteristic of the distinction between two modes of the same substance that the one mode can be
conceived or recognized without the other, but neither the one nor the
other mode can be conceived "without recognizing that both subsist in
one common substance". Thus, it is possible to perceive this particular
stone clearly without thinking of its form or movement, but it is not
possible to perceive either of these modes clearly and distinctly without
perceiving the stone, i.e., without the substance or thing to which they
are attributed and that they can be said to modify or diversify. In a
similar way, the mind can be conceived without its particular thoughts,
but none of these can be conceived without a mind, i.e., a thinking
substance (AT VIII, 29; HR I, 244).23
Finally, the distinction of reason is defined as a distinction created by
the thought, e.g., that
between a substance and some of its attributes without which it is not possible that we
should have a distinct knowledge of it, or between two such attributes of the same
substance. (AT VIII, 30; HR t, 245; cf. note 16 above)

Such attributes, according to Descartes, are the only two essences or

essential attributes of things that he admits, namely, thought, and extension, together with the common attributes as unity, being, duration, etc.
(cf. the transcendentals of the Scholastics). Duration can be distinct from
a substance only in thought, and this holds for all the other common attributes too: they are not separable from the substance in reality. They
can hence be conceived without a substance only by abstraction of the
mind.
The crucial difference between the real distinction on the one hand,
and the modal distinction and the distinction of reason on the other
hand, is thus that these latter concern entities which cannot exist by
themselves and which are therefore not separable with regard to their existence from the substances in whIch they subsist or inhere (of which they
are predicated), and which can therefore not be clearly and distinctly
perceived or conceived without the substance of which they are either ac-

232

LILLI ALANEN

cidental or permanent properties. And although such entities can be conceived without the substance in which they inhere by way of abstraction,
they can never be conceived or known clearly and distinctly in this way.
In so far as they are conceived by abstraction they are not conceived or
known as "complete" things. What Descartes has to prove, therefore,
in order to show that the mind is really distinct from the body, is that
the knowledge of the mind is not acquired by way of abstraction and that
the mind is consequently neither an accidental nor a permanent property
of the body to which it is united, but that it can on the contrary be known
as a complete or self-subsisting thing in itself.

6. DESCARTES'S PROOF THAT THE MIND IS REALLY DISTINCT FROM THE

BODY

Let us now consider Descartes's proof of a real distinction between the

mind and the body. The main steps of this argument, which is presented
in the Sixth Meditation, are the following:
(i)
(ii)

(iii)

(iv)
(v)

Whatever I understand clearly and distinctly can be created by

God as I understand it;
It suffices that I understand one thing apart from another in
order to be certain that the one is different from the other
(follows from (i;
I know that I exist while in the meantime I notice nothing else
to pertain to my essence excepting that I am thinking (i.e.,
while still doubting of all other things, the existence of my own
body included, as is shown by the Cogito-reasoning or what
corresponds to it in the Second Meditation, AT VII, 24 - 25;
HR I, 149 - 150);
My essence consists in this alone that I am thinking (supposed
to follow from the Cogito);
I have a clear and distinct idea of myself inasmuch as I am only
a thinking, not extended thing. . . .(c/aram et distinctam

habeo ideam meum ipsius, quatenus sum tantum res cogitans,

non extensa, ... );
Therefore,
(vi)

I am really distinct from my body, and can exist without it.

(AT VII, 78; HR I, 190)

ON DESCARTES'S ARGUMENT FOR DUALISM

233

The structure of the argument, I take it, is roughly the following: (ii) can
be said to follow from (i), which, as we have seen, was a commonly accepted principle. That (i) is essential to Descartes's proof is easy to
understand against the background of the considerations above (cf. Section 4). This explains why Descartes insists that the distinctness of mind
and body cannot be fully demonstrated until the existence of a veracious
and omnipotent deity has been proved, although the argument for the
other relevant premises iii) and (iv seems to be given already in the Second Meditation (cf. AT VII, 219; HR II, 96). On the basis of the reasoning there presented, (iii) at least can be accepted as - relatively - unproblematic. But (iv) is more controversial. It is, presumably, supposed
to follow from (iii), since there seems to be no other place in the Meditations where the question of what really belongs to his essence (i.e., the
essence of his mind) is explicitly discussed besides the passages subsequent to the Cogito reasoning in the Second Meditation. 24 However, the
argument (the Cog ito reasoning) for (iii) can hardly be considered as sufficient for establishing (iv) as it is here formulated. And it is certainly not
sufficient for inferring (v). The problem, then, which has often been
raised, is to understand what further argument Descartes can give to
justify this claim that the essence of mind consists in thought alone, in
a way, as Descartes also seems to assume, which excludes the body and
other things from its essence. The proof of the mind-body distinction
presupposes, as Descartes himself recognizes, a move from the order of
clear and distinct perception to the order of things, or, in other words,
from concepts to reality. But how is it possible to infer, as he seems to
do, from what merely looks like a subjective state of certainty (expressed
in (iii, to the knowledge of the essential nature of the self or the mind
((iv) and (v? For, as I have argued elsewhere, the clarity and distinctness
of the knowledge of the self or the mind and its nature acquired in the
Second Meditation, seem to consist of nothing more than the certainty
of the facts expressed in the propositions" I think" and" I exist" . 25 And
how can this knowledge justify any further conclusion about the objective nature of self or the mind? (Cf. the objection of Arnauld in AT VII,
203; HR II, 84.)
However, instead of trying to look for a justification of these controversial premises, one could raise the question to what extent Descartes
really needs such strong claims for the conclusion he wants. to draw. (iv)
and (v), it seems to me, could be reformulated in a way that renders them
more plausible without altering the force of the proof:

234

(iv ')
(v ')

LILLI ALANEN

Thinking or thought pertains to my essence;

I have a clear and distinct idea of myself only inasmuch as I
am a thinking and (possibly) unextended thing ....

For though it is problematic where and when Descartes can be said to

have proved (iv) and (v), the weaker claims (iv ') and (v ') can at least be
defended on the basis of the reasoning presented in the Second Meditation. Descartes himself actually insists that he did not intend to prove
anything else in that place. (It is, by the way, questionable whether (iv)
and (v) or the corresponding theses (iv / ) and (v') are independent
premises or not: for my part I tend to consider them as merely different
formulations of the same claim.) Thus, in commenting on the following
passage: "I do not now admit anything which is not necessarily true: I
am thus, in the strict sense, only a thing which thinks, that is to say a
mind or a soul. .. (Nihil nunc admitto nisi quod necessario sit verum;
sum igitur praecise tantum res cogitans, id est, mens, sive animus.
[myef"lphasis] AT VII, 27; HR II, 152), Descartes points out that:
... by using the words in the strict sense, only I do not mean an entire exclusion or negation, but only an abstraction from material things, I said that, in spite of that, I was not
sure that there was nothing corporeal in the soul, although nothing of such nature was
known to exist in it; ... (AT IX, 205; HR II, 133)26

Now this, I want to argue, is all Descartes actually needs, besides the
assumption of an omnipotent Deity, in order to conclude that the mind
is really distinct from the body and can exist without it (vi). In this conclusion the same point seems to be repeated twice, since to say that one
thing is really distinct from another, is, as we have seen, merely to say
that it can be posited, i.e., exist, by the power of God, without the other.
But by the same means Descartes can also be said to have proved (iv) and
(v). Because, as the term is used by Descartes, essence is the attribute
through which a real thing or being, i.e., a substance, is known. 27
Thus, to Arnauld, who wonders where Descartes has shown "how it
follows, from the fact that one is unaware that anything else (except the
fact of being a thinking thing) belongs to one's essence, that nothing else
really belongs to one's essence" (AT VII, 199; HR II, 81), Descartes affirms that this is demonstrated in the place where he proves that God exists, "that God, to wit, who can accomplish whatever I clearly and
distinctly know to be possible" (qui potest omnia quae ego clare et
distincte ut possibilia cognosco, AT VII, 219; HR II, 96). And, he
continues:
For although much exists in me of which I am not yet conscious (for example in that
passage I did, as a fact, assume that I was not yet aware that my mind had the power of

ON DESCARTES'S ARGUMENT FOR DUALISM

235

moving the body, and that it was substantially united with it), yet since that which I do
perceive (adverto) is sufficient for me to subsist with alone, I am certain that God could
have created me without those other things of which I am unaware, and that these other
things do not pertain to the essence of mind ( .. .Quia tamen id quod adverto. mihi sUfficit

ut cum hoc solo subsistam. certus sum me a Deo potuisse creari absque a/iis quae non
adverto. atque ideo ista alia ad mentis essentia non pertinere. AT VII, 219; HR II, 97)

The essence, Descartes explains in the following paragraph, comprises

nothing but that without which a thing cannot exist. To think is essential
to the mind in this sense, to be united to the body is not essential to the
mind (AT VII, 219; HR II, 97). As is made very clear in the passage
quoted, to say that the essence of mind consists in thinking (alone) is yet
to say nothing of the actual state or condition of the mind or its relation
to other things, such as the body for instance. In drawing this conclusion
concerning the nature of his mind Descartes does not know, clearly and
distinctly, whether there are any bodies at all. Consequently he does not
know whether he is (united to) or has a body. But he knows, clearly and
distinctly, that he exists in so far as he thinks, and he can therefore conclude that his essence consists in thinking alone. And this, I take it, is also
the point of the whole argument: namely, that he can know his essence
(i.e., the essence of his mind) without knowing whether or not he
also is or has a body. For whatever the way in which body or matter is
defined, the knowledge of the essence of mind and the certainty of its existence, as acquired in the Second Meditation, is "complete" in the sense
that it does not include or presuppose the knowledge of any other
thing. 28 What Descartes has shown in proving by the Cogito reasoning
that he exists as a thinking thing, while stiII doubting of the existence of
his (or any) body is that the knowledge of his self or mind is not dependent on the knowledge of any other thing, his body included, and thereby
he has also proved that his mind is logically independent of the body.
The scholastic definition of the mind as the form of the body is thus
indirectly refuted. For if the mind is defined as the form or the first actuality of the living body, it can neither exist nor be conceived without
the body that it informs: the concept of the mind entails or presupposes
the concept of the (living) body. Hence the mind cannot be considered,
in itself, as a complete or self-subsisting thing, i.e., a substance in the
proper sense of the word. But in defining the mind as a merely thinking
thing, and showing that the existence of the mind, as a thinking thing,
can be known before the existence of the body is known, and that the
mind can thus be "clearly and distinctly" conceived without the body,
Descartes can be said to have shown that the concept of mind does not
entail the concept of the body. Relying on the assumption that whatever

236

LILLI ALANEN

he conceives clearly and distinctly is true, Descartes can thus conclude

that the mind is a substance, i.e. a complete thing, which is logically independent from the body and which, by the power of God, can be
posited or created without the body. 29
7. HAS DESCARTES ANSWERED HIS OPPONENTS?

Descartes's proof of the real distinction between mind and body, as mentioned before, leaves his opponents unsatisfied. Let us now consider
some of the objections raised against it. The first to oppose Descartes is
Caterus who invokes Duns Scotus's formal distinction, understood by
Caterus as a distinction intermediate between a real distinction and a
distinction of reason. On Caterus's reading of Scotus two things which
can be conceived as distinct and separated from each other are not (by
that fact alone) really, but merely formally distinct, as is the case with
the Divine attributes justice and pity. They have, Caterus observes, concepts prior to any operation of the understanding, "yet it does not follow
that, because God's justice can be conceived apart from his pity (mercy),
they can also exist apart" (AT VII, 100; HR II, 8).
Descartes answers by identifying, mistakenly, as he later concedes,
Scotus's formal distinction with his own modal distinction, that "applies
only to incomplete entities, which", he says, "I have accurately demarcated from complete beings" (AT VII, 120; HR II, 22).30 He continues:
Thus, for example, between the motion and the figure of the same body the distinction is
formal, and I can quite well understand (inlelligere) the motion without the figure, and the
figure without the motion, and either when abstracting from the body: but I cannot
however completely understand the movement without the thing in which the motion is,
nor the figure without the thing in which the figure is, nor finally can I feign that the motion
can be in a thing lacking figure, nor the figure in a thing incapable of motion. Nor can I,
similarly, understand justice apart from a just being, or compassion from a compassionate,
nor may I imagine that the same being which is just cannot be compassionate. But yet I
understand, completely, what a body is, in thinking merely that it is extended, figured,
movable, etc., and by denying of it everyt!Jing which belongs to the nature of mind, and
conversely I understand that mind is a compl-:te thing which doubts. understands, wills,
etc., although I deny that there is anything in it of what is contained in the idea of body.
Which would not be possible if there were not a real distinction between mind and body.
(AT VII, 120- 121; HR II, 22-23)

Things which can be conceived apart from each other merely by abstraction of the intellect, are always conceived inadequately. Since they are inadequately conceived, they are not known as complete, self-subsisting
things or beings. As Arnauld understands him, Descartes hereby claims
to have proved not only that the mind can be conceived completely

ON DESCARTES'S ARGUMENT FOR DUALISM

237

without the body but also that it can be conceived adequately apart from
the body (AT VII, 200 - 201; HR II, 82). But, as we saw before,
Descartes has not proved that the knowledge of his mind is adequate in
the sense assumed by Arnauld: it is not a knowledge embracing all the
properties of the thing known (cf. above, Section 5). Such knowledge,
Descartes stresses, is unattainable for the human mind, which is created
and finite, and it is therefore not required. An adequate knowledge
presupposes that one knows not only all the properties which are adequate for a thing, but also that one knows that God has not given the
thing in question other properties than those of which one has
knowledge. In other words, it is necessary to know that the knowledge
one has is adequate, which would require an infinite capacity of
knowledge (AT VII, 220; HR II, 97). Such knowledge, i.e., a knowledge
which is entirely adequate, is to be distinguished from a knowledge which
"has sufficient adequacy to let us see that we have not rendered it inadequate by an intellectual abstraction" (ibid.). Similarly, in order to
understand that the mind is a complete thing, we need not have an (entirely) adequate knowledge of it, but we must be able to see that the
knowledge we have of it is not rendered inadequate by abstraction, i.e.,
that it is not incomplete.
How then, according to Descartes, are incomplete entities to be
distinguished from complete entities? A complete entity, in Descartes's
view, is recognized by the fact that it can be conceived as existing in itself,
i.e., it must be understood as a real thing or entity in itself, independently of any other entities. By a complete thing, Descartes explains to Arnauld, "I mean merely a substance endowed with those forms or attributes which suffice to let me recognize that it is a substance" (AT VII,
222; HR II, 98). Now certain substances, as Descartes recognizes, are
popularly called "incomplete substances" (e.g., the mind and the body,
or parts of the living body). But, he adds, if they are called incomplete
because they cannot exist of themselves (i.e., without inhering in some
subject); it is contradictory to call them substances. However, substances
can be called incomplete in another sense, namely, when referring to
some other substance together with which they form a single selfsubsisting thing:
Thus. the hand is an incomplete substance, which taken in relation with the body. of which
it is a part; but. regarded alone. it is a complete substance. Quite in the same way mind
and body are incomplete substances viewed in relation to the man who is the unity which
together they form; but. taken alone. they are complete. (AT VII, 222; HR 11.99)

As we saw above substances, according to Descartes, are not immediate-

238

LILLI ALANEN

Iy known or perceived: the knowledge of a substance is always inferred.

What we perceived are certain forms or attributes, and we know, by a
principle of Natural light, that these forms or attributes must inhere in
something in order to have existence. 31 Now if the form or the attribute
perceived is sufficient to infer the existence of a thing, without knowing
any of its other attributes or properties and without knowing any other
things to which it might be united, the thing in which this attribute inheres can be considered as a complete thing, i.e., a substance:
For, as to be extended, divisible, possessed of figure, etc. are the forms or attributes by
which I recognize that substance called body; so, to be a knowing, willing, doubting being,
etc. are the forms by which I recognize the substance called mind; and I know that thinking
substance is a complete thing, no less than that which is extended. (AT VII, 223; HR II, 99)

I know that the mind or the thinking thing is a substance because I know
with certainty that I exist (i.e., that my mind exists) by the sole fact that
I think. This is sufficient to refute the contention of the Scholastics that
my mind is related to my body as species to genus. 32 " .it can nowise
be maintained", Descartes emphasizes, "that ... body is related to mind
as genus is to.species; for, although the genus can be apprehended apart
from this or that specific difference, the species can by no means be
thought apart from the genus" (AT VII, 223; HR II, 99). Since the
knowledge that the mind, conceived only as a thinking thing, exists, does
not imply the knowledge that the body exists, the mind cannot be an attribute or property of the body but is a complete thing, i.e., a thing
which, by the power of God, can exist independently of other things.
The same argument is also supposed to prove that the body, the existence of which is not fully established until the Sixth Meditation (and
to which the mind, according to Descartes, is in fact "very closely", intimately or "substantially' united, cf. AT VII, 222 and 228; HR I,
190 and II, 102), is a real, complete thing or substance in the proper sense
of the word. The possibility that the body is merely an attribute or a
species of the mind is hereby excluded. 33
The main problem with this argument, if it is accepted, and Arnauld
seems actually to have accepted it since he never returned to the charge,
is (as Arnauld also points out in his objections) that it proves too much.
It seems therefore to give support to the opinion of Platonists according
to which "nothing corporeal belongs to our essence, so that man is hence
only a soul (animus), while his body is merely the vehicle of the soul"
and according to which man is defined as "a soul that makes use of his
body" (AT VII, 203; HR II, 84). This is, however, a doctrine that
Descartes categorically rejects. He answers, characteristically, that "in

ON DESCARTES'S ARGUMENT FOR DUALISM

239

order to prove that one thing is really distinct from another, nothing less
can be said, than that the divine power is able to separate the one from
the other" (my emphasis). He also claims that he actually proved the
substantial union between mind and body in the Sixth Meditation while
dealing with the distinction between mind and body, by employing
arguments the efficacy of which he says he cannot remember to have ever
seen surpassed (!). He adds:
Likewise, just as one who said that a man's arm was a substance really distinct from the
rest of his body, would not therefore deny that it belonged to the nature of the complete
man, and as in saying that the arm belongs to the nature of the complete man no suspicion
is raised that it cannot subsist by itself, so I think that I have neither proved too much in
showing that mind can exist apart from body, nor yet too little in saying that it is substantially united to the body, because that substantial union does not prevent the formation of
a clear and distinct concept of the mind alone as a complete thing. (AT VII, 228; HR II,
102 - (03)3.

The nature of man is hence not, according to Descartes, purely spiritual,

i.e., man is not a mind but a mind united to a body. When pressed on
the question of how this union between two really diverse and distinct
substances should be understood, Descartes invariably answers by referring to a third primitive notion, known through experience. 35 This notion
is precisely the Scholastic notion of the union of form and matter, or of
a substantial form (quality) and the particular body that it informs.
But this traditionally obscure notion of a substantial, or as it also was
called, "real" union between the soul and the body (which was difficult
to explain already for the Aristotelians and the Thomists) seems to be inconceivable for a Cartesian. 36 It presupposes that the soul or the mind
and the body, which can be clearly and distinctly conceived only when
considered separately, are at the same time conceived as one single and
as two distinct things. This, as Descartes is forced to admit, is contradictory (AT III, 693; Philosophical Letters, 142). Descartes's way out of this
dilemma is to distinguish between different kinds or "paradigms" of
knowledge. This distinction, when properly understood, seems to render
the whole question of how the union of the mind and the body can be
"clearly and distinctly" understood inappropriate. The knowledge we
have of the union and the interaction between the mind and the body,
even if this cannot be explained, is based on our daily experience, and
as such it is evident and indubitable. But it can neither be compared to
the knowledge we have of the mind regarded as a purely thinking thing,
nor to that we have of body when regarded as a purely extended thing:
it is a fact "shown to us not by any reasoning or cC'mparison with other
matters, but by the surest and plainest everyday experience". As such it

240

LILLI ALANEN

is "one of those things which we can only make obscure when we try to
explain them in terms of others" (AT V, 222; Philosophical Letters,
235).
The notion of the union of the mind and the body is therefore
characterized as "primitive": it is-given as such and cannot be rendered
more intelligible or clear by means of simpler or more primary notions. 37
Although Descartes accepts the notion of a "real" union of an immaterial form nd a corporeal body in his account of human nature, he
rejects the use of this same notion in the explanation of physical
phenomena. According to Descartes the familiar and daily experience
that we have as sentient and acting conscious subjects of being "closely
united to" and "intermingled" with our body is the only context where
the use of this notion can be considered as legitimate (AT VII, 81; HR
I, 192). For it is the only context where this notion is based on a clear
and immediate experience. Applied to external, physical bodies it is confused and unintelligible: we have no concrete or clear experience of any
immaterial forms attached to or operating in physical bodies, such as the
"real qualities", "substantial forms" or forces (e.g. heaviness),
postulated by the Scholastics. The assumption of such forms in the scientific explanation of nature is not only an illegitimate and "occult"
hypothesis, it is also superfluous in Descartes's view. 38
I will not discuss the difficulties of Descartes's dualistic doctrine here.
As the comparison of Descartes's argument for mind-body dualism and
the theory of distinctions on which it is based with the ontological distinctions discussed by the Scholastics shows, many of these difficulties are
connected to ambiguities inherent in the Scholastic terminology of
distinctions and also to Descartes's application of this terminology in a
conceptual and scientific framework which in many important respects
differs from that in which it was originally developed. Descartes's main
innovation can perhaps be said to consist in the definition be gives of the
traditional notions of the soul and the body in terms of the concepts of
thought and extension. 39 Relying on the view according to which a real
distinction requires a mutual separability of the distinguenda, and accepting, for instance, like Ockham and Suarez did, logical independence and
non-identity as a sufficient sign of mutual separability, Descartes has no
difficulty in proving the separability and hence the real distinction of the
soul and the body in his sense of these terms. At the same time Descartes
reduces all other properties and things, considered as "real" entities or
things by many Scholastics, Ockham and Suarez included, to modes in
Suarez's restricted sense of the term. 40 The only objects of knowledge

ON DESCARTES'S ARGUMENT FOR DUALISM

241

which according to Descartes can be considered as "real" things or beings in the proper sense of the word, are thinking and extended
substances. All individual as well as generic properties of things, other
than the attributes of thought and extension, are thus reduced to modes
of these two attributes. Far from removing the original ambiguity related
to the various uses and meanings of the term being and related concepts,
Descartes's restructuring of the traditional conceptual framework leaves
us with a host of unsolved and perhaps unanswerable problems. Among
these, besides those encountered above, as the problem of interpreting
the two apparently contradictory notions of a real distinction and a real
union between the mind and the body, is the difficulty of accounting for
individual things or substances in the framework of Descartes's theory
of distinctions. No wonder that Descartes's opponents and followers are
perplexed, or that the doctrine called Cartesian dualism is difficult for
modern readers to understand or interpret in a satisfactory way.
NOTES
I am greatly indebted to Erik Stenius, Georg Henrik von Wright, Simo Knuuttila, Norman Malcolm, and Karl Ginet for valuable remarks and useful criticism at different stages
of my work. I wish especially to thank Erik Stenius for helping me to structure the present
interpretation of Descartes's argument for the mind-body distinction by giving me in a
private communication a very clarifying analysis of the argument.

Following a current usage a double reference will be given to the works of Descartes: one
to the Adam & Tannery edition (1897 -1913), abbreviated AT, volume and page, and one
to the English translation of Haldane and Ross (1911, 1978), abbreviated HR, volume and
page. For Descartes's correspondence references will be given to the English translation by
A. Kenny (1970), cited as Philosophical Letters. I have endeavoured to give literal translations of the Latin text and have therefore, occasionally, departed from the Haldane & Ross
translation. For controversial or unclear passages, quotes are also given in Latin.
I This has been amply shown by the work of scholars like Etienne Gilson, Alexandre
Koyre, J. R. Weinberg and others. See, e.g., Gilson, Index Scolastico-Cartisien, 1912
(1964), and Gilson (1925), and (1930); Koyre (1922); Wei~berg (1977); Beck (1965), and
Wells (1965).
2 Cf. Gilson 1912 (1964), p. 87; Weinberg (1977), pp. 75 -77, and below Notes 16 and 22.
J See, e.g., Wilson (1978); Weinberg, J. R. 'Descartes on the Distinction of Mind and
Body', in Weinberg (1977), pp. 71 - 81; Williams (1978), pp.102 -129; and the articles by
M. Hooker, A. Donagan, F. Sommers in Hooker (ed.), Descartes, Critical and Interpretive
Essays (1978). As appears from these studies there is no general agreement on the premises
of Descartes' argument. It is not very clear either how the conclusion of the argument

242

LILLI ALANEN

should be interpreted or understood. Cf. Wilson (1978), pp. 180ff; Wilson 'Descartes: The
Epistemological Argument for Mind-Body Distinctness', in Noils (1976), pp . 3 - 15, and
Malcolm (1971), pp. 5ff. For other recent studies of Descartes' argument for dualism see
the references given in Alanen (1982), p. 92, Note 16.
4 Cf. Wilson (1978), p. 243, Note 14. It is true that Descartes does not give any detailed
exposition of his theory of distinctions before the Principles of Philosophy (AT VIII,
28 - 30; HR II, 244 - 245), written some years later than the Meditations on First
Philosophy where the proof of the mind-body distinction is found (AT VII, 78; HR I, 190).
It is, however, clear from the Objections and Replies, published with the Meditations (.e.g,
AT VII , 120, 169-170, and 220ff; HR II, 22, 59, and 97ff), that Descartes already relied
upon a definite theory of distinctions in developing his controversial argument, and that
the notion of a real distinction, discussed by the Scholastics, hence plays a central part in
it. The term distinctio realis appears also, e.g., in the title of the Sixth Meditation (AT VII,
71; HR I, 185). I have dealt, briefly, with Descartes's use of these terms in a previous study,
'Descartes on the Essence of His Mind and the Real Distinction Between Mind and Body',
in Acta Philosophica Fennica 33 (1982), 66 - 73.
S Descartes's argument for the mind-body distinction was severely criticized by Descartes's
contemporaries, and notably by Arnauld whose objections Descartes, according to many
later critics, was unable to meet. See Kenny (1967), p. 94, and Wilson (1978), p. 198f.
6 Thus, if Descartes's proof of a real distinction between mind and body is less problematic
than is usually thought, when it is considered in the light of the traditional uses of the notion of a real distinction, the interpretation of this notion remains an open question (cf.
below Note 29). As to the Scholastic theory of a substantial union between the mind and
the body that Descartes defends as the only correct view of the human nature it is neither
clear nor distinct in the framework of Cartesian dualism, as Descartes himself was forced
to admit. Also, in trying to explain the union of mind and body to Princess Elizabeth whose
question on this matter, as Descartes recognizes, is "the one which may most properly be
put to me in view of my published writings" (AT III, 663, Philosophical Letters, p. 137),
Descartes invokes a third "primitive" notion, and asks Elizabeth to forget about tne
arguments proving the distinction between mind and body "in order to represent to herself
the notion of the union which everyone has in himself without philosophizing" . AT III,
692ff; Philosophical Letters, 142. Cf. the Principles of Philosophy, I. 48, and Section
7, below, esp. Notes 35 and 37.
1 Cf. Emile Boutroux' remark, often quoted, concerning Descartes' ability of "pouring
new wine into old bottles". See, e.g., Wells (1965), p. 22.
8 See AT VIII 2, 347; HR I, 434; AT VII, 549; HR II, 335 . Cf. also AT VII, 3; HR I,
133-134; AT VII, 13 -14; HR 1,140-141; AT VII, 153-154; HR 11,47.
9 See Edwards (1974), pp. 1-2; Aristotle, Met. 1l7, 1017 6b33 and Topics, I, 7.
10 Whatever is, Aristotle says, is one thing, Met., G. 2 1003 b22. See also, e.g., F. Suarez,
Disputationes Metaphysicae (hereafter quoted as DM) VII, Opera Omnia, Vo!. XXV, p.
250, and Weinberg (1964), p. 245.
II C. Vollert (trans!.), Francis Suarez: On the Various Kinds of Distinctions (1947), Introduction, p. 12.
12 The comparison of Descartes's theory of distinctions to those of his predecessors is
therefore instructive in many ways. The account presented in this paper is preliminary and
tentative: I hope to examine the problems here discussed more thorougl1'ly in a larger study
on the same subject.
il See, e.g., Met. Il 1017blO - 25, and Categ., Ch. 5, 2a1lff. Cf. Lloyd (1968), pp.
114-115.

ON DESCARTES'S ARGUMENT FOR DUALISM

243

Edwards (1974), p. 10. Cf. also A. Wolters in Ryan and Bonansea (1%5), p. 45.
Cf. Wolter, op.cil., p. 45. See also Note 16 below.
:6 Suarez, for instance, divides the mental distinction into two kinds: (a) a distinction of
reasoning reason (distinctio rationis ratiocinantis) which "arises exclusively from the
reflection and activity of the intellect" and which is in this sense purely mental; and (b) a
distinction of reasoned reason (distinctio rationis ratiocinatae) which is defined as a mental
distinction preexisting in reality, and which, he says, requires "the intellect only to
recognize it, but not to constitute it". (F. Suarez, DM VII, Section I, 4; transl. by Vollert
(1947), p. 18.) Descartes makes a similar distinction but rejects the former kind: he does
not, he writes, admit "any distinction of reason rationis ratiocinantis - that is one which
has no foundation in things - because we cannot have any thought without a foundation ... " (AT IV, 349-350; Philosophical Letters, 188). The formal distinction assumed
by Scotus is reduced by Descartes to a distinction of reason of the latter kind, i.e., to a
distinction of "reasoned reason", the only kind of mental distinction that Descartes admits
and which he characterizes, like Suarez, as having some kind of foundation in reality. (See
the letter referred to above and Descartes's Principles, I, 62, AT VIII, 30; HR I, 245. Cf.
also Alanen (1982), pp. 68 -70 and below, Section 7.) According to Suarez's characterization of the objects of a distinction of "reasoned reason~', they are not to be considered as
merely mental or conceptual entities because they are not produced or created by the mind,
but "real entities, or rather, a single real entity conceived according to various aspects. ...
Hence it is not the objects distinguished but only the distinction itself that results from the
reasoning" (DM VII, Section 1, 6; Vollert (1947), p. 19, my emphasis).
17 See Wolter, op.cit., p. 46. Cf. also Wolter, A., in JP S9 (1%2), p. 726.
18 Cf. below, Section 6 and Note 19. Although Descartes seems to question this principle
in some contexts, and suggests that God's omnipotence cannot be subjected to any such
limitations (cf., e.g., AT IV, lI8; Philosophical Letters, 151, and AT V, 223f;
Philosophical Letters, 236) the consequences of such a radical view, destructive for all
knowledge, are not taken seriously by Descartes. Cf. Koyre, op.cit., pp. 20ff.
19 See, e.g., the Second Replies, Proposition III, Corollary and Demonstration:
" .. .Deus. . . potest efficere id omne quod clare percipimus, prout idipsum percipimus. . . Est autem in nobis idea tantae alicujus potentia, ut ab illo solo, in quo ipsa
est, coelum & terra & c. creata sint, ab eodem fieri possint" (AT VII, 169; HR II, 58- 59,
my emphasis). Note that nothing corresponds in this passage to the words doivent avoir
ete creees in the French translation (AT IX, 131). Haldane and Ross who follow the French
Translation render the latin fieri possint by: ... whatever is apprehended by me as possible
must be created by Him too" (HR II, 59). But this is not the view professed by Descartes.
What the text says is merely that whatever exists has been created by the power of God,
and whatever we conceive as possible can be created by him. Cf. the note of Alquie, F.
(ed.), Oeuvres phllosophiques de Descartes, Vol. II, Note 1 to p. 597. See also the letter
to Regius, June 1642, AT III, 567, and to Mersenne, March 1642, AT III, 544 - 545;
Philosophical Letters, 132.
20 Cf. Suarez's discussion ofthe criteria of a real distinction, DM VII, Section 2, 2-12;
Vollert (1947), pp. 40-49, and Weinberg (1977), pp. 75ff. For Ockham's application of
the principle of God's omnipotence, see e.g., Weinberg (1964), pp. 245ff.
The radical position of Ockham requires a special mention in this context. True to his
view that nothing other than individuals (ir.dividual things) exist in the extra-mental reality
(outside human consciousness) Ockham reduced all distinctions (at least in so far as created
things are concerned) having any basis in reality to real distinctions; and he consequently
held that anything that can be conceived as distinct is a real thing or unity which
14

IS

244

LILLI ALANEN

is really distinct from all other things to which it might be united. Ockham hence allows
for no other kind of real distinctions than numerical distinctions. Therefore, the only kind
of distinctions existing in nature, i.e., having any foundation in extra-mental reality, according to Ockham, are real, numerical distinctions. (Whether or not Ockham accepted any
forlT)al distinction is a subject of controversy which we need not go into here. If he admits
such a distinction he restricts it to supranatural things, e.g., the Trinity.) Whatever can be
conceived as .distinct is thereby also a singular thing, distinct from all other things, by the
principle of God's omnipotence. Any quality which can be distinctly conceived is a singular
thing and hence also really distinct from the substance in which it inheres. See Edwards
(1974), pp. 12-13 and pp. 180ff, and Weinberg (1964), pp. 248-249. See also Weinberg
(1965), p. 49.
It is interesting to note also how far Ockham carries this doctrine, which is based on purely logical arguments. Since matter and form, for instance, are distinctly conceivable parts
of a composite substance, according to the common view, this means, in Ockham's view,
that the particular matter of a given composite must be a singular thing which, logically,
can exist also by itself without any form. Matter, as such, is thus not mere potentiality, as
the Scholastics generally held, but has some actuality, being, nature, or whatever it is
called, in itself. The same holds, according to Ockham, for the substantial forms and the
qualities. See Weinberg (1965), pp. 51 - 52.
21 See Note 2 above and Alanen (1982) pp. 69ff. For Suarez's discussion of the distinctions
to be retained, see DM VII, Section I; Voller! (1947) pp. 16-39.
22 Cf. Suarez' discussion of the signs for discerning various grades of distinction, DM VII,
Section 2, and notably his discussion of the signs of a real distinction, ibid., 9 - 28;
Vollert (1947), pp. 46-6i. Cf. also Weinberg (1977), p. 75 and Alanen (1982), p. 78f.
2J Cf. Suarez, D., DM VII, Section I, 16 - 21, and Section 2, 6ff. - It is interesting
to note how Descartes, while retaining Suarez's criteria for the modal distinction, gives the
term mode (modus) a much wider application than Suarez. Suarez, in discussing different
uses of the term mode, restricts it to a particular aspect of a given attribute (e.g., quality
or quantity), namely the mode of inherence of the attribute in question. Thus, the inherence
of quantity, for instance, is called its mode by Suarez "because it is something affecting
quantity, and, as it were, ultimately determining its state and manner of existing, without
adding to it any new entity, but modifying a preexisting entity" (DM VII, Sc.;tion I, 17;
Vollert (1947), p. 28). The mode of being of the inherence is such "that it cannot exist
unless it is actually joined to the form of which it is the inherence" (ibid., 18; p. 29).
Suarez hence distinguished two aspects in quantity: the first, he says, is called the thing or
being of quantity "comprising whatever pertains to the essence of the individual quantity
as it is found in nature, and remains and is preserved even if quantity is separated from
its subject" (my emphasis); the second, "the inherence of quantity is called its mode
because it is something affecting quantity, and, as it were, ultimately determining its state
or manner of existing, without adding to it a new proper entity ... " (ibid., 17; p. 28).
A mode, hence, in Suarez's restricted sense of the word, is not a thing or entity in itself:
it has no being of its own. "Its imperfection is clearly brought out by the fact that it must
invariably be affixed to something else to which it is per se and directl)! joined without the
medium of another mode, as, for instance, sitting to the sitter, union to the things
united ... " (ibid., 19; p. 31). Having no being or essence of its own the mode "so
necessarily includes conjunction with the thing of which it is a mode that it is unable by
any power whatsoever to exist apart from that thing" (ibid., 20; p. 32, my emphasis).
Descartes, as we saw above, uses the term mode in a much wider and general sense: he
treats all the qualities and accidental attributes or properties of things as modes in Suarez's

ON DESCARTES'S ARGUMENT FOR DUALISM

245

restricted sense. Unlike Suarez Descartes makes no distinction between an individual quality, say this whiteness, and its actual mode of inherence, the whiteness of this paper. Thus,
what for Suarez holds for the mode of inherence of a particular quality or quantity, holds
according to Descartes for any particular quality or quantity: it has no being of its own and
can hence not be clearly conceived without the essential attribute of the thing, i.e., the
substance, in which it inheres. Since Descartes a<;lmits only two such essential attributes,
namely thought and extension, this means that all other properties of things are reduced
to mere modes of either thinking or extended substances.
24 AT VII, 24-25; HR 1,149-150. Cf. A1anen (1982), pp. 29ff. - In proving that he
exists as a thinking thing, Descartes takes it that he has proved that his mind or soul exists
(cf. ATVI, 32-33; HR 1,101; AT VII, 27; HR 1,152). In discussing his nature or essence,
in the Second Meditation, Descartes makes therefore no distinction between his essence
and the essence of (his) mind. This, however, does not mean that Descartes identifies his
essence (as a human being) with the essence of mind, which, strictly taken is a pure reason
or intellect. See AT VII, 78-79; HR I, 190-191; AT III, 479; Philosophical Letters,
125 - 126; AT III, 371; Philosophical Letters, 102, and below, Note 25 ...,
2S Alanen (1982), pp. 22 - 43, and pp. 63 - 64. Notice that the term to think (cogitare) is,
deliberately, used in a very wide sense by Descartes, and covers all kinds of acts or states
or consciousness, from the acts of the pure intellect to dreams and sensations. See, e.g.,
AT VII, 28; HR I, pp. 152ff; AT VII, 217; HR II, 52. Cf. Alanen (1982), pp. 115-116.
26 Cf. AT VII, 175; HR II, 63, and AT VII , 355; HR II, 209.
27 AT VIII, 24f, 30; HR I, 240, 245. Cf. AT VII, 219; HR II, 99. Descartes adheres to the
view that substances, as such, are not immediately known in themselves. See AT VII, 161;
HR II, 53; AT VIII, 8; HR I, 223; AT VIII, 25; HR I, 240.
28 As to the clear and distinct concept of matter, acquired partly through the analysis of
the wax in the Second Meditation, and partly through the considerations in the Fifth and
Sixth Meditations, it is not really indispensable for this conclusion, that the essence of mind
is thinking. Descartes's definition of matter as mere geometrical extension, and the conception of the body as a piece of a mechanically moved extended substance that this implies
(a conception that is, by the way, difficult to prove), can certainly - if it is granted - be
said to give it additional support. But it is not essential to the argument as it is here
understood. This is the reason for which I have omitted the second part of premise (v)
which states (v)/ I have a clear and distinct idea of the body inasmuch as it is only an extended and unthinking thing (AT VII, 78; HR I, 190). For a discussion of the part played by
this premise and Descartes's concept of body in the argument for a real distinction between
mind and body see, e.g., Williams (1978), pp. 213ff, Gueroult (1953) I, pp. 121ff, and II,
pp. 67ff, and Alanen (1982) pp. 45-65.
29 On this reading Descartes's argument is, it seems to me, both more simple and forceful
than I was earlier inclined to think. This, however, is not to say that the argument as here
interpreted is unproblematic, but I will leave the consideration of the probiems it raises
aside for the moment (cf. Alanen (1982), pp. 84ff). The conclusion of the argument,
notably, is perplexing. For what it states, as I want to stress, is the absence of a logical entailment between the concepts of mind and body, and the existence of the mind in separation from the body is therefore a mere logical possibility. What the implications of this
scholastic notion of a real distinction are on the ontological level is and remains an open
question. - For an interesting critical discussion of the import and consequences of
Descartes' distinction between the concepts of mind and body see Malcolm, 'Descartes'
Proof that He Is Essentially a Non-Material Thing, Thought and Knowledge (1977),
58 - 84.

246

LILLI ALANEN

In the Principles, Descartes corrects himself and assimilates Duns Scotus's formal
distinction with his own distinction of reason (cf. Note 16 above). This mistake or uncertainty concerning the classification of the formal distinction from Descartes' side is not important here. But it is quite interesting. For both the distinction of reason (i.e., reasoned
reason) and the modal distinction, as understood by Descartes, require in fact an abstraction of the mind and are therefore opposed to the real distinction, which, as will be seen,
is restricted by Descartes to entities or things which can be conceived in themselves as complete, i.e., self-subsisting things. Cf. Suarez's characterization of the distinction of reasoned reason: it "does not exist strictly by itself, but only dependently on the mind that conceives things in an imperfect, abstractive, and confused manner, or inadequately". DM
VII, Section I, 8; Vollert (1947), 20-21.
31 Cf. above, Notes 23 and 27.
32 Cf. Arnauld's objection to the effect that Descartes's argument that the body can be
completely understood merely by thinking that it is extended, figured, movable, etc., is of
little value: it does not exclude that the body might be related to the mind as genus is to
species, for, as is commonly agreed, the genus can be conceived without the species (AT
VII, 201; HR II, 82). See also the letter to (Mesland?), 2 May 1644, AT IV, 120;
Philosophical Letters, 152.
33 Note that the fact that he has a body with which he is "very intimately conjoined" is
fully established only in the Sixth Meditation, after the proof of the mind-body distinction
is given. What this proof can therefore be said to show is that if the self or the mind is united
to (or has) a body, then it is really distinct, i.e. can be separated from the body, and conversely. Cf. AT VIII, 28; HR I, 243-244.
34 According to Suarez and, I presume, most of the Scholastics, the soul and the body as
well as form and matter in general, although they are considered as separable by the divine
power and hence as really distinct in the sense given above, are regarded as incomplete and
partial beings in themselves, whether they are united or in a separate state. What Descartes
here says about the mind and the body would according to Suarez apply only to integral
parts, e.g., homogeneous parts of a continuum, which unlike form and matter are not of
themselves "ordained to the composition of another thing", i.e., to be parts of the union
or compound to which they actually belong. Cf. Suarez, DM VII, Section I, 23; Vollert
(1947), pp. 33 - 34.
35 See, e.g., the letter to Arnauld, 29 July 1648, AT V, 222f; Philosophical Letters, pp.
235 - 236. Cf. also the letter to Elizabeth, 21 May 1643, AT III, 667; Philosophical Letters,
p. 139, and references given in Note 6 above.
36 Cf. Spinoza, Ethica III, Proposition II, scholium, Vloten et Land (ed.), (1914), p. 124.
37 See, e.g., AT III, 665; Philosophical Letters, 138f. Cf. Alanen (1982), p. 87f.
38 See, e.g., Descartes's letter to Mersenne, 26 April 1643, AT III, 648; Philosophical Letters, pp. 135 - 136.
39 For Descartes's definition of matter, see, e.g., Meditations, V and VI and Principles,
I, 60, AT VIII, 28; HR I, 243 - 244; ibid., II, 22 and 64, AT VIII, 52 and 78 -79; HR
1,265.
40 See Note 23 above.
30

BIBLIOGRAPHY
Descartes, R.: Oeuvres de Descartes (AT), publies par Ch. Adam et P. Tannery, Leopold
Cerf, Paris, 1897 - 1913, 12 vols.

ON DESCARTES'S ARGUMENT FOR DUALISM

247

Descartes, R.: The Philosophical Works of Descartes (HR), transl. by E. S. Haldane and
G . T . Ross, London, 1911 (1978), 2 vols.
Descartes, R.: Philosophical Letters, transl. and ed. by A. Kenny, Clarendon Press, Oxford, 1970.
Alanen, L.: 'Studies in Cartesian Epistemology and Philosophy of Mind', in Acta
Philosophica Fennica, Vol. 33, Helsinki, 1982.
Alquie, F. (ed.): Oeuvres philosophiques de Descartes, Garnier Freres, Paris, 1%3 -1973,
3 vols.
Aristotle: The Works of Aristotle, Vols. I and VIII, ed. by W. D. Ross, Clarendon Press,
Oxford, 1928.
Beck, L. J.: The Metaphysics of Descartes, Oxford University Press, Oxford, 1965.
Donagan, A.: 'Descartes' "Synthetic" Treatment of the Real Distinction Between Mind
and Body', in M. Hooker (ed.), Descartes, Critical and Interpretive Essays, The Johns
Hopkins University Press, Baltimore, 1978.
Edwards, S.: Medieval Theories of Distinction, University of Pennsylvania, Ph .D., 1974.
University Microfilms International, Ann Arbor, Michigan and London, 1981.
Gilson, E.: Index Scolastico-Cartl!sien, Paris, 1912; repro Burt Franklin, New York, 1964.
Gilson, E.: Rene Descartes: Discoursde la methode. Texte et commentaire, J . Vrin, Paris,
1925 (1967).
Gilson, E.: Etudes sur Ie role de la pensee medievale dans la formation du systeme cartesien, J. Vrin, Paris, 1930 (1975).
Hooker, M. (ed.): Descartes: Critical and Interpretive Essays, The Johns Hopkins University Press, Baltimore and London, 1978.
Kenny, A.: Descartes, A Study of His Philosophy, Random House, New York, 1968.
Koyre, A.: Essais sur I'idee de Dieu et les preuves de son existence .chez Descartes, Ernest
Leroux, Paris, 1922.
Lloyd, G . E. R.: Aristotle: The Growth and Structure ofHis Thought, Cambridge University Press, Cambridge, 1968.
Malcolm, N.: 'Descartes' Proof That His Essence Is Thinking', in The Philosophical
Review (PR) LXXIV (1965), 315 - 338.
Malcolm, N.: Problems of Mind, Descartes to Wittgenstein, Harper Torchbooks, New
York, 1971.
Malcolm, N.: Thought and Knowledge, Cornell University Press, London, 1971.
Sommers, F.: Dualism in Descartes: The Logical Ground, in M. Hooker (ed.), 1987, pp.
223 - 233.
Spinoza, B. de: Opera, I - II, in l. van Vloten and J. P. N. Land, The Hague, MCMXIV.
Suarez, F.: Opera Omnia, Vol. XXV, Paris, 1866; repr. Georg Olms, Hildesheim, 1965.
Suarez, F.: Francis Suarez: On The Various Kinds of Distinctions, transl. by C. S. J.
Vollert, Marquette University Press, Wisconsin, 1947.
Weinberg, J. R.: A Short History of Medieval Philosophy, Princeton, New Jersey, 1964.
Weinberg, J. R.: Abstraction, Relation, and Induction. Three Essays in the History of
Thought, Madison & Milwaukee, 1965.
Weinberg, J. R.: Ockham, Descartes, and Hume, The University of Wisconsin Press,
Madison, Wisconsin, 1977.
Wells, N. J.: 'Descartes and the Modal Distinction', The Modern Schoolman XLII (1965),
1-22.
Williams, B.: Descartes, The Project ofa Pure Enquiry, Penguin Books, Hammonsworth,
1978.
Wilson, M. D.: 'Descartes: The Epistemological Argument for Mind-Body Distinctness',
Noils 10 (1976),3 -15.

248

LILLI ALANEN

Wilson, M. D.: Descartes, Routledge & Kegan Paul, London, 1978.

Wilson, M. D.: 'Cartesian Dualism', in M. Hooker (ed.), 1978.
Wolter, A. B.: 'The Realism of Scotus', The Journal oj Philosophy OP) 52 (1962),
725-736.
Wolter, A. B.: 'The Formal Distinction', in J. K. Ryan and B. M. Bonansea (eds.), John
Duns Sco/us, 1265 - 1965, The Catholic University of America Press, Washington,
1965.

Dept. of Philosophy,
University of Helsinki,
Unioninkatu 40B,
SF-00J70 Helsinki 17, Finland.

JAAKKO HINTIKKA

KANT ON EXISTENCE, PREDICATION,

AND THE ONTOLOGICAL ARGUMENT*

The subject matter of my paper can be discussed only against the

background of two larger issues. They loom large both systematically
and historically. They are the ontological argument for God's existence l
and the Frege - Russell thesis that "is" is multiply ambiguous,2 indeed
ambiguous between the "is" of existence, "is" of identity, "is" of
predication, and "is" of generic implication. These allegedly different
meanings can be illustrated by "God is" or (3x)(God = x), "Jack is John
Jr." or Jack = John Jr., "Jack is blond" or Blond(Jack), and "man is
an animal", or (x)(Man(x):::J Animal(x. Any discussion of Kant's treatment of being is thus being tacitly presided over by Anselm and Frege.
As the case is frequently, the crucial aspect of one's discussion of the
history of an interesting configuration of ideas turns out to be a topical
analysis of the conceptual situation which these ideas exemplify. 3 Hence
I will have to spend some time and care examining the logic and semantics of the ontological argument, as well as the logic of "is" and "being". My paper thus needs something like the familiar nineteenthcentury German subtitle 'Eine historisch-kritische Studie'.
My conclusions concerning Kant's relation to these two overreaching
issues are rather negative. Kant's criticisms of the ontological argument
are misplaced, not to say mistaken. Nor is he an early herald of the
Frege - Russell thesis, as has been claimed. 4 The mainstay of Kant's
discussion, viz. his thesis that "existence is not a predicate", can be
shown to be false.
These are big claims, perhaps bigger than I can hope to prove in a
single paper. I shall nevertheless spell out my results somewhat more fully and indicate why I think they are valid.
To put my first point bluntly, most of the criticisms of the ontological
argument have been misplaced, Kant's included. The mistake in the main
version of the ontological argument, or what I find by far the best rational reconstruction of the argument, is an operator-switch fallacy.
Consider first the sentence.
(I)

N(3x)[(y)(y exists

:::J

x exists)]

249
S. Knuullila and 1. Hintikka (eds.), Tne Logic oj Beini!. . 249--267.
<0 1981 by Dia/ecrica.

250

JAAKKO HINTIKKA

where "N" is the necessity operator. There are two things to be observed
about this sentence. First, it is trivially valid (logically true), and, second,
that it looks very much like a summary of the ontological argument. Let
me spell out these two points.
I have noted before that (1) is trivially valid completely independently
of what the "predicate of existence" used in it is. 5 I have also claimed
that it is the logical truth of (1) that makes the ontological argument so
perennially seductive. In fact, (1) might seem to express precisely the
desired conclusion, viz. the necessary existence of an existentially perfect
being. Indeed, it can readily be seen that (1) is closely related to the ontological argument in its actual historical versions. What (1) says is that,
necessarily, there is an individual such that if anything exists, it does,
which nearly says - or seems to say - that there is something which is
greatest with respect to existence ("pre-eminent in its mode of existence", to use Kant's words in A 586 = B 614). Thus the inside conditional in (1), viz. (y) (y exists => x exists), can be considered as a
characterization of god ( = x), conceived of as the most powerful being
with respect to existence. Kant asserts what is very similar to this inside
conditional of (1) when he says (A 588 = B 616) that "from any given existence ... we can correctly infer the existence of an unconditionally
necessary being", that is, a God. Thus, the whole of (1) seems to express
quite well the Anselmian idea that the most perfect being - a being
greater than which cannot be conceived of - must necessarily exist, with
the perfection in question restricted to perfection or maximal greatness
with respect to existence.
From (1) the defenders of the ontological argument in effect
fallaciously infer
(2)

(:lx)N[(y)(y exists => x exists)).

This quantifier switch is the crucial mistake in the most interesting versions of the ontological argument. For, appearances notwithstanding, it
is (2) and not (1) that Anselm, Descartes & Co. really want to establish.
Thus the logical truth of (1) helps them only if they could take the further
step from (1) to (2). But this further step is illegitimate. I shall later return
to the question as to what further premises might serve to validate the
step.
It remains to spell out more fully what is involved here. First why is
it (2) and not (1) that the ontological argument is calculated to prove?
This is seen easily by means of the obvious possible-worlds semantics of
(1) - (2). What (1) says is that in each world there is something such that

KANT ON EXISTENCE AND PREDICATION

251

if anything at all exists in that world, it does. The reason why this is not
enough is due to the fact that in different worlds such individuals can be
entirely different from each other. Indeed, the triviality of (1) is reflected
by the fact that any existing individual can be chosen as the value of the
existentially bound variable "x" in (1).6 We would be able to infer (2)
from (1) only if we could assume that all these individuals are (or can be
chosen to be) identical with each other.
In contrast, (2) attributes the status of existentially greatest being to
some one individual in all (nonempty) worlds. 7 This is obviously what the
argument is supposed to establish. In view of the validity of (1), the right
way of criticizing the ontological argument is hence to spell out the difference between (1) and (2) and to show how and wh~ the step from (1)
to (L) is fallacious.
A prolegomenon to such a criticism is to point out how natural
language tends to hide the differences between (1) and (2). Indeed, such
English sentences as
(3)

There necessarily is some individual which is existentially the

greatest

are ambiguous between (1) and (2). Moreover, consider the use of any
expressions which rely on grammatical cross-reference, e.g., "which" in
(3) and "it" in the fuller form of (3), viz. in
(4)

There necessarily is some individual which is such that if

anything exists, it does.

The use of such expressions usually presupposes that their reference is

well-defined in all the possible worlds which we are tacitly considering.
Such well-definedness is, we saw, just what is needed to move from (1)
to (2). How deep the sources of fallacy run here is illustrated by the
etymology of the: English existential quantifier word "some" as having
the same root as "same one".
What is even more pertinent to note here is that similar locutions
abound in philosophical discussions, not the least in discussions about
the ontological argument. For instance, Kant speaks of, as we saw, a being (some one being) whose existence can be inferred from any given
existence.
Apart from explaining the temptation to infer (2) from (1) (or, better,
to assimilate the two to each other) which ordinary discourse generates,
the best way of defusing the fallacy seems to me to be to expose the
general type of mistake that is exemplified by the fallacious derivation

252

JAAKKO HINTIKKA

of (2) from (I). Probably the most effective way of doing so is to appeal

to the informal ideas which are systematized in my game-theoretical

semantics. 8 We don't need thesystematization here except as a backup
line of defense for the appeal to certain "iconic" ideas.
In (I), malicious nature (who is trying to defeat me) chooses a possible
world w for which I will have to show that in it there is something such
that it exists in w if anything at all exists in w. In contrast, in (2) I have
to be able to choose (for (2) to be true) an individual x such that, no matter what world w is subsequently chosen by nature, the very same individual x will exist in w if anything at all exists in w. As you can see,
the difference is subtle but unmistakable. This is one of the many cases
where informal seman tical ideas can perform a tremendous service in
philosophical analysis and philosophical argumentation. It is perhaps
not at all surprising that philosophers who were thinking much more in
verbal than in semantical terms should have been confused about the interrelations of (I) and (2). The early medievals, including Anselm, would
presumably be cases in point.
Moreover, the informal use of the idea of possible worlds which I
relied on was deeply foreign to most of the medieval tradition. Aristotle
squarely refused to consider any world history different from out actual
one. For him, my informal explanation would not have made any sense,
because it employs crucially the idea of a mUltiplicity of possible worlds.
Admittedly Aristotle operated freely with assorted possibilia, but they,
too, had to prove their mettle in the course of the one actual course of
events. 9 Even in the absence of a detailed historical investigation, it
seems clear that Anselm had not disentangled himself fully from Aristotle's actualism. Hence the mistake is in a literal sense of the expression
doubly natural for someone in his historical position to make. Kant, unfortunately, has much less of an excuse in this respect.
Although the real fallacy of the ontological argument lies in the step
from (I) to (2) and not in (1) itself, most of the actual criticisms of the
argument have to be construed as criticizing (1) and not the transition
from (1) to (2). Kant's thesis that "existence is not a predicate" is a case
in point. \0 It amounts to an attempt to deny the legitimacy of the way
(I) is formulated in the first place. For (I) will not get off the ground
without some way of expressing the existence of individuals, i.e., of using
"existence as a predicate". Hence the Kanti:m gambit is admittedly
prima Jacie quite tempting, even though it is misplaced.
There is a tempting way of trying to smuggle into the ontological argument what amounts to the crucial quantifier switch. It is to try to

KANT ON EXISTENCE AND PREDICATION

253

strengthen the "material implication" in (1) into something like a logical

or analytical "strict" implication. This would indeed turn (1) into (2). It
may be that Kant was in effect trying to guard himself against this
manoeuver when he removed (Aristotle notwithstanding) all existential
force from essential predications. This is not the crux of the matter,
however. If we try to insert a necessity-operator to from the inner conditional of (1), we lose the trivial logical truth of (1). Why? Not, as Kant
(and others) seem to have thought, because existential presuppositions
are not fulfilled, but because we need what I have called a uniqueness
premise to justify the existential generalization which the step from (1)
to (2) in effect is. For suppose "g" (short for "a god") is a singular term
which in each world picks out the kind of individual which is asserted to
exist there by (1). In other words, it is true that
(5)

N(3x)(g == x)

Now we know from modal logic that this term "g" can help to establish
(2) only if we have at our disposal the additional premise
(6)

(3x)N(g == x),

where no existential force is being assumed. What (6) expresses is precisely the identity (g == x) of g with some one individual (the x in "(3X)" in
all the possible worlds (introduced by "N"). As we saw, the failure of
this identity is precisely the fatal flaw in the usual versions of the ontological argument.
The auxiliary premise (6) is analogous to the proposition
(7)

(3X) it is known that (g == x)

which says that it is known who (or what) God is. In fact it is (7) and not
(6) that we need as an auxiliary premise if we want to establish that it is
known that God exists. It is no wonder, in view of these observations,
that the literature on the ontological argument is full of considerations
of whether God necessarily is who he (or she) is, and whether we can
"conceive of" or "understand" who God is. As we know, Anselm and
Gaunilo were already discussing the latter issue. As far as the former is
concerned, the missing premise (6) seemed to be conveniently supplied
by Exodus 3: 13 - 14, where God says: "I am who I am", presumably
meaning that He necessarily is who He is. II It would divert our purpose
to explore these historically important lines of thought here, however.
I think it is the time to lay to rest the myth that "existence is not a
predicate". It is embarrassingly clear what Kant's grounds for maintain-

254

JAAKKO HINTIKKA

ing this thesis were. They were largely due to the paucity of the logics and
languages he was contemplating. He envisaged only two types of judgment relevant here, viz. what I shall here call judgments of "essential"
predication and judgments of existence. (Where contingent predication
was supposed to find a niche was not explained by the good Immanuel. 12
In the former, exemplified by "God is omnipotent", a necessary connection is asserted to obtain between the subject and the predicate, without
prejudicing the existence of either. As Kant puts it, "the omnipotence
cannot be rejected if we posit a Deity, for the two concepts are identical". But this judgment carries no existential import. " ... if we say,
'There is no God', neither the omnipotence nor any other of its
predicates is given; they are one and all rejected together with the
subject" .
The other kind of judgment Kant mentions is the existential one, e.g.,
"God exists". In neither one is existence a predicate, Kant says in effect.
A judgment of essential predication has no existential force, whereas in
an existential one we take a subject as it were all ready-made with its
essential predicates and simply assert that this particular complex of
predicates is in fact instantiated in reality. Here existence is not one of
the configuration of predicates; it is what is asserted of the configuration.
Nothing is wrong here. A faithful Aristotelian would have worried
about the total absence of existential import in a judgment of essential
predication, for on certain conditions Aristotle seems to have maintained
such an import. He went so far as to worry lest this would lend other instances of copula a similar existential force, so that we could fallaciously
infer from "Homer is a poet" that "Homer is", i.e., exists. 13
However, our worries are not Aristotelian. Kant's mistake is not that
he says something false, but that his philosophical diet is one-sided: he
nourishes himself on too few kinds of examples. In reality, there is a
tremendous multitude of forms of proposition which go way beyond the
ones Kant envisages. Among them, I suggest, we can safely assume to be
included some in which' 'existence is a predicate" in whatever reasonable
sense we can give to this phrase.
The following argument may indicate why this assumption is eminently natural - and also why the use of existence as a proper predicate has
met with such resistance among philosophers. This line of thought
would of course have been rejected by Kant, but I think that it would
have been appreciated by Leibniz.
Obviously, we attribute to actual individuals all the time predicates

KANT ON EXISTENCE AND PREDICATION

255

which turn on what they would be like in other possible worlds, for instance, what they could be or could do. Sometimes these predicates turn
on the existence or nonexistence of these individuals in those other circumstances. For example, speaking of the necessary conditions of life in
the case of some particular organism involves this kind of predication.
All that is needed to be able to use actual existence as a predicate (so as
to refute Kant) is then apparently a parity of cases. If we can take an individual in the actual world and assign to it a predicate which involves
existence or nonexistence in some other world, surely we ought to by the
same token be able to take a "merely possible individual", i.e., a denizen
of some other world, and attribute to it predicates definable in terms of
its actual existence, maybe the "predicate of (actual) existence" itself.
Basically, it seems to me that this argument is unanswerable. There certainly are concepts applied to actual individuals which can only . be
defined in general in terms of (merely) potential existence, e.g., the
biological concept of fertility. Even though we don't do it often, we surely can pick out one "merely possible" individual from others by specifying that it enjoys the dubious distinction of actual existence. Examples
are not very easy to come by, but speaking of the actual Hamlet seems
to be good enough. For many of us, Hamlet is first introduced as a merely fictional "possible individual", and we learn only subsequently that
the melancholy Dane has a real-life counterpart. (You didn't know that
Hamlet really existed? Yes, he did enjoy the predicate of existence!)
There are several different kinds of difficulties here which have led
some philosophers to deny the possibility of the sort of return of an individual from other possible worlds to the actual one which I am envisaging. Some philosophers have failed to see how we can individuate a merely possible individual. Doesn't the very possibility of considering some
one definite individual (to which predicates are to be ascribed) presuppose its actual existence? The fact that philosopher-logicians as eminent
as Montague and Kripke have maintained this presupposition shows that
we are not dealing with a mere idle worry. I cannot here discuss this complex of problems in its entirety. A good descriptive account of how merely possible individuals can enter into our discourse is given by David
Kaplan in 'Quantifying In' .14 In general, I believe that the denial of
merely possible individuals is based on an unrealistically narrow view of
how out language actually functions. IS
There is another reason why rec'.!nt logicians may have been wary of
the line of thought I just adumbrated. In it, we took an individual which
had been considered qua citizen of another world and began to consider

256

JAAKKO HINTIKKA

it also as a member of the actual world. A moment's reflection shows

that in the conventional languages of modal logic there is no way of doing so. To consider an individual initially in an alternative possible world
means having its linguistic representative within the scope of a modal
operator. But then we cannot any longer consider it as a member of the
actual world, for that would presuppose that its representative occurs
outside the scope of all modal operators. But that would mean having
introduced it initially qua an inhabitant of the real world. In brief, in
conventional modal logic, you can have only one-way tickets from the
actual world to its alternatives; no round trips are possible. In less
metaphoric terms, anaphoric back reference to worlds considered earlier
is impossible in the usual notation of modal logic. This feature of the
modal languages most logicians are familiar with seems to have
discouraged them from thinking that individuals introduced as members
of the domains of other worlds can subsequently be considered also as
denizens of the actual world. 16
David Kaplan used to illustrate this point by means of the following
formula:
(8)

M[(x)(A(x)]

B(x

where "M" is the possibility operator. Of course (8) is ill-formed in traditionat modal logic. Yet it expresses a perfectly good semantical sense. It
says that there is some alternative possible world such that everything
that is there A is in fact (i.e., in the actual world) B. The remarkable thing
is that this sense cannot be expressed by any well-formed formula of conventional modal logic. 17
This is nevertheless merely a limitation of one particular kind of notational systems. One can, as Esa Saarinen has done, introduce special
"backwards looking" operators which effect just the kind of return
journey I was envisaging. 18 What is more interesting, Saarinen has
shown convincingly that the kind of anaphora which these operators are
calculated to facilitate does occur frequently and importantly in ordinary
discourse. He has thus removed one important obstacle from the way of
vindicating existence as a genuine predicate, and incidentally illustrated
how important metaphysical dogmas can be embodied in a perfectly
innocent-looking formalism. Although further arguments are still
needed, I hope to have persuaded you at least that there is no mistake
in considering existence a predicate. The mainstay of Kant's criticism of
the ontological argument is simply wrong.
But doesn't Kant deserve at least the honor of anticipating the Frege

KANT ON EXISTENCE AND PREDICATION

257

distinction? I am not sure that such precursorship necessarily qualifies

as an honor. I have shown that the Frege - Russell distinction is not
indispensable. 19 There are correct alternative ways of dealing with the
semantics of "is", ways which may even be preferable to the Fregean
method for the purpose of understanding natural language and
arguments conducted in natural language, including the arguments conducted by virtually all pre-Fregean philosophers. But even apart from
this devaluation of Frege, Kant does not rate the non-honor of paving
the way for Frege. It is amply evident from Kant's own words that he
does not think of our words for being as exhibiting the Frege - Russell
ambiguity. Seeing this is nevertheless made somewhat more difficult by
the fact that Kant does not speak of just existence (Dasein, Existenz), being (Sein), and "is" (ist), but also of "position" (setzen). What is this
"positing", anyway? Kant's pre-critical essay on proofs of God's existence shows unmistakably that it is merely another expression for being
("Sein"), and that it is unambiguous ("einJach"):
Der Begriff der Position oder Setzung ist viillig einfach und mit dem vom Sein iiberhaupt
einerlei. (Academy ed., Vol. 2, p. 73.)

In the Critique oj Pure Reason (A 598 = B 626), Kant says that" 'being'
is obviously [sic] not a real predicate .... It is merely the positing of a
thing, or of certain determinations, as existing in themselves".
The reason why Kant introduces the term "setzen" is probably a desire
to have a term which sits more happily with the cases in which "is" apparently has a merely predicative function. "God is omnipotent" could
according to Kant be true even if there were no God. It merely expresses
a necessary relation between the subject and the predicate. "God is omnipotent" does not logically imply for Kant that "God is", even though
the step might seem tempting. In order to avoid this temptation, it seems,
Kant uses his terminus quasi technicus "setzen" for positing something
as being - in any sense of being.
The explanation Kant gives of the difference between' 'God is omnipotent" and "God is" nevertheless shows that we are dealing with the same
"is" in both cases. In both cases, we are "positing" something. The only
difference is that in the former case the positing is relative but in the latter
case absolute. Otherwise, it is the same old positing.
lIn] the proposition, 'God is omnipotent' .... the small word 'is' adds no new predicate,
but only serves to posit !he predicate in its relation IKant's italics] to the subject.

Here positing clearly means predication. But Kant continues:

258

JAAKKO HINTIKKA

If, now, we take the subject (God) with all its predicates ... and say 'God is', or 'There
is a God', we attach no new predicate to the concept of God, but only posit the subject
in itself with all its predicates ...

Here positing clearly means to assume existence. The relation of existence to predication is thus merely that of an absolute positing to a
relative one. This observation is confirmed by further passages; witness
e.g. the following:
[In an existential jUdgment) ... nothing can have been added to the concept, which expresses merely what is possible, by my thinking its object (through the expression 'it is')
as given absolutely [my italics).

In the Beweisgrund, Kant puts the same distinction as follows:

Nun kann etwas als bloss beziehungsweise gesetzt, oder besser, bloss die Beziehung (respectus logicus) von etwas als einem Merkmal zu einem Dinge gedacht werden, und dann ist
das Sein, das ist die Position dieser Beziehung, nichts als der Verbindungsbegriff in einem
Urteil. Wird nicht bloss diese Beziehung, sondern die Sache an und fUr sich selbst gesetzt
betrachtet, ~o ist dieses Sein soviel wie Dasein. (Academy ed., Vol. 2, p. 73.)

Thus Kant clearly thinks of the "is" of predication (the copula) and the
"is" of existence as two uses of the same notion. Occasionally he even
seems to consider the copulative "is" (at least in necessary judgments)
as a variant of the "is" of identity. He thinks of a: necessary judgment
like "God is omnipotent" as expressing the identity of a God and an omnipotent God. "The omnipotence cannot be rejected if we posit a deity
... , for the two concepts are identical" (A 595 = B 623). Hence meaning differences between the first three elements of the Frege ambiguity
are rejected by Kant.
As to the fourth alleged sense of "is" apud Frege and Russell, Kant's
assimilation of it to other senses (especially to the "is" of predication)
is seen from his failure (or refusal) to distinguish the subsumption of one
concept to another from the application of a concept to a particular (in
other words, this particular's failing under the concept). This is particularly striking in the schematism chapter of the Critique oj Pure
Reason, as has been often remarked. 20
The insight that Kant did not assumm the Frege - Russell distinction
enables us to make further observations. Among other things, it follows
that Kant's main thesis is expressed somewhat inaccurately - and in
any case very narrowly - when it is said that according to him existence
is not a predicate. What he maintained, and frequent[y said, is that being
is not a real predicate. This applies to both existential and predicative
uses of "is"; predication is accordingly for Kant as little a predicate as

KANT ON EXISTENCE AND PREDICATION

259

existence is. This parity of the two is of course just a corollary to Kant's
failure (or refusal) to distinguish the different Fregean senses of "is"
from each other.
This helps to put certain puzzling-looking statements of Kant's
perspective. For instance, one of Kant's main pronouncements on our
topic in the first Critique runs as follows:
'Being' [SeinJ is obviously not a real predicate .. . . In its logical use lim logischen
Gebrauchl it is merely the copula of a judgment. (A 598 = B 626; Kant's emphasis.)

Here Kant makes his claim about being in general and then goes on to
apply it to predication rather than existence. Indeed, this predicative use
is precisely what he means by the "logical use" of being. In other words,
Kant's distinction between the logical use and other relevant uses of "is"
is the same as his contrast between the relative and absolute positing
discussed above. (This is, among other items of evidence, shown by our
latest displayed quote from the Beweisgrund; see especially the words
respectus logicus.) Philosophers have been puzzled by Kant's remarks as
to what happens to "is" in its merely logical use, and declared it irrelevant to Kant's main thesis that existence is not a predicate. 21 Kant'~
remarks are indeed not directly relevant, but only because they pertain
to a different but parallel case of his more general claim that being is not
a predicate.
There is one superficial aspect of Frege's and Russell's formalism
which misleadingly encourages the idea that Kant's thesis "existence is
not a predicate" is an anticipation of Frege. In the most literal sense, existence is not a predicate for Frege, either, viz. in the sense of being an
explicit predicate of individuals. We cannot take a free singular term
(Frege's "proper name"), say "b", and go on to assert "b exists".
However, this is a merely contingent feature of Frege's notation. What
is more, it partially hides one of the most fundamental features of his
treatment of existence, viz. that existence is expressed only by the existential quantifier.
In fact, the reason why Frege can get along without a predicate of existence is that he assumes that all proper names (free singular terms) are
nonempty. This is reflected by the validity of existential generalization
in Frege's system: from any proposition F(b) containing "b" we can infer (3x)F(x). This obviously presupposes that b exists. If we do not make
this assumption, we have to amplify the rule of existential generalization
and formulate it as saying that from the two premises
(9)

F(b) and b exists

260

JAAKKO HINTIKKA

we may infer
(10)

(3x)F(x).

From certain eminently natural assumption one can show (as I have
demonstrated)22 that the extra premise "b exists" must be equivalent
with
(11)

(3x)(b = x)

Indeed, all we need for this purpose is in effect that the logical constants
have their customary semantics and that the "predicate of existence" ,
whatever it is or may be, is subject to the same substitutivity principles
as other expressions of first-order logic .
This result shows that in a Frege - Russell logic it is in the last analysis
the existential quantifier alone that need carry existential assumption,
contrary to the misleading appearance created by Frege's notation. This
idea can be considered an integral part of Frege' s distinction between the
"is" of existence and other senses of "is". Indeed, this privileged position of the existential quantifier seems to me to be a much more important feature of the overall Frege - Russell approach to logic than the
alleged impermissibility of asserting the existence of an individual in
Frege's canonical notation. We can now see that Frege's distinction does
not presuppose that "existence is not a predicate". On the contrary, the
full import of Frege's approach cannot be spelled out without a
"predicate of existence". Hence Kant's thesis does not make him into a
precursor of Frege and Russell.
Thus we can likewise see that in the last analysis we could, and should,
have "a predicate of existence" also for the extremely simple languages
to which Frege (and mutatis mutandis also Kant) restricted his attention.
Consequently, the reasons for having such a predicate in one's language
are not applicable only to the rich languages envisaged above, but apply
also within the present-day Frege - Russell languages.
This observation nevertheless need not drive a wedge between Frege
and Kant. One way of expressing our result concerning Frege might be
to say that for Frege existence was a predicate, but not a normal or
"real" predicate. In its primary use, existence is a second-order
predicate, saying that a certain first-order predicate is instantiated. The
question whether this second-order predicate can be extended to the
trivial first-order predicates of the form "(b = x)" is of little interest to
Frege. But if so, there is after all a partial agreement between Frege and
Kant. For Kant frequently formulates his point by saying, no! that ex-

KANT ON EXISTENCE AND PREDICATION

261

istence is not a predicate, but that It IS not a real predicate. 23 Indeed, Kant
must obviously allow us to express not only the nonemptyness of common nouns but also the nonemptyness of singular nouns. Then Kant's
injunction that existence is not a real predicate might per haps be interpreted as saying merely that it must not be used in the definition of
anything. 24 This is an interesting point, but on the reconstruction of the
ontological argument presupposed here it is neither necessary nor sufficient for a refutation of the ontological argument. However, it is far
from clear what the precise import of Kant's locution is when he speaks
of a "real predicate", and it is not obvious a priori that his exclusion
of existence from the definition of anyone thing cannot itself be turned
into a line of defense for his critical claims. We shall return to these
points later.
That Kant's criticism of the ontological argument is largely beside the
point can also be seen in terms of his own system. It is largely a Fremdkorper in the body of his own transcendental philosophy. Earlier, I
quoted Kant as saying (in A 599 = B 627) that the concept "expresses
merely what is possible [my italics]". Elsewhere, too, he clearly thinks
of what I have called essential judgments as expressing possibilities. 25 in
an existential judgment, this possibility is asserted to be actualized,
without adding anything to the concept itself. Now this is precisely what
Kant could not say as his definitive opinion in the case of God. For if
God were in the fullest sense of the word possible for Kant, in the sense
of being empirically possible (possible in experience), he would
presumably be sometimes actual, and hence (since we are dealing with a
putatively necessary being) always actual. This possibility of restoring
something like the ontological argument by means of the additional
premise that God is possible had been exploited by Leibniz. Even though
Kant presumably would have rejected Leibniz' argument for other
reasons, he could scarcely afford to admit God's possibility.26 (In order
to see this, we may for instance recall Bill, where Kant says that
"necessity is just the existence which is given through possibility itself".)
It is true that the step I envisaged a moment ago from God's experiential possibility to his actuality is not backed by any outright assertions in
the Critique of Pure Reason of what Lovejoy called the Principle of
Plenitude, that is to say, of the principle that each genuine possibility is
actualized in the long run. Howevec, a closer examination of Kant's position shows that he could not really countenance vi~lations of the Principle
of Plenitude among full fledged experiential possibilities. This examination I have attempted, jointly with Heikki Kannisto, in an earlier

262

JAAKKO HINTIKKA

paper. 27 If instant evidence is desired, suffice it to recall Kant's

characterization of "the schema of modality and its categories" as "time
itself as the correlate of the determination whether and how it belongs
to time" and of necessity as "existence at all times" (A 145 = B 185).
The last point comes very close to asserting a version of the Principle
("what is always true is necessary"). Further evidence for the presence
of this version in Kant is obtained from the Prolegomena 18 - 19 and
from B 183. All told, Kant's ambivalence notwithstanding, his commitment during his critical period to something like the Principle of
Plenitude (for experiential possibilities) was too deep to allow him to rest
content with an unqualified statement that the concept of God is a merely
possible concept. In general, there is a (somewhat hidden but nevertheless unmistakable) connection between Kant's relative degree of
adherence to the Principle of Plenitude and his willingness to infer God's
existence from His possibility. Indeed, when Kant argues that from mere
concepts (conceptual possibilities) we can never infer actual existence, he
comes very close to criticizing one version of the Principle of Plenitude.
But as I have shown (with Kannisto, see Note 27 above), such criticism
is much more characteristic of Kant's interim philosophy (in the sixties
and early seventies) than of his mature "critical" philosophy.
God's possibility had to be banished from the realm of genuine
transcendental (experiential) possibilities to the outer darkness of the attenuated possibility enjoyed by mere ideals of reason. This is a move we
might have expected from the beginning that Kant the Copernican
revolutionary would make. Without it, his criticism of the ontological
argument does not do the whole job. It is the main link between Kant's
criticism of Anselm's and Descartes' arguments and his own
transcendental vantage point.
Kant was fully conscious of the necessity of this further move in his
campaign against God's existence on the level of experience (intuition)
or even understanding. In A 60 1 - 602 = B 629 - 630 he writes:
The concept of a supreme being is in many respects a very useful idea; but because it is a
mere idea, it is altogether incapable, by itself alone, of enlarging our knowledge in regard
to what exists. It is not even competent to enlighten us as to the possibility [Kant's italics)
of any existence beyond that which is known in and through experience.

Kant was likewise aware of the specific arguments that are lurking
here, ready to jump in at a mere admission of God's possibility:
And thus the celebrated Leibniz is far from having succeeded in what he prided ~imself on
achieving - the comprehension u priori of the possibility [my italics) of the sublime ideal
being. (A 602 = B 630)

KANT ON EXISTENCE AND PREDICA nON

263

Without taking up the details of Kant's line of thought in this direction, we can see the main idea clearly enough. God is a mere ideal because
He is an ens realissimum, the sum total of all reality. Few philosophers
or theologians would quarrel with this idea, and Kant is perhaps not so
very wrong in seeing in this definition a ground for assigning to God
merely ideal existence. The only remark I want to make here is that
Kant's move is beside the point as a criticism of the original ontological
argument, on the rational reconstruction offered above. For on this interpretation, nothing turned on God's being ens realissimum. God was
not assumed (or defined) to embody all perfections. For my
reconstructed argument, it sufficed merely to assume that God has one
particular perfection. It suffices to assume that He is the most powerful
being existentially, in precisely the sense asserted by Kant. (Cf. the quote
above from A 588 = B 616.) For this is what the inside conditional of (I)
expresses. Hence the move that Kant uses to undercut Leibniz-style
arguments for God's existence is completely beside the point as an objection to the original ontological argument. Once again, we see how Kant
misses the key fallacy in the argument.
Perhaps there nevertheless is a charitable interpretation of Kant's position which makes his transcendental vantage point relevant to a valid line
of criticism of the ontological argument. Very roughly, my initial
diagnosis of the ontological argument may be expressed by saying that
the trouble with the argument is not that existence is not a predicate, but
that we don't know who God is in the sense that the designation ("definition") of him as existentially greatest being provides us with no grounds
for concluding that the existentially greatest beings in different worlds
are identical. We may try to interpret Kant's position (and perhaps even
his slogan that existence is not a predicate) as emphasizing this aspect of
the situation. Kant's slogan is applicable in that Kant's saying that existence is not a "real" predicate may perhaps be taken to amount to saying that it does not help us to determine God's identity in the sense of
bringing us to know who God is. This idea is related to my earlier observation that according to Kant existence cannot be a part of the definition
of any entity. This would make Kant's discussion relevant to a valid line
of criticisms against the ontological argument. For it amounts to saying
that Kant was in the last analysis denying the indispensable auxiliary
premise illustrated by (6).
This highly interesting line of thought is the thesis of Hans Wagner's
paper (above, Note 5). If there is a kernel of truth in Kant's criticism of
the ontological argument, this undoubtedly is it. In spite of the evidence

264

JAAKKO HINTIKKA

Wagner marshals, I have nevertheless not quite been convinced that we

can actually attribute this idea to Kant. The best primajacie evidence for
this defense of Kant becomes from the analysis of his claim that existence
is not a real predicate, which was indicated briefly above. Undoubtedly
a closer examination of the texts could yield some evidence for this view.
It finds some support for instance in C. C. E. Schmid's Worterbuch zum
leichteren Gebrauch der Kantischen Schriften (vide Bestimmung), where
we find a contrast between merely logical predicates or accidents and real
predicates which can be "a determination of a thing" (my italics). All
told, it nevertheless seems to me that this line of thought in Kant is so
hopelessly entangled with other ideas that there is little hope of reaching
a viable defense of his views. Before a sharp distinction is made between
(l) and (2), the need of the auxiliary premise is not obvious, and the true
target of Kant's criticism is impossible to establish.
It is nevertheless worth asking what precisely was meant by Kant when
he said - as he occasionally did - that existence was not a real
predicate. One passage in which he does so is A 598 == B 626, partially
quoted above. A fuller version of the passage runs as follows:
'Being' (Seinl is obviously not a real predicate; that is, it is not a concept of something that
could be added to the concept ora thing. It is merely the positing of a thing, or of certain
determinations, as existing in themselves.

Kant indicates that this positing can be of two kinds:

In its logical use, it [sc. beingl is merely the copula of a judgment. ... [Then I the small
word 'is' adds no new predicate but only serves to posit the predicate in its relation to the
subject.

In contrast,
if, now, we ... say "God is" or "There is a God", we ... only posit the subject in itself
with all its predicates.

Then I am
thinking its [my concept'sl object (through the expression 'it is') as given absolutely.

Here we can also see what the force of Kant's little word "real" (in
saying that existence is not a "real predicate") really is. Kant is merely
following his customary contrast between what is logical and what is real
and identifying the "real" use of "being" to its existential use. In other
words, what is "unreal" about the purely predicative use of "is" is the
absence of existential presuppositions. In fact, the quoted passage continues immediately:
Otherwise stated, the real contains no more than the merely possible.

We have already seen that Kant brackets together the merely possible,

KANT ON EXISTENCE AND PREDICATION

265

the predicative (copulative) use of being, and logic (i.e., the world of
concepts).27 By contrast, the real or actual should go together with the
existential use of being. If I may turn Kant's point into a tautology, what
he is saying is that the predicative use of "is" is not its existential use.
If this suggestion is correct, then the burden that many philosophers
have tried to put on "real" in "real predicate" is largely misplaced.
Some philosophers have for instance tried to find links between Kant's
criticism of the ontological argument and his discussion of reality as one
of the modal categories in the Transcendental Analytic, and assumed
that they are what is highlighted by the word "real". Others have
thought that they could perceive in Kant a contrast between being as a
"real" predicate (as a predicate of individuals) and its "merely logical"
use (as a higher order predicate). There is no foundation in the text for
either view, and the passage we just examined suggests that what Kant
intended was something much simpler. 28
It might seem at this point that one part of the Frege - Russell distinction does after all playa major role in Kant, viz. the distinction between
the existential and the predicative uses of "is". It is true that Kant puts
a premium on this distinction, but we have already established beyond
all doubt that it is for him a difference in use and not a difference in
meaning.
NOTES
The writing of this paper was made possible by a fellowship from the John Simon Guggenheim Memorial Foundation. In writing it, I have profited greatly from discussions with
Merrill B. Hintikka, Russell Dancy, and Robert Beard, and esrecially from conversations
and correspondence with Robert Howell. I also profited greatly from the discussion of an
early version of this paper at the Fourth International Colloquium in Biel, May 1-4, 1980,
and I .would like to thank all the participants in that discussion.
1 Cf. e.g. The Many-Faced Argument, ed. by John Hick and Arthur C. McGill, the Macmillan Co., New York, 1967; The Ontological Argument, ed. by Alvin Plantinga and
Richard Taylor, Doubleday, Garden City, N.Y., 1965; Dieter Henrich, Der ontologische
Gottesbeweis, J. C. B. Mohr, Ttibingen, 1960 (second ed. 1967); W. L. Gombocz, Uber
1: Zur Semantik des xistenzpriidikates und des ontologischen Argumentes, Verband der
wissenschaftlichen Gesellschaften bsterreichs Verlag, Wien, 1974; Jonathan Barnes, The
Ontological Argument, Macmillan, London, 1972. All these give further references to the
literature. For a contemporary journalistic view, see Time, April 7, 1980, Pi>. 65 - 68.
2 Cf. my paper ' "Is", Seman tical Games, and Semantical Relativity', Journal of
Philosophical Logic 8 (1979), 433 - 468, which provides further references to the literature.
3 Cf. here my AP A presidential address 'Gaps in the Great Chain of Being: An Exercise
in the Methodology of the History of Ideas', Proceedings and Addresses of APA 49
(1975 - 76), reprinted in S. Knuuttila (ed.), Reforging the Great Chain of Being, D. Reidel,
Dordrecht, 1981, pp. 1 -17.
4 See e.g. Jonathan Bennett, Kant's Dialectic, Cambridge University Press, Cambridge,

266

JAAKKO HINTIKKA

1974, pp. 228 - 240. Bennett even speaks in the title of this 72 of "the Kant - Frege view".
See my essay, 'On the Logic of the Ontological Argument', in laakko Hintikka, Models
for Modalities, D. Reidel, Dordrecht, 1%9. This essay prompted a perceptive attempt to
show that my criticism of the ontological argument is related to Kant's; see Hans Wagner,
'Uber Kants Satz, das Dasein sei kein Priidikat', Archiv fur Geschichte der Philosophie 53
(1971), 183 - 186.
6 Cf. 'On the Logic of the Ontological Argument' (loc. cit.)
7 The difference between (I) and (2) is essentially a de dicto-de re contrast. Further discussion of the relation between the two constructions in Kant is found in Robert Howell's
paper in Dialectica 35 No. I (1981).
8 See my essays collected in Esa Saarinen (ed.), Game-Theoretical Semantics, D. Reidel,
Dordrecht, 1979.
9 See here my books Time and Necessity, Clarendon Press, Oxford, 1973; and (with Simo
Knuuttila and Unto Remes) Aristotle on Modality and Determinism (Acta Philosophica
Fennica 29, No. I), North-Holland, Amsterdam, 1977.
10 Kant's thesis has implications beyond what is being discussed here. It can be construed
as criticizing the medieval and neo-Platonic idea that existence qua existence carries with
itself interesting attributes of which we can profitably theorize. The problems goes back
to Aristotle's aporia concerning a science of being qua being. By and large, Aristotle was
l1luch more wary of such a science than were his followers. Kant's denial that existence is
a predicate may hence be viewed as the end of a long neo-Platonic and scholastic detour.
(Cf. Notes 13 and 24 below.)
II On the history of the interpretations of this pronouncement, cf. E. Gilson, History of
Christian Philosophy in the Middle Ages, Random House, New York, 1955, pp. 3, 69 -70,
92,149,216,253,293,368,371,438-439,579, and 591.
12 Some indications are nevertheless found in Kant, Logik, Academy Edition, Vol. 9, pp.
60-61.
13 Cf. Russell M. Dancy, Sense and Contradiction: A Study in Aristotle, D. Reidel, Dordrecht, 1975. (See esp. Appendix II, pp. 153-155.)
14 In Donald Davidson and laakko Hintikka (eds.), Words and Objections, D. Reidel,
Dordrecht, 1969, pp. 178-214.
IS Cf. also lerome Shaffer, 'Existence, Predication, and the Ontological Argument', Mind
71 (1962), 307 - 325. Shaffer maintains that "the most that the Ontological Argument
establishes is the intensional object, God ... ". Apparently Shaffer does not see any problem in the uniqueness of that "intensional object".
16 This observation helps in understanding other features of the literature on the ontological argument. For instance, why do most of the recent formal or semi-formal discussions of the ontological argument presuppose S5? Because the alternativeness relation is
symmetric in S5, and thus allows for an attenuated form of "return journeys", which
brings a merely possible individual back to the actual world.
17 Even some of the medievals seem to have been aware of the need of the kind of return
trip logic exemplified by (8); see Simo Knuuttila and Esa Saarinen, 'Backwards-Looking
Operators in Buridan', in I1kka Niiniluoto et al. (eds.), Studia Excel/entia: Essays in
Honour of Oiva Ketonen (Reports from the Department of Philosophy, University of
Helsinki, 1977, No.3.), pp. 11-17.
18 See Esa Saarinen's own contributions to Game-Theoretical Semantics, ed. by Esa
Saarinen, D. Reidel, Dordrecht, 1979.
19 See Note 2 above.
20 Cf. e.g. Henry Allison, 'Transcendental Schematism and the Problem of the Synthetic
A Priori', Dialectica 35 (1981); Gerold Prauss, Erscheinung bei Kant, Berlin, 1971, p. 103.
5

KANT ON EXISTENCE AND PREDICATION

267

21 See, e.g ., S. Morris Engel, 'Kant's "Refutation" of the Ontological Argument", in

Kant: A Collection of Critical Essays, ed . by Robert Paul Wolff, Doubleday, Garden City,
N.Y., 1967, pp. 189-208, esp. pp. 193-194.
22 See my essay, 'On the Logic of Existence' , in Modelsfor Modalities (Note 5 above), and
cf. my 'On the Logic of Existence and Necessity', The Monist 50 (1966), 55 - 76. For Frege,
see, e.g., Rainer Stuhlmann-Laeisz, 'Freges Auseinandersetzung mit der Auffassung von
"Existenz" als ein Priidikat der ersten Stufe und Kants Argumentation gegen den ontologischen Gottesbeweis', in Christian Thiel (ed.), Frege und die moderne Grundlagenforschung, Anton Hain, Meisenheim am Glan, pp. 119-133.
23 For this issue, cf. Gerold Prauss, Einfuhrung zur Erkenntnistheorie, Wissenschaftliche
Buchgesellschaft, Darmstadt, 1979.
24 Somewhat surprisingly, Kant is here completing a full circle in the history of philosophy
and returning back to the views of Aristotle, who similarly held that "being is not the
essence of anything" (An . Post. 11,7,92 b 13-14).
2S Cf. A 599 = B 627; "My financial position is . . . affeeted very differently by a hundred
real thalers than it is by the mere concept of them (that is, by their possibility)" . (Kemp
Smith translates this instead as " .... .. of their possibility". This is also grammatically
possible, but seems to me to water down Kant's unmistakable point.)
26 Some reasons why Kant thought that the ontological argument cannot introduce God
as a possible but individuated being (so that we could move from (I) to (2) can be gathered
from Robert Howell's interesting papers. See his contribution to Dialectica 35 (1981) as
well as his paper, ' Intuition, Synthesis, and Individuation in the Critique of Pure Reason',
NoUs 7 (1973), 207 - 232.
27 See laakko Hintikka and Heikki Kannisto, 'Kant on "the Great Chain of Being" "
Philosophic Exchange 2 1976), 69 - 86, esp. 75 - 78 reprinted in Knuuttila (Note 2 above),
pp. 287 - 308.
28 In any case, it does not help in trying to vindicate existence as a predicate (if not a 'real'
predicate) apud Kant to appeal to his use of existence as one of the modal categories. Kant
maintains that none of the modal categories is a predicate, either. Indeed, he says (in
A 74 = B 99 - 1(0) that the distinguishing characteristic of modality is that it "contributes
nothing to the content of judgment ... but concerns only the value of the copula in relation
to thought in general". Thus modality is basically also as a matter of different uses of the
copula, i.e., of "being" .

Dept. of Philosophy,
Florida State University,
Tallahassee, FL 32306-1054, U.S.A.

LEILA HAAPARANTA

ON FREGE'S CONCEPT OF BEING*

1. INTRODUCTION

One of the doctrines which Frege emphasizes in his writings is the thesis
that words for being, such as the English word is, are ambiguous. A large
part of his philosophy can be seen as an attempt to make us realize the
importance of keeping the different meanings of is apart and to catch the
philosophical mistakes brought about by our failure to see the ambiguity. Jaakko Hintikka has recently argued that, except for John Stuart Mill
and Augustus De Morgan, the ambiguity claim did not play any major
role in philosophical thinking before Frege and Russell. l What Frege and
Russell accomplished was to make the ambiguity of is a cornerstone of
modern first-order logic. Therefore, as Hintikka has pointed out,
"anyone who uses this logic as his or her framework of semantical
representation is thus committed to the Frege - Russell ambiguity thesis"
(Hintikka, 1983, p. 449). Hintikka has shown that in an alternative
seman tical representation, namely, in game-theoretical semantics, no
ambiguity claim need be made, which, of course, is not to deny that there
are different uses of is. The operative question is whether the differences
between these different uses have to be accounted for by assuming that
one particular verb for being is ambiguous, i.e., has several altogether
different meanings.
Independently of Hintikka, Charles Kahn (1973) and Benson Mates
(1979), among others, have called attention to the ambiguity doctrine
and partly challenged the validity of what Frege considered as an "eternal truth" concerning the verb to be.
But how is the verb is ambiguous in Fregean logic? Frege distinguishes
from each other the following meanings of is:
(1)

(2)
(3)

the is of identity (e.g., Phosphorus is Hesperus; a= b),

the is of predication, i.e., the copula (e.g., John is a
philosopher; P(a)),
the is of existence
(i) expressed by means of the existential quantifier and the
symbol for identity (e.g., God is; (3x)(g = x)),

269
S. Knuuttila and 1. Hintikka (eds.), The Logic of Being, 269-289.
1986 by D. Reidel Publishing Company.

270

(4)

LEILA HAAPARANTA

or
(ii) expressed by means of the existential quantifier and the
symbol for predication (e.g., There are human beings/There
is at least one human being; (3x)H(x,
and
the is of class-inclusion, i.e., generic implication (e.g., A horse
is a four-legged animal; (x)(p(x) ::::l Q(x))).

As is shown in the brackets, each putative meaning of is has its own formalization in first-order logic.
Since the Fregean view of is has held such an indisputable position in
modern first-order predicate calculus, few philosophers have tried to
find out what actually led Frege to the ambiguity thesis. This paper is an
attempt to give some hints of an interpretation of the Fregean distinction. I shall concentrate on the following questions: (1) How did Frege
arrive at the distinction between the is of existence and the is of predication? (2) What is the philosophical background which motivated Frege's
distinction between the is of identity and the is of predication?
2. FREGE ON 'IS'

Frege discusses the ambiguity thesis and the different meanings of is on

various occasions, but he does not present all four meanings in anyone
single text. His view of is can be gathered from pieces taken mainly from
'Dialog mit Ptinjer tiber Existenz' (written before 1884, published in
NS), Die Grund/agen der Arithmetik (1884), and 'Uber Begriff und
Gegenstand' (1892). There are also several remarks on the subject in the
rest of his writings.
The distinction between first-order and second-order concepts
(Begriffe erster und zweiter Stuftr), to which the ambiguity thesis is
related, was one of the central innovations in Frege's logic. In various
places Frege himself points out that he believes himself to be the first one
ever to present the distinction sufficiently clearly. 3 It is true that closely
similar suggestions had been made in the Middle Ages. 4 However, it was
undoubtedly Frege's new grasp of the scope and the limitations of the
first-order he created that made him sensitive to the contrast between
first-order and higher-order concepts. According to Frege's terminology, an object literally speaking falls under (jallt unter) a firstorder concept (Begriff erster Stufe) , in which case we use the is of
predication (copula), while a first-order concept falls in (jallt in) a
second-order concept (Begriff zweiter Stufe). S

ON FREGE'S CONCEPT OF BEING

271

Existence is, for Frege, a second-order concept. He expresses his view

most clearly in discussing the ontological argument for God's existence
and in criticizing Hilbert. As early as in the Begriffsschrift (1879) he hints
at his account of the is of existence, but a detailed argumentation is
presented only later in 'Dialog mit Piinjer iiber Existenz'. Frege continues his discussion in Die Grundlagen der Arithmetik, where he writes:
Affirmation of existence is in fact nothing but denial of the number nought. Because
existence is a property of concepts the ontological argument for the existence of God breaks
down. (GLA, 53f

In 'Uber Begriff und Gegenstand' he argues along the same lines:

I have called existence a property of a concept. How I mean this to be taken is best made
clear by an example. In the sentence 'there is at least one square root of 4', we have an asser
tion, not about (say) the definite number 2, nor about -2, but about a concept, square root
oj 4; viz. that it is not empty. (KS, p. 173; Geach and Black, p. 49.)

Later, in a letter to Hilbert (1900), Frege criticizes inferences from

essence to existence, which were used in the ontological argument for
God's existence. According to Frege, we are not allowed to infer that if
the propositions "A has the property <I> ", "A has the property 'If", and
"A has the property X" together with all their consequences do not contradict one another, there is an object which has the properties <1>, 'If and
X (BW, p. 75).
In Frege's logic, any concept which is not primitive can be decomposed
into its characteristics (Merkmale) , i.e., into concepts to which the
original concept is subordinate (untergeordnet). The is of class-inclusion
is used to express this subordination (Unterordnung), which is a relation
between two concepts of the same order . According to Frege, the object
to which the proper name of a given singular sentence refers is related
to the concept to which the remainder of the singular sentence refers in
a special way, which Frege calls subsumption (Subsumtion). Sub sumption is the same as first-order predication, and Frege takes it to be totally
different from subordination (NS, p. 210). For Frege, the is of classinclusion, which expresses the subordination, has no existential force
(NS, p. 207).
Frege regards the characteristics of a concept as its logical parts ('Uber
die Grundlagen der Geometrie', KS, p. 271). He stresses that the expressions Eigenschaft and Merkmal must be sharply distinguished from each
other: The concepts under which an object falls are called its properties
(Eigenschaften). An object has all the characteristics (Merkmale) of a
first-order concept as its properties if, and only if, it falls under the con-

272

LEILA HAAPARANTA

cept. For Frege, existence is a property, not a characteristic, of a firstorder concept. 7

The distinction between the is of identity and the is of predication is
presented in the article 'Ober Begriff und Gegenstand'. I shall repeat
Frege's own words:
Surely one can just as well assert of a thing that it is Alexander the Great, or is the number
four, or is the planet Venus, as that it is green or is a mammal? If anybody thinks this, he
is not distinguishing the usages of the word 'is'. In the last two examples it serves as a
copula, as a mere verbal sign of predication. (In this sense the German word ist can
sometimes be replaced by the mere personal suffix: cf. dies Blatt ist grUn and dies Blatt
grUnt.) ... In the first three examples, on the other hand, 'is' is used like the 'equals' sign
in arithmetic, to express and equation. In the sentence 'The morning star is Venus', we have
two proper names, 'morning star' and 'Venus', for the same object. In the sentence 'the
morning star is a planet' we have a proper name, 'the morning star', and a concept-word,
'planet'. So far as language goes, no more has happened than that 'Venus' has been
replaced by 'a planet'; but really the relation has become wholly different. (KS, pp.
168-169; Geach and Black, pp. 43-44.)

The roots of Frege's distinctions can, so it seems, be found quite easily.

One obvious starting-point for Frege is Kant's criticism against the ontological argument for God's existence, which is perspicuous in Frege's
early article on existence but which also appears in the Grundlagen.
Another origin is, undoubtedly, the mathematical usage, as to the sign
of identity. However, what we must still detect are the philosophical
tenets which are implicit in Frege's ambiguity thesis.
3. THE DISTINCTION BETWEEN 'SINN' AND 'BEDEUTUNG'

Before discussing the meanings of is in detail, I shall clarify the concepts

of sense (Sinn) dnd reference (Bedeutung) in Frege's philosophy. Frege
divides linguistic expressions into two groups: there are complete expressions, or proper names (Eigennamen), and incomplete expressions, or
function-names (Funktionsnamen) (GGA I, pp. 43 - 44). A functionname is ungesiittigt, i.e., it has a "gap" which can be filled with other
expressions. For example, "( ) is a capital" and "( ) is ( )'S son" are
function-names. A proper name is gesiittigt, i.e., it has no gap. For example, "Walter Scott", "the present king of France", and "Helsinki is the
capital of Finland" would be proper names according to Frege.
Function-names include concept-words, which have one argument place,
e.g., "( ) is a capital", and relation-words, which have two or more empty places, e.g., "( ) is ( )'S son". By filling gaps with complete expressions, we can form sentences, which are themselves complete expres-

ON FREGE'S CONCEPT OF BEING

273

sions. Besides concept-words and relation-words, there are functionnames like "the capital of ( )" from which we do not derive sentences
by filling their gaps with complete expressions.
Proper names can refer to objects, e.g., particular things, numbers,
classes, etc.; and sentences, which are -included by Frege among proper
names, refer to truth-values, the True and the False. Function-names
refer to functions: concept-words to concepts, relation-words to relations, and other function-names to other functions. 8 The values of functions named by concept-words and relations-words are always truthvalues. A concept, when filled with an object, will turn into an object,
namely, into a truth-value, and the same applies to a relation (GGA I,
pp. 7 - 8). When a function which is neither a concept nor a relation is
filled with an object, it will also turn into an object, but not a truth-value.
Names stand for their references (Bedeutungen) and express their
senses (Sinne). 9 Frege tells us quite explicitly that the sense of a sentenceis a thought. It is, however, much more difficult to see what he actually
means by the sense of a proper name which is not a sentence. In 'Uber
Sinn und Bedeutung' he remarks that the sense of a proper name is a way
in which the object to which this expression refers is presented (die Art
des Gegebenseins), or a way of 'looking at' this object. Furthermore, he
states that the sense expressed by a proper name belongs to the reference.
In other words, for Frege, senses are not primarily senses of names but
senses of references. Hence, it is more advisable to speak about senses
expressed by names than senses of names. Frege also gives some
examples of Sinne, like the Evening Star and the Morning Star as senses
of Venus, and the teacher of Alexander the Great and the pupil of Plato
as senses of Aristotle ('Uber Sinn und Bedeutung', KS, p. 144).
On the basis of Frege's hints, it is obvious at least that his concept of
Sinn is throughout cognitive. I suggest that the Sinne are nothing but
complexes of individual properties of objects. Frege regards it as possible
for an object to be given to us in many different ways. For him, our
knowledge of an object determines what sense, or what senses, the name
of the object expresses to us. Furthermore, he argues that complete
knowledge of the reference would require us to be able to say whether
any given sense belongs to the reference but that such knowledge is
beyond our reach ('Uber Sinn und Bedeutung', KS, p. 144).
As far as the suggested interpretation of the concept of Sinn is correct,
it is Frege's view that we know an object completely only if we know all
its properties, which is not possible for a finite human being. It also
follows that, according to Frege, each object can in principle have an in-

274

LEILA HAAPARANTA

finite number of names which correspond to the modes of presentation

of the object. None the less, objects may not be bundles of properties for
Frege, since he does not maintain that we know an object if, and only
if, we know all its senses, which leaves us the possibility that there is
something else than properties in an object. I shall return to this problem
later on.
Accordingly, Frege does not hold the position that knowing some arbitrary property or complex of properties of an object constitutes knowing the object completely since, for Frege, a necessary condition for
knowing an object would be knowing all the facts, or true thoughts, of
that object. 10 Nevertheless, according to him, in a weaker sense we know
an object precisely by knowing some properties of that object. It is true
Frege's weaker sense of knowing an object is not free from problems,
either, even if it is more natural than the stronger sense. For Frege does
not explain which properties of an object one must know in order to
know the object.
So far, I have limited my considerations to the senses of proper names.
In the 1950s and 1960s Frege scholars held conflicting positions on
function-names' having senses and references. II After Frege's
posthumous writings were published in 1969, it could be easily
documented that Frege attaches both senses and references also to
function-names (NS, p. 135). Frege decomposes the thought expressed
by a sentence, removing the senses of complete expressions, and we are
left with the Sinne of the function-names, which are simply incomplete
parts of the thought (NS, p. 187).
4. EXISTENCE AND PREDICATION

Frege's view of being is reminiscent of Kant's position in Kritik der

reinen Vernunft, where Kant states:
Being' is obviously not a real predicate; that is, it is not a concept of something which
could be added to the concept of a thing. (A 598/8 626)

It is far from surprising that Frege's view of is, as not being a common

first-order concept, originates from Kant's philosophy once we keep in

mind that Frege repeatedly refers to Kant's writings. Admittedly he
sometimes adopts a critical standpoint, but - what is particularly striking - even then he often pays homage to Kant's work. 12 However, the
similarities between Kant's and Frege's texts must be carefully scrutinized if we wish to advance our understanding of how Frege came to

ON FREGE'S CONCEPT OF BEING

275

regard existence as a second-order concept. This is in fact the position

that led him to the distinction between the is of existence and the is of
common first-order predication. What follows is merely a preliminary
inquiry into the connections between Kant and Frege.
Let us first look at Frege's argumentation in 'Dialog mit Piinjer iiber
Existenz', which has also been analysed in an interesting manner by
Rainer Stuhlmann-Laeisz (1975). According to Frege, such sentences as
"Leo Sachse is" and "Leo Sachse exists" are self-evident. Piinjer opposes this view. He suggests that the word is carries the same meaning
as is something that can be experienced (ist erfahrbar). He considers the
set of objects of experience (Gegenstiinde der Erfahrung) to be a subset
of the set of objects of ideas (Gegenstiinde von Vorstellungen). Frege
argues that Piinjer's account results in a contradiction: If "A is not"
means that A is not an object of experience, then there is something that
is not an object of experience. This is an existential statement, which, according to Piinjer, ought to mean that there is an object of experience
which is not an object of experience, which is a contradiction (NS, pp.
71-72).
In the afterword of the dialogue Frege continues his discussion of
"Leo Sachse is". He assumes for the sake of argument that this sentence
is not self-evident. That amounts to saying that being is a real property
of an object. If it is a property of an object, it is a characteristic of a firstorder concept. Frege proceeds with his argument as follows: If the word
being has some content and the sentence "A is" is not self-evident, then
the negation of the sentence "A is" could be true in some circumstances,
i.e., there could be entities which do not have the property of being.
From this it follows that the expression there are (es gibt) cannot be
replaced with being, i.e., "There are B's" is not equivalent in meaning
to "Something that has being (einiges Seiende) falls under the concept
B". For if the sentence "There are entities which do not have the property of being" means the same as "Something that has being falls under
the concept of not-being (der Begriff des Nichtseienden) " , it is evidently
a contradictory sentence. Frege concludes that if the sentence "There are
B's" is equivalent in meaning to the sentence "Something that has being
is B" , then the concept of being is completely self-evident. It is in no way
a determination (Bestimmung) of a thing, for it does not help us to
distinguish between any two objects (NS, p. 73).
Frege holds the view that we are forced to regard being as a concept
which is superordinate to every concept and which therefore lacks all
content. As Stuhlmann-Laeisz also remarks (p. 126), Frege's argumenta-

276

LEILA HAAPARANTA

tion actually shows that being can, after all, be used as a logical firstorder concept, although Frege himself does not propose anything of that
kind.
Frege's argument for the claim that sentences like "A is" or "A exists" are self-evident is based on the impossibility of negating these
sentences without contradiction. But it is not conclusive. Why does he
not suggest that the sentence "Something that has being is not" or
"Something that has being falls under the concept of not-being" means
that something for which it is possible to exist does not exist in the actual
world? Interpreted this way, the contradiction seems to disappear.
Precisely that kind of procedure can be used in possible worlds semantics. As laakko Hintikka argues, possible worlds semantics allows us "to
take a 'merely possible individual', i.e., a denizen of some other world,
and attribute to it predicates definable in terms of its actual existence,
maybe the 'predicate of actual existence' itself" (Hintikka, 1981 a, p.
134).
That kind of step is, however, impossible for Frege, due to his overall
conception of language and world. Frege points out on several occasions
that his aim in the Begriffsschrift was, in Leibniz's terms, not only to present a calculus ratiocinator but to create a lingua characteristica. 13 The
idea of logic as language seems to be incorporated in many of the doctrines maintained by Frege, as has been emphasized by lean van Heijcnoort (1967) and laakko Hintikka (1981b). For Frege, the conceptual
notation is a proper language, which must be learnt by means of suggestions and clues. Frege is committed to the doctrine that we cannot step
outside the limits of our language in order to consider the seman tical relations between language and reality. Moreover, in the Fregean framework
we cannot talk about changing universes. As van Heijenoort points out,
"Frege's universe consists of all that there is, and it is fixed" (van Heijenoort, 1967, p. 325). Frege cannot pick out an individual from some
possible world and attribute to it the property of actual existence for the
simple reason that, for him, there are no alternative worlds. That kind
of position prevents Frege from regarding the is of existence as an expression of a meaningful first-order concept.
Frege attaches the meaning of an existential judgement to the form of
a particular judgement. He states:
Every particular judgement is an existential judgement that can be converted into the 'there
is' form. E.g. 'Some bodies are light' is the same as 'There are light bodies'. (NS, p. 70;
Long and White, p. 63.)

ON FREGE'S CONCEPT OF BEING

277

The same point is repeated in 'Aufzeichnungen fur Ludwig Darmstaedter', which Frege wrote in 1919 (NS, pp. 274 - 275). A particular
judgement serves to express a connection between two concepts. Because
it always involves two concepts, it is, Frege notes, difficult to turn a
sentence like 'There are men' into the form of a particular judgement.
The problem can be solved, however, if the concept in question is
definable by means of two concepts, as the concept man is definable by
means of the concepts rational and living being.
Frege chooses the concept being identical with itse/f(sich selbst gleich
sein) as the most general concept of the hierarchy of concepts. It is a concept superordinate to all concepts and having no content since its extension is unlimited. Frege remarks that in natural language it is the copula
that is purported to express the most general concept without content.
He argues:
This makes it possible to say: inen = men that have being; 'There are men' is the same as
'Some men are' or 'Something that has being is a man'. Thus the real content of what is
predicated does not lie in 'has being' but in the form of a particular judgement. (NS, p.
71; Long and White, p. 64.)

Frege's view of being is also intertwined with his distinction between

Sinn and Bedeutung. That distinction throws some new light on Frege's
view of existence as an empty first-order concept. In 'Uber Sinn und
Bedeutung' he writes:
Idealists or sceptics will perhaps long since have objected: 'You talk, without further ado,
of the Moon as an object; but how do you know that the name 'the Moon' has any
reference? How do you know that anything whatsoever has a reference?' I reply that when
we say 'the Moon' , we do not intend to speak of our idea of the Moon, nor are we satisfied
with the sense alone, but we presuppose a reference. (KS, p. 147; Geach and Black, p. 61.)

Thus, Frege's view is that if we talk about an object and then state that
it exists, that statement does not add anything to what we have said so
far. We presuppose the existence of objects; we do not say that they exist. For Frege, every predication carries with itself the claim for existence.
None the less, Frege must admit that we talk about objects, for example,
fictional entities, which do not exist. In 'Uber Sinn und Bedeutung' he
states that a name may have a Sinn, although it lacks a Bedeutung and,
accordingly, a sentence may express a thought, although it lacks a truthvalue. In Frege's view, a sentence is deprived of a truth-value if it contains a name which has no bearer (KS, p. 148).
Even if Frege takes being, or existence, to be included in every predication, he regards it as a proper concept in itself, since he paraphrases it

278

LEILA HAAPARANTA

as a specific relation of an object to itself, i.e., self-identity. Frege also

suggests us a rendering of the statements" A is" and" A exists", which
makes them even meaningful statements. In 'Kritische Beleuchtung
einiger Punkte in E. Schr6ders Vorlesungen tiber die Algebra der Logik'
he states:
We must keep well apart two wholly different cases that are easily confused, because we
speak of existence in both cases. In one case the question is whether a proper name
designates something; in the other, whether a concept comprehends objects under itself.
I f we use the words 'There is a -', we have the latter case. (KS, p. 208; Geach and Black,
p. 104.)

According to Frege's clue, the sentences "A is" and "A exists" can be
paraphrased as the metalinguistic sentence "the name 'A' has a
reference". However, if Frege is consistent in his view of language as the
universal medium, the limits of which we cannot exceed, he does not accept such statements as talk about the expressions of our language.
In 'Dialog mit Ptinjer tiber Existenz' Frege concentrates on discussing
the concept of being as an unanalysed univocal concept and insists on the
Kantian claim that being is not a real predicate. Only at the end of the
dialogue does he come to the idea which he later emphasizes and
develops, i.e., taking existence to be a property of a concept:
The existence expressed by 'there is' cannot be a characteristic mark of a concept whose
property it is, just because it is a property of it. In the sentence 'There are men' we seem
to be speaking of individuals that fall under the concept 'man', whereas it is only the concept 'man' we are talking about. (NS, p . 75; Long and White, p. 67.)

Frege maintains that we can use existence as a proper second-order concept which expresses that a concept is instantiated, that an object falls
under a concept. Therefore, it is, after all, possible for Frege to say
meaningfully that an object exists in the sense of saying that a certain
bundle of properties is instantiated. Thus, existence is here something
that is asserted of the bundle of properties.
Now we are able to see more clearly the ties between Kant's and Frege's
positions. In Der einzig mogliche Beweisgrund zu einer Demonstration
des Daseins Gottes Kant argues:
Consider any arbitrary subject, e.g., Julius Caesar. Combine all his conceivable predicates,
including those of time and space; then you will soon realize that it is possible for him to
exist or not to exist with all these determinations .... Existence, which occurs in everyday
speech as a predicate, is not a predicate of a thing itself but rather a predicate of the thought
which we have of the thing (Kant's Gesammelte Schriften, Band 1/, p. 72.)

Frege's view of existence implies that we cannot talk about objects and
their existence directly, i.e., independently of the properties of objects.

ON FREGE'S CONCEPT OF BEING

279

At the same time, that is also one of Frege's basic semantical tenets. Between the name and the Bedeutung there is always the Sinn. An object
is for us always an object that falls under some concept.
The above extract from Kant's text shows how close Frege's view of
existence as a property of a concept actually comes to Kant's statements.
For Kant, existence is a predicate of a thought concerning an object. The
same holds true of Frege since, in his view, objects exist for us only as
subsumed under concepts, which means that we can know an object only
by knowing some thought or some thoughts concerning the object. Thus,
it seems as if Kant's views on existence influenced Frege's ideas at least
in two ways: first, Kant argues that being is not a real predicate, which
is the starting-point of Frege's discussion concerning existence, and
secondly, the idea that being is a property of a thought is developed by
Frege in the view that existence is a property of a concept.
5. IDENTITY AND PREDICATION

There are several discussions of Frege's concept of identity in the

literature, although hardly any suggestions have been made as to the
background of the distinction between the is of identity and the is of
predication. 14 Frege deals with the problem of identity (Gleichheit)
already in the Begriffsschrift. He begins his treatment with the remark
that sometimes a name seems to stand for its content (Inha/t), sometimes,
namely, in an identity statement, for itself. For Frege, it is, however, a
false impression that an identity statement concerns names only. According to the Begriffsschrift, an identity statement expresses that two signs
have the same content (Inhalt) but the content is determined in two different ways (zwei Bestimmungsweisen) (BS, 8).
In 'Uber Sinn und Bedeutung' Frege continues as follows his discussion of identity statements in connection with the distinction between
Sinn and Bedeutung:
What is intended to be said by a = b seems to be that the signs or names 'a' and 'b'
designate the same thing so that those signs themselves would be under discussion; a relatIOn between them would be asserted. But this relation would hold between the names or
signs only in so far as they named or designated something .... A difference can arise only
if the difference between the signs corresponds to a difference in the mode of presentation
of that which is designated. (KS, pp. 143 - 144; Geach and Black, pp. 56 - 57.)

Here the argument is essentially the same as in the Begriffsschrift,

although Frege changes his terminology: he replaces the Inhalt of a name
(but not that of a sentence) by Bedeutung and Bestimmungsweise by die

280

LEILA HAAPARANTA

Art des Gegebenseins or Sinn. In the article 'Funktion und Begriff' he

explicitly identifies the concept of Inhalt with that of Bedeutung (KS, p.
126).
Frege gives his account of identity in the Begriffsschrift as follows:
Now let I- (A == B)
mean: the symbol A and the symbol B have the same conceptual content, so
that we can always replace A by B and vice versa. (BS, 8.)

In the Grundlagen Frege not only suggests, but gives an explicit formulation of, Leibniz's law concerning the substitutivity of identicals (GLA,
65).15 Again, in 'Uber Sinn und Bedeutung' he repeats the same law
(KS, p. 150). In the Grundgesetze (Vol. I) he establishes his view of identity as the substitutivity of identicals, not only by referring to the
substitutivity of the names of identical objects, but by stating that if two
objects are identical, they fall under the same concepts, i.e., they have
exactly the same properties (GGA I, 20). Nevertheless, Frege does not
regard his account as a definition of identity, as he unquestionably indicates in 'Rezension von: E. G. Husserl, Philosophie der Arithmetik I'
(KS, p. 184). For Frege, identity is indefinable but Leibniz has offered
a valuable tool for understanding the identity relation. In this paper, I
shall not discuss the similarities and dissimilarities between Leibniz and
Frege in more detail.
We can, however, conclude that Frege gives four different accounts
of identity. First, he regards an identity statement as a rule for
substitutivity of names in different contexts. Secondly, he takes it to be
a metalinguistic statement concerning the number of senses and
references of two names. Thirdly, he considers identity to be a relation
between two objects. Fourthly, he takes it to be a relation of an object
to itself, which occurs in sentences like "a = a". As we saw above, the
fourth meaning is related to his view of existence.
As far as objects are concerned, the contexts where the names occur
express properties of the objects named. If two objects are said to be
identical, they are claimed to have the same properties. Considered this
way, identity can be regarded as a border-line case of predication. When
paraphrased in the manner Frege suggests, i.e., by means of the expression no other than, it predicates of the object under discussion all the
properties that it in fact has ('Uber Begriff und Gegenstand), (KS, p.
169). Nevertheless, I shall argue that Frege does not reduce objects to
their properties. 16 Nor does he reduce the identity of objects to the
sameness of their properties. What I seek to show is that those very tenets

ON FREGE'S CONCEPT OF BEING

281

of irreducibility are what makes Frege speak so eagerly in favour of the

distinction between the is of identity and the is of predication.
Frege takes special pains to distinguish identity from predication in his
article 'Uber Begriff und Gegenstand', where he deals with, and endorses, the crucial difference between objects and concepts. The importance of not blurring the distinction between objects and concepts is also
strongly emphasized by Frege in the Grundlagen, so much so that he adds
it to the list of his basic principles (see Einleitung, p. X). J7 In order to
attend to that principle, Frege considers it necessary to realize that the
is of identity differs from the is of predication and, moreover, that the
difference is a reflection of how things actually are (see 'Uber Begriff und
Gegenstand', KS, p. 169).
In contrasting objects with concepts Frege comes up with a decisive innovation compared with traditional Aristotelian logic. Indeed, he
himself recognizes some of the differences between his own treatment
and traditional logic. He remarks:
In fact, it is one of the most important differences between my mode of interpretation
I A uffassungsweise I and the Boolea~ mode - and indeed I can add the Aristotelian mode
- that I do not proceed from concepts, but from judgements. But this is certainly not to
say that I would not be able to express the relation of subordination between concepts.
(,Ober den Zweck der Begriffsschrift', BS, 1964, p. 10\; Bynum, p. 94.)

Frege's analysis of sentences differs from Aristotelian analysis and

results in a new treatment of the relation of subordination between concepts. In the beginning of the Begriffsschrift Frege points out:
A distinction between subject and predicate does not occur in my way of representing a
judgement. (BS, 3.)

Also, in the article 'Uber die Begriffsschrift des Herrn Peano und meine
eigene', Frege notes that subordination is alien to predication (KS, p.
244). What is reflected in the distinction between subordination and
predication is precisely the principle of always keeping in mind the difference between objects and concepts. Frege would also reject analyses
like that given by Aristotle in Analytica Posteriora, in which man is said
to be identical with animal, since man is identical with a species of animal
(An. Post. 83 a 23 - 35).
The innovation of Fregean logic that I am here aiming at is, however,
related to the epistemological import of the distinction between objects
and concepts. It amounts to an attack on the general philosophical doctrines implicit in the received logical analysis of Frege's day. Frege's view
is that we cannot reach the essence of an object by means of our concepts
and our cognitive capacities. To say, for instance, that Plato is a man,

282

LEILA HAAPARANTA

meaning that Plato is identical with a man, is not possible, according to

Frege. He expresses his standpoint incontestably in 'Logik in der
Mathematik' (1914):
Since therefore we need inferences and a second premise, the thought that Cato is mortal
is not included in what is expressed by the sentence 'All men are mortal', and so 'man' is
not an ambiguous word which amongst its many references has that which we designate
by the proper name Plato'. On the contrary, a conceptword simply serves to designate
a concept. And a concept is quite different from an individual. If I say 'Plato is a man',
I am not as it were giving Plato a new name - the name 'man' - but I am saying that
Plato falls under the concept man. (NS, p. 231; Long and White, p. 214.)18

Traditional Aristotelian logic blurs the distinction between individuals

and properties, since it allows the identification of individuals with their
essential properties. Therefore, Frege's distinction proves to be a notable
turning-point in the history of logic. 19 Frege does not give separate
treatments of essential and accidential properties. What the Fregean
distinction seems to bring about is a chasm between objects in themselves
and objects as falling under the concepts that we are able to harness.
What are objects, then, for Frege? In order to answer the question, I
must try to grasp the basic elements of the Fregean universe. In 'Ober
Begriff und Gegenstand' Frege writes:
What is simple cannot be decomposed and what is logically simple cannot have a proper
definition. (KS, p. 169; Geach and Black, pp. 42-43.)

When the reference to be assigned is logically simple, we cannot, according to Frege, give a proper definition but we must comply with
elucidatory propositions, by means of which we give hints concerning the
wanted reference (BW, p. 63).
It is obvious that the logically simple elements of Frege's world can be
concepts and, more generally, functions. That is implicitly expressed by
Frege himself in 'Ober die Grundlagen der Geometrie I - III' (KS, pp.
288 - 289). Concepts can be decomposed into more primitive concepts,
which are their characteristics. Still, we are not able to decompose for
ever. The most primitive units that we reach are the Bedeutungen of the
concept-words that can no more be defined.
Eike-Henner W. Kluge argues that the concept of logisch einfach is important in Frege's ontology and that logisch einfach is, for Frege, always
a functjon and never an object (Schirn II, 1976, pp: 60 - 65). Kluge's
argument is as follows: For Frege, a thing without properties is an Unding. Now if an object were logisch einfach and had only one property,
it ought to coincide with that propt;rty. For if it did not, it would be composed of this property and something that has this property, which is in-

ON FREGE'S CONCEPT OF BEING

283

consistent with the assumption that the thing is logisch einfach. Consequently, if it coincides with the property, it is a property itself and, as
a property, incomplete, which is inconsistent with its nature as an object.
Since there are no other possibilities, it follows that objects cannot be
logisch einfach.
Kluge's argument is based to a considerable extent on the fact that
Frege regards an object having no properties as an Unding. It is true
Frege maintains that objects without properties are impossible or even
absurd, but his wordings do not warrant the conclusion that the concept
of Unding applies to ontological considerations. Frege formulates his
claim ironically as follows:
First, things are regarded as similar, so similar to each other, that they can no more be
distinguished from each other; that is, all properties by which they are distinguished would
be demolished ... .Is there still something left of those things? Certainly! There are still
left the natureless things I die naturlosen Dinge I .... Nonetheless, it will be difficult to find
the natureless pieces of wood, or, rather, not-pieces of wood, in the set of the natureless
things. But even the gre~test difficulties are overcome by good will. The best means is
always this: they are totally ignored. ('Uber die Zahlen des Herrn Schubert', KS, pp.
247 - 248.)

Frege continues:
I have been led astray because I know the names of fixed stars but I do not know the names
of natureless things. (KS, pp. 250 - 251.)

In these passages we only encounter a version of Frege's doctrine according to which it is via a'sense and only via a sense that a proper name
is related to an object (NS, p. 135). Frege does not think that it is possible
to talk meaningfully of an object without referring to the properties of
the object. This is precisely the view that makes Frege an antihaecceitist
(see, e.g., Kaplan, 1975, p. 725).20 For Frege, it is impossible to know
objects in themselves or to talk about them. But if we wish to understand
the structure of Frege's ontology, objects can be taken to exist without
having any properties to carry. They are not known to us as simple,
however, and strictly speaking we cannot even talk about their existing
or not existing without properties. That is exactly what I emphasized in
the preceding section when discussing existence. According to hege, we
can say what an object is like, but we cannot say what it is.
But it does not suffice to recognize Frege's concept of Unding as having an epistemological force, for his concept of logisch einfach is not applicable to ontological considerations, either. For Frege, a reference is
logically simple if it cannot have a proper definition. But we can con-

284

LEILA HAAPARANTA

sistently assume that a reference has a complex structure even if we cannot define the reference by means of our concepts. 21 Nothing can be
logically simple from an ontological point of view, for if one calls a
reference logically simple, one already considers it within the limits of
human knowledge.
Frege's epistemological view comes up in Die Grund/agen der
Arithmetik as follows:
It is this way that I understand objective to mean what is independent of our sensation,
intuition and imagination, and of all construction of mental pictures out of memories of
earlier sensations, but not what is independent of the reason - for what are things independent of the reason? To answer that would be as much as to judge withoUI judging, or to
wash the fur without wetting it. (GLA, 26.)

It is, perhaps, symptomatic that in the beginning of this paragraph Frege

refers approvingly to Kant's ideas.
Now I am able to suggest an answer to the question concerning the
distinction between the is of identity and the is of predication: As I have
argued, Frege did not regard it as possible to find out what objects really
are in themselves. He relies on the distinction between what is accessible
to human knowledge and what is beyond our cognitive capacities. He
wants to contrast objects as metaphysical entities with objects as we
know them through a finite number of concepts, and therefore he also
stresses the sharp distinction between objects and concepts. In Frege's
epistemology, we can approach objects from different points of view by
means of the concepts that are available to us. But, for Frege, an object
is neither identical with any property nor with any combination of properties that we can know it to have. The best we can do, he believes, is
to handle identity, as it were, within the limits of our reason, i.e., to
regard identity statements as statements which claim two objects to have
precisely the same properties. However, according to Frege's
philosophy, identity statements claim more than our reason is able to
grasp. An identity statement concerns objects in themselves, while
predication belongs to the sphere of our reason.
6. CONCLUDING REMARKS

What I have argued for in this paper requires some reservations.

Although I have hinted at certain Kantian tenets, the connections between Kantian epistemology and Frege's ambiguity thesis are far from
being completely clarified. One open question is, no doubt, Frege's relation to what Hintikka calls Kant's Aristotelian mistake, which is the doc-

ON FREGE'S CONCEPT OF BEING

285

trine that we become aware of individuals and their existence by means

of passive sense-perception. 22
Another reservation I want to make is related to Aristotle and
Aristotelian logic. Here I have not carefully examined what Frege actually means in criticizing Aristotle. I have relied on the interpretation that
he means the logic prevailing in his own day, that is to say, I have not
discussed Frege's relation to Aristotle's metaphysics and epistemology;
not even his relation to Aristotle's logic. I have focussed my discussion
on Frege's critical standpoint concerning what he regarded as the errors
of Aristotelian logic and of the epistemology which seems to be codified
in that logic.
Given these additional remarks, what I can conclude on the basis of
the above discussion is the conception that Frege's thesis concerning the
ambiguity of the word is is epistemologically motivated and that the
epistemological tenets are, if not purely Kantian, at least strongly influenced by Kant's philosophy.
NOTES
I wish to thank Prof. laakko Hintikka, Dr. Lilli Alanen and Prof. Ernest LePore for
helpful comments.
I See, e.g., Hintikka (1979) and (198Ia).
2 In GLA ( 53) Frege uses the term Begriffe erster und zweiter Ordnung; in his later
writings he talks about Begriffe erster und zweiter Stufe (see, e.g., Funktion und Begriff,
KS, p. 140, and 'Uber die Grundlagen der Geometrie II', KS, p. 269).
3 See BW, pp. 73 -74, and 'Uber die Grundlagen der Geometrie II', KS, p. 269.
4 Angelelli (1967) discusses the subject to some extent (see pp. 192 - 200). I shall not go
into the problem in this connection.
5 Cf. 'Uber Begriff und Gegenstand', KS, p. 174, and 'Uber die Grundlagen der Geometrie
II', KS, p. 271.
6 I have taken the quotations from the English translations mentioned in the bibliography.
When no translation was available, I have referred only to the German text and proposed
my own translation.
7 See 'Uber Begriff und Gegenstand', KS, p. 175, and 'Uber Begriff und Gegenstand', NS,
pp. 106-121. See also BW, pp. 150-151.
8 Cf. GGA I, Vorwort, p. X; see also 'Uber die Grundlagen der Geometrie 1- III', KS,
p.285.
~ See 'Uber Sinn und Bedeutung', KS, p. 147. Frege formulates the distinction between
Sinn and Bedeutung in that very article.
10 For Frege's view of facts, see 'Der Gedanke', KS, p. 359.
II See, e.g., W. Marshall, 'Frege's Theory of Functions and Objects'; M. Dummett, 'Frege
on Functions: A Reply'; and W. Marshall, 'Sense and Reference: A Reply', in Klemke
(1968), pp. 249 - 267, 268 - 283, and 298 - 320.
12 See, e.g., GLA, 89.
13 See 'Uber den Zweck der Begriffsschrift', BS, 1964, p. 98, and 'Uber die Begriffsschrift

286

LEILA HAAPARANTA

des Herrn Peano und meine eigene', KS, p. 227. See also' Anmerkungen Freges zu: Philip
E. B. Jourdain, The Development of the Theories of Mathematical Logic and the Principles of Mathematics', KS, 341. Frege uses the expression lingua characterica instead of
lingua characteristica. For this, see G. Patzig's footnote 8 in Gottlob Frege, Logische
Untersuchungen, ed. by G. Patzig, p. 10.
14 See Schirn (1976).
IS Austin's translation of this passage is the following: "Now Leibniz's definition is as
follows: 'Things are the same as each other, of which one can be substituted for the other
without loss of truth'. This I propose to adopt as my own definition (Erkliirung) of identity .... Now, it is actually the case that in universal substitutability all the laws of identity
are contained." I think it is misleading to use the word definition as the translation of
Erkliirung, because Frege argues elsewhere that the concept of identity is indefinable.
10 This was originally suggested to me by Prof. Jaakko Hintikka.
17 For the sharp distinction between objects and concepts, see also 'Ober die Begriffsschrift des Herrn Peano und meine eigene', KS, p. 233, and 'Ober die Grundlagen der
Geometrie II', KS, p. 270.
18 I have changed P. Long's and R. White's translation to the effect that I have adopted
the word reference as the translation of the German word Bedeutung, while Long and
White use the word meaning in this connection.
19 A similar idea concerning Frege's central role in the history of logic has been put forward by Ignacio Angelelli. See Ange1elli (1967), pp. 253 - 254.
Aristotle's view is discussed by Jaakko Hintikka in the article 'The Varieties of Being
in Aristotle' (this volume).
20 Different kinds of haecceitism have been discussed by R. M. Adams (1979). He defines
haecceitas as the property of beillg identical with a certain particular individual (Adams,
1979, p. 6). For the details of Frege's doctrine, see Haaparanta (1985).
21 For this point I am indebted to Prof. Jaakko Hintikka.
22 Cf. Hintikka (1981 b) and Hintikka, 'The Paradox of Transcendental Knowledge'. See
also Haaparanta (1985).

BIBLIOGRAPHY
Frege, G., Begriffsschrift, eine der arithmetischen nachgebildete Formelsprache des reinen
Denkens, Verlag von L. Nebert, Halle a. S., 1879; repro in Frege (1964), pp. 1-88.
(Referred to as BS.)
Frege, G., 'Ober den Zweck der Begriffsschrift', in Sitzungsberichte der Jenaischen
Gesellschaft fiir Medizin und Naturwissenschaft fiir das Jahr 1882, Verlag von G.
Fischer, Jena, 1883, pp. 1-10; repr. in Frege (1964), pp. 97 -106.
Frege, G., Die Grundlagen der Arithmetik: eine logisch mathematische Untersuchung iiber
den Begriff der Zahl, Verlag von W. Koebner, Breslau, 1884; repr. and transl. by J. L.
Austin in The Foundations of Arithmetic/Die Grundlagen der Arithmetik,. Basil
Blackwell, Oxford. 1968. (Referred to as GLA.)
Frege, G., Funktion und Begri'f. Vortrag. geliaIten in der Sitzung vom 9. Januar 1891 der
Jenaischen Gesellschaft fiir Medizin und Naturwissenschaft, H. Pohle, Jena, 1891;
repro in KS. pp. 125 - 142.
Frege, G., 'Ober Sinn und Bedeutung', Zeitschrift fiir Philosophie und philosophische
Kritik 100 (1892), 25 - 50; repro in KS, pp. 143 - 162.

ON FREGE'S CONCEPT OF BEING

287

Frege, G., 'Uber Begriff und Gegenstand', Vierteljahrschrift fur wissenschaftliche

Philosophie 16 (1892), 192-205; repro in KS, pp. 167 -178.
Frege, G., Grundgesetzeder Arithmetik, begriffsschriftlich abgeleitet, I. Band, Verlag von
H. Pohle, Jena, 1893. (Referred to as GGA I.)
Frege, G., 'Rezension von: E. Husserl, Philosophie der Arithmetik, Erster Band, Leipzig,
1891, Zeitschrift fur Philosophie und philosophische Kritik 103 (1894), 313 - 332; repro
in KS, pp. 179 - 192.
Frege, G., 'Uber die Begriffsschrift des Herrn Peano und meine eigene', Berichte uber die
Verhandlungen der koniglich siichsischen Gesellschaft der Wissenschaften :;u Leipzig,
Mathematisch-Physische Klasse 48. Band (1896), 361 - 378; repro in KS, pp. 220 - 233.
Frege, G., Ober die Zahlen des Herrn H. Schubert, Verlag von H. Pohle, Jena, 1899; repro
in KS, pp. 240 - 261.
Frege, G., Grundgesetze der Arithmetik, begriffsschriftlich abgeleitet, I/. Band, Verlag
von H. Pohle, Jena, 1903.
Frege, G., 'Uber die Grundlagen der Geometrie', Jahresberichte der Deutschen
Mathematiker- Vereinigung, 12. Band (1903), 319 - 324; repro in KS, pp. 262 - 266.
Frege, G., 'Uber die Grundlagen der Geometrie II', Jahresberichte der Deutschen
Mathematiker- Vereinigung, 12. Band (1903), 368 - 375; repro in KS, pp. 267 - 272.
Frege, G., 'Was ist eine Funktion?', in Festschrift Ludwig Boltzmann gewidmet zum
sechzigsten Geburtstag, 20. Februar 1904, Ambrosius Barth, Leipzig, 1904, pp.
656 - 666; repro in KS, pp. 273 - 280.
Frege, G., 'Uber die Grundlagen der Geometrie I-III', Jahresberichte der Deutschen
Mathematiker- Vereinigung, 15. Band (1906), pp. 293 - 309,377 -403,423 - 430; repro
in KS, pp. 281 - 323.
Frege, G., 'Anmerkungen zu: Philip E. B. Jourdain, The Development of the Theories of
Mathematical Logic and the Principles of Mathematics', The Quarterly Journal of Pure
and Applied Mathematics 43 (1912), 237 - 269; repro in KS, pp. 334 - 341.
Frege, G., 'Der Gedanke, eine 10gische Untersuchung', Beitriige zur Philosophie des
deutschen Idealism us, Band I (1918), 58-77; repro in KS, rp. 342-362.
Frege, G., 'Die Verneinung, eine logische Untersuchung', Beitriige zur Philosophie des
deutschen Idealism us, Band I (1918), 143 - 157; repro in KS, pp. 362 - 378.
Frege, G., 'Logische Untersuchungen, dritter Teil: Gedankengefiige', Beitriige zur
Philosophie des deutschen Idealism us, Band III (1923), 36-51; repro in KS, pp.
378 - 394.
Frege, G., Translations from the Philosophical Writings of Gottlob Frege, ed. by P. Geach
and M. Black, Basil Blackwell, Oxford, 1952.
Frege, G., Begriffsschrift und andere Aufsiitze, ed. by I. Angelelli, Georg Olms,
Hildesheim, 1964.
Frege, G., The Basic Laws of Arithmetic, Exposition of the System, transl. and ed. by M.
Furth, University of California Press, Berkeley and Los Angeles, 1964.
Frege, G., Logische Untersuchungen, ed. by G. Patzig, Vandenhoeck & Ruprecht, G6ttingen, 1966.
Frege, G., Kleine Schriften, ed. by I. Angelelli, Wissenschaftliche Buchgesellschaft, Darmstadt and Georg Olms, Hildesheim, 1967. {Referred to as KS.)
Frege, G., Nachgelassene Schriften, ed. by H. Hermes, F. Kambartel, and F. Kaulbach,
Felix Meiner Verlag, Hamburg, 1969. (Referred to as NS.)
Frege, G., On the Foundations of Geometry and Formal Theories of Arithmetic, transl.
by E.-H. W. Kluge, Yale University Press, New Haven and London, 1971.

288

LEILA HAAPARANTA

Frege, G., Conceptual Notation and Related Articles, transl. and ed. by T. W. Bynum,
Clarendon Press, Oxford, 1972.
Frege, G., Wissenschaftliche Briefwechsel, ed. by G. Gabriel, H. Hermes, F. Kambartel,
C. Thiel, and A. Veraart, Felix Meiner Verlag, Hamburg, 1976. (Referred to as BW.)
Frege, G., Logical Investigations, ed. by P. T. Geach, Basil Blackwell, Oxford, 1977.
Frege, G., Posthumous Writings, transl. by P. Long and R. White, Basil Blackwell, Oxford, 1979.
Frege, G., Philosophical and Mathematical Correspondence, abridged for the English edition by B. McGuinness and transl. by H. Kaal, Basil Blackwell, Oxford, 1980.

Adams, R. M., 'Primitive Thisness and Primitive Identity', The Journal of Philosophy 76
(1979), 5 - 26.
Angelelli, I., Studies on Gottlob Frege and Traditional Philosophy, D. Reidel, Dordrecht,
1967.
Angelelli, I., 'Friends and Opponents of the Substitutivity of Identicals in the History of
Logic', in Schirn (1976), Band II, pp. 141-166.
Aristotle, Analytica Posteriora, in The Works of Aristotle, Vol. I, ed. by W. D. Ross, Oxford University Press, London, 1928.
Bell, D., Frege's Theory of Judgement, Clarendon Press, Oxford, 1979.
Dummett, M., Frege: Philosophy of Language, 2nd ed., Duckworth, London, 1981. (First
published in 1973.)
Dummett, M., The Interpretation of Frege's Philosophy, Duckworth, London, 1981.
Forgie, J. W., 'Frege's Objection to the Ontological Argument', NoUs 6 (1972),251- 265.
Grossmann, R., 'Structures, Functions and Forms', in Schirn (1976), Band II, pp. 11- 32.
Haaparanta, L., Frege's Doctrine of Being (Acta Philosophica Fennica, vol. 39), Helsinki,
1985.
van Heijenoort, J., 'Logic as Calculus and Logic as Language', Synthese 17 (1967),
324-330.
Hintikka, J., '''Is'', Semantical Games, and Semantical Relativity', Journal of
Philosophical Logic 8 (1979), 433-468.
Hintikka, J., 'Kant on Existence, Predication, and the Ontological Argument', Dialectica
3S (1981), 127-146. (Referred to as Hintikka, 1981a; in this volume.)
Hintikka, J., 'Wittgenstein's Semantical Kantianism', in E. Morscher and R. Stranzinger
(eds.), Ethics, Proceedings of the Fifth International Wittgenstein Symposium, HolderPichler-Tempsky, Vienna, 1981, pp. 375-390. (Referred to as Hintikka, 198Ib.)
Hintikka, J., 'Semantical Games, the Alleged Ambiguity of "Is", and Aristotelian
Categories', Synthese 54 (1983), 443 - 468.
Hintikka, J., 'The Varieties of Being in Aristotle', this volume, p. 81 -114.
Hintikka, J., 'The Paradox of Transcendental Knowledge', forthcoming in the Proceedings of the 1981 Cambridge Conference on Transcendental Argumentation.
Hugly, P., 'Ineffability in Frege's Logic', Philosophical Studies 24 (1973), 227 - 244.
lshiguro, H., Leibniz's Philosophy of Logic and Language, Duckworth, London, 1972.
Kahn, C., The Verb 'Be' in Ancient Greek, D. Reidel, Dordrecht, 1973.
Kahn, C., 'On the Theory of the Verb "To Be" " in M. K. Munitz (ed.), Logic and Ontology, New York University Press, New York, 1973, pp. 1 - 20.
Kant, I., Der einzig mogliche Beweisgrund zu einer Demonstration des Daseins Gottes, in
Kant's Gesammelte Schriften, Band II, Vorkritische Schriften II, 1757 - 1777, G.
Reimer, Berlin, 1905, pp. 63 - 163.

ON FREGE'S CONCEPT OF BEING

289

Kant, 1., Kritik der reinen Vernunft, 1781, 2nd ed. 1787, in Kant's Gesammelte Schriften,
Band III, O. Reimer, Berlin, 1904; trans\. by N. Kemp Smith, The Macmillan Press,
London and Basingstoke, 1929.
Kaplan, D., 'How t() Russell a Frege-Church', The Journal of Philosophy 72 (I975),
716-729.
Kauppi, R., 'Substitutivity salva veritate in Leibniz and in Modern Logic', Ratio 10 (1968),
141-149.
Klemke, E. D. (ed.), Essays on Frege, University of Illinois Press, Urbana, Chicago, and
London, 1968.
Kluge, E.-H. W., 'Freges Begriff des Logischeinfachen', in Schirn (I976), Band II, pp.
51-66.
Kneale, W. and Kneale, M., The Development of Logic, Clarendon Press, Oxford, 1962.
Mates, B., 'Identity and Predication in Plato', Phronesis 24 (I979), 211-229.
Schirn, M. (ed.), Studien zu FregelStudies on Frege /- III, Frommann-Holzboog, Stuttgart - Bad Cannstatt, 1976.
Sluga, H. D., 'Frege as a Rationalist', in Schirn (1976), Band I, pp. 27-47.
Sluga, H. D., Gottlob Frege, Routledge and Kegan Paul, London, Boston and Henley,
1980.
Stuhlmann-Laeisz, R., 'Freges Auseinandersetzung mit der Auffassung von "Existenz"
als ein Priidikat der ersten Stufe und Kants Argumentation gegen den ontologischen
Oot~esbeweis', in C. Thiel (ed.), Frege und die moderne Grundlagenforschung, Anton
Hain, Meisenheim am Olan, 1975, pp. 119-133.
Woods, M. J., 'Substance and Essence in Aristotle', Proceedings of the Aristotelian Society75 (I974-75)., 167-180.

Dept. oj Philosophy,
University oj Helsinki,
Unioninkatu 40B,
SF-OOI70 Helsinki 17, Finland

INDEX OF NAMES

Abelard, Peter xii, xiii, 5, 8, 23,123,141,

145-156,158-163,166-176,178-180
Ackrill, J. L. 78,100-103,112
Adams, R. M. 286, 288
Alanen, Lilli 242-246
Albertus Magnus 126, 141, 214, 220
Albritton, Rogers 112
Alexander of Aphrodisia 116, 121, 125,
186, 187
Alexander the Great 97, 272-273
Allan, K. 77
Allison, Henry 266
Alquie, F. 243, 247
Ammonius 118-121
Angelelli, Ignacio 285, 286, 287, 288
anonymus Cantabrigiensis 125
anonym us Moraux 119, 141
anonymus Pragensis 132
anonymus Taran 121
Anselm 249, 250, 252-253, 262
Apelt, O. 103, 112
Aquinas, Thomas xiii, xiv, 80, 178-208,
210,212-216, 218-222, 224
Aristophanes I I
Aristotle x, xi, xiii, xiv, I, 3, 5, 8-9, 13,
17,20-22, 24, 25, 27, 29, 33,49-56,
58-59, 61, 63-90, 92-98, 100-108,
110-117,119-121,123,125,141,142,
145-146, 149-150, 152, 154, 160,
162-163,168,173-178,180,181,190,
194-196,199,201,204-205,213,215,
219,221,222,224-227,229,242,247,
252-254,266,267,274,281,285,286,
288
Arnauld, A. 218, 223, 233-234,
236-238, 246
Asclepius 121, 134
Ashworth, E. J. 140, 141
Augustine 193
Aureoli, Peter xiv, 207
Austin, J. L. 78, 89, 287
A vicenna 188

Bach, E. 77
Bacon, Robert 123, 131, 141
Bacon, Roger 129, 141, 142
Bailey, C. 142
Balme, D. M. 86, 112
Bambrough, R. 25, 74, 113
Barnes, Jonathan 25, 26, 265
Barth, T. A. 216
Beck, A. L. 219,241
Beck, C. H. 75
Beck, L. J. 247
Bekker, Immanuel 116, 141
Bell, David 274, 288
Benardete, Seth 82, 112
Bennett, Jonathan 265
Benveniste, Emile 23, 24, 100, 103, 112
Black, M. 77
Blass, F. 74
Bluck, Richard S. 81, 112
Boehner, P. 142,218,219,222
Boethius, de Dacia 135
Boethius, A. Manlius S. 115, 117-121,
127,132-134,141,142,145,160-161,
176-178
Bonansea, B. M. 216,219,221,222,243,
248
Bonaventura 224
Bonitz, Hermann 77,79,100-101,112
Borgnet, A. 126, 141
Boutroux, Emile 242
Braakhuis, H. A. G. 123,128,141,142,
180
Brentano, F. 79
Bresnan, Joan 109-112
Brito, Radulphus 132
Brody, B. A. 77
Brown, S. F. 142,216,218, 219, 222
Brugmann, K. 23, 45, 47
Brunschwig, J. 79
Buchanan, E. 79
Buridan, John xv, 135, 136, 138-142,
211,218,.220

291

292

INDEX OF NAMES

Burley, Walter 141

Bury, R. G. 142
Bynum, T. W. 288
Cargile, J. 76
Cassin, B. 116, 141
Catan, J. 2\3, 221
Caterus 229, 236
Cathala, M.-R. 80
Chatillon, J. 179
Cherniss, Harold 34-35,45,47,81
Clarke, Francis P. 221
Code, A. 75
Conington, Richard 216
Cornford, Francis M. 44,47,81
Couturat, L. 45, 47
Dal Pra, M. 179
Dancy, Russell M. x, xi, 83, 95, \05, 112,
213,219,266
David (pseudo-David) 119,141-142
Davidson, Donald 76, 266
de Alvernia, Petrus 135
de Faversham, Simon 127
De Finance, J. 216,219
De Libera, A. 141
De Raeymaeker, L. 216, 221
de Rijk, L. M. 123,141,142,172-180
215,221
Debrunner, A. 23
Decker, B. 222
Del Punta, F. 142
Delbruck, B. 23
Descartes xv, 27, 223-225, 228-248,
250
Dexippus 118
Diels, H. 142
Doig, James C. 199,213,219
Donagan, A. 241,247
Dummett, Michael 76-77, 196, 199,
286, 288
Duns Scotus, John 201, 204, 207-213,
216-222,224,226-227,236,243,246,
248
During, I. 25
Ebbesen, Sten xii, xiii, xv, 117-118, 126.
128,131-132,140-143
Edwards, A. 242. 243, 244

Elias (pseudo-Elias) 118-119,141-142

Elizabeth, Princess 242, 246
Engel. S. Morris 266
Ephesius, Michael 120
Etzkorn, G. I. 142
Eustratius 118
Fabro, C. 215,219
Farler, C. Sam 172
Fillmore, C. J. 77
Fleming, N. 74
Fobes, F. H. 74
Foot, P. 75
Forgie, J. W. 288
Forster, E. S. 79
Frankfurt, Harry G. 112
Fredborg, K. M. 128, 142
Frede, Dorothea 197
Frede, Michael ix, x, 75, 78, 112
Frege, Gottlob xiv, xv, 60, 77, 78, 186,
188,192,196, 198, 199,256,259-260,
265-267,269-277,279-289
Furth, Montgomery 96, 112, 287
Gabriel, G. 288
Gal, G. 142, 218, 222
Galantiere, L. 215,220
Galileo 113
Gallop, D. 25-27
Gambatese, A. 142
Geach,PeterT. 45,47,75-77,196-199,
277-278,280,282,287,288
Geiger, L. B. 215,219
Gerth, B. 23
Geyer, B. 23,173, 178, 179
Giles of Rome 133
Gilson, E. 215,219,241,247,266
Goethe, J. W. 45
Gomboez, W. L. 265
Gomez-Lobo, A!fonzo 25, 81, 87-89,
112
Goodwin, W. W. 45,47
Gorgias 115-116,121
Gosling, J. 74
Graeser, Andreas 112
Graham, A. C. 3
Griffin, N. 77
Grossmann. R. 288
Gueroult, M. 245

INDEX OF NAMES
Guthrie, W. K. C. 81
Hager, F. P. 79
Haldane, E. S. 241-247
Hamlyn, D. W. 195,198,199
Happ, H. 213,219
Haring, N. M. 172, 179
Harlfinger, D. 121, 142
Harman, G. 76
Harms, R. T. 77
Heidegger, M. 63
Helias, Petrus 177
Henrich, Dieter 265
Hermann, G. 23
Hermes, H. 288
Herodotus 24, 25
Hertz, M. 179
Hick, John 265
Hilbert, D. 271
Hintikka, Jaakko x, xv, xvi, 23, 26, 78,
82,87,90, 101-102, 106, 112, 113,
181, 197, 199,215,218,265-267,269,
276, 285, 286, 288
Hoeres, Walter 212,217-219
Homer x, 11-12,66-67,85, 115-116,
128, 157, 165,254
Honnefelder, L. 216,217,219,220
Hooker, M. 241,247,248
Howell, Robert 266, 267
Hubien, H. 218,220
Hugly, P. 288
Hume, David 3, 247
Hunt, R. W. 177
Husserl, E. G. 280, 287
lamblichus 118
Irwin, T. 74, 78
Ishiguro, H. 289
Jacobi, K. 172, 174, 179
Jacobs, W. 74,78,79
Jolivet, 1. 172, 177, 179
Jourdain, Philip E. B. :86
Judy, A. G. 142
Kaal, H. 288
Kahn, Charles ix, x, 47, 74, 78, 79, 81,
96, 100, 103, 113, 196, 199, 269, 289
Kambartel, F. 288

293

Kaminski, S. 216,220
Kamiah, W. 75
Kamp, 1. A. W. 75
Kane, W. H. 214,220
Kannisto, Heikki 261-262, 267
Kant, l. xv, xvi, 249-267, 272, 275, 279,
284-285, 288, 289
Kaplan, David 255-256, 283, 289
Kaulbach, F. 288
Kauppi, R. 289
Keil, H. 179
Kelley, F. E. 142
Kemp, Smith 267, 289
Kennan, E. L. 75
Kenny, A. 179,180,199,214,220,241,
242,247
Kilwardby, Robert 124-128, 130-\31,
142
Kirwan, Christopher 74, 75, 80, 84, II?
Klowski, J. 24
Kluge, Eike-Henner W. 282, 288, 289
Kneale, M. 289
Kneale, W. 289
Kneepkens, C. H. 141, 180
Knuuttila, Simo 113,214,217-220,241,
266
Koehler, K. F. 1\3
Koyre, Alexandre 241, 243, 247
Krakow, Bjag 138
Kranz, W. 142
Krapiec, M. A. 216,220
Kretzmann, N. 170-172,179,180,214,
220-222
Kripke, S. 76
Kuhner, R. 23,45,47, 74
Kurdzialec, M. 216, 220
Laeisz, R. 267, 289
Land, J. P. N. 246, 247
Langston, D. C. 220
Larkin, T. 79,80
Laupp, H. 79
Lee, E. N. 77
Leibniz, G. W. 36, 70, 106, 112,
212-213, 218-221, 254, 261-263,
276, 280, 286, 289
LePore, Ernest 285
Leroux, Ernest 247

294

INDEX OF NAMES

Lewis, Frank 76
Lewry, O. 122-124,127-128,142
Lloyd, G. E. R. 242, 247
Lockwood, M. 76, 78
Loemker, L. E. 218, 220
Lombard, Peter 189
Long, P. 286, 288
Louis, R. 179
Loux, Michael J. 113,213,218,220
Lovejoy, A. O. 261
Lucretius 117, 120, 142
Luschei, E. C. 45, 47
Lyons, John 23, 24, 77
Lyttkens, H. 215,220
Magentinus, Leo 116
Mahoney, E. P. 214,220
Maier, Heinrich 79,81,83,100,107,113
Maieru, A. 178, 179
Malcolm, John 196, 199, 242, 245
Malcolm, Norman 241, 247
Mandonnet, P. 197,198,222
Manheim, R. 77
Mansion, S. 79
Maritain, J. 215
Marshal, David J. 200
Marshall, W. 286
Mates, Benson x, 23, 78, 82, 113, 220,
269,289
Matthen, Mohan 25, 27
Matthew, M. 217,221
Mazzarella, P. 143
McCabe, Herbert 196, 199
McGill, Arthur C. 265
McGuinness, B. 288
McInerny, R. M. 214,215, 220
Mei, Tsu-Lin 77
Meillet, A. 23
Meiser, C. 117,141,178
Melissus 11,13-14,17-18,25
Mersenne, M. 243, 246
Mesland, D. 246
Mill, John Stuart 4, 23, 60, 269
Miller, Barry 195, 196, 199
Minio-Paluello, L. 178, 179
Mohr, J . C. B. 265
Montague, Richard 75, 255
Moody, E. A. 178,179
Moos, M. F. 198

Moraux, Paul 141

Moravcsik, J. M. E. 78,81,97,113
Morscher, E. 288
Mourelatos, A. P. D. 26, 78
Munitz, Milton K. 23. 76, 113, 289
Neal, Gordon C. 81, 112
Niiniluoto, IIkka 218,220
Nuchelmans, G. 172, 176, 177, 179
Ockham, W. 100,
113, 135-136,
139-140,142,162,211,217-219,221,
222, 224, 227, 229, 240, 243, 244, 247
O'Donnell, J. R. 179
Oesterle, Jean T. 196, 198
Osterr, M. 123
Owen, G. E. L. x, xi, 22, 25, 49-50, 59,
74,78-81,86,101,113,190,215,221
Owens, Joseph 76, 213, 221
Parmenides 1,4,7, 12, 14-17,21-22,
25-27,29-33,38,44, 82, 116, 142
Patzig, G. 198, 286, 288
Peano, G. 281,286-287
Pelletier, F. J. 75
Pera, C. 222
Perlmutter, D. M. 77
Petrus, de Hibernia 134
Phelan, G. 215,220
Philoponus 116-118, 120
Pinborg, Jan 133, 140, 142, 178-180,
214, 220, 221
Pirotta, A. 222
Plantinga, Alvin 265
Plato ix, x, I, 4-5, 12, 14-23, 25-30,
32-36,38-43,44-47,50-54,60-63,
74-78,81,101,112,113,116,120,
134,202,213,215,221,273,282,289
Platts, M. 75
Porphyrius 118-119
Poser, H. 218,221
Prauss, Gerold 267
Prester, John 51,61
Priscian 145, 149, 179
Protagoras 11- 15, 25, 45-46
pseudo-Scotus 211
Punjer 275
Quine, W. V. 24,44,47,75,200

INDEX OF NAMES
Rashdall, H. 141
Reimer, G. 289
Reina, M. E. 141
Remes, Unto 266
Richard of Conington 219
Roberts, L. N. 140, 142
Rorty, R. 77,78
Ross, G. T. 247
Ross, J. F. 213,215,221,241-246
Ross, W. D. 25,44,47,75-77,79-81,
85,101,113,178,219,242,288
Russell, Bertrand 4, 23, 33-34, 45, 47,
106, 289
Ryan, J. K. 216,219,221,222,243,248
Ryle, Gilbert 81, 106, 113
Saarinen, Esa 102, 113, 218, 220, 256,
266
Schaffer, Jerome 266
Schepers, H. 218, 221
Schim, M. 76, 286, 288, 289
Schmaus, M. 216, 221
Schmidt, C. C. E. 45, 264
Schmidt, F. 47
Schmidt, R. W. 221
Schwyzer, E. 23,45, 47
Scott, T. K. 141
Sextus Empiricus 116
Shircel, C. L. 216, 222
Shorey, P. 28, 221
Siger 214, 220
Simplicius 118-120
Sluga, H. D. 289
Smyth, H. W. 45,47
Soames, S. 77
Socrates xiii, 7, 9,10,14,16,27,29-33,
38,40,43-46,50-56,58-61,64-65,
67, 70-72, 78-80, 83-84, 86, 94, 97,
106,123,125,156-157,159-166,177,
182-183, 188, 196, 197,200,202
Sommers, F. 241,247
Sorabji, Richard 25
Spiazzi, R. M. 80, 179,222
Spinoza, B. 246, 247
Stead, C. 80
Stenius, Erik 241
Stephanus 118
Stranzinger, R. 288
Strawson, P. F. 196, 199

295

Stuhlmann-Laeisz, Rainer 267,

275-276, 288
Suarez,
Francisco 212,
217-219,
222-224, 230, 240, 242-247
Taran, L. 26, 141
Taylor, Richard 265
Thiel, C. 267,288,289
Thomas, R. 179
Thomason, R. H. 75
Thompson, M. 78,79
Thorp, J. W. 79,80
Thucydides 24
Trendelenburg, Adolf 103, 113
Trentman, J. A. 218,222
Tugendhat, Ernst 23
Tweedale, M. M. 172-179
Ulger, V. 178
van Heijenoort, Jean 277-278, 288
Veatch, Henry 196, 197, 199
Veraart, A. 288
Verardo, R. A. 222
Veres, Torno 197, 199
Verhaar, John W. M. 3,22,23, 199
V1astos, G. 25, 31, 34-35, 44-47,
74-75, 81
Vollert, C. S. J. 242-244, 246
von Fritz, K. 79, 80
von Wright, Georg Henrik 241
Wagner, Hans 263-265
Waltz, T. 79
Walzer, Richard 23
Wedberg, Anders 44, 47
Wedin, M. V. 79
Weidemann, Hermann xiii, xiv, 81,114,
196, 198, 200, 215, 222
Weinberg, Julius R. 106, 114, 241-244,
247
Wells, N. J. 241,242,247
Westerink, L. G. 142
White, F. C. 74
White, R. 286, 288
Whorf, Benjamin Lee 77
Wieland, Georg 195
Wiggins, D. 76
William of Sherwood 141, 179

296

INDEX OF NAMES

Williams, B. 247
Williams, C. J. F. 23,78,241,245
Wilson, M. D. 241,242,247,248
Wittgenstein, L. 45, 47, 247, 288
Wolff, Christian 23
Wolff, Robert Paul 266
Wolter, A. B. 216, 217, 218, 222, 243,
248

Wolterstorff, N. 74
Woods, M. J. 94. 114, 289
Zdybicka, Z. J. 216,220
Zeno 25
Zimmermann, Albert 178, 180, 198,200

INDEX OF SUBJECTS

Thomas Aquinas 181-195,201-207;

William Ockham 135-136 (see also
copula, existence, einai, esse, 'is')
body 223, 226, 228, 232-241

abstract entities 39-40

abstract terms 39, 43, 119, 121, 182
accident, accidental 19, 21, 58, 64,
68-73, 83, 93-95, 107, 124, 150,
152-153,
157,
181-184,
188,
191-194, 197, 204, 2i5, 225-228,
231-232,264
actuality 5, 17, 68, 95, 128, 133-134,
182-187, 190-191, 195, 201-204,
208-213,217,255
actuality sense of 'is', see 'is'
adequate knowledge 237
adjectives 7, 33, 39, 41, 43, 54, 63, 123
ambiguity xi, 4, 29, 34, 38, 40, 78, 81-82.
85,148-149,155,181,186-189,191,
195,240, 269-270
ampliation 138
anaphora 256
antihaecceitist 283
aorist 7
apodeictic syllogisms 88
ars sermocinalis 167
aspect of the verb 4
attribute 32, 75, 208, 210, 224, 230-231,
238,241

calculus ratiocinator 276

categorematic words 148, 156
categorial content 148, 152-153
categorial proposition 120, 166
categories xi, xiii, 5, 14, 17, 21, 68,
71-73, 78-80, 85, 90-92, 99,
Ill,
118,
147-150,
101-108,
181-182,186-189,192,208,225,262
characteristics (Merkmale) 271- 272, 278,
282
chimera xiii, 115-116, 118-123, 125,
128-129,131-140,157-158,160,216
clear and distinct concept 230, 232, 234,
236,239
complete entities 232, 236-239
complete expressions 159,272-274
compossible 211-212
concept 117, 135, 148, 159-160, 172,
202-203, 225-226, 281-284
concept-words 273, 282
conceptual beings 224
conceptual existence 117, 227
consigni/icatio 146, 149-154, 161, 190
convertible notions 208, 224
copula xii, 1-12, 17-21,23,26,74,94,
101, 107, 122-125, 139, 147-148,
152-158, 163-165, 168-169, 171,
184-185,189,191,194,195,258-259,
270, 272, 277
count noun 111

barbara 87-88
Bedeutung 272-273, 277, 279-280, 282
Begriffe erster und zweiter Stufe 270, 285
being, philosophical concept of: Abelard
149-152, 155-162; Aristotle 17,
49-50, 55-59, 64-73, 81-96,
100-111,200-201; Boethius of Dacia
132-134; Descartes 224-234; Frege
269-285; John Buridan 138-139;
John Duns Scot us 207-213; Kant
249-265; Leibniz 212-213; Melissus
13-14,
17-18;
modists
131;
Parmenides 14-18, 25-27; Plato
16-21, 34-39, 42-44, 51-55; Protagoras 12-14; Robert Kilwardby
124-131; Roger Bacon 129-131;

de re 211-222
deep structure 155, 160, 163
definition 18-20, 76, 84, 102, 120, 129,
135, 139, 193,204,282
dictio 127, 158,205
dictum 159-162

297

298

INDEX OF SUBJECTS

disjunctive attributes 208, 210

disjunctive concept 207
distant vs indistant composition 137
distinction of reason 230-231, 236
divine: attributes 236; intellect 209-210,
217; omnipotence 204, 209, 217, 243,
254,257; omniscience 212; possibilities
203-204,209; will 209, 217
dualism: Cartesian 223-224, 232-241;
Platonic 20

einai: hap/os 49,59,67, 89; kath' hauto

56, 64-68, 70-73
enuntiatio 159
equivocity 39, 124, 130, 150, 164-165,
168,186-188,204-205; pros hen xi,
xii, 3, 205
esse, ens, est 12:'-124, 128-133, 135,
138, 145-152, 155-158, 162-166,
172-178, 182-195, 198, 204-208,
212-217
essence, essential 18-21,55-59,64,67,
71-73, 83, 85-86,90-95, 126-128,
132, 155-156, 182-184, 192, 194,
201-204, 206, 210-211, 224-225,
227,231-235,271,282
essentialism 50, 55, 57, 61, 63, 94
existence 1, 3-4, 9-17, 21, 61, 65-67,
71,75-78,81-82,85-92,96,99,101,
105,107,124,127-128,131,156-157,
181,183-188,192,202,206,210-211,
227,238,257,259,269-272,275-276,
280, 284; as a predicate 210, 249,
252-254, 256, 258-259, 263-264,
276, 278-279; degrees of 54, 131, 206,
250; 'is' of existence, see 'is'
existential: generalization 253, 259; import xii-xiii, 9, 79, 91-93, 152,
155-158,165,253-254,271; quatifier
10,59,251,259-260,269-270
extensional 154, 167, 204
fallacy: of consequent 125, 127; of
equivocation 17, 127; of the magnifying glass 64-65, 73; secundum quid et
simpliciter xiii, 125-126, 136, 138
fictitious entities 135; see chimera
first-order concepts 271-272, 275, 277
first-order logic 92,96-97,260,269-270

focal meaning xiv, 3,107-108,190

form 95, 105, 108, 110-111, 167, 197,
213, 224, 229, 238
forman distinction 226-227, 236
Forms (Plato) 4,19-21,32,41,52
Frege - Russell distinction ix-xi, xv,
81-87, 89-90, 92-93, 96-98, 101,
107- 108, 110, 112, 186, 188, 249,
257-260, 265, 269, 272, 285
function-names
(Funktionsnamen)
272-274
game-theoretical semantics 102- 103,
252,269
generic implication 81-82, 249, 270
genus 14, 71, 73, 90-92, 99, 101, 122,
129, 182, 238
great chain of being 202
haecceitas 212
Hermann's rule 50
hierarchy of being 206
higher order predicate 265
homonymy 82, 108; see equivocation
idea 29, 31,42; see Forms (Plato)
identifiability 208, 210-212
identity 29, 34, 63-64, 76, 78, 81-83,
85-86, 88-89, 92-96, 99, 101-102,
104,107, 167, 181, 188,210-212,269,
272, 279-280
identity theory of predication 167, 188
impossible
beings
120,
133-134,
136-138, 140, 216-217
incomplete expressions 158-159,272
incomplete entities 237
indistant composition 137
individual 92, 94, 105, 107, 110-111,
128, 210-213, 226, 229, 243, 255, 282
individuation 94-95, 110, 202
inesse principle 212
infinite words 117,119,127,129-130
inherence theory of predication 167, 188,
197
instantia naturae 209
instantiation 27, 96-97, 99, 102, 104
intellectus 146-147,190,192,214
intensional 54, 56- 57, 167, 209- 211

INDEX OF SUBJECTS
'is': accidental 56, 64, 66, 68-71; actuality sense of 181-187, 191; copulative
(see copula), ellipsis hypothesis of
(86-92); of existence ix, 1-4,6,9-17,
27,53,59,61,65-67,71,73,81-82,
84-91,93,96,99,101,105-107, 124,
183-188,249,260,265,269; of generic
implication ix, 81-82, 249, 270; of
identity 35-36,38-39,63-64,81-84,
88-89,93-94,99,101,183,186,188,
249, 258, 269, 281, 284; of location
3-4,7-11, 15, 17; of predication ix,
1-10, 17, 20-22, 26-27, 35-36,
38-39,53,59,61,63-64,81-84,86,
89,93-94,99,124,152,183-188,195,
249, 265, 270, 281 (see also copula);
there-is sense of 182-187; veridical
3-4, 8-9, 12-13, 16-17, 20-22,
26- 27, 184. (See also being, einai, esse,
existence.)

lingua characteristica 276

linguistic relativism 3 -4, 21
logic of tense 169
logical: functors 148; possibilities 204,
209, 212; predicates 264; truth 40;
types 104-107
logically simple 282-284
many-sorted logic 97-98, 110
many-sorted quantification theory 97
mass terms III
matter 94-95, lOS, 108, 110-111, 202,
213,226-227
mind 223, 232-241
modal: categories 262, 265; concepts 94,
203-204; distinction 228, 230-231;
logic xiv, 211, 256; operator 41,
249-250; syllogisms 88; theoTY 167
modes: of being 131, 205; of intellection
205; of signifying 206
modists xiii, 131-132, 135
natural kinds 117, 201, 203
necessary beings 132-133, 250
necessity 88, 147, 161,234,249-254,262
nominal definition 120, 129, 135, 139
nominalists xiv
numerical identity 88, 102, 224, 229

299

(to) on 6,8-9, 11-14, 18-20,27-28,68,

82,91,101
ontological argument xv, 249-253, 256,
261-263,271
ontological relativity 76
opinable 116, 119-125, 127-131, 136,
138
oratio 158
ousia 18, 71, 91, 93, 107, III
Pauline predication 34, 39
possibility 204, 208-213, 234, 243,
252-258, 261-264
possible beings 132-139, 210-213, 234,
255, 261-263, 276
possible world semantics xiv, 250, 276
potentiality 5, 17,68,94-95,210,255
predication 1-10, 12, 17,20-22,27,29,
35-39,53,59,61,63-64,81-86,89,
92-95, 9~ 105, III, 124, 132-138,
162-171,
181-189,
249,
259,
269-272, 275, 279-284
principle of individuation 202
principle of plenitude 203,261-262
privation 17, 127, 182, 195
proper name 60, 76, 259, 272-274,
282-283
propositon: two-part analysis vs threepart anaiysis 163-171
quality 33,147,149-150,153,201,205,
207,226,229-231,240
quantification theory 92, 96, 104, 110
quantifier phrase 99, 110-111
quantifying in 255
quantity 33, 106, 149, 205, 207
question: types 103; words 104; and
answers 101-102
quidditative: beings 132, 206, 210; definition 129
quiddity 203, 213

ratio: communis 205-207; propria

205-206
real distinction 206, 226- 232, 236,
240-241
real predicate 259, 264-265, 275, 278
reality 6, 19-20,22
reference, see sense-reference distinction

300

INDEX OF SUBJECTS

relations 33, 106, 162-163, 209

relative pronouns 98, 104
res signijicata 206
Russell's Antinomy 39

suppositio 134-137,162,167,188,197
surface structure 155
syllogism 87-88, 90-92
syncategorematic words 148. 156, 162

scibile 119, 132

scientific knowledge 87-91, 120, 139
second-order concepts 270-271,275, 278
secundum adiacens xii-xiii, 122-127,
164-165
self-predication 29
sense-reference distinction 60, 76, 195,
272-274, 277-280
sermocinalis scientia 168
signification
135-139,
147-155,
158-167, 184, 187-191, 204-206
simplification 57-58, 65-67, 70
Sinn 272-274, 277-280
soul, see mind
species 73,86, 119, 122, 125, 128-129,
201-204,238,281
square of opposition 78
stative-durative 4
stative-mutative 20
substance, substantial 71-72, 79, 85,
93-94, 102, 104- 105, 107- 108,
110-111,
120,
127-128,
147,
149-152, 157, 181-187, 191-193,
201-202,204-207, 224-231,236-240
substitutivity principle 260
sUbsumption 271

temporal: copula 139; co-signification

150-154; meaning 149; predicates 106
terminist logic 204
tertium adiacens xii, xiii, 122-124,
164-165
there-is sense of 'is', see 'is'
Third Man Argument 29-34, 43
(to) Ii en einai 20, 56-57, 85, 95
transcendental philosophy 261
transcendentals 208, 231
transformational grammar 62, 103
truth 8-22, 159-163, 184, 186, 189,
192-196
universals 33-35, 41,105,122,127,135,
225-226, 267
univocation 17, 131,204-207,212,230,
278
veridical sense of 'is', see 'is'
well-definedness 251
wh-words xi, 98-99, 102
what-questions 98, 104- 105