Professional Documents
Culture Documents
(Synthese Historical Library 28) Charles H. Kahn (Auth.), Simo Knuuttila, Jaakko Hintikka (Eds.) - The Logic of Being - Historical Studies-Springer Netherlands (1986) PDF
(Synthese Historical Library 28) Charles H. Kahn (Auth.), Simo Knuuttila, Jaakko Hintikka (Eds.) - The Logic of Being - Historical Studies-Springer Netherlands (1986) PDF
(Synthese Historical Library 28) Charles H. Kahn (Auth.), Simo Knuuttila, Jaakko Hintikka (Eds.) - The Logic of Being - Historical Studies-Springer Netherlands (1986) PDF
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
ACKNOWLEDGEMENTS vii
INTRODUCTION ix
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
IX
logic of being'. As shown in the papers of Klaus Jacobi and Sten Eb-
besen, 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 ter-
tium 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 posi-
tions. 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 discus-
sion 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 ge-
nuine 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 Eb-
besen, 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
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 discus-
sions) 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 argu-
ment for the mind-body distinction is based on a specific interpretation
of this traditional doctrine, and that the early discussions of the argu-
ment 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 Hin-
tikka 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 'ex-
istence is not a predicate'. Instead, he finds the crucial flaw in the on-
tological argument elsewhere, viz. in its tacit dependence on our being
able to identify God (the being of whose existence is to be proved) be-
tween the different possible worlds presupposed in the argument. This
provides a new perspective on historical as well as contemporary discus-
sions 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 in-
gredients in Frege's treatment of being is his idea that existence is a
second-level concept (property of a concept). Haaparanta sees the foun-
dation of this assumption in Frege's ideas about the identification (in-
dividuality) and existence of individuals (objects), incorporated in
Frege's treatment of the senses by means of which we can grasp an in-
dividual 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
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 pre-
theoretical 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.
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 in-
fected one another, since the linguists have borrowed their notions of ex-
istence 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
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 pre-
judice, 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 regard-
ed 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 mean-
ing" 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, loca-
tion 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 l'-C",'flL: tnr rune·
tipn." or 'to ht" as ('opld3 depend lIPon a grarnmatical rule t'O!' thl...' i:-" ,~-:atinn ,)\" the
sentence. and it would b::.- merely a ·:elDcid'.:r~'(' ~r one found an\'thi:,,(~ Icscnl~)l;-:,' 11 in (:
ir-tr!!!uagc without this rule. S - - .
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 ver-
balizer," 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 func-
tion (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 func-
tion in a sentence of the form X is Y.
6 CHARLES H. KAHN
(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 cor-
responding 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 con-
notation 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 - be-
tween 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 "ex-
istential" 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 horse-
nourishing 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 secon-
dary function, which I would analyze as introducing a subject for further
10 CHARLES H. KAHN
where Fx stands for the relative clause and an initial esti functions as ex-
istential 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 for-
mula 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 con-
struction. 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 explana-
tion 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 "existen-
tial" connotations in virtue of its locative association and its rhetorical
function of introducing a subject for predication. Given these connota-
tions, it is natural that the existential function of (2) becomes one of the
values esti can have when used alone, without nominal br locative com-
plements. (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:
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 mid-
dle 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 Indo-
European. 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
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 ques-
tion 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).
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 pre-
philosophic 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 presen-
tation 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 pre-
sent, 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 starting-
point. 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 ra-
tional cognition (noesal) to travel on (fr. 2.2). Where is this inquiry sup-
posed 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 con-
text. With these clues Parmenides makes quite clear that what the god-
dess 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 ar-
ticulated, 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 inter-
preters, 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 disad-
vantage 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 objec-
tively 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 assump-
tion 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 substrate-
privation-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 re-
quired 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 per-
sistently 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 on-
tological vocabulary in the preliminary and mature statements of his
theory of Forms.
(1) In the so-called Socratic dialogues, the first philosophically rele-
vant 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 in-
to 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 no-
tion 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 preci-
sion 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, bet-
ween 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) doc-
trine in which, for the first time, Plato provides his specimen Form the
Beautiful with a definite ontological status, based upon the Eleatic op-
position 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 con-
texts, 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 predomi-
nant 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 defini-
tional 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 con-
stituted not by a fusion of copula and existence but by the union of
timeless-invariant Being (in contrast with Becoming) and cognitively-
reliable, veridical Being (in contrast with Appearance), both of them ex-
pressed or expressible in copula predications, but most rigorously dis-
tilled 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 in-
teresting ways. First, by accepting an intermediate "mind" reality be-
tween Being and Not-Being as object for the cognitive state of opinion
(doxa) between knowledge and ignorance, Plato has provided an on-
tological 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 nega-
tion of Being as what is not in any way (to medame on); for this is in-
describable 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 pro-
legomena to Greek ontology and not as a history of the subject. In con-
clusion, 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 ex-
istence 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 confu-
sion 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
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
3 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 ap-
peals 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 fac-
tual 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 epistemolo-
gy 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 Ob-
ject, 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 Rund-
schau 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.
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" .
20 'Be', p. 332, Note 2.
21 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, in-
tentionally), 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 en-
campment" (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', Ar-
chiv 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).
THE VERB 'TO BE' AND THE CONCEPT OF BEING 25
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 argu-
ment 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 com-
plex (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 correspon-
ding (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 cor-
responds to nothing in the Greek data. Also, our difference in "logical form" between pro-
positional (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 ade-
Quate and complete grasp of to on at 486E 2.
Dept. of Philosophy,
University of Pennsylvania,
Philadelphia, PA 19104, U.S.A.
BENSON MATES
Among the Platonic statements that have most agitated his commen-
tators, 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 category-
mistake 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 "self-
predications" 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
"Only a part."
"Then the forms themselves would consist of parts, 0 Socrates, and the things par-
ticipating 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 ap-
proached 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 correspon-
ding 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 ex-
istence apart from the things of which they are properties. There are in-
deed 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 con-
cluded 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 in-
dividual 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 quan-
tity, 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 difficul-
ty, 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 commen-
tators have problems with his assertions that largeness is large, beauty is
34 BENSON MA TE~
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 conjunc-
tion. 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 con-
vey 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 begin-
ning 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 cer-
tain 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 in-
ferences. 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 dis-
juncts 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 dis-
juncts 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 con-
sidered 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
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 beau-
ty (or, let us say, Beauty) is an abstract entity, eternal, changeless, ex-
isting 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 at-
tributes 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
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 eter-
nal" 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 ac-
count would at least replace (i) - (vii) above by corresponding principles
for Plato's Greek, and difficult problems of word order and the place-
ment of the article would have to be dealt with. Still further complica-
tions will result from Plato's use of the abstract noun and other expres-
sions as apparently synonymous with the corresponding adjective-plus-
article. 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 doc-
trine 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
NOTES
1 A Swedish translation of an earlier version of this paper is included in a privately publish-
ed 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 con-
sidered. 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 cor-
responding 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 for-
mulating 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
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". Fur-
ther, 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 pseudo-
problem 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 con-
tradictions 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
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 ad-
jectival 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.
BIBLIOGRAPHY
Dept. of Philosophy,
University of California, Berkeley,
Berkeley, CA 94720, U. S.A.
RUSSELL M. DANCY
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
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 First-
year 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 pronun-
ciation 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 ex-
ists'. 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 ex-
ists', 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) Socrates is pale.
(3) 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 con-
demn too many philosophers too quickly to the asylums - but because
that replacement severs the contrast from the argument for it. What con-
demns 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
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 re-
quirement 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 sup-
plied that will answer the question 'a good what?'. And, although in a
specific conversational context it may be clear that the noun to be sup-
plied 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 ex-
istence 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 conversa-
tion, 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 pro-
nouncing (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 pro-
vided 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 Aristot-
le, the democratic attitude is a problem for a different reason.
ARISTOTLE AND EXISTENCE 55
'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
ou
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.)
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
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
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 cor-
responding 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 con-
nection 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 perdi-
tion, 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 in-
stance, a predicative 'is', and only becomes existential by cancellation of
the predicate.
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 av-
t?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 permit-
ted 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 Aristotle·thinks '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 ap-
parent 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 un-
wanted conclusion just as easily.)
68 RUSSELL M. DANCY
e.g., we say the just is cultivated, the man [is] cultivated, and
the cultivated [is a] man.
We may write:
(21) The just [one] is [a] cultivated [one].
(22) The man is [a] cultivated [one].
(23) The cultivated [one] is [a] man.
The English is pretty awful. The bracketed material has nothing to cor-
respond 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.
ARISTOTLE AND EXISTENCE 69
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 posi-
tion. 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) Socrates is a man.
(I) 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 truth-
conditions 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
ARISTOTLE AND EXISTENCE 71
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 ac-
cidental '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 Of
n
71'Otfiv ?raUXftv, Ta Of ?rov, Ta Of 71'OTf, IxauTVJ TOVTWV TO dVaL TauTo CTf/P.c<LVft.)
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 essen-
tial 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
[ouuLav), 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.
Here Aristotle unhesitatingly moves from the point that "all premisses
(1I'QoTauHs) . ..signify either what [it] is or how-big or what-like or one
of the other categories" (b25 - 27), where the sentences under considera-
tion are ones like our old friends (3) and (2) and
(28) Socrates is a cubit tall
72 RUSSELL M. DANCY
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 con-
taining '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 transfer-
ring 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 dif-
ferentiae 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), Appen-
dix 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 arti-
cle 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 argu-
ment 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" adop-
ting 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 vir-
tue 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 (Pon-
tifical 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 transla-
tion 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 yet·1 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 Univer-
sity 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, Syntac-
tic 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.), Univer-
sals 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', Foun-
dations 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 predica-
tion) 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, Darm-
stadt, 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, Greif-
bares, 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
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 gram-
mar 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 connec-
tion (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
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, Dor-
drecht, 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 relative-
to 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 Man-
sion, 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). Sup-
posing 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
60 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. Con-
trast 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).
Dept. of Philosophy,
Florida State University,
Tallahassee, FL 32306-1054, U.S.A.
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 Kir-
wan's):
fi o~ TO OV )(al TO EV mUTOV )(al p.ia q,llCm,:r~ (noAoullftV aAA~AOtS wune I aex~ )(W Cit-
nov 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 in-
dicated 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.
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 essen-
tially 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 species-
characteristic 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' exemplify-
ing 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
I mean if one is or is not simpliciter and not if [one isl white or not (Post. An. B 89 b 33).
TO 0' fifCTTLV ij 1'1, Q'JI"AWi Ai-yW, ahA' oux d AWXOS ij I'~.
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 ellipsis-
hypothesis. 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 Ai~fWi
XCX! I't){eOV Otcxq,ieHV TO flvcx; TL Toii flvCXI,.xcx! TO 1'1, dvcx! n Toii 1'1, flvCXI.
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
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 think-
ing seems to be confirmed by Met. H 6, 1045 a 34 - b 8.
A fourth observation may likewise be in order. Philosophers', dif-
ficulties 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 bet-
ter 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 cor-
related - at least roughly - with an ontological distinction between dif-
ferent 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 other-
wise strange practice of Aristotle's in his syllogistic theory of scientific
explanation, where he treats the being (existence) of what looks like in-
dividuals 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 contem-
porary post-Fregean philosophers.
THE VARIETIES OF BEING IN ARISTOTLE 93
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 com-
parison 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 Aristot-
le'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 distinc-
tion 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
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 }
Y + wh-word + Z
some
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 pro-
nouns 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 rele-
vant subdomain consists of persons, if "where", of locations in space,
if "when", of moments (and/or periods) of time, etc. Further subdo-
mains 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-and-
ready correspondence certainly obtains.
The main discrepancy here is the fact that "what" covers several dif-
ferent subdomains. By asking, "what is X?", we can mean at least three
different things, to wit:
(i) Which partic4lar entity is X?
(ii) What kind of entity is X?
(iii) What does X consist of?
(What is it made of?)
as illustrated by the following sample questions:
(i) What is Sirius?
(ii) What is a gnu?
(iii) What is cordite?
Similar things can be said of Aristotle's Greek word esti: This ambigui-
ty 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 lex-
ical 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 deal-
ing with, when using quantifiers in English, i.e., the largest classes of en-
tities 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 han-
dle serves to betray the relevant sort (subdomain).
Furthermore, since each instantiation step (witness (3» introduces an
occurrence of "is", these correlated classifications are likewise cor-
related 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) Certain wh-words (and phrases with wh-words).
(ii) Different kinds of simple predicates.
(iii) The largest classes of entities we have to recognize in the logic
of our language as domains of quantification.
(iv) Certain semantically different uses of "is". (In them we of
course cannot distinguish from each other the ises of identity,
existence, and predication.)
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
6. ARISTOTLE RECONSTRUCTED
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 ques-
tions. The correlation depends crucially on an analogy between relative
pronouns and question words, and this analogy (or near identity) can ob-
viously 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 dif-
ferent 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 cor-
nerstone of Aristotelian categories. It seems to me that their thesis re-
mains unproven. Admittedly, the importance of the dialectical question-
ing games practiced in the Academy for Aristotle can scarcely be exag-
gerated. However, there is little evidence in the Topics or elsewhere that
the theory of categories was developed for (or from) such games.
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 distinc-
tion (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 in-
dividuals 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 af-
fections, 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 emphatical-
ly 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 dif-
ferent categories as not being homonymous but instances of focal mean-
ing (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 on-
tology. (Concerning Aristotle's attempted Aufhebung of category
distinctions along these lines, see Owen (1960) and (1965).)
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 ex-
pounded 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:
NP
QUANT ...............NP
I
some
every
each
THE VARIETIES OF BEING IN ARISTOTLE 109
This is too simple to be realistic, however. Indeed, Joan Bresnan has pro-
posed the following more refined analysis:
NP
/~
/'"
QP PP
Det Q P
/"'"NP
/~ I I
some NUM of
every I
e:~h r;~
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 ingre-
dients, 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 reconstruc-
tion of Aristotelian categories presented above can only be an approxi-
mation 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 ap-
proximative character are very closely related to the reasons why the sim-
ple picture of categories so far adumbrated did not satisfy Aristotle,
either.
Thus we are beginning to see what light do these systematic observa-
tions 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 treat-
ment 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 in-
dividuality 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 ques-
tionable 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 in-
stantiation 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 mat-
ter 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
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.), Man-
chester 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', Stan-
ford 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.), Aristo-
tle: 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 Com-
mentary, 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
Dept. of Philosophy,
Florida State University,
Tallahassee, FL 32306- /054, U.S. 4.
THE CHIMERA'S DIARY
Edited by
STEN EBBESEN
[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 cen-
taur ({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.)
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.
('PQoP€iat?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 (n€QL 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.
4.50 B.C.
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 'skindap-
sos' 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 con-
fused 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 con-
cept 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 Por-
phyry 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
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.)
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 adjec-
tival one; this, then, must be a case of equivocal ion - or else the said dislinclion is void.]
[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, Kilward-
by 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 pro-
positions '(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 sub-
jective 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 Alex-
ander 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
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 de-
struction 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-and-
absolutely. But when the opinable is taken in its proper sense, no fallacy of consequent oc-
curs. 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-and-
absolutely, 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-and-
absolutely. But he introduces the paralogism for the sake of its vicious in-some-respect-
and-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"). Ac-
cordingly, 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.]
esse ut nunc), i.e. a way of being which consists in entering into a rela-
tionship. 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 pro-
position, and so it will be demolished below when I shall talk about propositions.l
[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 be-
ing has a finite meaning and may be infinitized so as to leave unspecified
being, whereas unspecified being has no finite meaning and cannot be in-
finitized (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
as well as of things which are as of things which are not - the very pro-
perty 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 revolu-
tionary 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
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 inter-
course 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 ac-
tualized, 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 dicen-
dum, 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 under-
stand 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 in-
stance 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).)
8. AD 1340
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 [Ex-
positio super libros Elenchorum, p. 264] accepts the traditional assign-
ment 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 re-
quiritur 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 ter-
minus 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 sup-
positionem impossibile est chimaeram esse, similiter impossibile est chimaeram posse esse,
sim!1iter impossibile est chimaeram fuisse, similiter impossibile est chimaram fieri. Con-
similem 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 in-
tellegibile' 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 presup-
posed 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 im-
possible 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 presupposi-
tion 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 self-
evident. 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 presup-
position 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 sup-
pone 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 in-
stance, 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
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 ac-
cepted 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 ex-
pressions, 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 proposi-
tion 'the chimera is nameable' is true or that 'a chimera is an animal com-
posed 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 con-
cept. 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 occa-
sions 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 assump-
tion 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 ex-
tramental referents.
140 STEN EBBESEN
[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 (unpublish-
ed 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, 1955-
LM = 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.
THE CHIMERA'S DIARY 141
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 Interpreta-
tione (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.
Boethius, Anicius Manlius Severinus: 1877, Int. ed. r = Commentarii in librum
.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 mit-
telalterliche 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. Bonaven-
ture University, St. Bonaventure, N.Y.
Ockham, Guillelmus de: 1977 -1979, Scriptum in librum primum Sen ten tiarum, Or-
dinatio, 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 Vor-
sokratiker, 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.
1. DISCUSSIONS
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 in-
dispensable 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 under-
standing 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 discus-
sion 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 in-
volved 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 con-
ception to be a part of the understanding of the full statement. 11
According to the thesis thus proposed, the copula 'est' is merely a con-
necting 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
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 seman-
tic 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 ad-
jectives, 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 counter-
evidence, 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 in-
deed 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 am-
biguity. 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
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 ac-
cidents 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 existen-
tial 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 com-
mentators 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 pro-
position 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 in-
dividual theses which Aristotle advocates. The correctness of the content
of Aristotle's remark is scarcely a subject of contention among the inter-
preters, however: "Verbs were above all invented to be predicated,
nouns that something be predicated of them". The "complete" proposi-
tion 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 defini-
tion. 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 at-
tributed 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 tem-
poral 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
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 namely-
rider: '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 vI'.
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 ex-
horts 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 ex-
istential 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 ex-
istential 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 ex-
plain 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.
PETER ABELARD ON THE SPEECH SIGN 'EST' 157
In a first special case, the name of a person stands in the subject posi-
tion of the proposition as an expression for a concrete substance. An ex-
pression for an accident stands in the predicate position. Take for exam-
pIe 'Socrates est albus'. This predicate joins two items to the subject, first
the whiteness (albedo) as an accidental form (in adiacentia), and second-
ly 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
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 can-
not 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
position. And then they bring about that the things signified by the word string in the sub-
ject position are conceived of in a certain way, just as an interposed verb ('est') or a con-
junction ('if. .. , then') whose coupling indicates necessity serve to do 50. 98
'. . . is the case' and '. . . is not the case' are syncategorematic ex-
pressions. 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 dif-
ferent '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 discus-
sion 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 proposi-
tions. 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
(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 con-
troversial 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 predica-
tion 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 sup-
port 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 parti-
ciple seen as a predicate noun, just as language does to make certain
tenses and passive forms, is manifestly easier than to conceive of all com-
binations made with the copula as substitutes for verb forms which the
language has failed to "invent".
164 KLAUS JACOBI
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 un-
solved, 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 some-
one were to succeed in constructing a theory which explained the rela-
tionship 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 copulative-
ly, 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 refor-
mulating 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 Pro-
positions about the chimaera are to be interpreted as treating something
166 KLAUS JACOBI
III. REFLECTIONS
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 in-
terrelation 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 posi-
tion 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 be-
ing 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 con-
front 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 attemp-
ted 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 func-
tioning 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 pro-
positions. Problems of the logic of tense and of propositions whose sub-
ject 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 ex-
pression 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 proposi-
tion. Is the two-part analysis indeed more elegant when seen in the con-
text of logical theory as a whole and not merely of particular detail prob-
lems, assuming that the occasional sentence and a logic of tense cor-
responding 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 specify-
ing 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 con-
tained 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 affir-
mative proposition can be transformed into a particular affirmative pro-
position. 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 ab-
sorbed 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 Ex-
pressed 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
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 oppor-
tunity 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 in-
tended 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 reap-
pear 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. Com-
pare 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 consli-
tuant, 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 separa-
tionem 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 totius·proposilionis. 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' 173
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 in-
venta, 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 inven-
tionem 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 praesen-
tialiter albedine est informatum, sicut et 'currens' in quodam praesentialiter cursum par-
licipat. 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 tan-
tumdem '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.
positions. 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 ascrib-
ed 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 adiec-
tivis 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
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'
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 be-
ing 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;
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
Philosophisches Seminar II
Albert-Ludwigs-Universitiit
Werthmannplatz
D-7800 Freiburg im Breisgau, B.R.D.
HERMANN WEIDEMANN
Being acquainted with the familiar distinction between the "is" of ex-
istence, 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 over-
shadowed in his writings by another distinction between two semantical-
ly 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.
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 ex-
istens, 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 ex-
ists 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 ex-
amples 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 univer-
sale de particulari: ibid.). A predication of the latter kind, which in con-
tradistinction 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 distinc-
tion 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 actual-
ity 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
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 existen-
tial 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 Commen-
tary 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 ab-
solutely, 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'~ rejec-
tion 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 dif-
ferent 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 dif-
ference, 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
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 in-
tellect 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 intellec-
tus 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 pro-
positional 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, secun-
dum 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 ge-
nuine 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 ef-
fected 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 (proposi-
tional) combination" (or "composition"), Aquinas remarks that, accor-
ding 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 com-
monly, 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 ac-
tu esse; et ideo significat per modum verbi. Quia vero actualitas, quam principaliter
significat hoc verbum 'est', est communiter actualitas omnis formae 25 vel actus substan-
tialis vel accidentalis, inde est quod cum volumus significare quamcumque formam vel ac-
tum 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 con-
sideration 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
IV
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), com-
mits him to excluding what we would call facts from his ontology, reserv-
ing 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 will-
ing 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 to·which "the true is that
which is" (verum est id quod est), we are told by Aquinas, can be ac-
counted 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 no-
tion 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 fun-
damentum in re et non secundum id quod ratio veri completur in adaequatione rei ad in-
tellectum. - 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 componen-
tis, 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.).
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 ac-
tum 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 be-
tween 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 con-
ceptiones 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 ge-
nuinely 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 am-
biguous" and "systematically ambiguous" expressions (1975, p. 346).
2 The question to what extent Aquinas's account of this dichotomy is faithful to the posi-
tion 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 so-
and-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 "s 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 une-
quivocal 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
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 com-
positio intellect us a compositione rei; nam ea quae componuntur in re sunt diversa; com-
positio 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 composi-
tionem formantis") and De ver. q. I, a. I, ad I (" ... secundum id quod ratio veri com-
pletur 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 substan-
zialen 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 con-
dition 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.)
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. Mar-
shall 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
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 be-
ing in singular things and as a concept in the mind. 4 The essences ex-
emplified 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 in-
stances 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 in-
cludes 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 real-
ity, 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
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 ele-
ment of the conceptual correlate of the extramental reality. Correspon-
dingly, 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 intro-
duced in Scholasticism; at all events, it seems to have been commonly ap-
plied 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 con-
sequenti. 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 pro-
pria and ratio communis of an analogous name. The ratio propria is the
206 SIMO KNUUTTILA
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., Dor-
drecht, 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 essen-
tia 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.
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
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. Dur-
ing 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, Ox-
ford 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.
46 Ord. IV, d. 8, q. I, n. 2 (ed. Vives XVII), cf. Quod/. q. 3, n. 2 (ed. Vives XXV).
47 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 Interna-
tionalis 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 Fran-
ciscan 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 in-
fluence 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 MiI-
te/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 dis-
junctive 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 in-
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 doe-
trina 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 Fran-
ciscan 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 Seven-
teenth 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 Ox-
oniensis), 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 Commen-
tary 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', Fran-
ziskanische Studien 48 (1966), 297 - 307.
Brown, St. F., 'Scotus's Univocity in the Early Fourteenth Century', in De doctrina loan-
nis 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. Ac-
ta 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 Uniwer-
sytetu 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., Dor-
drecht, 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 Sup-
plementa 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 ex-
positio, 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 ex-
positio, 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.
1. INTRODUCTORY REMARKS
Scholastics, aligns himself with the views that inform this terminology,
and when he is on the contrary innovating, giving new senses to tradi-
tional 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 con-
tributes 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 com-
monplace 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.
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 sup-
posed 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 combina-
tion 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 ra-
tionis) , 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 in-
termediate class were often defined as conceptual (or formal) but they
were not regarded as created by the intellect. Although they were suppos-
ed, 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 in-
separable 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 as-
sumed by these authors, a real distinction, for the later Scholastics,
228 LILLI ALANEN
Before discussing this thesis and Descartes's argument for it, I will pre-
sent, 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 in-
terprets 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 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 ac-
cepted principle. That (i) is essential to Descartes's proof is easy to
understand against the background of the considerations above (cf. Sec-
tion 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 Se-
cond Meditation (cf. AT VII, 219; HR II, 96). On the basis of the reason-
ing there presented, (iii) at least can be accepted as - relatively - un-
problematic. But (iv) is more controversial. It is, presumably, supposed
to follow from (iii), since there seems to be no other place in the Medita-
tions where the question of what really belongs to his essence (i.e., the
essence of his mind) is explicitly discussed besides the passages subse-
quent to the Cogito reasoning in the Second Meditation. 24 However, the
argument (the Cog ito reasoning) for (iii) can hardly be considered as suf-
ficient 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 objec-
tive 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 con-
troversial 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 LILLI ALANEN
Descartes's proof of the real distinction between mind and body, as men-
tioned 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, con-
cepts 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 demar-
cated 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 abstrac-
tion of the intellect, are always conceived inadequately. Since they are in-
adequately 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 ade-
quate 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 inade-
quate by an intellectual abstraction" (ibid.). Similarly, in order to
understand that the mind is a complete thing, we need not have an (en-
tirely) 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, independent-
ly of any other entities. By a complete thing, Descartes explains to Ar-
nauld, "I mean merely a substance endowed with those forms or at-
tributes 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 self-
subsisting 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
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 substan-
tially 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.
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 im-
material 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 con-
fused 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 scien-
tific 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 distinc-
tions 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 accep-
ting, 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
NOTES
• I am greatly indebted to Erik Stenius, Georg Henrik von Wright, Simo Knuuttila, Nor-
man 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 transla-
tions 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 no-
tion 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), In-
troduction, 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
14 Edwards (1974), p. 10. Cf. also A. Wolters in Ryan and Bonansea (1%5), p. 45.
IS 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 founda-
tion ... " (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 characteriza-
tion 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 per-
cipimus. . . 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
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, accor-
ding 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 pure-
ly 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 quali-
ty, 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 concep-
tion 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 extend-
ed 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 en-
tailment between the concepts of mind and body, and the existence of the mind in separa-
tion 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
30 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 uncer-
tainty concerning the classification of the formal distinction from Descartes' side is not im-
portant 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 abstrac-
tion 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 com-
plete, i.e., self-subsisting things. Cf. Suarez's characterization of the distinction of reason-
ed reason: it "does not exist strictly by itself, but only dependently on the mind that con-
ceives 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 con-
versely. 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 Let-
ters, 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.
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, Ox-
ford, 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 carte-
sien, J. Vrin, Paris, 1930 (1975).
Hooker, M. (ed.): Descartes: Critical and Interpretive Essays, The Johns Hopkins Univer-
sity 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 Univer-
sity 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
Dept. of Philosophy,
University of Helsinki,
Unioninkatu 40B,
SF-00J70 Helsinki 17, Finland.
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 on-
tological 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 exis-
tence", to use Kant's words in A 586 = B 614). Thus the inside condi-
tional 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 ex-
istence ... 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 ver-
sions 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 dif-
ference 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 be-
ing (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 mat-
ter what world w is subsequently chosen by nature, the very same in-
dividual 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 in-
terrelations 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 Aristot-
le's actualism. Hence the mistake is in a literal sense of the expression
doubly natural for someone in his historical position to make. Kant, un-
fortunately, 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 argu-
ment what amounts to the crucial quantifier switch. It is to try to
KANT ON EXISTENCE AND PREDICATION 253
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 judg-
ment 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 connec-
tion 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 iden-
tical". 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 con-
figuration.
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 in-
stances 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 eminent-
ly 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 in-
stance, what they could be or could do. Sometimes these predicates turn
on the existence or nonexistence of these individuals in those other cir-
cumstances. 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 in-
dividual 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 cer-
tainly 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 sure-
ly can pick out one "merely possible" individual from others by specify-
ing 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 mere-
ly 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 in-
dividual from other possible worlds to the actual one which I am envisag-
ing. Some philosophers have failed to see how we can individuate a mere-
ly possible individual. Doesn't the very possibility of considering some
one definite individual (to which predicates are to be ascribed) presup-
pose 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 com-
plex of problems in its entirety. A good descriptive account of how mere-
ly 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
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" ap-
parently 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 om-
nipotent" 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 omnipo-
tent" 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.
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 ...
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 om-
nipotent God. "The omnipotence cannot be rejected if we posit a deity
... , for the two concepts are identical" (A 595 = B 623). Hence mean-
ing 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 par-
ticularly 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 irrele-
vant 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, ex-
istence 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 existen-
tial quantifier.
In fact, the reason why Frege can get along without a predicate of ex-
istence 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 in-
fer (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 posi-
tion of the existential quantifier seems to me to be a much more impor-
tant 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 com-
mon nouns but also the nonemptyness of singular nouns. Then Kant's
injunction that existence is not a real predicate might per haps be inter-
preted 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 suffi-
cient 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 Fremd-
korper 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 experien-
tial 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 posi-
tion shows that he could not really countenance vi~lations of the Principle
of Plenitude among full fledged experiential possibilities. This examina-
tion I have attempted, jointly with Heikki Kannisto, in an earlier
262 JAAKKO HINTIKKA
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 distinc-
tion 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 Gug-
genheim 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 Mac-
millan 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".
5 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 discus-
sion 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, Dor-
drecht, 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 pro-
blem in the uniqueness of that "intensional object".
16 This observation helps in understanding other features of the literature on the on-
tological argument. For instance, why do most of the recent formal or semi-formal discus-
sions 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.
KANT ON EXISTENCE AND PREDICATION 267
Dept. of Philosophy,
Florida State University,
Tallahassee, FL 32306-1054, U.S.A.
LEILA HAAPARANTA
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 ambigui-
ty. 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 "eter-
nal 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) the is of identity (e.g., Phosphorus is Hesperus; a= b),
(2) the is of predication, i.e., the copula (e.g., John is a
philosopher; P(a)),
(3) 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 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
(4) 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 for-
malization 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 distinc-
tion. I shall concentrate on the following questions: (1) How did Frege
arrive at the distinction between the is of existence and the is of predica-
tion? (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'
tion actually shows that being can, after all, be used as a logical first-
order concept, although Frege himself does not propose anything of that
kind.
Frege's argument for the claim that sentences like "A is" or "A ex-
ists" 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 seman-
tics. 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 pre-
sent a calculus ratiocinator but to create a lingua characteristica. 13 The
idea of logic as language seems to be incorporated in many of the doc-
trines maintained by Frege, as has been emphasized by lean van Hei-
jcnoort (1967) and laakko Hintikka (1981b). For Frege, the conceptual
notation is a proper language, which must be learnt by means of sugges-
tions 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 rela-
tions 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 Hei-
jenoort, 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 expres-
sion 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
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 ex-
ist. 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 truth-
value. In Frege's view, a sentence is deprived of a truth-value if it con-
tains a name which has no bearer (KS, p. 148).
Even if Frege takes being, or existence, to be included in every predica-
tion, he regards it as a proper concept in itself, since he paraphrases it
278 LEILA HAAPARANTA
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. Be-
tween 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.
In the Grundlagen Frege not only suggests, but gives an explicit formula-
tion 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 iden-
tity 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 in-
dicates 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 expres-
sion 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
consistent with the assumption that the thing is logisch einfach. Conse-
quently, 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.)
6. CONCLUDING REMARKS
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 Prin-
ciples 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 ident-
ity .... 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 Begriffs-
schrift 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 for-
ward 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., 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, Ox-
ford, 1979.
Frege, G., Philosophical and Mathematical Correspondence, abridged for the English edi-
tion 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, Ox-
ford 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, Holder-
Pichler-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 Pro-
ceedings 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 On-
tology, 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, Stutt-
gart - 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 Socie-
ty75 (I974-75)., 167-180.
Dept. oj Philosophy,
University oj Helsinki,
Unioninkatu 40B,
SF-OOI70 Helsinki 17, Finland
INDEX OF NAMES
291
292 INDEX OF NAMES
297
298 INDEX OF SUBJECTS
'is': accidental 56, 64, 66, 68-71; actuali- (to) on 6,8-9, 11-14, 18-20,27-28,68,
ty sense of 181-187, 191; copulative 82,91,101
(see copula), ellipsis hypothesis of ontological argument xv, 249-253, 256,
(86-92); of existence ix, 1-4,6,9-17, 261-263,271
27,53,59,61,65-67,71,73,81-82, ontological relativity 76
84-91,93,96,99,101,105-107, 124, opinable 116, 119-125, 127-131, 136,
183-188,249,260,265,269; of generic 138
implication ix, 81-82, 249, 270; of oratio 158
identity 35-36,38-39,63-64,81-84, ousia 18, 71, 91, 93, 107, III
88-89,93-94,99,101,183,186,188,
249, 258, 269, 281, 284; of location Pauline predication 34, 39
3-4,7-11, 15, 17; of predication ix, possibility 204, 208-213, 234, 243,
1-10, 17, 20-22, 26-27, 35-36, 252-258, 261-264
38-39,53,59,61,63-64,81-84,86, possible beings 132-139, 210-213, 234,
89,93-94,99,124,152,183-188,195, 255, 261-263, 276
249, 265, 270, 281 (see also copula); possible world semantics xiv, 250, 276
there-is sense of 182-187; veridical potentiality 5, 17,68,94-95,210,255
3-4, 8-9, 12-13, 16-17, 20-22, predication 1-10, 12, 17,20-22,27,29,
26- 27, 184. (See also being, einai, esse, 35-39,53,59,61,63-64,81-86,89,
existence.) 92-95, 9~ 105, III, 124, 132-138,
162-171, 181-189, 249, 259,
lingua characteristica 276 269-272, 275, 279-284
linguistic relativism 3 -4, 21 principle of individuation 202
logic of tense 169 principle of plenitude 203,261-262
logical: functors 148; possibilities 204, privation 17, 127, 182, 195
209, 212; predicates 264; truth 40; proper name 60, 76, 259, 272-274,
types 104-107 282-283
logically simple 282-284 propositon: two-part analysis vs three-
part anaiysis 163-171
many-sorted logic 97-98, 110
many-sorted quantification theory 97 quality 33,147,149-150,153,201,205,
mass terms III 207,226,229-231,240
matter 94-95, lOS, 108, 110-111, 202, quantification theory 92, 96, 104, 110
213,226-227 quantifier phrase 99, 110-111
mind 223, 232-241 quantifying in 255
modal: categories 262, 265; concepts 94, quantity 33, 106, 149, 205, 207
203-204; distinction 228, 230-231; question: types 103; words 104; and
logic xiv, 211, 256; operator 41, answers 101-102
249-250; syllogisms 88; theoTY 167 quidditative: beings 132, 206, 210; defini-
modes: of being 131, 205; of intellection tion 129
205; of signifying 206 quiddity 203, 213
modists xiii, 131-132, 135
ratio: communis 205-207; propria
natural kinds 117, 201, 203 205-206
necessary beings 132-133, 250 real distinction 206, 226- 232, 236,
necessity 88, 147, 161,234,249-254,262 240-241
nominal definition 120, 129, 135, 139 real predicate 259, 264-265, 275, 278
nominalists xiv reality 6, 19-20,22
numerical identity 88, 102, 224, 229 reference, see sense-reference distinction
300 INDEX OF SUBJECTS