Escolar Documentos
Profissional Documentos
Cultura Documentos
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
TABLE OF CONTENTS
ACKNOWLEDGEMENTS
vii
INTRODUCTION
ix
29
49
81
115
145
181
201
223
249
269
INDEX OF NAMES
291
INDEX OF SUBJECTS
297
ACKNOWLEDGEMENTS
vii
INTRODUCTION
INTRODUCTION
Most philosophers and most classicists were not immediately convinced by Kahn and Frede. It seemed to them that to say that Plato (or
any other Greek philosopher) did not show the Frege - Russell distinction amounts to accusing him of a logical howler. Isn't that distinction
a purely logical one, an indispensable part of our logical apparatus? This
question seems to have haunted even those students of Plato and Aristotle who had found the Frege - Russell distinction less a research tool than
a Procrustean bed. They seem to have been reluctant to claim in so many
words that the greatest Greek philosophers did not really operate with the
Frege - Russell distinction. In the case of Plato, it took a scholar like
Benson Mates, who combines a high degree of logical expertise with
historical scholarship, to come out of the closet and argue expressly that
to acknowledge that Plato Ia,cked the distinction is not to accuse him of
any mistake; for there are alternative logical treatments of 'is' (or 'esti')
which do not presuppose any irreducible ambiguity in these verbs between the allegedly different Frege - Russell senses of 'is'. Mates's article
'Identity and Predication in Plato' appeared in 1979, and is reprinted in
this book. He argues there that it is a radical mistake to try to project
the distinction onto Plato.
In the same spirit as Mates, but with an entirely different alternative
logical framework in mind, laakko Hintikka argued in his paper' "Is",
Semantical Games and Semantical Relativity' (1979) that the
Frege - Russell distinction is not only dispensable but indeed a mistaken
representation of the logic of natural language. In the first half of Hintikka's paper 'The Varieties of Being in Aristotle', appearing for the first
time here, he argues that Aristotle did not operate with the
Frege - Russell distinction any more than Plato did (according to Frede
and Mates), and traces some of the consequences of this insight.
In some ways, these developments are only the tip of an iceberg. Much
of the best work on Greek philosophy in the last twenty years has been
inspired or otherwise influenced by the late G. E. L. Owen. He made it
entirely clear that Aristotle handled the concept of being in a way essentially different from what the Frege - Russell logic leads us to expect.
One of the many stimulating suggestions Owen made was that the purely
existential uses of 'esti' in sentences of the type 'Homer is', in the sense
'Homer exists', are construed by Aristotle as being in the last analysis
elliptical for 'Homer is a man', more generally 'Homer is what he essentially is'. Owen never argued for his position, however, and he may never
have adopted it. In his paper, 'Aristotle and Existence', reprinted here,
Russell M. Dancy sets Owen's idea against the background of Aristotle's
INTRODUCTION
xi
xii
INTRODUCTION
logic of being'. As shown in the papers of Klaus Jacobi and Sten Ebbesen, early medieval inquiries into the logic and semantics of 'is' were
a part of an investigation of the nature of predication. It was usually
thought that the standard logical form of an affirmative proposition
could be thought of as a three-part form consisting of a subject term, a
predicate term, and an interposed copula. In the Aristotelian manner,
the three-part form was conceived of as an explanatory reformulation of
a two-part form, in which a noun in the nominative case is combined with
an inflected verb. On this approach, the question of the properties of the
copulative 'is' became the main problem of the theory of predication.
One of the difficulties was to understand the relation between 'is' as tertium adiacens, that is, as a copula, and 'is' as secundum adiacens, that
is, without additions.
Abelard's attempts to solve the question and his reports of the theories
of his contemporaries are discussed by Klaus Jacobi in his contribution
to this volume. In Abelard's time there were two main alternative positions. According to one theory, the function of 'is' as a copula is to join
the semantic content of the predicate term to that of the subject term.
It can exercise this function only when it has no semantic content of its
own. No connection was seen between the copula and the 'is' used as
secundum adiacens. The opponents of this equivocation theory argued
that 'is' used as secundum adiacens expresses that the thing under discussion exists, but that, when 'is' is used copulatively, the predicate serves
to determinate the manner in which the subject exists. According to this
view, the actual multiplicity of uses if 'is' is not accompanied by any genuine multiplicity of meaning. The non-copulative use was thought of as
an existential one, and it was suggested that the copulative propositions
with non-existent subject terms can always be translated into forms
having existential import.
In many places Abelard seems to hesitate between these alternatives.
This hesitation was connected with the fact, Jacobi argues, that both
alternatives were based on the three-part analysis of a proposition,
whereas Abelard's main interest was to develop a theory of predication
in which a two-part form is preferred. From this vantage point he tried
to interpret the copula as an auxiliary verb, which in conjunction with
a predicate noun does duty for verbs which often are not yet invented.
Abelard's ideas did not win any adherents, and as stated by Sten Ebbesen, in the thirteenth century the equivocation theory met with a
general disapproval. In his paper Ebbesen delineates the ancient and
medieval discussion of the problems of non-existent things, as ex-
INTRODUCTION
xiii
xiv
INTRODUCTION
INTRODUCTION
xv
the uses of 'is' are fully inderendent. Sten Ebbesen mentions John
Buridan as a representative of this trend. However, fourteenth-century
discussions of these questions (and the later significance of these discussions) are still only partially known.
One of the medieval conceptual tools which has also been used in the
modern period is the doctrine of the different kinds of ontological
distinctions, e.g., real, mental, and formal distinctions. These are
discussed by Lilli Alanen. She shows that Descartes's well-known argument for the mind-body distinction is based on a specific interpretation
of this traditional doctrine, and that the early discussions of the argument were largely concentrated on the peculiarities of Descartes's way of
drawing the ontological distinctions just mentioned. The role of these
issues concerning the logic of being has not received its due attention in
the earlier literature.
It is sometimes said, or implied, that the Frege - Russell distinction
between the allegedly different senses of 'is' goes back to Kant and to
Kant's idea that 'existence is not a predicate'. In his paper, 'Kant on
Existence, Predication, and the Ontological Argument', Jaakko Hintikka argues that Kant does not in the last analysis assume anything like
the contemporary Frege - Russell distinction. All we find in Kant is a
contrast between relative and absolute uses of the concept of being.
As a part of Hintikka's argument, he offers a largely new diagnosis of
the fallacy in the ontological argument, denying Kant's dictum that 'existence is not a predicate'. Instead, he finds the crucial flaw in the ontological argument elsewhere, viz. in its tacit dependence on our being
able to identify God (the being of whose existence is to be proved) between the different possible worlds presupposed in the argument. This
provides a new perspective on historical as well as contemporary discussions of Anselm's argument.
All these different investigations naturally raise the question: What is
the origin of the Frege - Russell distinction? What is its background? In
her paper, 'On Frege's Concept of Being', Leila Haaparanta discusses
Frege's treatment of being in its historical setting. One of the crucial ingredients in Frege's treatment of being is his idea that existence is a
second-level concept (property of a concept). Haaparanta sees the foundation of this assumption in Frege's ideas about the identification (individuality) and existence of individuals (objects), incorporated in
Frege's treatment of the senses by means of which we can grasp an individual object. These were according to her inspired by Kant's ideas,
especially by Kant's distinction between the predicative and existential
xvi
INTRODUCTION
uses of 'is'. Even though Kant did not subscribe to or even anticipate the
Frege - Russell Jistinction, he thus seems to have inspired it.
SIMO KNUUTIILA
JAAKKO HINTIKKA
CHARLES H. KAHN
When I began work on the Greek verb to be in 1963, in the project that
took shape in the article 'The Greek Verb "to be" and the Concept of
Being,l and eventually resulted in a book on the Greek verb 'to be' in
1973, 2 my aim was to provide a kind of grammatical prolegomena to
the study of Greek ontology. I wanted to give a description of the
linguistic facts concerning the ordinary use and meaning of the verb,
apart from its special use by the philosophers, in order to clarify the pretheoretical point of departure for the doctrines of Being developed by
Parmenides, Plato, and Aristotle. I thought at the time (and still think)
that the ancient use of the verb estileinail on was poorly understood, and
that much of the modern discussion is vitiated by false assumptions, in
particular by an uncritical application of the notions of existence and
copula to the interpretation of ancient texts. I take the present occasion
to summarize the results of my work both for the theory of the verb and
for the interpretation of some of the early philosophical texts, referring
to earlier publications for more detailed exposition and defense of the
views outlined here.
I. THE DISTORTING INFLUENCE OF THE TRADITIONAL VIEW
As I see it, confusion reigns both in the traditional account of the verb
given by linguists and philologists, and also in much of the philosophical
exegesis of ancient theories of being. The two lines of confusion have infected one another, since the linguists have borrowed their notions of existence and the copula from philosophy (and from rather superficial
philosophy at that), while philosophers have in turn made use of
linguistic doctrine as a basis for their own account of Greek ontology,
and in some cases as the weapon for a general attack on the Greek notion
of Being. I begin by stating what I take to be the principal errors in the
standard view, by which I mean the views prevailing twenty years ago and
still to be found in many handbooks and commentaries.
(1) It was generally assumed that the uses of einai could be classified
either as (a) meaning 'exists', or (b) serving only as copula. 3 But this
S. Knuullila and 1. Hintikka (eds.), The Logic oj Being, 1-28.
1986 by D. Reidel Publishing Company.
CHARLES H. KAHN
the copula uses of einai are overwhelmingly more frequent than any
other use in the earliest texts. The idea that the existential uses are
somehow more fundamental or more primitive seems to be a mere prejudice, a prejudice based in part upon a mistaken view of existence as an
ordinary predicate, taken together with an empiricist theory of meaning
which assumes that the original sense of any word must have been
something concrete and vivid, something like what Hume calls an
"impression" .
Hence I propose a modest Copernican revolution: to reinstate the
copula at the center of the system of uses of einai. I do not claim that
the copula uses are older, since for that claim also there is no evidence.
In purely synchronic terms I propose that the copula uses must be regarded as more fundamental in three respects: (1) they are statistically
predominant, (2) they are syntactically elementary, whereas other uses
(existential, veridical, potential) are grammatically "second-order",
operating as functors on a more elementary sentence, and (3) they are
conceptually prior and central to the whole system of uses of the verb,
in a sense that remains to be clarified, but which bears some analogy to
the unifying role of a central term in Aristotle's scheme of "focal meaning" or pros hen legomena. Thus if we take the copula uses as given, we
can see why the same verb may serve in other ways, for existence, truth,
possession and the rest. But if we take any of the other uses as primary,
the way back to the copula becomes difficult, if not impossible [Q
understand. 6
(3) It was correctly noted by a number of linguists that the existence
of a verb to be in our sense, which is at once a verb of predication, location and existence (to name only three of its functions), is a peculiarity
of Indo-European. 7 As we can see from the monograph series on "the
verb 'be' and its synonyms" edited by .I, W. M, Verhaar, the topic of
be can itself be defined only by reference to Indo-European verbs from
the root *es-. But why should a historical peculiarity of this kind be of
any general significance, and how can a concept based upon the
parochial usage of an Indo-Eurorcan verb provide a genuine topic for
philosophical theory? Thus A. C. ~:(2harn has claimed, a rropos of the
very different situation in Chinese. thJ.t
there is no concert of Being which languages ar~ wei! Of iii cquiprcd to
tipn." or 'to ht" as ('opld3 depend lIPon a grarnmatical rule t'O!' thl...'
sentence. and it would b::.- merely a :elDcid'.:r~'(' ~r one found an\'thi:,,(~
l'-C",'flL:
tnr rune
,)\" the
i:-" ,~-:atinn
Icscnl~)l;-:,'
11
in (:
CHARLES H. KAHN
1. The Copula
In order to see this as the central focus of the system of uses of einai, we
must have a more adequate account of the copula itself. Linguists often
speak of the copula as a dummy verb, as a merely formal "bearer" of
the verbal marks of tense, person, and mood. 11 Hence a recent author
can describe the entire function of the copula as "simply to act as a verbalizer," to convert "an adjective like 'cunning' into a verb-phrase 'is
cunning', which is of the same category as 'snores' " or of any finite
verb. 12 Abelard, who was either the inventor or at least the codifier of
the classical theory of the copula, rightly associated the copulative function (vis copulativa) with all finite verbs. He saw is as distinctive in that
it provides only the predicative link and not also the predicate (copulat
tan tum et non copulatur); other finite verbs do both.13 Abelard's theory
has the merit of focussing attention on the verbal function as such, and
not simply on is as the verb for nominal predicates. What he saw is that
the copula separates out the specific function of the verb, which is
obscured in the case of other verbs such as runs, sleeps, just because they
combine the information content of the predicate (running, sleeping)
with the verbal form. This general verbal function is what one might
identify as the propositional tie or mark of assertion; what we, following
Abelard, call the copula is simply the canonical expression of this function in a sentence of the form X is Y.
CHARLES H. KAHN
In sentences of this form, the copula expresses what is otherwise indicated by the verb-ending: in either case we have a "sign of predication". What is meant by predication here has both a syntactic and a
semantic aspect. (I) Syntactically, the copula or verb-ending serves to
make a grammatical sentence out of two terms that would otherwise only
form a list: John and running (or to run). It thus indicates not only tense,
mood, and person but something more fundamental: sentencehood. In
many cases the finite verb does this alone, without a noun, both in impersonal verbs like UH (pluit, "it is raining") and where the subject is
understood from the context: 7f!x.H (currit, "he is running"). (2)
Semantically, if we take the indicative-declarative form as basic, the verb
or copula gives formal expression to the truth claim of the sentence. (We
need not attend here to the ways in which this truth claim may be
modified, by interrogative or conditional sentence structure or by the
various moods; in some but not all of these cases, the modification will
be reflected in the verb ending.) Limiting ourselves to indicative forms
used in declarative sentences we can say: tbe semantic function of is as
copula or sign of predication is to bear the mark of sentential truth Claim,
to serve as focus for the claim of the whole sentence. (The truth claim
of a sentence corresponds roughly to the fact that it can have a truth
value because it does have truth conditions.)14 This basic assertive function of the copula - more precisely, the intimate connection between the
copula and the assertive function of the sentence - shows up when we
stress the verb in pronunciation: "Margaret is clever, I tell you!" "The
cat is on the mat after all." This semantic role of the copula as sign of
sentential truth claim permits us to understand one of the most important
special uses of einai in Greek, the so-called veridical, where the verb by
itself (both in the third person indicative and in the participle) expresses
the notions of truth and reality. If we lose sight of this connection between the copula and the truth claim that is fundamental for all
declarative discourse, the fact that "being" in (;reek (to on) may mean
reality will become a mysterious anomaly, quite independent of the
predicative function of the verb. Hence philolog.ists have often tended to
overlook the veridical use or to conflate it with the existential, despite the
fundamental differences in sentence structure. This is a principal support
for my claim that the copula use is fundamental; neither veridical nor existential use can be explained on the basis of the other, but both can be
understood on the basis of the copula.
Before considering these special uses of einai, I call attention to two
features of the copula verb which are often on:rlooked.
CHARLES H. KAHN
The Lexicon recognizes, what every good Hellenist knows, that in many
cases the verb esti and its participle on must be translated by "is true" ,
"is so", "is the case" or by some equivalent phrase: esti tauta "these
things are so" (cf. French c'est cela "that's right"), legein to onto "tell
the truth", "state the facts" .19 Comparative grammar shows that this is
a pre-historic use of*es- in Indo-European 2o ; and Aristotle himself notes
this as one of four basically different uses of einai (Met. 117). All I have
done is give this use a name, "the veridical," and correlate it with a
definite sentence form. The veridical construction proper is characterized by three syntactic features: (i) the understood grammatical subject
of esti is not a noun form (like man or hunting) but a sentential structure,
as represented in English by a that-clause (that the man was hunting); (ii)
the construction of the verb is absolute, i.e. there is no nominal, locative
or other adverbial complement, except the comparative "so" (houtos),
which introduces (iii) a comparative clause of saying or thinking. which
is expressed in the full veridical construction and implied in every case:
These things are as you say, fCTn raimx ovrws 'WCT1I"EQ CTU Af'YHS.
Without claiming that the veridical construction is derived from the
copula use, either historically or transformationally, I nevertheless
believe that it is easy to see how the two uses logically and naturally
belong together, as long as we keep in view the semantic function of the
copula as the mark of truth claim. A properly veridical use of to be (as
in "Tell it like it is") simply makes general and explicit the truth claim
that is particularized and implicit in every declarative use of the copula
("The cat is on the mat"), as in every declarative sentence. Because the
great flexibility of the copula construction makes it possible to produce
an Sis P sentence that is roughly equivalent to every noun-verb sentence,
the copula tends to serve as verb par excellence, that is, as representative
for the finite verb as such and for its predicative force, what Abelard
called its vis copulativa. It is because of this very general function of the
copula as sign of predication and sentencehood that the very same form
(esti, "is") can serve to express the veridical idea as such: to bring out
the implicit truth claim ("This is how I say things are") and the corresponding notion of reality ("This is how things really are").
Despite my general reluctance to decide when a different use becomes
a different sense, I am inclined to speak here of a veridical meaning or
connotation of einai, in cases where the Greek verb cannot be adequately
translated by the copula or by an idiomatic use of is alone. This is most
conspicuous when the participle (to) on is used to mean "truth" or "the
fact of the matter", and when it may be replaced in Greek by a word like
aletheia or (to) alethes. 22 And there are clear cases of the veridical connotation attached to a copula construction, as in the example which
Aristotle cites of einai meaning "is true": esti S6krates mousikos, "that
Socrates is musical, i.e. that this is true". 23 In English as in Greek, this
force of the verb is typically brought out by a contrast between Being and
Seeming: "He wants not to seem (dokein) but to be (einOl) the noblest"
(Aeschylus Septem 592). Here again a basic philosophical contrast - between appearance and reality - is fully prepared in the pre-philosophical
usage of the verb.
3. Existential Uses
I briefly describe three uses of einai that we intuitively recognize as "existential" and are inclined to render by there is or even (in the third case)
by exists.24
(1) The existential copula:
(a)
"There is a city (esti polis) Ephyre in a corner of horsenourishing Argos". (Iliad VI. 152)
(b)
Perhaps the most common of all "existential" uses of the verb in Greek
are sentences such as these, where esti seems to functions twice: to assert
the existence of a subject ("There is a city ... ") and then to say
something about it: "The cityEphyre is in the corner of Argos". In most
instances the predicative use will be locative, as in (a); (b) is one of the
rare examples where a purely nominal copula takes an existential force.
It is clear that the underlying syntax of the verb in such sentences is that
of the copula, but that this construction has been overlaid with a secondary function, which I would analyze as introducing a subject for further
10
CHARLES H. KAHN
11
where Fx stands for the relative clause and an initial esti functions as existential quantifier: "there is someone/something such that ... ". The
verb esti serves precisely to affirm or deny (the existence ot) a subject for
the following clause, to assert that the set specified by the following formula is or is not empty. There is no trace of the copula construction, nor
any way to derive this form logically or syntactically from the copula construction. But there is a logical overlap with type (1), which can be seen
as a copula construction overlaid with the existential function of type (2).
This properly existential use is relatively rare: I found only 4 examples
out of 562 occurrences of einai in the first twelve books of the Iliad. I
think it would be unreasonable to suppose that (2) somehow represents
the original, prehistoric value of*es-. Can we offer a historical explanation of this use of esti? My suggestion is that it arose out of the copula
use by way of sentences of type (1), where the copula acquires "existential" connotations in virtue of its locative association and its rhetorical
function of introducing a subject for predication. Given these connotations, it is natural that the existential function of (2) becomes one of the
values esti can have when used alone, without nominal br locative complements. (The veridical, possessive, and potential uses represent other
values esti may possess when it appears outside of the copula
construction.)
It is on the basis of the existential force of the verb in (1), where this
force is secondary, and in (2) where it is primary but serves directly as
the basis for ensuing predication, that we can understand the appearance
of a new sentence type, in which esti itself becomes the grammatical
predicate.
(3) The existential predicate:
(a)
(b)
Type (2) is rare in Homer, but type (3) does not occur at all. My earliest
specimens are from Melissus, Protagoras, and Aristophanes in the middle and second half of the fifth century B.C., and they clearly show the
influence of philosophical speculation. 27 Sentences of this kind are
sometimes cited as exhibiting the oldest meaning of *es- in IndoEuropean. On the contrary I regard this as a fifth-century innovation,
based upon the existential force of the verb in the older types (1) and (2),
but focussing attention on existence as such (i.e. on the question whether
or not there is such a thing), as a result of philosophical speculation,
12
CHARLES H. KAHN
The outcome of my linguistic survey has been to underline the fundamental role of einai as copula verb and at the same time as verb of state and
station, characterized both by locative and by durative-stative values.
Among the non-copula uses I have called attention to the veridical expression for truth and fact, and I have insisted upon the very limited
range of early existential uses, bound to a specific locative or predicative
context. Thus while existential sentences of type 1 and 2 are well attested
in Homer, the stripped-down "absolute" use of type 3, in which einai
appears alone as existential predicate, is not found before the fifth century, and then only in contexts where philosophic or sophistic influence
is clear. 30 My suggestion is that for understanding the early philosophical
usage, both in Parmenides and in Plato, the veridical notion (whether or
not it is the case that p) turns out to be more important than the idea of
existence (whether or not there is such a thing as X), although both notions are present.
To illustrate the new, quasi-technical use of the verb as a pure existential (my type 3), we may cite what is probably the earliest unambiguous
example, the welI-known statement of Protagoras that, concerning the
13
gods, he does not know "either that they are (has eisin) 0;- that they are
not (hos ouk eisin), or what they are like in form" (fr. 4). The contrast
furnished by the last clause guarantees that einai here refers to the question of the gods' existence; and the verb itself might properly be
translated as "exists". 31
However, in another even more famous quotation from Protagoras
the natural reading of the verb to be is veridical: "Man is the measure
of all things, of the things that are (ton onton), that they are (has estin) ,
and of the things that are not, that they are-not (has ouk estin)" (fr. 1).
Here we have prefigured, in a slight modification of the old idiom for
truth ("tell it like it is"), the distinction between ~he intentional being-so
of judgment and statement (hOs estin) and the objective being-so of the
way things stand in the world (Tex 11m). This intuitive distinction between
the ways things are and the way they are judged to be, which Protagoras
recognizes only to deny its validity, is precisely what we find in the two
terms of Aristotle's definition of truth, where the participle (ta onta) is
used for the facts of the case, as in Protagoras fr. 1, while the finite
verb in Protagoras' formula is replaced in Aristotle by the infinitive, for
the asserted einai of thought and statement. 32 The parallel being so exact,
it is no accident that Protagoras' book was called "Truth". 33
Another early example from the philosophical literature shows how
veridical and existential values can intersect in a single occurrence, or
how an author can oscillate between the two. Melissus is conditionally
assuming what he wants to deny: the reality of phenomenal diversity.
If there really is (ei eSlI) earth and water and air and fire and iron and gold, and living and
dead, and black and white and all the other things men say are true or real (hosa phasin
... einai alethe), if these things really are (ei tauta estl), and we see and hear correctly ...
(Melissus fro 8.2).
14
CHARLES H. KAHN
15
contrasts define the sense in which for Parmenides esti means "it exists",
where the durative and locative values of the verb give some definite
shape to the claim of existence. But there is no reason, neither in the prephilosophic usage nor in the context of the poem nor in the later echoes
of the Parmenidean thesis in Protagoras (fr. 1, cited above) and Plato
(Rep. V, 476E - 477 A, cited below), to suppose that in the initial presentation of his thesis in the one-word sentence esti in fr. 2, the sense of
Parmenides' claim can be adequately captured by the translation "it
exists" . 38
How are we to construe this claim? The contrasts just cited require
that, if the argument is to be coherent, the content of what is claimed
must be such that it is (a) something rather than nothing, (b) already present, and (c) guaranteed to endure. But that gives us no clue as to where
the argument begins, or how we are to understand Parmenides' initial
presentation of the thesis so as to provide him with a plausible startingpoint. For this we must look at the context in the poem and above all at
the preceding context: the allegorical proem. 39
Parmenides' thesis (that it is and that it cannot not-be)is introduced as
the acceptable member of a pair of alternative "ways of inquiry" for rational cognition (noesal) to travel on (fr. 2.2). Where is this inquiry supposed to lead? Obviously, to knowledge and to truth, as is clear both
from the proem and from the words immediately following the thesis
(2.4: "it is the path of Persuasion, who follows on Truth"). In the
allegorical proem the voyager on the right road is a "knowing mortal",
transported by wise horses and clever escorts, the daughters of the Sun
who are leading him "to the light" (fr. 1.10). When he arrives, a goddess
promises to instruct him in everything, but first of all in "the unshaken
heart of persuasive Truth" .40
This is what Parmenides gives us as a background for understanding
his thesis that it is, and that it cannot not-be. To interpret the thesis we
must be able to say: what is the subject of the claim esti? And what is
the content of that claim? I think the subject can be specified with some
confidence, on the basis of clues from the proem and the immediate context. With these clues Parmenides makes quite clear that what the goddess holds out is a promise of knowledge, and that the path of it is must
lead to truth. Hence the understood "it" which the goddess is referring
to in the thesis must be located in the region of knowledge and truth; it
can also be identified as the goal of inquiry and the object of that quest
that began in the first verse of proem, where the horses are said to carry
the youth "as far as his desire can reach". So we C:in ciescribe the subject
16
CHARLES H. KAHN
referred to by the goddess as what our youth has come to find out, and
what will be made hIOwn to him in the revelation of persuasive truth.
Such an initial characterization of the subject as "the object of inquiry"
or "the knowable" can be taken for granted before the thesis is articulated, though any fuller characterization remains to be spelled out in
the course of the argument. 41
What does esti say about the object of inquiry that (a) can be taken
for granted as a point of departure, and (b) can justify the immediate,
categorical rejection of the negative way, that it is not? Recent interpreters, looking ahead to see what content is given to the thesis later in
the argument, propose to read esti as "it exists". But in addition to the
dubious procedure of reading a poem backwards, this view has the disadvantage of saddling Parmenides from the outset with an essentially
anachronistic notion of to be. If my interpretation of the linguistic
evidence sUPlmarized above is even approximately correct, then it is
highly unlikely that either Parmenides or his readers would understand
a bare unadorned esti as meaning primarily or predominantly "it exists".
Of course the parallel to sentence types 1 and 2 guarantees that "there
is such a thing" will be there as a background meaning for Parmenides
to rely upon. But the primary idiomatic sense of an unqualified esti in
the early fifth century can only be the veridical, in this case taken objectively for the reality as known: "it is so" or "this is how things stand".
And the logic in support of the initial thesis then becomes unassailable:
what is known or knowable must be the case and cannot not be so. "For
you could not know what is not (so)" (fr. 2.7).42 Thus Plato, when he
echoes this argument in Rep. Y, 476E-477A, has Socrates ask: "Does
a knower know something or nothing? ... Something which is (on) or
which is not (ouk on)?" To which the interlocutor replies: "Something
which is (on); for how could anything which is not (me on tl) be known?"
Plato adopts Parmenides' starting-point here precisely because he wants
to make the premises of his argument as plausible as possible. For the
ancients as for the moderns, knowledge entails truth: what is known
must be really SO.43
It is the veridical use, then, which not only provides the idiomatic
background for understanding Parmenides' stark initial esti, but also
provides the conceptual grounds for granting it,; necessary truth (best
understood as necessity of the consequence: if p is knowable, then,
necessarily, p is the case). Once this starting point has been granted, on
the basis of a veridical what is (so), Parmenides will go on to unfold the
richer implications of an esti whose full content will depend on other uses
17
of the verb, including the locative associations which justify his assumption that to eon is spatially continuous, indivisible, and sphere-like if not
spherical. Among these other properties of Parmenides' being there will
surely be some that depend upon the copula use (Being is unchanging,
for if it is F, for any F, then it can never be not-F without falling into
Not-being), and some that reflect the existential use in the sense specified
above: if Being is, it is not nothing; if it is ungenerated it is already there;
if it is imperishable it will persist. Whether Parmenides' move to these
richer senses of einai necessarily involves him in a fallacy of equivocation
is not entirely clear. We might well suspect something of this sort, in view
of the astonishing nature of his conclusions. Plato is at pains to show,
against Parmenides, that something can be X and also not be Y without
falling into nonentity; whereas Aristotle distinguishes being not only in
terms of the categories but also in terms of potency-act and substrateprivation-form (in Physics I) in order to avoid the conclusions which
Parmenides draws by taking to on as univocal. Here I suggest that
Plato's diagnosis cuts deeper into the actual structure of Parmenides'
argument. Some unclarity but no radical incoherence results if
Parmenides takes to on (I) veridically, as the objective state of affairs required for truth and knowledge, then (2) existentially, as a real, enduring
object which is the "subject" of this state, and also (3) copulatively as
being F for various F's, as well as (4) locative, i.e. spatially extended. 44
Fallacy enters only with negation, and the assumption that what-is-not
in any respect must be a Non-being pure and simple. The inference from
(I), "there is something which is the case, which is determinately so" to
(2) and (3) "there is'something which exists as an enduring subject, and
which is F" requires for its validity only the reasonable (if not inevitable)
assumption that for a state of affairs to be definitely so there must be a
definite subject with definite properties. The undeniable category-shift
from a propositional entity that is the implied subject of esti in (I) to a
substantial or thing-like entity for (2) and (3) is precisely parallel to the
shift between "if these things are true (einai alethe), ' and "if these things
exist" (ei tauta estl) in the text of Melissus cited above, and parallel also
to the cat ego rial ambiguity of einai alethe in the same text: "if they are
true" and "if they are real". Since similar shifts and ambiguities between
propositional and substantial entities occur in Plato and Aristotle too,45
it would be surprising indeed if the paradoxical esti of the earliest Greek
ontology were quite unequivocal in this respect. 46
This interpretation of Parmenides' thesis also provides a natural
historical explanation for the paradox of false statement and false belief,
18
CHARLES H. KAHN
which seems to have been popular with some sophists and which persistently recurs in Plato's Cratylus (4290), Theaetetus (189A 10-12)
and Sophist (236E, 237E), with an early variant in the Euthydemus
(283E - 284C). If speaking falsely is saying what is not (the case), and
what is not is nothing at all, then speaking falsely is saying nothing and
hence not speaking at all. There may be other dimensions to this
paradox, but the crucial move is clearly the slide from what is/what is
not as the object of true and false statement and belief to what is not as
that which is nothing at all, the non-existent - a slide precisely parallel
to the one we have identified in Parmenides' argument, and which has
its counterpart in Melissus' oscillation between doubts that the multitude
of phenomenal things really exist, on the one hand, and claims that we
do not see or hear correctly, on the other hand, or that there are not as
many things "as men say are true". Thus "true being" (to on alethinon)
for Melissus (in fro 8.5) is both (a) what really exists and (b) what is true,
as the content of true statement and belief. This ambiguity is relatively
harmless in the affirmative case, where (a) and (b) coincide (given the
failure to distinguish between facts and things). But the corresponding
negation leads to fallacy and paradox, if the denial of (b) in a reference
to the object of falsehood is also taken as a denial of (a).
Since I have treated Plato's use of esti and to on at length in a recent
article, I will here simply list my chief conclusions concerning Plato's ontological vocabulary in the preliminary and mature statements of his
theory of Forms.
(1) In the so-called Socratic dialogues, the first philosophically relevant use of einai is its occurrence in connection with the What-is-X?
question of Socratic definition. Examples: Laches 190B7-C6: "we ought
to possess knowledge of what virtue is" eidenai hoti pot' estin arete); "we
say we know what it is" (eidenai auto hoti estin); "But if we know, then
we can say what it is" (ti estin). In the Euthyphro we find the contrast
between such a "whatness" and other attributes of a thing hardening into a terminological distinction between ousia, "essence", i.e. the content
or correlate of a true answer to the what-is-it? question, and pathos, any
other property or attribute (Euthyphro IlA7-8). Here ousia is simply a
nominalization for the verb estin in the what-is-it? question (to hosion
hoti pot' estin, IlA7). In such contexts the verb is syntactically the
copula, but logically or epistemically strengthened by the context of use
into what we might call the definitional copula or the is of whatness,
which aims at locating the true, proper, deep or essential nature of the
thing under investigation.
19
(2) The first trace of a more technical use, growing directly out of the
definitional copula, appears in the Lysis, where the attempt to explain
what makes something dear, friendly, or beloved (phi/on) leads to the notion of "that which is primarily dear (ekeino ho esti prOton phi/on), for
the sake of which we say that all other things are dear" (219C 7), these
other things being potentially deceptive "images" (eidola) of "that
primary thing, which is truly dear" (ekeino to prOton, ho hos alethos esti
philon (219D 4). What is new here is (a) the use of the definitional copula
as a kind of proper name for the concept under discussion, or for its
primary instance: what is (truly, primarily) X, prefiguring the canonical
reference to the forms in later dialogues as to ho esti X, and (b) the
veridical strengthening of the copula in "what is truly (alethos) dear" cf.
tpC)..OIl Of T~ olin at 220 BI and B4), by contrast with the "images" which
are only "verbally" dear (220B 1), i.e., said to be dear because of their
relation to the primary case (219D 1). Just as the Euthyphro adds precision to the is of whatness by a version of the essence-accident distinction,
so the Lysis reinforces the metaphysical import of a privileged use of this
formula by introducing a contrast between Reality and Appearance, between what really is F and what is only an image or a putative instance
of F.
(3) In Plato's first explicit statement of his mature doctrine, in the
speech of Diotima in the Symposium, the Beautiful itself is announced
as the goal of a process culminating in a final study "which is (ho estin)
the study of nothing but that Beautiful itself", where the student will end
by knowing "that which itself is beautiful (auto . .. ho esti kalon, 211 C
8). This is the formula of the Lysis, with its veridical force (' 'he will know
what is truly beautiful") again underscored by contrast with appearance
and images (2IIA 5, 212A 3). But in this case the formula unmistakably
refers to the Form. For here we have a new (or newly formulated) doctrine in which, for the first time, Plato provides his specimen Form the
Beautiful with a definite ontological status, based upon the Eleatic opposition between eternal, unchanging Being (aei on) and inconstant,
perishable Becoming (211A I - 5). In this context the participle on is used
both existentially ("it is forever") and as copula ("it is not beautiful in
one respect, ugly in another,,).47 It is precisely in such Parmenidean contexts, where Being is contrasted with Becoming, that it seems most
natural to regard to on in Plato as existential, though the aspectual value
is that of the stative copula.
(4) In the Phaedo and Republic, where the doctrine for Forms is
systematically developed, the philosophical uses of einai become too
20
CHARLES H. KAHN
diverse for cataloguing here. I would emphasize only that (a) the veridical
overtones of to on, used roughly as a synonym for "truth", are predominant in the initial presentation of the Forms in both dialogues: in Phaedo
65B - 67B, to on and ta onta occur together with aletheia and to alethes
for the "true reality" which the philosopher's soul desires and pursues
(and which is identified as Forms at 650); in Republic V, 476A the Forms
are introduced by the contrast of Reality and Appearance: each of them
is one but appears (phainesthal) as many (A7); and their ontological
status is again expressed by a use of to on for the object of knowledge;
that which is wholly real (pantelos on) is wholly knowable; that which
is in no way real (me on medame) is in every way unknowable" (477A
3). The veridical-epistemic contrast between Being and Seeming
(phainesthal) serves to distinguish the Forms and "the many"
throughout this passage (cf. 479A 7 - BIO).
(b) The formula auto to ho esti (ison), "that itself which is (equal)",
familiar from the Lysis and the Symposium, is gradually developed in the
Phaedo from idiomatic phrases into a semi-technical designation for the
Forms (notably at 7502 and 7804, recalled at 9308), with a parallel use
of ousia for the distinctive being, essence, or reality of the Forms. 48 The
same designation is used to reintroduce the Forms into the central
epistemological passage of the Republic: the Beautiful itself and the
Good itself and the other unique entities, "each of which we call what
it is" (ho estin hekaston prosagoreuomen, 507B 7). In this designation
the predicative form is Fis taken separately, independent of all subjects,
and made itself the target of the question 'what is it?' Thus ho esti serves
in Plato, like to ti en einai in Aristotle, for the objective essence or definitional content given in a correct answer to the question "what is Fl" for
a given predicate F. The syntax of the verb is still that of the copula, but
its predicative role is reinforced now not only by the definitional search
for the true nature of a thing but also by the ontological dualism of
Plato's neo-Parmenidean opposition between Being and Becoming, the
One and the Many, the Intelligible and the Visible. The specifically
Platonic use of einai in the doctrine of the middle dialogues thus consists
in a convergence between (i) the definitional copula from the what-is-it?
question, (ii) the veridical Being that contrasts with Seeming, and (iii) the
stative-invariant Being that contrasts with Becoming and Perishing. An
unqualified use of to on, einai, or ousia may bear any and all of these
connotations. 47 The predicative syntax is always latent if not manifest.
The existential value appears above all in (iii), but even here the copula
use, on which the stative-mutative contrast of Being-Becoming is found-
21
ed, may reappear at any moment. so The Platonic concept of Being is constituted not by a fusion of copula and existence but by the union of
timeless-invariant Being (in contrast with Becoming) and cognitivelyreliable, veridical Being (in contrast with Appearance), both of them expressed or expressible in copula predications, but most rigorously distilled in the frozen auto ho esti version of the definitional is of whatness
in application to the Forms.
(c) In the Eleatic introduction to the doctrine of Forms at the end of
Republic V, Plato has moved beyond Par men ides in a number of interesting ways. First, by accepting an intermediate "mind" reality between Being and Not-Being as object for the cognitive state of opinion
(doxa) between knowledge and ignorance, Plato has provided an ontological basis for change and becoming, which was simply the domain
of error and falsehood for Parmenides: Plato thus accounts for the
possibility of true opinion short of knowledge by giving it an object of
its own. In the second place, by his development of the copula use for
parallel designations of Forms (as "what is F") and particulars (as "what
is and is not F"), Plato opens the way to a philosophical analysis of
predication and the diverse uses of to be which he will pursue in the
Sophist and elsewhere, and which will lead eventually to Aristotle's
theory of categories and his distinction between essential and accidental
predication. On the other hand, where Plato in the Republic has not
moved substantially beyond Parmenides is in his conception of the negation of Being as what is not in any way (to medame on); for this is indescribable and unintelligible, as Parmenides had insisted and as Plato
in the Sophist will finally agree. The paradox of false statement and false
belief will haunt Plato until he works out a way to negate the "being"
of truth without falling into this hopeless region of blank non-entity.
This is far enough to pursue a project that began as linguistic prolegomena to Greek ontology and not as a history of the subject. In conclusion, I want to say a word against the charge of linguistic relativism,
in so far as it claims that ancient ontology was vitiated or distorted by
the accidental possession of a verb that combines the functions of existence and predication. It is certainly true that the verb einai serves a
multitude of functions that are rarely combined in languages outside of
Indo-European. And if Greek-ontology had begun with a radical confusion between existence and the copula, then its first task should have been
to distinguish the two, a task that neither Plato nor Aristotle undertook.
On the contrary, both of them systematically subordinate the notion of
existence to predication; and both tend to express the former by means
22
CHARLES H. KAHN
23
The earliest clear statement of the dichotomy known to me is that of J. S. Mill in his
System of Logic (IB43), I. iv. I, who attributes it to his father, James Mill, in the Analysis
of the Human Mind{IB29). But the philologists were already using this dichotomy as early
as G. Hermann in IBOI. (See the quotation in 'Be', p. 420, Note I.) Hermann in turn appeals to "what logicians call the copula", and is apparently dependent on the logic of
Christian Wolff, ('Be', p. 423 with Note 5).
4 References to Brugmann, Delbriick, Meillet, Kiihner-Gerth, and Schwyzer-Debrunner in
'Be', p. 199, Note 21. Compare John Lyons, Introduction to Theoretical Linguistics
(196B), p. 322: "Even in the Indo-European languages the copulative function of 'the verb
to be' appears to be of secondary development".
5 For the argument see 'Be', pp. 199 - 207.
6 Thus to explain the stative value of copula*es- we would have to posit an original sense
"to stay, remain" which is unattested, and turns out to be only a projection of the 'be''become' contrast for the copula. The priority of the copula uses is partially clarified
below; for fuller discussions see Chapter VIII of 'Be', especially pp. 395 - 402,407 -409.
For methodological remarks on the claim of priority here, see 'On the Theory of the Verb
To Be', in Logic and Ontology ed. by M. K. Munitz (New York, 1973), pp. 17 - 20.
7 See in particular E. Benveniste, 'Categories de pensee et categories de langues' and
, "~tre" et "avoir" dans leurs fonctions linguistique', in Problemes de linguistique
generale, pp. 63 -74 and IB7 -193.
8 , "Being" in Classical Chinese', The Verb 'be' and Its Synonyms, ed. by J. W. M.
Verhaar, Part I (Reidel, 1967), p. 15. Compare Lyons, Introduction to Theoretical
Linguistics, pp. 322f.
9 See the reference to Mill's Logic in Note 3, above.
10 There has recently been a noticeable trend away from the Mill- Russell view that "is"
has different senses, which the Greek philosophers should have distinguished. See, e.g.,
Benson Mates' suggestion that Plato's different uses of "is" can all be understood on the
basis of a single, univocal use of the copula: 'Identity and Predication in Plato',
Phronesis 24 (1979), 211- 229. And compare Jaakko Hintikka's paper in this volume. In
my opinion, the question whether "is" has different meanings or only different uses cannot
be answered without confronting certain very deep problems in the theory of meaning,
which is ultimately a part of the theory of knowledge. For example, are senses of a word
distinguishable as a matter of logical form and conceptual truth, independently of any factual question as to the kinds and natures of the things to which the word is applied? Up
to a point, linguistics can settle questions of syntax and sentence structure. But epistemology and metaphysics must be called in to decide how linguistic "meanings" are related to
the nature of things or to our "conceptual scheme".
II See, e.g., Lyons, Introduction to Theoretical Linguistics, pp. 322f.
12 C. J. F Williams, What is Existence? (Oxford, 19BI), echoing Quine, Word and Object, pp. 96f. This view of the copula ignores the distinction between an 'is' of identity and
of predication, a logical distinction which is not reflected in the syntax of the verb and is
not plausibly regarded as a difference in meaning for 'is'. My argument for this view (in
'Be' p. 372, Note I and p. 400, Note 33) is defended by C. J. F. Williams, op. cit., pp.
10-12. For criticism of this view on philosophical rather than linguistic grounds, see Ernst
Tugendhat, 'Die Seinsfrage und ihre sprachliche Grundlage', in Philosophische Rundschau 24 (I977), 164.
13 Logica 'Ingredientibus', ed. by Geyer, p. 351, cited with other passages from Abelard
in 'On the Terminology for Copula and Existence', in Islamic Philosophy and the Classical
Tradition. Essays presented . .. to Richard Walzer (Cassirer, 1972), pp. 146 - 149.
3
24
CHARLES H. KAHN
14 For the notion of truth claim, see 'Be', pp. 186f; 'Theory of the Verb', pp. Ilf. Compare
Quine's statement: "Predication joins a general term and a singular term to form a
sentence that is true or false according as the general term is true or false of the object, if
any, to which the singular term refers" (Word and Object, p. %). This makes clt:ar the
sense in which predication is more than a syntactic notion.
15 For the stative-mutative contrast see' Be' , pp. 194 - 198, following Lyons, Introduction
to Theoretical Linguistics, pp. 397ff. Compare Benveniste, Problemes de linguistique
generale, p. 198: "etre' ... est ... un verbe d'etat, ... est meme par excellence Ie verbe
d'etat" .
The durative aspect emerges as a distinct "sense" of the verb in the Type I ("vital") use
with persons, where einai means "continue (in life), survive": eti eisi "they are still alive",
theoi aiei eontes "the gods who live forever". See 'Be', 241 ff.
16 For the verbs of posture as static be-replacer, see' Be', pp. 217 - 219.
17 See my own exposition of this view in 'Be', pp. 225f, 375 - 379, with the work of J.
Klowski cited there (p. 375n.) from Archiv fur Geschichte der Philosophie 49 (1967), 138ff.
18 'Be', pp. 201ff.
19 Cf. Thucydides VII. 8.2: tpo{3ov,uvos 010 1'17 oi 1rtl'1rO,uVOL ... ou Ta DVTCl Cx1ra-y-yi>V.w(lIv
"(Nicias) fearing that this messengers might not report the facts (sent a written letter)".
For additional examples, see 'Be', pp. 335 - 355. The veridical "is" appears in
Shakespeare, e.g., King Lear IV.vi.l41: "I would not take this from report: it is,/ And my
heart breaks at it". The idiom is still alive and well in contemporary speech: "Tell it like
it is" .
20
21
25
33
26
CHARLES H. KAHN
27
poses that knowledge of the truth is possible is too weak to support Parmenides' argument,
since "a sceptic might well respond, that no one knows anything, precisely because there
is no truth to be known" (ibid). This objection seems to me to mistake Parmenides for
Descartes. Why should Parmenides be thought of as arguing against a sceptic? His argument is about how to get to the truth and what one will find there, not about whether there
is any such thing.
Gallop's third objection is more substantial: if we start by rejecting "what is not the
case" as an object for knowledge, how do we get to the rejection of non-existence that is
required to disprove generation and perishing? (ibid., pp. 67 and 72). I agree that we must
find in to me eon a sense of not-being which is equivalent to "nothing at ail", and if this
is what is meant by an existential eon, then Parmenides' Being must be existential. But, as
suggested above, if to eon as a determinate state of affairs is understood to contain or imply
a real ("existing") subject and definite attributes, and if the negation (to me eon) is
understood as denying everything contained in or implied by to eon, then "what is not"
must be construed not as a well-defined, unrealized state of affairs but rather as a blank
non-entity: no subject ("what does not exist") with no attributes ("is not P' for every Fl.
It is obvious that Not-being so understood must turn out to be not only unknowable but
indescribable. Plato will defuse Parmenides' argument precisely by distinguishing this
hopelessly unqualified Not-being from the more precisely defined not-being-F for various
values of F.
46 The most enlightening explanation known to me for the easy shift from propositional
to existential and copulative construals of einai in Greek 'philosophy is the notion of a
"predicative complex" proposed by Mohan Matthen in an unpublished paper. Matthen
defines a predicative complex as "an entity formed from a universal and a particular when
that particular instantiates the universar'. Thus artistic Coriscus is such an entity, which
"exists when and only when Coriscus is artistic". In grammatical terms, a predicative complex (or rather, its linguistic expression) is the attributive transform of an ordinary copula
sentence: corresponding to X is Y we may assume the existence of a logically equivalent
predicative complex, the YX exists. Thus for (I) Socrates is healthy we have the corresponding (2) The healthy Socrates exists, where the truth conditions for (I) and (2) are assumed
to be identical. Furthermore, truth conditions will also be the same for the veridical
transform of (I), namely (lA): It is the case that Socrates is healthy. Aristotle in Met. 1!t..7
slides effortlessly between (1) and (lA). (See 1017" 33 - 35, as interpreted in Phronesis
(l9SI), 106f) Now if (I) is transformed as (2), we see how the copula-veridical-existential
slide can seem so natural in Greek, since all three formulations are logically equivalent. I
believe this construal (following Matlhen) captures something quite deep, and quite strange
to us, in the use of einai by the Greek philosophers. And it shows why our conventional
dichotomy between existence and copula imposes a choice upon the interprett:r which corresponds to nothing in the Greek data. Also, our difference in "logical form" between propositional (fact-like) and substantial (thing-like) entities as subject of einai will reappear
in this conception simply as a difference in formulation between (I A) and (2).
47 For the double construal of on at Symposium 21 IA I, see Phronesis (l9SI), lOR.
48 First in an idiomatic variant at 65D 13 - E I, then progressively from 74B 2 to 75D and
7SD. See Phronesis (1981), 109- Ill.
49 For a convergence of veridical (Being versus Seeming) and "existential" values (Being
versus Becoming), see Rep. VI, 50SD 5 - 9, where to on is first paIred with atetheia, as
radiating the light of rational cognition, and then contrasted with to gignomenon te kai
apollumenon, the source of darkness and inconstant opinion. This is of course compatible
with a slightly different contrast in the following section (50SE - 509B), where the role of
28
CHARLES H. KAHN
the Good as cause of truth and knowledge is distinguished from its role as source of Being
(to einai te kai ten ousian) for the objects known. There Being for intelligible things is
presented as parallel to generation (genesis) and growth for visible things and must refer
to the stable existence of the Forms as appropriate objects for knowledge. (Shorey renders
einai ... ousia as "existence and essence", thus recognizing that both ideas are in play;
but it would be a mistake to look for any fundamental difference in sense between the verb
and the noun. At 479C 7 he renders ousia alone as "existence or essence", again rightly.)
so See, for example, the double syntax of aei on cited in Note 47. The stative, and hence
potentially predicative rather than strictly existential force of expressions like (TO) ov ad
is clearly indicated by the alternative formula with hELv: e.g. TO aft Xo/Ta
TO/UTa WUo/VTWS fXOV for Forms as object of knowledge at Rep. VI, 484B 4 (cf. Phaedo 79A
9, 80B 2, etc.), which is immediately picked up by TO OV with veridical overtones (T~ ovn)
at 484C 6: "those who are veritably deprived of the knowledge of the veritable being of
things" (Shorey). Thus at 485B;2 ExfLVT/ ;, ouuLO/I;, ad oJu<x is first of all the object of the
philosophical eros for knowledge, and at the same time contrasted with "ousia" that is made
to wander by coming-to-be and passing-away". Of course this convergence of veridical and
stative-existential values for einai is systematically motivated by Plato's theory: the Forms
are reliable objects for knowledge and truth just because they are eternally invariant. (And
in technical contexts this invariant being-what-they-are will be expressed by strong copula
uses such as to ho esti.) It is just this union of" true reality" plus "eternally stable reality"
that is conveyed by to on and ousia throughout Rep. VI - VII, e.g. (following on the
passages just cited) "the spectacle of all time and all ousia" at 486A 8, thought which is
naturally led t1f't rilv TOU OVTOS iOi<xv lxaurov at 486D 10, the soul which is to have an adeQuate and complete grasp of to on at 486E 2.
Dept. of Philosophy,
University of Pennsylvania,
Philadelphia, PA 19104, U.S.A.
BENSON MATES
Among the Platonic statements that have most agitated his commentators, from Aristotle's time down to the present, are those in which he
seems to be saying (and with great confidence, too, as though there were
no question about it) that beauty itself is beautiful, justice itself is just,
largeness is large, piety is pious, and the like. On the one hand, these
statements are considered by many to involve some sort of categorymistake or serious ambiguity: beauty itself, they say, is not the sort of
thing that can be beautiful, at least not in the same sense in which people,
statues, paintings, or pieces of music are beautiful. And likewise with
justice itself, largeness itself, and the other Ideas. On the other hand,
though, there is the awkward fact that these so-called "selfpredications" cannot be lightly dismissed as mere lapsus linguae on the
part of our author, for they seem essentially related to his doctrine that
each Idea is a paradigm or perfect exemplar for the particulars that fall
under it; beauty itself is said not only to be beautiful, but to be the most
beautiful thing of all.
In recent times this situation has been analyzed on the basis of the
assumption that the verb "to be" has at least two senses, viz., the
predicative sense, as in "Socrates is human", and the identity sense, as
in "Socrates is the husband of Xanthippe". Plato's critics castigate him
for being unaware of the distinction, while his defenders believe that he
was perfectly well aware of it and that the allegedly self-predicative
statements are to be understood as assertions of identity. In this paper
I wish to investigate the possibility that the assumption is false, and that
consequently neither the attacks nor the defenses that are based upon it
are well-founded. J
1. THE THIRD MAN ARGUMENT
29
s. Knuuttila and J. Hintikka (eds.), The Logic oj Being,
1979 by Van Gorcum, Assen, The Netherlands.
29-47.
30
BENSON MATES
Now the first thing to observe here is that the point of Parmenides'
argument is not, as has often been erroneously said, that certain assumptions lead to an 'infinite regress' (for there is nothing per se wrong with
an infinite regress, anyway), but simply that Socrates' admissions are inconsistent with the principle:
(1)
31
"Only a part."
"Then the forms themselves would consist of parts, 0 Socrates, and the things participating in them would participate in parts, and in each of them there would no longer
be the whole but only a part of each form."
"So it seems."
"Are you willing, then, Socrates, to say that our one form really has parts and still is
one?"
"Not at all", he said.
form "A is one", mean for Plato? In my opinion this is a very difficult
question, which has never been satisfactorily answered and can be approached only through a careful study in which one would notice, among
other things, what conclusions Plato is willing to draw from such
statements and what other statements he regards as implying them. I
presume, with most others who have considered the passage we are
studying, that when Plato says that each form is one, he does not intend
merely to express the apparent triviality that each form is one form. In
particular, I presume that (3) does not mean that the form of largeness
is one form. On the other hand, the best I have to offer toward an ac-
32
BENSON MATES
count of what these statements do mean is the claim that part of the
meaning of, e.g. (3), is this:
(4)
It is not the case that there are two different forms Fand F',
such that something is large by virtue of F and something is
large by virtue of F' .6
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
33
may be, it is clear that there is indeed a fairly conspicuous gap, which
cannot be filled by the addition, as a "suppressed premise", of any thesis
for which Plato argues elsewhere. This gap is at the point where it is concluded that "then there is another form by which all these are large";
there is no apparent reason why the first form cannot be one of the things
that are large by virtue of it, i.e., cannot be large by virtue of itself. Of
course, everything here depends on the sense of "by virtue of", or, more
precisely, on that of the datives we thus translate. If, to mention but one
possibility, 'x is <p by virtue of ex' is analyzed along the lines of 'x is <p and
if there were no such thing as ex it would be impossible for x to be <p',
Platonic doctrine would justify not only such assertions as "The Mona
Lisa is beautiful by virtue of beauty itself" but also "Beauty itself is
beautiful by virtue of beauty itself". 12
The jump from "there is again a form by which ... " to "there is
another form by which ... " is, in my opinion, the only gap in
Parmenides' argument. All other aspects of it are consonant with Plato's
views and should pass inspection by logicians. 13
Now many influential commentators, from ancient times down to the
present, have in effect located the difficulty at a different place, namely,
at the point where it is assumed that largeness is large. Aristotle says that
the proof that there is a "third man" distinct from Man and from individual men rests on the fallacious assumption that "Man", like the
proper name "Callias", denotes an individual substance, whereas in fact
every such general term denotes either a quality, or a relation, or a quantity, or something of that kind. 14 Applied to the argument as given in our
passage, this evidently amounts to the claim that largeness is not the sori
of thing (i.e., an individual substance) that can be large.
In model n times Russell has made essentially the same point, using
much more drastic language:
In the first place, PlalO has no understanding of philosophical syntax. I can say "Socrates
is human," "Plato is human." and so on. In all these statements, it may be assumed that
the word "human" has e.xactly the same meaning. But whatever it means, it means
something which is not of the same kind as Socrates, Plato, and the rest of the individuals
who compose the human race. "Human" is an adjective; it would be nonsense to say
"human is human". Plato makes a mistake analogous 10 saying "human is human". He
thinks that beauty is beautiful ... He fails altogether to realize how great is the gap
between universals and particulars ... He himself, at a later date, began to see this difficulty, as appears in the Parmenides, which ~ontains one of the most remarkable cases in
history of self-criticism by a philosopher.1 5
And not only Plato's critics but also the more sympathetic commentators have problems with his assertions that largeness is large, beauty is
34
BENSON MA TE~
Of
35
(b)
36
BENSON MATES
be true of those sticks; and whenever either of the latter is true, the
former will be true. But obviously this does not suffice to show that "is
no longer than" is here ambiguous, having sometimes the sense of
"matches" and sometimes that' of "is shorter than". When I say 'A is
no longer than B' of a couple of sticks that happen to be equal, I am using
the phrase "is no longer than" in exactly the same sense as when I apply
it to a couple of which the first is shorter than the second. Note further
that to say 'A matches B' amounts to saying 'A is no longer than B' and
something more; and likewise for 'A is shorter than B'. If this 'something
more' were obvious from the context, I could communicate the fact that
two sticks match by simply stating the first component of the conjunction. Thus, if the context makes 'B is no longer than A' obviously true,
I can, as a practical matter, employ simply 'A is no longer than B' to convey the information that the two sticks match. But, again, this would not
show that "no longer than" sometimes means "matches" and the rest
of the time means "is shorter than".
To spell out the intended analogy between "is no longer than" and
"is" is probably unnecessary, but I hope that the reader will forgive my
doing it anyway. The point is that perhaps the "is" of identity and the
predicative "is", so-called, can both be defined in terms of a more
primitive "is", in a manner similar to that in which "matches" and
"shorter than" were defined above in terms of "no longer than". In fact
Leibniz 21 and, if I am not mistaken, certain Polish philosophers beginning with LeSniewski,22 have done just that. Leibniz defines 'A is the
same as B' as 'A is Band B is A', and 'A is (a) B' as 'A is B but B is not
A'. Analogously to the situation with the sticks, we have the result that
whenever 'A is B' is true either 'A is the same as B' or 'A is (a) B', but
not both, will be true, and each of the latter implies the former. Thus,
the fact that in "Scott is the author of Waverley" we can replace "is"
by "is the same as" and get a true sentence, while if we replace "is" by
"has as a property" we get a sentence that is false or nonsensical, in no
way shows that in this sentence "is" means "is the same as". We can also
carryover the point about what happens when the truth of one of the
conjuncts is part of the background infQrmation or is in some other way
too plain to need stating.
Leibniz was defining "same" in terms of "is" for a sort of regimented
Latin, where (because of the lack of a definite article and because of certain features of the regimentation) the indicated types of transformation
work better than they do in English. I do not wish to claim that in Plato's
Greek tad behaves in relation to TD:VTOIJ in exactly the way Leibniz sug-
37
gests for est and idem. Nevertheless the relation may well be similar
enough to justify suspicion that the sort of evidence usually adduced in
support of the multiple sense hypothesis for f.aTL does not at all rule out
the possibility that that verb may be used in a single sense everywhere.
We shall return to this matter in connection with (i) - (vii) below.
In determining whether a word or other expression has more than one
sense, the unwary may be tempted to make still other fallacious inferences. In modern introductions to logic, for example, one often finds
it said that there are two senses of the connective "or"; the "exclusive"
and the "inclusive" senses. Sometimes, we are told, "or" is used in a
sense that excludes the possibility that both disjuncts are true, while in
other occurrences it has a sense that allows such a possibility. (Then one
is usually informed that for reasons of simplicity, etc., logicians have
decided to use the word, or a corresponding symbol, in the inclusive sense
only: a disjunction counts as true if and only if at least one of the disjuncts is true.)
Now it turns out that finding indisputable cases of the exclusive sense
of "or" in the natural language is not quite so easy as might be thought.
If I tell you that I shall either go to the concert or stay home and read
a good book, it is clear enough that I am not allowing the possibility that
I might both go to the concert and stay at home; but it is also clear that
we do not need to postulate an exclusive sense of "or" to account for
the exclusion, for the content of the disjuncts suffices to eliminate the
possibility that both might be true. (Note that even after the logician has
given his "inclusive" sense to the symbol "v", he uses it, without change
of sense, in disjunctions like' P v -P' , where it is impossible that both disjuncts be true). So, in order to have critical cases before us we must look
for disjunctions which are such that (a) the whole disjunction will be considered false if both disjuncts are true, and (b) it is at least possible that
both disjuncts be true. But even in these cases we must beware of such
contribution as the context or background information may make to the
inferences the hearer will draw from the disjunction. For example, if my
daughter has been expressing a wish to go to the concert and also to buy
a recording of the symphony that will be performed there, and I have
responded that it's certain we cannot afford both of these, and even
doubtful whether we can afford either, then, when I finally say, "All
right, you may go to the concert or buy the record", it will be obvious
to her that the possibility of both is excluded. But again the responsibility
for the exclusion need not be pinned on the "or"; rather, it seems more
properly attributable to the background information. In short, the fact
38
BENSON MATES
39
Let us next consider assumption (b), that there is something wrong with
Platonic sentences like "Beauty is beautiful". Why are so many
philosophers and other scholars ready to teJl us that such sentences, if
taken literaJly, are "sheer nonsense"? It seems that the principal reason
- and this is surely paradoxical - is that the Platonic metaphysics has
been swaJlowed, hook, line, and sinker, and has then been interpreted in
such a way as to rule out part of itself. That is, one first agrees that beauty (or, let us say, Beauty) is an abstract entity, eternal, changeless, existing or subsisting in a world apart, while particular beautiful things alf
belong to the world of sights and sound, and, indeed, are beautiful
precisely because of how they look, sound, or in other ways affect the
senses . Then one infers that things so utterly different as these, belonging
even to different categories (whatever that means), cannot have attributes in common; e.g., that neither "is beautiful", "is good", nor any
other predicate can be true of the abstract entity Beauty if taken in the
same sense in which it is true of particular concrete objects.
These notions are occasionaJly reinforced by the mistaken idea that
unless we subscribe to some sort of theory of types, which would declare
it nonsensical to attribute a property to itself, we shaJl inevitably faJl into
Russell's Antinomy and related contradictions. But, as is well known,
type theory is not the only device, nor even the preferred one, for
avoiding the fundamental antinomies; so that if Plato wishes to make
statements like "Beauty is beautiful" he is thus far in no particular
danger from the side of logic.
Plato formulates his puzzling reflexive assertions in various ways. The
most common of these is of special interest. Instead of using a standard
abstract noun in the subject position, he employs the adjective with the
article, thus producing what appear to be literal counterparts of the
English sentences "the large is large", "the beautiful is beautiful", "the
just is just", "the holy is iJoly", etc. What do these statements mean?
40
BENSON MATES
41
42
BENSON MATES
(iii)
'A is the same as B' ('A = B') is true if and only if 'A is B'
and 'B is A' are true;30
'A isp B' ("Pauline predication") is true if and only if (a) for
all terms D, 'D is B' follows from 'D is A' and (b) for some
terms D, E, 'D = E' does not follow from 'D is A and E is
A',3l
'A is (a) B' ("ordinary predication") is true if and only if 'A
is B' is true but 'A = B' and 'A isp B' are not.
'A is' is true if and only if, for some term D, 'A is D' is true;32
If Plato's usage is more or less along these lines, then we can expect that,
for him, whenever 'A is B' is asserted, then 'A is the same as B' or 'A
isp B' or 'A is (a) B' could also be asserted. Of these three, the last is inconsistent with each of the first two, though the first two are consistent
43
with each other. As examples of sentences that come out true according
to the above scheme, we have "Socrates is identical with the teacher of
Plato", "Socrates is a man", "The just isp good", "The just is eternal"
(but not "Justice isp eternal"). In each case, the corresponding sentence
with the primitive "is" will be true; i.e." "Socrates is the teacher of
Plato", "Socrates is a man", "The just is good", and "The just is eternal" are all true in the same sense of "is". We shall also have such results
as that if "The statue is large" is true, then "The statue is similar to the
large" and "The large is similar to the statue" will also be true.
I hasten to acknowledge, however, that the matter is very much more
complex than these suggestions might indicate. A more satisfactory account would at least replace (i) - (vii) above by corresponding principles
for Plato's Greek, and difficult problems of word order and the placement of the article would have to be dealt with. Still further complications will result from Plato's use of the abstract noun and other expressions as apparently synonymous with the corresponding adjective-plusarticle. So the most that can be claimed for the above scheme is that it
shows one way in which the copula could be used univocally everywhere
and yet give rise to the kinds of texts that have made scholars consider
it ambiguous. 33
Returning in conclusion to the Third Man Argument, we may note that
it is fortunate for Plato that there is another way out besides that of
declaring that "is large" is ambiguous. For, as has often been noted, that
sort of ambiguity would render almost unintelligible his important doctrine that the particulars are likenesses of their corresponding ideas. The
text most clearly illustrating this is in the Symposium,34 where Socrates
describes a hierarchy of beautiful things; there are beautiful bodies, but
more beautiful than these are the beautiful souls, and the beauty of the
laws and of the various branches of knowledge ranks still higher. Most
beautiful of all, he says, is beauty itself. Then he goes on to explain in
detail exactly why beauty itself is more beautiful than anything else. 35
Unlike the other beautiful things, it is eternal, neither corning to be nor
passing away; unlike them, it is not beautiful in one respect and ugly in
another; nor beautiful from one point of view and ugly from another;
and so on. If the predicate "is beautifui" were not used in a single sense
throughout this comparison, the passage would be very dark indeed; for
to say that beauty itself is more beautiful than a beautiful soul, but in
a different sense of "is beautiful", would be like saying that light travels
faster than sound, but in a different sense of "fast".36
Thus Plato cannot very well join those who would save him from the
44
BENSON MATES
14
45
46
BENSON MATES
Platonic statements of the form 'the", is '" are sometimes Pauline, sometimes ordinary. But
it seems to me that whenever 'the", is rp' is asserted in the dialogues it is put forward on
the same basis. To suppose with Vlastos (p. 265) that "justice is just" in the Prolagoras
is Pauline, while "beauty is beautiful" in the Symposium is not (pp. 262 - 3), should be
a last resort; far better to suppose that Plato uses "is" univocally but has not thought out
what to do about the difficult cases.
30 This condition for identity may be too weak. In the Prolagoras, in a discussion initiated
by the question "whether virtue is one, and justice, temperance, and piety are parts of it,
or whether these things that I have just now mentioned are all of them names of the same
one thing" (329C - D), it is concluded from "justice is pious" and "piety is just" that
"justice is either the same as piety or maximally similar to it" (33IB, cpo 333B); and from
considerations indirectly establishing that every temperate act is wise and every wise act is
temperate it is concluded that temperance and wisdom are one (333B). (Cp. Vlastos (1973),
pp. 243 - 6). Following this, Socrates begins what is plainly an attempt to show that
temperance and justice are one; and it looks as though his argument, never completed, was
going to involve establishing that every temperate act is just and every just act is temperate.
Thus, he seems to be trying to show, perhaps only to discomfit Protagoras, that
"wisdom", "temperance", "justice", and "piety" all name the same thi~. Whether he
or Plato actually believed this, is irrelevant; the crux of the matter is whether the course
of the argument sho"'s what he thinks would have to be the case if the various identity
statements were true. However, the references to similarity suggest that perhaps the truth
of 'A is the same as B' requires something more than that of 'A is B' and 'B is A', at least
when A and B are names of ideas.
31 Clause (b) is designed to eliminate the possibility that A is a name or description of a
particular. Otherwise, since e.g. "If anything is Socrates, then it is eo ipso a man" is true,
we should have "Socrates is a man" as a Pauline predication.
32 Thus two possibilities suggest themselves for 'A is not': (I) for no term B is 'A is B' true,
or (2) for some term B, 'A is not B' is true.
33 As emphasized in the text, there is in general no hope of finding simple, exact rules to
cover the usage of a given author writing in a natural language. The following may help
to indicate at least a significant subset of the Greek examples I seek to catch with clauses
(i) - (vii). In forming substitution-instances of a given clause:
(I) Any adjective, count noun, or proper name, prefixed by the definite article, may be
substituted for a variable in subject position.
(2) Any adjective or noun, with or without the article, may be substituted for a variable
in predicate position.
(3) Where, in the given clause, the same variable occurs both in subject and in predicate
positions, it is to be replaced in subject positions by an expression with the article if and
only if it is replaced in predicate positions by the same expression without the article.
(4) An abstract term (e.g., ~ Ol){atOUUP'I) is interchangeable with the corresponding adjectival phrase (TO M){atOP) or a fun phrase (a fun M){atOP).
Cases involving complex terms may, it is hoped, be treated by analogy with the foregoing
principles.
Some examples:
Of (i): TO OL){CnOp ){aL TO autOp mUToP fun is true iff TO M){atOP amop fun and TO OUtOP
M){atop fun are true. 0 Ew){ear"s ){aL 0 otOaU){aAOS IIAaTwPos mUToP fun is true iff 0
Ew){ear"s otOaU){aAOS IIAaTwPos fun and 0 otOaU){aAOS IIAaTwPos Ew){ear"s fUTL are
true.
Of (ii): TO M){atOP Cx-yalJol' fun (as a Pauline predication) is true iff (a) for all terms D,
47
'D &."(CiIJOV Eun' follows from 'D ot)(CltOV Eun' and (b) for some terms, D, E, 'D = E'does
not follow from 'D ot)(CltOV Eun )(Cit E o[)(CltOV Eun'.
Of (iii): TO O[)(CltOV &')([V'1TOV Eun (or ~ O!)(CiWUUV'1 &')([V'1TOV Eun as an ordinary
predication, is true iff TO ot)(CltOV &')([V'1TOV Eun is true but not as a Pauline predication,
and TO ot)(CltOV )(Cit CiUTO TO &')([V'1TOV TCiUTOV Eun is not true.
34 Symposium, 210fr.
35 Op. cit., 21IA-B.
36 Perhaps I should state explicitly that, according to the interpretation I am advancing,
the sentences "Beauty is beautiful" and "The soul is beautiful" are both true fOf Plato
when the "is" is taken in what I am calling its "primitive" sense. When it is taken in the
identity sense, only the first sentence is true, and when it is taken in the sense of "ordinary"
(to us) predication, only the second sentence is true. As I have indicated in (v) and (vi),
"The soul resembles Beauty" will be true on the basis of the two sentences mentioned, with
"is" understood in the primitive sense in both cases.
BIBLIOGRAPHY
Brugmann, K. and E. Schwyzer: 1966, Griechische Grammatik, 3rd ed., Munich.
Cherniss, H.: 1957, 'The Relation of the Timaeus to Plato's Later Dialogues', American
Journal of Philology 78, 225 - 66.
Cornford, F. M.: 1939, Plato and Parmenides, Routledge & Kegan Paul, London.
Couturat, L.: 1903, Opuscules et fragments inedits de Leibniz, Presses Universitaires de
France, Paris.
Geach, P. T.: 1956, 'The Third Man Again', The Philosophical Review 65, 72 - 82.
Goodwin, W. W.: 1963, A Greek Grammar, London.
Kahn, c.: 1973, The Verb 'Be' in Ancient Greek, D. Reidel, Dordrecht.
Kiihner, R.: 1898-1904, Ausfiihrliche Grammatik der griechischen Sprache, Hannover
and Leipzig.
Luschei, E. c.: 1962, The Logical Systems of Lesniewski, North Holland, Amsterdam.
Quine, W. V.: 1947, Mathematical Logic, Harvard University Press, Cambridge.
Ross, W. D.: 1951, Plato's Theory of Ideas, Clarendon Press, Oxford.
Russell, B.: 1903, Principles of Mathematics, Cambridge University Press, London.
Russell, B.: 1945, A History of Western Philosophy, Simon & Shuster, New York.
Schmidt, F.: 1960, Leibniz: Fragmente zur Logik, Akademie-Verlag, Berlin.
Smyth, H. W.: 1956, Greek Grammar, Harvard University Press, Cambridge.
Vlastos, G.: 1973, Platonic Studies, Princeton University Press, Princeton.
Wed berg , A.: 1955, Plato's Philosophy of Mathematics, Almquist & Wiksell, Stockholm.
Wittgenstein, L.: 1922, Tractatus Logico-Philosophicus, Routledge & Kegan Paul,
London.
Dept. of Philosophy,
University of California, Berkeley,
Berkeley, CA 94720, U. S.A.
RUSSELL M. DANCY
Aristotle 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
49
S. Knuuttila and J. Hintikka (eds.), The Logic oj Being, 49-80.
1986 by D. Reidel Publishing Company.
50
RUSSELL M. DANCY
Ie's reflections on Plato's mistakes, and fitting it into Aristotle's essentialism. This occupies Section 1 below.
The theory is, in the first instance, a theory about the Greek word
'lvo:t' whose bearing on the English translation, 'to be', of that word
is not obvious. But I think that, mutatis mutandis, it does have
something to tell us about Anglo-Saxon being, and, ultimately, perhaps,
something about Being (or even Sein: but I do not want to commit
sacrilege). Accordingly, after stating the theory, I try to mutate the things
that need mutation to bring it to bear on English. Here I hope to irritate
any Heideggerians that are still with us. I also complain a bit about the
theory's fringe benefits, and it is probably at about this point that my
disagreement with Owen begins to set in (but I shall not be pursuing this
disagreement beyondproviding relevant footnotes). At any rate, I make
as many enemies as I can.
Lastly, although it is a bit of a letdown after twitting the Heideggerians
and so on, I turn to some texts with which this theory is to help us. On
this occasion I consider only two: 4 one in De interpretatione 11 that
famously conflicts with a passage in Categories 10, and one in
Metaphysics A 7, the chapter on 'being' in Aristotle's 'dictionary'. Even
with these texts, I shall be more cursory than the subject ultimately will
allow. But brevity may at least yield clarity of outline, and in a
preliminary statement such as this, that is perhaps best.
1. WHAT IT IS TO BE IN GREEK
The Greek that Plato and Aristotle wrote and probably spoke does not
possess separate words, as English does, for 'to exist' and 'to be,.5 Firstyear Greek textbooks 6 tell the student that the single word EaTtV is to
be accented differently depending on whether it means 'exist' or just 'is',
but this orthographic convention, known as 'Hermann's rule', has no
foundation in the writing of Greek or in what we can tell of its pronunciation in ancient times, and may have no foundation before 1801, when
Hermann ruled it 7. The Greek word l:~{aT(Xat'}o:t, which, in the sense 'to
be separated from', 'to stand out from', or just 'to stand out', is destined
to travel through its Latin cognate into French and thence to appear in
English as 'to exist', is not used to mean 'exist', at least, not in the time
of Plato and Aristotle.
So, for Plato and Aristotle, the claim that Socrates exists would be
represented by a sentence shaped like this one:
(1)
Socrates is. 8
51
Now, in fact, the sentence 'Socrates exists' has a peculiar ring to it. It
helps here to change tenses and subjects: "So Prester John really existed,
after all". Here the extra words, the tense, and, most importantly, one's
background knowledge about the use of the quasi-proper name 'Prester
John' help to give an air of realism to the enterprise. But for starters, it
is easier to ignore friction and wind resistance; anyway, the theory itself
will involve us in restoring some of the realism lacking in (1), or in its
more normal but still rather peculiar English equivalent, 'Socrates exists'. In particular, we shall attend to the relevance of background
knowledge. So I shall stick with (1).
In this and subsequent formulations I am deliberately ignoring one
subtlety. It has frequently been said that the English 'existential prefix',
"there is (such a NP as) x", is not used to make existential claims, 'x exists', for a variety of reasons. 9 I do not want to deny this claim. On the
other hand, I take it that, on occasion, the sentences "So Prester John
really existed, after all" and "So there really was such a person as Prester
John, after all" are interchangeable. And that admittedly sloppy
equivalence is enough for present purposes: our focus is not, ultimately,
on the English verb 'exists' anyway ..
So a sentence like (I) translates the English 'Socrates exists', and the
Greek word in it that corresponds to the word 'is' is the same word that
appears in the Greek for these:
(2)
(3)
Socrates is pale.
Socrates is [a] man [i.e., a human being].
52
RUSSELL M_ DANCY
S is P
->
Sis,
->
S is not.
And what makes (E) and its negative counterpart plausible to him is the
idea that the consequent in each case is merely a simplification of the
antecedent, just as in
(P)
->
Socrates is pale.17
For this sort of simplification to work, the residue after the simplifica-
53
tion must have the same significance it had before: its contribution to the
unsimplified original must be just what is left. Failure to meet this requirement is obvious where there is a gross change in the sense of one or
another word, as, perhaps, in
(L)
-+
Socrates is lousy,
-+
Socrates is good.
Here we need not, and, I think, should not, say that 'good' has changed
sense from antecedent to consequent. 'Good' is what Geach once 18
called an 'attributive' adjective: a noun or noun phrase needs to be supplied that will answer the question 'a good what?'. And, although in a
specific conversational context it may be clear that the noun to be supplied is 'cobbler' ("The cobblers in Athens are all terrible." - "No,
Socrates is good.' '), we are talking about justified patterns of inference,
and that means we need a conclusion we can, so to speak, carry away
with us. But once we have carried away 'Socrates is good', the noun we
should supply to complete the sense is no longer 'cobbler', but,
presumably, 'man'.
All this applies to (E) (and to (NCE. At the very least, the 'is' in its
antecedent must be the same 'is' as that in its consequent. In particular,
there can be no shift from an alleged predicative sense of 'is' to an alleged
existential sense of 'is'. Since, as I shall be arguilig in the next section,
there are no such senses of 'is', this is not a problem. But also there must
not be any of the more subtle problems such as beset (G). And here, as
we shall see, trouble may arise. Aristotle is going to abandon (E), as aid
Plato, but in neither case does it have to do with shifting from a
predicative to an existential sense of 'is'. In Aristotle's case, it is not so
much a matter of abandoning (E) as of restricting its range of operation
while preserving the idea that the consequent in the allowable cases is a
simplification of the antecedent, and so contains the same 'is'.
But still, I must part company with those whose distaste for degrees
of existence leads them to read it out of Plato. As I have said, "Socrates
is" is, as far as I can tell, the way a Greek goes about asserting the existence of Socrates. Aristotle himself may be found using this format in
Cat. 10, 13 b 16-17,19, etc. For realism, one must imagine a conversation, say, between Xenophon and Cyrus. Xenophon has been talking
about Socrates, and Cyrus denies that there cuuld be anyone that
strange; Xenophon replies "Oh, Socrates exists, all right; I was at a party
with him just last month". Here a Greek might say (the Greek for) "Oh,
54
RUSSELL M. DANCY
Socrates is, all right; .... " So Plato is, as far as I can tell, committed
to degrees of existence, as we might phrase the matter; when he says that
the forms 'purely' and 'entirely' are, what he means can be put into
English by saying that their hold on existence is absolute because they are
what they are irrefragably, while Helen's hold is relative and weak,
because some of what she is, if not all of what she is, she also is not.
Then what is it that condemns (E)? I have already said something
about that, although in the 'deep' terminology; before putting it more
austerely, there are a couple of obvious problems with it that are not
directly covered by the theory I shall have Aristotle adopting, which
ought to be mentioned so as to avoid raising false hopes.
First, there is a problem recognized and dealt with by both Plato and
Aristotle: the notorious problem of what is not. Unicorns (a modern
favorite) and goat-stags (Aristotle's favorite) do not exist: they are not.
But goat-stags are believed in by some and used as examples by others.
If we apply (E) to that last sentence, we end up having them in the zoo
that is reality. The restriction on (E) needed here is not the one I am
primarily after, although it is close by, in Aristotle's mind. As I see it,
it would involve isolating the predicates, like 'believed in' and 'used as
examples', that give trouble, call them 'intensional predicates', and pronouncing (E) unfit for use with intensional predicates. Aristotle does not,
I think, see it quite that way. But I am not going to worry any more about
that here.
A second interesting problem not covered by the theory to come is provided by certain adjectives such as 'fake', 'mythical', and, particularly
for Aristotle, 'dead'. A fake diamond is no diamond. So we do not want
it to follow from 'that diamond is fake' that that diamond is (exists), and
Aristotle does not want it to follow from 'that man is dead' that that man
is. Here again we may simply label the problem and shelve it: call such
adjectives 'adjectives alienantia' or 'alienating adjectives'; 19 (E) does not
work for alienating adjectives as predicates.
Now what is the trouble with (E)?
As Plato uses it and (NeE) in the Republic, every predicate true of S
contributes equally to its existence, and every predicate false of its
detracts equally from its existence (ignoring intensional predicates and
adjectives alienantia). I shall refer to this as the 'democratic' attitude
toward S's predicates.
For Plato, the democratic attitude is a problem because there are many
things a form is not, yet its existence must be unadulterated. For Aristotle, the democratic attitude is a problem for a different reason.
55
Socrates is a man
is one that holds of the subject 'by virtue of itself', 'in its own right', 'of
itself', or per (or propter) se. These are varying translations of the single
phrase' xed)' a,vro' .20 Let us adopt the first one: the predicate in (3) holds
of the subject by virtue of itself. By contrast, the predicate in
(2)
Socrates is pale
56
RUSSELL M. DANCY
~P
57
'essence' is a curious one,23 and what is of interest here is that its core
is an occurrence of the infinitive 'to be'. So, instead of the conventional
"the essence of each thing", as above, I shall adopt the more literal
"what it is24 for each thing to be". I take this to amount to things like
'what to be is for each thing', 'what each thing's being is, or consists in',
etc. Then the above passage will be retranslated, along with its immediate
sequel, as follows (1029 b 13 -16)
... what it is for each thing to be is what it is said [to be) by virtue of itself. For it is not
[so) that for you to be is for [you) to be educated,zs for you are not by virtue of yourself
educated. Therefore, [it's) what [you are) by virtue of yourself. (lcrTL TO TL ~V
elvw l}(ixcrT~ 0 >.i-YfTW }(CII')' aUTO. ob -yixe lcrTL TO crOL dva, TO JLovcr,)(~ dvw.
-yae }(aTa
cravTov d JLovcr,}(os. 0 aea }(aTa cravTov.)
ou
If we take this to give us different senses of 'to be' for Socrates on the
one hand (for whom 'to be' means 'to be a man') and Bucephalus on the
other (for whom 'to be' means 'to be a horse'), we are committing
another blunder of the type mentioned above and to be dealt with below.
We need not. We might just take it that filling in the predicate tells us
under what conditions 'Socrates is' (i.e., 'Socrates exists') is true, and
leave open the a)ternative possibility (which is, in fact, the right one) that
what determines which predicate to fill in is, not 'is', but 'Socrates'. For
the remainder of this section, I shall take it that this alternative possibility
has not been ruled out, and leave further discussion for the next.
Consider once more the two inferences
(E)
S is P
-+
S is
and
(NCE) S is not P
-+
S is not.
It is not that the essentialist view we have just handed Aristotle requires
him to reject the unrestricted use of (E) (again, leaving intensional and
alienating predicates out of the picture). In the first instance, what he
must reject is its converse, by restricting its application to cases in which
'P' is a predicate essentially true of S: cases in which S is by virtue of itself
p.26 And so he must reject (NeE): only where S is P by virtue of itself,
essentially, does it follow from'S is not P' that S is not.
But all this has, given the background, an indirect impact on (E) as
well. For the justification for the inference (E) was supposed to be simp)e
simplification, the sort of move that justifies
(P)
-+
Socrates is pale
58
RUSSELL M. DANCY
-+
Socrates is lousy.
For (P), that meant that 'pale' had to have the same role in antecedent
and for
(G)
-+
Socrates is good.
(L) and (G) were ruled out on the ground that the simplified residue had
changed character in the course of the simplification. (P) was to be allowed because its residue was constant. And so, if (E) is to survive, it
must show a constant residue.
Consider, then, the instance of (E) that reads
(el)
Socrates is pale
-+
Socrates is.
Socrates is a man
-+
Socrates is.
-+
We should not count 'that pale thing is a man' as a straightforward essentialpredication, and, in fact, Aristotle would not so count it. To the extent that he is willing to count it as a predication at all, he thinks of it
as accidental: he thinks of it as grounded in, or having for its truth conditions, 'that man is pale' (see Section 4 below). And, as we have noted,
in the Posterior Analytics he would prefer to legislate it away: "either it
doesn't predicate at all, or it predicates not simply, but accidentally"
(~TOL p:"ocxp.ws xcxrrnoeeiv, 7} xcxrq'Yoeeiv p.Ev 1'1, a1rAWS, XCXTCx
CTvp.{3e{31/xos oE xcxrq'Yoeeiv, 83 a 15 - 17).
For simplicity of outline, let us allow him the legislation he tries to pass
in the Posterior Analytics. We shall expand our horizon slightly in Section 4 below.
59
Socrates is.
Here we have just 'is', not 'is' + Pred. Such occurrences mark existential
sentences: if there were an existential sense of 'is' in Aristotle, this would
be it. But there is plainly no special existential sense here; there is a merely
syntactic difference between 'to be' followed by a predicate and 'to
be #'. The latter results flom truncating a predication. So even 'to be #'
is, implicitly, to be something.
But, as we have seen, not just any predication may be so truncated:
the predicate must be restorable. Under the simplifying assumption that
such locations as 'that pale thing is a man' can be ruled out, the type of
predication that allows truncation is an essential predication. So while to
be# is to be something, it is not to be any old thing. In fact, (1) is elliptical for
(3)
Socrates is a man.
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
Without the parenthetical material, properly filled in, this does not even
scan.!7 And properly filling in the blank requires some fairly substantial
covering noun or noun phrase: just 'thing', for example, will only work
in quite special cases ("My good man, there is such a thing as decency,
you know", and so on). 'Plato' is the name of a computer language as
well as that of a philosopher, and, for that matter, I could use it to name
my pet cat, my pet theory, my favorite number, the national debt, the
coin Washington skimmed across the Delaware, or the value of that coin
in real terms on Christman Day 1776 given in terms of the Laspeyres index. In the absence of a covering phrase that has some bite to it, we do
not know what is being talked about.
It is not altogether clear, in that case, what pigeon-hole to place the
ignorance in. Some will, no doubt, feel the urge to argue that it is not
linguistically relevant. The pattern for such an argument might be this:
invoke the sense-reference distinction and couple it (although this would
not have made Frege very happy), with Mill's doctrine of proper names.
According to that doctrine (translated into Fregese), proper names do
not have sense. So when someone says 'I'm studying Plato' (or 'I'M
STUDYING PLATO'), and it is not clear to you whether she is trying
to grasp the theory of recollection or the theory of programmed learning,
that does not mean you have failed to get the sense of what she is saying:
you have got all the sense there is to get if you know the rest of her words
and, perhaps, the minimal fact that 'Plato' (or 'PLATO') is a proper
name.
This seems to me one of the many places where the sense-reference
distinction and its alliance with Mill's theory of proper names lead us
astray. It seems to me clear that there is something about the sense of
what she said that we are not getting, and somethin associated with the
proper name in it. 28 But I am not able to layout a proper account of the
matter, at least not here. So I shall simply assert that (4) and (5) require
for their understanding, for grasping their sense, some fairly substantial
covering noun or covering noun-phrase, and admit that I do not know
whether the 'explication' (I shudder at the word) of making sense belongs
to the science of semantics, that of pragmatics, or necromancy.
61
(6)
62
RUSSELL M. DANCY
( 7')
(10')
(13' )
At least, this would generate (7) - (13) from (7') - 13'). And, just
among others, it would generate 'There is such a man as Plato' from
'Plato is a man'.
I said I would not pretend to correctness. And (T) is, accordingly, not
correct. It would generate some things we - or at least I - do not want.
E.g., starting from
(14')
And it is not clear where the stuff represented by 'X' in (T) should go:
for the move from (13') to (13), (T) seems all right as it stands ('next to
me' is carried along with the verb), but with
(15')
63
But for Heidegger, this represents a lamentable loss: apparently, for the
Greeks before Plato, Erven was chockful to bursting with meariing. And
about that, Heidegger is wrong. But fifty percent is not bad.
So far all we have done is to show that the part of Aristotle's theory
that seems at first provincial to Greek may, in fact, have something corresponding to it in English. But, of course, that is not all there is to the
theory. For one thing, there is also its essentialism. This is not, on the
fact of it, a linguistic matter: the doctrine of essentialism is as plausible
or implausible in English, I take it, as it is in Greek. But it scmetimes
sounds as if the doctrine might not be plausible in Chinese,34 or
Nootka,35 or Rortyspeak .36 Here I am going to leave these languages,
along with others of which I am innocent, on one side.
But there is one thing about Aristotle's employment of his essentialism
that must be mentioned, that is closely connected with the presence or
absence of different senses of 'is' in English or in Greek.
To get at this, first consider another distinction that some have alleged
to be pertinent to English and Greek being: that between identity and
predication. 37 There is certainly a distinction here: it is between claims
like
(16)
64
RUSSELL M. DANCY
or
(18)
on the other.(16) states an identity; (17) and (18), I think,38 do not, and
I shall characterize both as predications.
Philosophers are, notoriously, myopic, and when they apply their
magnifying glasses to sentences that do different things, they usually can
only manage to bring into focus a single word. 39 Here the word is 'is'.
They profess to spot the 'is' of identity in (16) and the 'is' of predication
in (17), and then fall to arguing over the character of the 'is' in (18).
I shall call this habit of supposing that every difference in character
from one sentence to the next must be locatable in single ambiguous
words the 'fallacy of the magnifying glass'. We do not need to pull the
word 'is' out of those sentences and go into Angst over the meaning of
being. The situation is completely described by saying that 'is' is followed
in (16) by a singular definite noun phrase, in (17) by an adjective, and
in (18) by an indefinite noun phrase. Sentences that show the structure
of (16) state identities; the others are predications. But this has nothing
to do with the occurence of 'is' in different colors: it is what comes next
that counts. 40
Aristotle does not commit the fallacy of the magnifying glass in connection with 'dvOIL', identity, and predication. 41 But he is prone to the
fallacy; conceivably he is its inventor; probably42 he is its most influential
perpetrator. And if he falls into it anywhere, he falls into it with 'is'. He
thinks that the fact that 'man' and 'pale' relate differently to Socrates
makes for a different 'is' in
(3)
Socrates is a man
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
65
were two varieties of 'is'. In the sequel, I shall occasionally go along with
Aristotle's talking that way.
But I do it under protest. If this way of talking is not to be explained
away merely as a way of talking, but is to be taken as marking a genuine
distinction in senses of 'is', or concepts of being, or whatever, it is a case
of the fallacy of the magnifying glass. It would be a great relief to get
Aristotle off this particular hook. I do not at present know how to do it.
Rephrased in these terms, and still leaving 'that pale thing is a man'
out of account, the theory of Section 1 becomes this. The 'is' of (3) is
a by-virtue-of itself 'is', and permits cancellation of the predicate. The
'is' of the residue, 'Socrates is', is then a by-virtue-of itself 'is', and also
an 'is' &'1I'~Ws, 'is#', since it no longer has an attached predicate. But
these are two different characterizations: roughly, the latter is syntactic
and the former semantic. To go one step farther along the road to perdition, we might say: the existential 'is' is, paradigmatically (see Section 4),
a by-virtue-of itself 'is', but the by-virtue-of itself 'is' is, in the first instance, a predicative 'is', and only becomes existential by cancellation of
the predicate.
3. DE INTERPRETA TIONE II. 21" 25-28.
So Ammonius (in De into 210. 17 - 20)43 rightly says. Aristotle has just
been pointing out, for example, that the argumentative rule we might call
'Addition':
(A)
S is Adj. & S is NP
-+
S is Adj. NP
does not always work: from 'Socrates is good', where we have "simply
'good' " or 'good #' (b1l'~ws ayat'Jos, 21 a15) and 'Socrates is a cobbler',
you cannot use (A) to get 'Socrates is a good cobbler' (20 b 35 - 36).44
Now he notes that the rule we have been calling 'Simplification' also
needs restricting. He is very brief about this: he mentions all three of the
exceptions we listed in Section 1 above, namely things that are not (see
21 a 32 - 33), dead men (see 21 a 23), and inferences to existence based on
the wrong predicate, and it is not easy to see what he is trying to say about
them. I have excused myself from dealing with the first two here. What
he has to say about the third is this.
21 a 18 -21 tell us:
66
RUSSELL M. DANCY
It is true to speak of the particular [more literally, 'of the something'] simply as well, for
example, [to say that] the particular man [is a] man, or the particular pale man [is] pale,
but not always ... (ah'1t?~~ 0' fUTiv fl1f'fiv xCiTa TOU TtVO~ xCii (X1f'hW~, olov TOV Ttva avt?eW1f'OV avt?eW1f'OV ij TOV Ttva hfUXOV avt?eW1f'OV hfUXCIV' oux afi Of. ... ).
It is not clear what the first of these examples is telling us we are permitted to do,45 but the second is plainly allowing us the inference labeled
'(P)' in Section 1 above. 46
Aristotle then points out that simplification will not work with 'dead
man': in cases of this kind, where there is a contradiction between the
two terms antecedent to the simplification, the rule never works. (There
is a puzzle here: apparently Aristotlethinks 'this is a dead man' can be
true, although 'dead man' involves a sort of contradiction.) But in other
cases, sometimes it works and sometimes it does not;
for example, Homer is something, e.g., a poet; well, is he, therefore, or not? For 'is' is
predicated accidentally of Homer; for because he is a poet, not by-virtue-of itself, 'is' is
predicated of Homer. (21 a 25 - 28: wune OJL'1eO~ fUT! Tt, oiov 1f'OI'1rT,~' &e' oOV
XCi! EUTlV, ij ou; xCiTa UUJL(jf(j'1XO~ ,),ae XCiT7J')'OefiTCiI TO EUTIV TOU 'OJL~eOU' OTt ,),ae
1f'OI'1rT,~ iUTlV, aM' OU XCit?' CiUTO, XClT7J,),oefiTClI xCiTa TOU 'OJL~eOU TO EUTlV.)
The traditional way of taking these lines is as saying that it does not
follow from
(19)
Homer is a poet
that
(20
Homer is.
And along with this, often the denial that 'is' by-virtue-of itself applies
to Homer is understood to be the denial that the existential 'is' applies
to Homer (and then Aristotle must have had advanced views about the
'Homeric question').
In my view, neither of these things is correct. That is, to be as explicit
as possible: Aristotle is not denying that (20) follows from (19), and
"'is' by-virtue-of itself" is not the 'is' of existence. 47 I take up these
points in reverse order. Both bring in the theory of Section 1. The latter
is a simple application of it.
The theory of Section 1 tells us that the 'is' of 'Homer is a man' is a
}(cxf)' CXUTO 'is', a by-virtue-of itself 'is', yielding 'Homer is #', and that
this is the existential claim, 'Homer exists'. And the theory tells us that
the 'is' of (19) is a }(CXTa (JUP.{3f{3.,.,}(O~ 'is', an accidental 'is', and so
simplification may not be performed on (19). That is what Aristotle is
here telling us. Since he is giving as a reason for prohibiting simplifica-
67
tion the claim that 'is' is not here used xat'}' aUTO, he cannot mean by
this claim that 'is' is not here used existentially. That, after all, is the
question. He is simply saying: since (19) is not an essential predication,
and its 'is' is therefore not xat'}' aUTO, the predicate may not be canceled
by simplification to yield 'is' a7rhWS, the "existential 'is'''.
This way of putting it yields considerable ground to the fallacy of the
magnifying glass, but it certainly shortens the work.
So much for the second point: simplification cannot be applied to (19)
because its 'is' does not apply to Homer xat'}' aUTO. But then the first
point is clear: Aristotle is not saying that (20) does not follow from (19),
but that it does not follow by simplification. He puts forth the question:
"Homer is ... a poet; well, is he, therefore, or not?" In its context it is
not hard to read this, not as asking "does it follow, by some devious
means or other, that Homer is?", but "does the move we are talking
about apply here?"
This may, at first sight, seem unnecessary. But there are two good
reasons why we should not take Aristotle to be saying that it does not
follow from 'Homer is a poet' that 'Homer is'.
The first is that we know from elsewhere that, when it comes to the
question whether (20) follows from (19), he should say that it does. This
is a point he considers in Categories 10, and quite unambiguously
decides: when there is no Socrates, 'Socrates is healthy', 'Socrates is
sick', 'Socrates is blind', etc., are false, and their negations true. The apparent conflict of that passage with De into 11 is a notorious cruX. 48 In
my view, there is no conflict.
The second reason for avoiding making Aristotle deny the entailment
is that there is a very simple argument in favor of it, based in part on the
De into passage itself. Homer is a poet. But poets are, after all, human;
so they are men. So Homer is a man. But there the 'is' is a xat'}' aUTO
'is', and we can simplify. So Homer is. This argument is, admittedly,
phrased in terms of my interpretation of the passage. But there is very
little about it that is speci fie to that interpretation. On virtually any
understanding of the passage, the inference has to fail because the 'is' of
(19) is not xm'}' aUTO applied to Homer, where the 'is' in 'Homer is a
man' is applied to him xat'}' aUTO. So all we need is the concession that
poets are men, and we are away. It is extremely difficult to see how tha!
could be denied. (It may be worth noting that it would not help to say:
well, for all (19) has to say, Homer might have been a god. Then we
would have another alternative, all right, but it would get us to the unwanted conclusion just as easily.)
68
RUSSELL M. DANCY
e8 -
e.g., we say the just is cultivated, the man [is] cultivated, and
the cultivated [is a] man.
We may write:
(21)
(22)
(23)
The English is pretty awful. The bracketed material has nothing to correspond to it in the Greek: it serves to remind us that, where in English
the adjective 'cultivated' occurring as predicate is easily identified as a
predicate because it is an unsupplemented adjective, in Greek the
predicate adjective '/lovau(os' needs no supplementation in order to be
treated as a noun phrase.
These examples are to be construed as making reference to particular
people in each case, the fellow holding up that lamppost over there, say:
Aristotle is not talking about a maxim he and his friends like to utter to
the effect that the just man is a cultivated man.
69
nO - 11):
This imports the needed 'to be'. Aristotle is plainly not saying that (24)
shows accidental 'to be', but its parapnrase, which he seems to think is
more fundamental, does. So even (24) rests for its truth on accidental
being.
Aristotle's comment on (24) leads him to state the truth conditions for
his examples: the general form for aJl, including the paraphrase for (24)
(but not (24) itself) is "that this is this signifies that this is-accidental to
this" (al3 - 14: TO -rae TOOf: dvcn TOOf: aT/p.CtLVH TO aup.{3f:{3T/XEvcn Tct>O(;
TOOI':). More particularly, in cases like (21), both terms, 'just' and
'cultivated', are-accidental to the same thing e16) - namely, it appears,
to the existent thing designated by the subject term; and in cases like (23),
the subject term is-accidental to the predicate term (al7 - 18).
These are paraphrases of sentences employing 'is' in predicative position. They do not themselves employ 'is' in that role, but in one of them,
the paraphrase for (22), 'is' does occur. When Aristotle :.ummarizes in
a19 - 22, it becomes clear that this occurrence is important:
Well then, things said to be accidentally are so said either because both belong to the same
[thing that) is, or because that [i.e., the predicate-term)so belongs to [a thing that) is, or
because that, to which that of which it is predicated belongs, is.
These are occurrences of 'is # ': the claims are existential claims, and they
are taken to justify the occurrence of 'is' in the predications (21) - (23).
He is assuming stating (21) commits one to the existence of something
both just and cultivated. And presumably he thinks that this requires the
existence of, say a man who is both just and cultivated. 51
In a 18 - 19 he had made an incidental comment that pointed that same
way: just as (23) is aJlowed because the cultivated belongs to the man,
so also the not pale is said to be, because that to which it is-accidental is (OUTW Of 'Af-yf701L
xed TO p.~ 'Awxov dvOlL, OT!
c;,
70
RUSSELL M. DANCY
So
(25)
Socrates is a man.
Socrates is # .
(I) is to follow from (3) in the prescribed way. Assume that it is Thursday
night, after Socrates' appointment at the tanning parlor, and
(26)
The using this and something like Leibniz' Law, we might get (25) from
(I). If we did, the 'is # ' of (25) would be that of (I), and so, ultimately,
that of (3), and so, again, a )Cod)' cxilTo 'is'.
But the immediately preceding context suggests an alternative. The
parenthetical comment of al8 - 19 is attached to a statement of the truthconditions for (23), 'The cultivated one is a man': this is so because the
cultivated is accidental to the man, and so, Aristotle adds, even the not
pale is said to be.
That sounds as if Aristotle had the following derivation in mind. Start,
as before, with (3), and assume (26); this yields
(27)
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)
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 accidental 'is' there is a by-virtue-of itself 'is'.
What, then, about this latter 'is'? Aristotle has this to say
(1017 a 22 - 27):
As many things are said to be by-virtue-of themselves as the figures of predication signify,
for in as many ways as [they) are said, in so many ways 'to be' signifies. So, since of things
predicated, some signify what [itl is, some what-[it)-is-like, some how-big, some relative-to
what, some to do or to undergo, some where, some when, 'to be' signifies the same things
as each of these. (xali' aUTa Of dVaL hf-YfTat ouang Uf/p.aLVft Ta C1X~p.aTa rij5
xarrnogLa5' OUC<XW5
-yag
hi-YfTat,
TouaUTaXW5
TO
dvat
Uf/P.C.LVft.
inL
ouv
TWV
Of
About the only thing that is agreed on heie 52 is that Aristotle thinks that
'is' varies somehow from one category to the next.
Our theory tells us that the 'is' he has in view is that occurring in essential predications, 53 the paradigmatic source for occurrences of 'is #'. So
whatever variation there is will show up in existence-claims as well.
But what essential predications are at stake? The passage recalls
another, Top. A 9. There Aristotle lists his categories, naming the first
one 'what [it] is', as here, and then says (103 b 27 - 39):
It is clear from them S4 that one who signifies what [it) is sometimes signifies [a) substance
sometimes [a) how-big, sometimes [a) what-like, sometimes one of the categories
[xaTfl-yogLwV). For when, with a man set out, one says that what is set out is a man or an
animal, he says what [it) is and signifies a substance; when, with a pale color set out, he
says that what is set out is pale or a color, he says what it is and signifies [a) whatlike. And,
similarly, if, with a magnitude of a cubit set out, he says that what is set out is a magnitude
of a cubit, he says what it is and signifies [a) how-big. And similarly in the other [cases):
each such [term), both if it is said about itself [lav Tf aUTO neL aUToii) and if its genus is
said of it, signifies what [it) is, but whenever [it is said) about another [thing). it signifies,
not what [it) is, but how-big or what-like or one of the other categories.
[ouuLav),
Here Aristotle unhesitatingly moves from the point that "all premisses
signify either what [it] is or how-big or what-like or one
of the other categories" (b25 - 27), where the sentences under consideration are ones like our old friends (3) and (2) and
(1I'QoTauHs) . ..
(28)
72
RUSSELL M. DANCY
and
(31)
The sentences on this latter list are answers to 'what is it?' questions;
those on the former list are answers to various question, 55 one of which
is 'what is it?' (asked of the entities that Aristotle will call 'substances,).56
In the 'what is it?' list, the question is cut free from the first category,
that of substances, and ranges over the entities that are signified in
sentences on the variable list. (3), 'Socrates is a man', occurs on both
lists. 57 It shows an 'is' by-virtue-of itself, as we have been saying. So do
the other sentences on the 'what is it?' list. Here Aristotle describes those
sentences by saying that in each something is either "said about itself"
or "its genus is said of it"; the former terminology is familiar from An.
post. A 22.83 a24ff,58 and is, for present purposes, merely a variation on
the 'by-virtue-of itself' terminology.
The transition in Top. A 9 from the variable list to the 'what is it?' list
is fairly abrupt, but not as abrupt as that in Met . .:1 7, for Aristotle does
not there even bother to give examples. 59 He merely sketches the variable
list ("of things predicated, some signify what [it] is, some what-like,
... " 1017a24 - 27) and says that 'to be' will have a single force (to use
as neutral a word as I can think of) for each entry on the list (a27). But
he is discussing 'to be' }lod}'
so it looks (at least to me)60 as if he
is making the same transition.
He follows it with a comment that has caused consternation 61
e27-30):
auro,
for there is no difference between 'a man is flourishing' and 'a man flourishes', or between
'a man is walking' or 'cutting' and 'a man walks' or 'cuts', and similarly in the other cases
(OVt'JfP "yare Ola",eeEl TO apt?eW1rO~ U"y1a[pwp fUTIP ij TO apt?eW1rO~ U"y1a[PEI, OV& TO apt?eW1rO~ l3aOllwP lUTIP ij Tep.pwp Toii apt?eW1rO~ l3aOlIEl ij Tep.PfI" bp.o[w~ Of xal f1rl
TWP aAAwp).
On any reading, this is going to be elliptical. The most natural one seems
to me this: just as, at 1017 a lO-12, he had introduced an example
(sentence (24), "the cultivated one builds houses") that did not employ
'is', and tried to show that it rested on an occurrence of 'is' anyway, so
here he notices that on his list of categories there are two at least that are
easily invoked in predications without using 'is': to do and to undergo;
73
and others are easily imaginable, even if the category-name does not, as
it does in these cases, make it obvious. 62 So he points out that here, too,
there is an 'is' in the offing: where you have a sentence containing a verb
other than 'to be', it may be replaced by a periphrastic verb phrase containing 'to be'. 63 The examples employ 'is' accidentally, but this would
not have bothered Aristotle, for the 'figures of predication' apply in the
first instance to the sentences on what we were just calling the 'variable'
list, where all the occurrences of 'is' but the first are accidental anyway.64
I have passed by the important and difficult question of why Aristotle
thinks that the 'is's of (3), (30), and (31) must all be different while those
of (29) and (30) (e.g.) are the same. I have, in fact, very little to say here.
Btlt there is, I think, a little more to it than one might think from what
has so far been said.
For, at this point, one might be tempted merely to re-invoke the fallacy
of the magnifying glass: Aristotle is, one might think, merely transferring a distinction that properly pertains to the predicate to the 'is' that
precedes (in English). 6S
But, in fact, Aristotle has an argument available, and, ~lthough a full
examination is beyond the scope of this paper, I should at least like to
state it, and then stop.
First, recall that our theory dictates that differences in the force of
essential 'is' carry with them differences in the force of existential 'is'.
This is to some extent usable as a two-way street: differences in the force
of 'is #' should show differences in the force of by-virtue-of itself 'is',
if the latter simply is the former with predicate uncanceled. 66
Notoriously, Aristotle denies that there is a genus of beings, that is,
a genus labeled 'that which is' (An. post. B 7. 92b I3-14). Less
notoriously, he provides an argument for that claim, in Met. B 3.
998 b 22-28 (see also Top. ~ 1. 121 a 14-19 for a related, but different,
argument): 67 a genus, he says, cannot be predicated of something unless
one of its species is predicated of that thing, and no species can be
predicated of its own differentia; so, if beings formed a genus, its differentiae could not be beings, that is, they would not exist, which is
absurd.
I do not at present see how to make anything convincing out of this.
But, as they say, tomorrow is another day.
74
RUSSELL M. DANCY
NOTES
I G. E. L. Owen, 'Logic and Metaphysics in Some Earlier Works of Aristotle', in
I. During and G. E. L. Owen (eds.), Aristotle and Plato in the Mid-Fourth Century (Studia
graeca ct latina gothoburgensia, XI, Giitenborg, 1960), pp. 163-90. The quotations are
from p. 165.
2 Particularly W. Jacobs, 'Aristotle on Nonreferring Subjects', Phronesis 24 (1979),
282 - 300. On pp. 297 - 98 (Note 6) Jacobs is critical of my note on De into II . 21 "25 - 27,
Appendix II to Sense and Contradiction: A Study in Aristotle (Reidel, Dordrecht, 1975),
pp. 153 - 55. Others have found that appendix obscure, but it was only Jacobs's note that
made me aware how grossly I might be misunderstood. The present effort (see Section 3
below) mayor may not help.
1 Especially in 'Aristotle on the Snares of Ontology', in R. Bambrough (ed.), New Essays
on Plato and Aristotle (Routledge & Kegan Paul, London, 1965), pp. 69 - 95.
4 The present paper is part of a larger project; I hope in carrying that out to remove some
of the more outrageous demands on the reader's credibility that the present paper presents.
S There are different verbs for 'to be' in the Greek of the period: for example, 'iJ1ro,QXfLv'
is sometimes used in Aristotle where we might expect 'dvcn' and translate 'to exist' (see
Bz. indo 788 b43ff). But it is also used occasionally as a copula (see ibid.). It could not be
used to 'disambiguate' 'dvO/l' if there were any ambiguity to disambiguate (see below).
6 E.g., F. H. Fobes, Philosophical Greek: An Introduction (University of Chicago Press,
1957), p. 51 Note I. Of course, this is not simply an artefact of introductory texts: it will
also be found in, for example, R. Kuhner and F. Blass, Ausfiihrliche Grammatik der
griechischen Sprache, Ier Teil (Hahn, Hannover, 1966) [reprint of ed. 3, 1890), Vol. I, p.
344, 90.2.
7 See here C. H. Kahn, The Verb 'Be' in Ancient Greek (Reidel, Dordrecht, 1973), Appendix II, 'On the Accent of lUTi and Its Position in the Sentence', pp. 420 - 34, esp. pp.
422-24.
8 For Greek examples (a great many) see Kahn, The Verb 'Be', Ch. VI.
9 E.g., N. Fleming and N. Wolterstorff, 'On "There is" " Philosophical Studies (U.S.)
11 (1960), 41 - 48; G. Vision, 'Existentials and Existents', Theoria47 (1981), 1- 30; Y. Ziv,
'Another Look at Definites in Existentials', Journal of Linguistics 18 (1982), 73 - 88.
10 G. Vlastos, 'Degrees of Reality in Plato', in R. Bambrough (ed.), New Essays on Plato
and Aristotle (Routledge & Kegan Paul, London, 1965), pp. 1 - 20, reprinted in Vlastos's
Platonic Studies (Princeton University Press, 1973, 1981). My disagreement with this article should not mask the fact that I am greatly indebted to it.
II Here I pass by an interpretative possibility given a fair run for its money by J. Gosling,'
Republic, Book V: Ta lI'o~a xcx}..o" etc.' , Phronesis 5 (1960), 116 -128. See also F. C.
White, 'J. Gosling on Ta 1/'o~a xcx}..o,', Phronesis 23 (1978), 127 - 32, 'The "Many" in
Republic 475a - 480', Canadian Journal of Philosophy 7 (1979), 291 - 306, and Gosling's
'Reply to White', idem 307 - 14. I am not convinced by Gosling, but I think the issue has
little bearing here.
12 The treatment of what might as well be a static situation (Helen's being simultaneously
beautiful and not beautiful) as if it involved change (Helen's vacillating between one and
the other) is characteristic: see Aristotle, Met. A 6. 987"32 - bIO, etc., and T . Irwin, 'Plato's
Heracleiteanism', Philosophical Quarterly 27 (1977), 1- 13 .
" For a review of the evidence with a rather different conclusion, see C. Kirwan, 'Plato
and Relativity', Phronesis 19 (1974),112 - 29. I seem to be adopting what Kirwan calls the
75
'relativist interpretation' of R. 479, and, worse, I am going to reject the one thing that he
says "is surely right in this interpretation, that it does not attribute to Plato any doctrine
about existence": see below. As for the 'relativist interpretation' in general, I cannot follow
Kirwan's emphasis on the formulation of the conclusion in 479b6 -7: there it is said that
big things are no more big than small, and Kirwan wants to say that that does not mean
they are both big and small. But they are picked out as big things, so they are at least that,
and if they are no more big than small, it follows from that that they are small. I am not
sure how much of an issue there is between us here. I certainly do not want to say that Plato
at the time of writing the Republic was aware that 'big', 'beautiful' and other relational
predicates were somehow special, and that is one of Kirwan's primary targets.
14 It is here especially that I am in agreement with Vlastos (see Note 10 above). The
disagreement comes when I say, below, that the concept of being that operates in this argument just is Plato's concept of existence. Here I am siding with Owen, 'Aristotle on the
Snares of Ontology', p. 71. Owen partially retracts this in 'Plato on Not-Being', in G.
Vlastos (ed.), Plato: A Collection oj Critical Essays, Vol. I (Doubleday Anchor Books,
Garden City, N.!Y., 1970), pp. 223 - 67 and 265 - 67, but his retraction concerns only the
Sophist, as far as I can tell, and I am saying as little as I can about that dialogue here (see
the next paragraph).
15 On the Sophist's treatment of 'is' see M. Frede, Priidikation und Existenzaussage:
Platons Gebrauch von " .. .ist . .. " und" . .. ist nicht . .. " im Sophistes (Vandenhoeck
& Ruprecht, G6ttingen, 1967). Particularly relevant to my theme above is Chapter II, pp.
37ff.
16 W. KamIah, Platons Selbstkritik im Sophistes (C. H. Beck, Munich, 1963 [= Zetemata
33)) has much to say that is relevant here (see esp. Chapters V & VI), but he does not seem
to see in Sph. 256e5 - 6 the flat rejection of R. v 476 - 80 that I do.
17 I ignore here complications that have come to light in recent discussions of adjectives
(see, e.g., M. Platts, Ways oj Meaning [Routledge & Kegan Paul, London, 1979) pp. 16lff
on 'Adjectival Constructions' and the literature there cited, esp. 1. A. W. Kamp, 'Two
Theories about Adjectives', in E. L. Keenan (ed.), Formal Semantics oj Natural Language
[Cambridge University Press, 1975) pp. 123 - 55). I am, for purposes of expositiOl" adopting the simple-minded view that 'pale man' is merely a concatenation of 'pale' and 'man':
that, in Montague's terminology, 'pale' denotes an intersection function (see 'English as
a Formal Language', in Formal Philosophy: Selected Papers oj Richard Montague, ed.
by R. H. Thomason [Yale University Press, New Haven & London, 1974), pp. 188 - 221,
at p. 211). Roughly: the pale men are the pale things that also happen to be men. This, of
course, is false: what counts as a pale man varies with race, location, time of year, and so
on. (So far, then, 'pale' might be an attributive adjective in the terminology of the next
paragraph. In one of Kamp's two theories about adjectives, all adjectives are: see his The
Verb 'Be'). It is this that I am ignoring, and it does not i!ffect the point at issue.
18 P. Geach, 'Good and Evil', Analysis 17 (1956),33-42 (reprinted in P. Foot (ed.),
Theories oj Ethics [Oxford University Press, 1967), pp. 64 - 73 ab init.
19 See Geach, 'Good and Evil'. Kamp (The Verb 'Be', p. 125) calls them 'privative
adjectives' .
20 One of the faults of Ross's generally reliable (Oxford) translation of the Metaphysics
is the astonishing variety of renderings it shows for the crucial phrase 'xed}' aUTo': 'in virtue of itself' (0 18), 'propter se' (Z 4), 'in itself' (Z 3. 1029"20), 'of itself' ("24), 'in virtue
of its nature' (Z 5. 1030b 19 - 20), 'self-subsistent' (Z 6. 1031"28fl), and, no doubt, other
things as well.
21 There is a great deal more to be said here. See A. Code, 'Aristotle's Response to Quine's
Objections to Modal Logic', Journal oj Philosophical Logic 5 (1976), 159-86; F. 1.
76
RUSSELL M. DANCY
Pelletier, 'Sameness and Referential Opacity in Aristotle', Nous 13 (1979), 283 - 311, and
my 'On Some of Aristotle's First Thoughts about Substances', Philosophical Review 84
(1975), 338 - 73 (esp. 340 - 42, 365 - 68) and 'On Some of Aristotle's Second Thoughts
about Substances: Matter', Philosophical Review 87 (1978), 372 - 413. Frank Lewis finds
all of us objectionable in 'Accidental Sameness in Aristotle', Philosophical Studies
42 (1982), I - 36.
22 'Defines' should be taken with a grain of salt here, for in A 18. 10228 27 - 28 Aristotle
'defines' the phrase 'by virtue of itself by means of 'essence'. The point is that the two
go together, and there is no understanding the one without the other.
23 For a review of the literature, and an interpretation of the Phrase with which (I think)
I do not agree, see J. Owens, The Doctrine of Being in the Aristotelian Metaphysics (Pontifical Institute of Mediaeval Studies, Toronto, 1951, edition 2, 1%3), pp. 180ff.
24The Greek have the imperfect, 'was'. The explanation for this is not of importance in the
present context, but the one 1 opt for is one rejected by Ross, Aristotle's Metaphysics
(Clarendon Press, Oxford, 1924) 1 127: the imperfect is the 'philosophicM imperfect'. This
occurs in English as well: 1 get you to agree to'S is P', we continue talking, and 1 later want
to remind you of that, so 1 say'S was P' wasn't it?'. Here it could be that'S was P' was
'Two and two made four': there is no real implication of past ness about the tense. Often,
in Plato, the 'philosophical imperfect' is employed in appealing to a previously stated
definition. Ross (ibid.) objects that it "is used only when there has been an actual previous
discussion of the subject in hand, which is the case in but few of the passages in which TO
Tt 1}v flvm is used". But that is Aristotle for you: his (and the Academy's in general)
technical terminology is arrived at by detaching terms from their dialectical contexts. The
original context would have been an explicit definition, appeal to which would be made by
the phrase 'what it was for virtue to be', and then that just becomes a label for whatever
it is that virtue is correctly defined to be.
2S Alternatively: "for [something) educated to be", 1 have, admittedly, picked the translation that most favors my overall interpretation. But it could be done either way (and, in
fact, 1 arrived at it thinking in terms of the alternative translation).
26 'Aristotle on the Snares of Ontology', p. 71. For further examples, see OED s. v. 'be',
B 1 I (Vol. I, p. 717, col. 3).
27 This consideration plays a part in the literature cited in Note 9 above.
28 1 do not want to reject the sense-reference distinction, and yet1 do not find myself quite
comfortable with saying that proper names have sense. Among those who are comfortable
with saying that are P. Geach (Mental Acts [Routledge & Kegan Paul, London, 1957), pp.
66fO; M. Dummett (Frege: Philosophy of Language [Duckworth, London, 1973, 1981),
index 2nd edition only) s. v. 'proper names, sense of); and D. Wiggins ('Frege's Problem
of the Morning Star and the Evening Star', in M. Schirn (ed.), Studien zu Frege II: Logik
und Sprachphilosophie [Frommann - Holzboog, Stuttgart, 1976), pp. 221 - 55). S. Kripke
hedges here: see 'Naming and Necessity', in D. Davidson and G. Harman (eds.), Semantics
of Natural Language (Reidel, Dordrecht, 1972), pp. 253 - 355, at p. 322 (in the reprint,
Naming and Necessity [Harvard University Press, Cambridge, Mass., 1980), p. 127). See
also M. Lockwood, 'On Predicating Proper Names', Philosophical Review 84 (1975),
471 - 98, and J. Cargile, Paradoxes: A Study in Form and Predication (Cambridge University Press, 1979), Ch. 2, 'MiII's Theory of Names'.
29 P. Geach, 'Identity', Review of Metaphysics 21 (1967/68), 3 -12 (in Logic Matters
[University of California Press, Berkeley, 1972), pp. 238 - 47) and 'Ontological Relativity
and Relative Identity', in M. K. Munitz (ed.), Logic and Ontology (New York University
Press, 1973), pp. 287 - 302; D. Wiggins, Sameness and Substance (Harvard University
77
Press, Cambridge, Mass., 1980; a revision of Identity and Spatio- Temporal Continuity
[Blackwell, Oxford, 1967)); N. Griffin, Relative Identity (Oxford University Press, 1977),
and B. A. Brody, Identity and Essence (Princeton University Press, 1980), esp. Ch. I.
30 The following sketch took its departure from S. Soames and D. M. Perlmutter, Syntactic Argumentation and the Structure of English (University of California Press, Berkeley,
etc., 1979); see esp. pp. 46 - 52, some of the claims of which I shall implicitly be
challenging.
31 See also C. J. Fiilmore, 'The Case for Case', in E. Bach and R. T. Harms (eds.), Universals in Linguistic Theory (Holt, Rinehart & Winston, New York, etc., 1968), pp. 1-88,
esp. pp. 44ff; J. Lyons, 'A Note on Possessive, Existential and Locative Sentences', Foundations of Language 3 (1967), 390 - 96; 'Existence, Location, Possession and Transitivity',
in B. van Rootselaar and J. F. Staal (eds.), Logic, Methodology and Philosophy of Science
III (North-Holland, Amsterdam, 1968), pp. 495 - 504; Introduction to Theoretical
Linguistics (Cambridge University Press, 1968), pp. 388 - 90; K. Allan, 'A Note on the
Source of THERE in Existential Sentences', Foundations of Language 7 (1971), 1-18.
Allan's rl":ction of the attempt (made by Fillmore and Lyons) to treat 'there' as a locative
is correct, I think, but the above is independent of this issue.
32 This is a Fregean view: see 'Begriff und Gegenstand', Vierteljahrsschrift fiir
wissenschaftliche Philosophie 16 (1892), 192-205, at p. 94, where the copula is said to
serve "als blosst'~ Formwort der Aussage" (in Translations from the Philosophical
Writings of GOlllob Frege, ed. by P. Geach and M. Black [Blackwell, Oxford, 1960], p.
43 "as a mere verbal sign of predication"). The view is heartily endorsed by Geach: see
Reference and Generality (Cornell University Press, Ithaca, N.Y., 1962, 1968), p. 34,
where Geach ascribes the view to Aristotle as well, on the strength of An.pr. A I. 24 b l7 - 18
(where, unfortunately, Ross would delete the words that make Geach's case: see Aristotle's
Prior and Posterior Analytics [Clarendon Press, Oxford, 1949, 1957], 290 ff: the parallel,
De into I. 16"16 - 18, which Ross cites, seems to me to make the case for the deletion quite
strong). For Frege, see also M. Dummett, Frege, p. 214. Both Frege and Geach want to
retain special senses of 'is': Frege, the 'is's of identity and existenc~ (Ioc. cit.), and Geach,
that of existence as opposed to predication (see 'Assertion', Philosophical Review 74
(1965),449 - 65 [= Logic Mailers 254 - 69): on p. 460 (265) he rails against those who "two
thousand years and more after Plato's Sophist, will wantonly confuse [the 'is' of predication) with the existential 'is' "). See here the references in Note 37 below. For Aristotle,
see also H. Bonitz, 'Uber die Kategorien des Aristoteles', Sitzungsberichte der Kaiserlichen
Akademie der Wissenschaften, Phil.-his!. Klasse, 10 (1853), 591 - 645, p. 601 (available as
a separate reprint with the original pagination, Wissenschaftliche Buchgesellschaft, Darmstadt, 1967).
J3 "Aber das Sein bleibt unauffindbar, fast so wie das Nichts oder am Ende ganz so. Das
Wort 'Sein' ist dann schliesslich nur ein leeres Wort. Es meint nichts Wirkliches, Greifbares, Reales. Seine Bedeutung is ein unwirklicher Dunst." Einfiihrung in die Metaphysi~
(Max Niemeyer Verlag, Tiibingen, 1957), p. 27. In R. Manheim's English translation
(which I have departed from in the above), An Introduction to Metaphysics (Yale, New
Haven, 1959), the passage occurs at the bottom of p. 35.
14 Cf. Tsu-Lin Mei, 'Subject and Predicate: A Grammatical Preliminary', Philosophical
Review 70 (1961), 153 -75.
35 Cf. B. L. Whorf, 'Languages and Logic', in Language, Thought, and Reality: Selected
Writings of Benjamin Lee Whorf, ed. by J. B. Carroll (MIT Press, Cambridge, Mass.,
1956), pp. 233-45.
36 Cf. R. Rorty, 'Genus as Matter: A Reading of Metaphysics Z-H', in E. N. Lee,
78
RUSSELL M. DANCY
79
80
RUSSELL M. DANCY
So also, perhaps, Larkin, Language in the Philosophy of Aristotle, p. 87 with Note 20,
but I find pp. 87 - 88 very confusing (see also last note).
61 Cf. Note 59 above, and Ross, Aristotle's Metaphysics I 307 - 308; Owen, 'Snares', p.
82, Note I; and Thorp, Aristotle's Use of Categories', p. 250.
62 In fact, Aristotle refers 'flourishing' (irYlcdvfLV) to the category of quality in Soph. el.
4. 166b 16 - 19; cf. also Cat. 8. 9" 14 -16.
63 So far, I agree withThorp. As he points out (Aristotle's Use of Categories, p. 250), this
interpretation fits Aristotle's use of the same examples in De into 12. 21 b 5 -10,10. 20"3ff.
64 Here I leave Thorp (Aristotle's Use of Categories, 252 - 254) for Ross (Aristotle's
Metaphysics I 307 - 308).
6S This is certainly the way many presentations make it sound, e.g. von Fritz, loc. cit. Contrast C. Kirwan, Aristotle's Metaphysics, Books r, .a., E (Clarendon Press, Oxford, 1971),
pp. 142f.
66 But it is still wrong to say, as does Owen, that J(a~' aUTO ov is "the (or an) existential
use of the verb" ("Snares" p. 82).
67 The argument of Met. B is reviewed by Thomas Aquinas in his comments on .a. 7; see
In duodecim fibros Metaphysicorum Aristotelis expositio, ed. by M.-R. Cathala and R. M.
Spiazzi (Marietti, Turin, 1950), p. 238 (889).
60
Dept. of Philosophy,
Florida State University,
Tallahassee, FL 32306-1054, U.S.A.
JAAKKO HINTIKKA
(0) J 986
82
JAAKKO HINTIKKA
83
84
JAAKKO HINTIKKA
it. Fortunately, even though Aristotle does not say in so many words that
the word est; does not exhibit the Frege - Russell ambiguity, he does say
as it were in the material mode of speech that the entities that would be
differentiated from each other in the Frege - Russell distinction (if
Aristotle had made it) in reality are one and the same thing. Thus in Met
r 2, lOO3 b 22 - 32, he writes (the text is Ross's and the translation is Kirwan's):
fi o~ TO OV )(al TO EV mUTOV )(al p.ia q,llCm,:r~ (noAoullftV aAA~AOtS wune I aex~ )(W Citnov aAA' oux ws IVI AO'YI" O"l/Aovp.fva (Otaq,feH Oi oullfv ouo' (xv bp.oiws iJ1rOAa/3wp.fP, aAAa
)(al 1reO fe'YOU p.aAAov) mUTo 'Yae ds av'oewr.os [)(al Orvllew1ros], )(al WV Ctvllew1ros
)(al Orvllew1f"os, )(al oux fTfeOV n O"l/AOt )(aTa T~V Af~tV f1raVQ[H1f"AOVP.fPOV TO fis OrvlleW1f"Os
)(QI ElS WV OrvlleW1f"Os (OijAOV 0' on ou xweil;TQt OUT' f1r1 'YfVfUfWS OUT' f1r1
q,lIoeas), bp.oiws Of )(QI f1f"t TOU ivas, wun q,QVfeOV on " 1reaull(1tS fV TOVTOtS mUTo
O"l/AOt, )(QI OUOEV fneOV TO EV 1rQea TO OV, ....
Suppose it true, then, that that which is and that which is one are the same thing - i.e.
one nature - in that each follows from the other as origin and cause do, not as being indicated by the same formula (though it makes no difference even if we believe them to be
like that - indeed it helps). For one man and a man that is and a man are the same thing;
and nothing different is indicated by the reduplication in wording 'he is one man' and 'he
is one man that is' (it is plain that there is no distinction in [the processes of) coming to
be or destruction); equa\1y in the case of that which is one. It fo\1ows obviously that the
addition indicates the same thing in those cases, and that which is one is nothing different
apart from that which is.
Although the text of the lines 1003 b 28 - 29 is especially messy, it is important to note that Aristotle is in the quoted passage employing est; in
what I shall later in this paper argue to be a purely existential use.
Nonetheless he is emphatically assuring us that this use is not different
from the identity sense of est; "is one and the same man". As to the
predicative sense, by the priority of the three cases announced a couple
of lines earlier, "he is a man" will be a further synonym for the two
phrases Aristotle mentions.
An even blander assertion to the same effect is found in De Soph. E/.
6, 169 a 8 - 10: "For the same definition (horos) applies to 'one single
thing' and to 'the thing' hap/os; the definition e.g. of 'man' and 'one
single man' is the same, and so, too, with other instances".
This confounds the first three members of the four-fold distinction of
Frege's and Russell's. As for the fourth it is clear that there is no
Frege - Russell type difference in meaning for Aristotle between the different occurrences of ;s in "Socrates is a man" and "a man is an
animal". If further evidence is needed for the total absence of the
Frege - Russell ambiguity thesis in Aristotle, it is easily forthcoming.
85
86
JAAKKO HINTIKKA
be the most popular reaction to the data that can be adduced against the
presence of the Frege - Russell ambiguity in Aristotle. According to this
competing view, esti is unambiguous because it has basically always the
predicative sense. Where it apparently does not, e.g. in the existential
uses listed above, we must understand the usage as being elliptical:
"Socrates is" on this view basically means "Socrates is something or
else". )There may be important restrictions as to what this "something
or else" can be, but they need not detain us here.)
This view seems to have been suggested by G. E. L. Owen, and it has
recently cropped up in slightly different variants. There is a sense in
which it probably comes close to being a true representation of what
things are like according to Aristotle's last and final conclusions.
Roughly, for any entity to exist is for it to be what it is, i.e., what it essentially is.
However, admitting this does not mean that in the force of the term
esti in Aristotle's actual argumentation is tacitly predicative. For one
thing, the identification just offered is probably only an approximate
one, anyway. It is not clear that for Socrates to exist is (apud Aristotle)
for him to be a man. Rather, on a closer look it seems (as Balme has
shown) very much as if for Socrates to exist is not so much for him to
exemplify (more generally, to develop towards exemplifying) the speciescharacteristic form of man, but rather to exemplify (more accurately,
develop towards exemplifying) the particular nature which consists in his
likeness to his parents. And it is not clear at all that Socrates' exemplifying this particular form is a predicative relation rather than an identity.
Be this as it may, even if the elliptical character of fun a1l'AWS is
perhaps a conclusion of Aristotle's arguments for his metaphysical
theory, it cannot for this very reason be a part of what he assumes in
them. When I reject the ellipsis theories, it is thus as a claim of what the
basic semantical force of esti and its con gates are for Aristotle, and not
as a possible feature of his ultimate metaphysical doctrine. However, in
the former sense I do reject it tout court, and hence also reject the
mistaken idea that it is somehow implied by the absence of the
Frege - Russell ambiguity assumption from Aristotle.
This puts on me the onus of commenting on the recent denials of any
purely existential uses of verbs for being in Aristotle. Suffice it here to
deal with one of the most recent putative arguments for the absence of
the existential uses in Aristotle or in certain parts of the Aristotelian
Corpus.
The ellipsis hypothesis has not been defended by its reputed originator
87
Every B is simpliciter
Every C is B
Hence: Every C is simpliciter
88
JAAKKO HINTIKKA
I shall not discuss here what kind of treatment of existence is presupposed in (*) and (**).
My interpretation gains further credence from the fact that, according
to Aristotle, necessity is "carried downwards" in a syllogistic chain in the
same way as I have argued existence is. (Cf. Aristotle's theory of apodeictic syllogisms in Pro An. A 8 - 12, especially 9.) In the same way as in (**)
it is only the major premise that has to carry an existential force in order
for the conclusion to do so, in the same way we can obtain a necessary
conclusion from a barbara type syllogism if and only if the major premise
is a necessary one. In fact, it seems to me that Aristotle's treatment of
existence and necessity in the context of a syllogism are related to each
other very closely. 80th of them are based on the presence of an element
of identity in Aristotelian copula, whether or not it is in fact expressed
by him in terms of esti or not. For if the minor premise of a syllogism
like (**) expresses a numerical identity between each C and some 8 then
we must be able to say all the same things of each C as are said of each
8, for the former literally are among the latter. Hence the validity
(among other modal syllogisms) of the following form of barbara:
(***)
Every 8 is necessarily an A
Every C is (identical with) a 8
Hence: Every C is necessarily an A.
8e this as it may, there is plenty of collateral evidence that my construal of the role of existence in Aristotle's syllogistic theory is what
Aristotle in fact meant. Since the whole argumentative structure of
Gomez-Lobo's paper is thus mistaken, there is little that needs to be said
of the rest of his paper.
89
TO 0' fi
How could Aristotle possibly have explained more clearly by the means
he had at his disposal that he was presupposing a purely existential use
of fi EUTL? It seems to me that we have to realize that Aristotle, like J. L.
Austin, ordinarily means what he says.
Ironically, Aristotle's very usage in Post. An. B I - 2 provides us with
further counter-examples to the ellipsis thesis. When Aristotle there asks
whether a middle term is (d fUTL piuo II , cf. 89 b 37 - 38,90 a 6), he cannot
but mean whether the middle exists, for he contrasts this question in so
many words with the question as to what it is.
There is elsewhere, too, excellent direct evidence against the ellipsishypothesis. In discussing in De Soph. EI. 5 the importance of
distinguishing the absolute and the relative uses of a term from each
other Aristotle writes (167 a 4 - 6):
OU -yae TCiUTO TO 1'1, flvcx! TL ){CX! Q'JI"AWi 1'1, dvCXI. q,CX!VfTCXt Of Ota TO 'JI"aef-Y-YVi TijS
XCX! I't){eOV Otcxq,ieHV TO flvcx; TL Toii flvCXI,.xcx! TO 1'1, dvcx! n Toii 1'1, flvCXI.
Ai~fWi
For it is not the same thing not to be something and not to be simpliciter, though owing
to the similarity of language to be something appears to differ only a little from to be, and
not to be something from not to be.
One can scarcely ask for more direct evidence. At the same time the
passage shows that, in spite of their differences, the predicative and the
absolute (existential) uses of esti are not unrelated, for they are the
relative and absolute uses of the same notion. The quoted passage hence
also offers evidence against ascribing the Frege - Russell ambiguity to
Aristotle.
What we have found deserves a few additional comments. First, my
defense of the presence of a purely existential use of esti in Aristotle is
squarely based on the absence of a purely existential meaning as
distinguished from its alleged predicative meaning and identity meaning.
For, in order for (**) to do the double duty of both establishing a
predicative link between C and A and at the same time carrying existence
assumptions downwards from higher wider terms to lower (narrower)
ones, the (**) must carry (for a Fregean) both a predicative sense and the
existential one. Morc generally, it is undoubtedly the illegitimate
90
JAAKKO HINTIKKA
91
Every C is B
where B is a part of (or the whole) of the essence of C, for (hat would
presuppose an irreducible existential force of "is" in the latter and
thereby make being (existence) part of the essence of C. As an ultimate
(atomic) premise of an Aristotelian science (' ') thus does not have an existential import. Rather, Aristotle's point is that the existence of the C's
must be demonstrated, as in (*) or (**). It is not implied by C's essence
alone. Hence there is a sense in which the existl!nce of the C's is not due
to their being what they essentially are, but their falling under (as
established by a potentially quite long chain of appropriate scientific
syllogisms) the genus which characterizes the science in question. This
genus is what we do assume to exist in the science in question. As Aristotle continues,
And that is what the sciences as a mailer of fact do; for the geometer assumes what triangle
signifies but proves that is.
Even though Aristotle docs not press the point in the quoted passages,
if we proceed tov,'ard morc and more general sciences we never reach one
92
JAAKKO HINTIKKA
where existence is part of the essence of the genus, so that it need not be
assumed. That is what the quoted passage amounts to. Elsewhere
Aristotle indicates that the upward chain comes to an end with one of the
different categories. Hence the only case in which we are allowed to
assume (according to Aristotle) the existential import is "every C is D"
where D is a term for the category to which C belongs. In this sense, even
if we believe that such attributions of existence as (') are elliptical, the
omitted term is not the essence of C but the category of C.
This view of the role of existential presuppositions in Aristotle thinking seems to be confirmed by Met. H 6, 1045 a 34 - b 8.
A fourth observation may likewise be in order. Philosophers', difficulties in understanding all the different things that are going on in an
Aristotelian syllogism like (**) illustrate a more general methodological
moral. The Frege - Russell ambiguity assumption is built into all the better known formalisms of first-order logic (quantification theory, lower
predicate calculus). Since this assumption is completely foreign to
Aristotle, virtually all applications of modern logic to Aristotle are partly
anachronistic, and have to be viewed with considerable caution. This
flaw does not automatically invalidate them, however, nor does it make
such historical applications of formal techniques inferior to the work of
informal analysts of Aristotle's work, for most of the latter have likewise
been relying on the Frege - Rw;sell ambiguity thesis, as was pointed out
above.
A further general observation is the following: The Frege - Russell
distinction between different meanings of "is" and its cognates is correlated - at least roughly - with an ontological distinction between different kinds of entities. The is of identity equates in its clearest instances
particulars, modern philosophers' "individuals". In contrast, the is of
predication expresses the being of facts. Thus my thesis of the absence
of the Frege - Russell distinction in Aristotle is not without consequences
for the rest of his ontology. For instance, it is connected with the otherwise strange practice of Aristotle's in his syllogistic theory of scientific
explanation, where he treats the being (existence) of what looks like individuals and the being (occurrence) of facts or events on a par without
any apologies or explanations. (For examples, see, e.g., Met. Z 17, 1041
a 14 - 16, b 4 - 5.) Furthermore, another part of the same syndrome is
the important fact that, appearances notwithstanding, Aristotle does not
really have as sharp a notion of an individual (particular) as contemporary post-Fregean philosophers.
93
It is not clear whether ousia here means essence or substance, but a comparison with such passages as Met. r 4, 1007 a 20 - 33 shows that the
former possibility cannot be excluded here. This is also shown by Aristotle's reference to incidental predication at a 27 - 28, complete with
Aristotle's stock example (AfV)(OV) of an accidental predicate. Hence
what we have found is that one important element in Aristotle's distinction between essential and accidental predication is that the former is an
assertion of identity whereas the latter one is not. In so far as predication
is expressed by means of a verb for being, this means that in essential
predication the verb is used to express identity whereas in an accidental
predication it is used to express predication. Once again, this difference
in use does not mean that Aristotle is thinking of esti as having different
senses or meanings.
94
JAAKKO HINTlKKA
This partial construal of the essential vs accidental distinction is further confirmed by such passages as Met. ~ 18, 1022 a 26 - 27 and Post.
An. A 4, 73 b 5 - 10. It is by any token an important element in Aristotle'~ ideas of essence and essential predication. Further arguments for the
same view have been presented by M. J. Woods, who argues that
"Aristotle held that a staement like 'Socrates is a man' was, despite appearances, to be construed as a statement of identity". In defense of this
view, Woods refers inter alia to Met. ~ 18, 1022 a 26-27; Z 4, 1029 b
28; Z 7, 1032 b 1 - 2; Z 8, 1034 a 8. As was noted above, the unmistakable
(albeit not exclusive) presence of this idea in Aristotle's mind is also
witnessed by the question he raises and discusses in Met. Z 6 as to
whether each thing is identical with its essence. Woods's thesis is essentially that Aristotle answers his own questibn affirmatively.
One intriguing feature of these observations is that they seem to turn
the contrast between essential and accidental predication into a prima
facie nonmodal distinction, whereas it has in recent discussions of
"Aristotelian essentialism" been treated as an almost paradigmatic example of a modal distinction. It is of course true that in spelling out the
distinction in systematic terms we soon get entangled with modal concepts, and it is equally true that in his discussion of the matter in Met.
Z-H Aristotle likewise resorts in the end to the notions of potentiality and
matter. Nevertheless it seems to me that the initial impression of a nonmodal distinction points to an interesting truth (and to an interesting
flaw in recent discussions). Even I cannot elaborate the point here, the
way in which the essential-accidental distinction involves modal concepts
is not through any direct appeal to them, but through a modal element
which there is in the very notion of an individual (for Aristotle, in the
notion of substance). Individuation, in short, is a process which inextricably involves a modal element.
My observation concerning the link in Aristotle between the essentialaccidental contrast and the difference between identity and copula - as
well as Woods's thesis - nevertheless represent only the first step in
understanding Aristotle's views on essence and substance. A telling indication of the problems one encounters in this direction is the fact that
Aristotle occasionally coulitenances also "accidental unities" like "pale
man" or "pale Socrates". In other words, in one of his moods he
somehow construes "Socrates is pale" and "man is pale" also as identity
statements , and not predications contrasted to " Socrates is man" or
"man is an animal " , which alone were supposed to be identities. Aristotle's reasons are nevertheless clear. As was noted earlier, Aristotle's key
95
96
JAAKKO HINTIKKA
Suffice this as an indication of one line of thought opened by our observations. To return to the main theme of this paper, notice that from the
absence of the Frege - Russell ambiguity it does not follow that there
might not be other ambiguities about esti, over and above the nonambiguous differences in use which Frege and Russell mistakenly raised
to the level of ambiguities. Further light can be thrown on this question,
too, by means of recent topical insights. There is in fact another major
way in which recent logical and seman tical work on the concept of being
puts Aristotle in an interesting new perspective. In order to see what it
is, we have to stray temporarily away from Aristotle and discuss certain
topical problems in the logic of natural languages. I shall discuss them
in the case of English, even though similar things can be said of other
languages, including ancient Greel(.
These problems are as close to t.1e heart of all Sprachlogik as we can
hope to get. Any logician knows that the lifeblood of virtually all interesting logical techniques in that basic part of logic which is variously
known as first-order logic, quantification theory, or lower predicate
calculus are the rules of instantiation (i.e., rules for substituting names
or name-like terms for quantified variables). Now suppose w~ want to
deal with the logic of natural languages directly, without first attempting
the dubious and by this time largely discredited translation to formal
languages. Then our first task is to formulate. likewise directly for
natural languages, rules of instantiation far the quantifier phrases which
take over the role of quantified variables in natural languages. How can
we do that? How are we to deal with, say, a quantifier phrase like "every
97
white horse which Alexander rode" or "some small town where Socrates
lived", occurring in a context X - W? (We take here the general form
of these quanti fer phrases to be
every }
some
Y + wh-word + Z
X - b - W if b is a white horse
and Alexander rode b
or, respectively,
(2)
Socrates lived in d,
where" b" and" d" are the respective instantiating terms. In general, the
output of an instantiation step is of the form
if
(3)
X - b - W
} b is a Y and Z'
and
where b is the instantiating term and Z' is like Z except that the trace has
been replaced by "b" with the appropriate preposition. (We have been
assuming that Y and Z are here singular.)
The details need not detain us here. What is of interest to us here is
an important difference between the situation in formal first-order
languages and natural languages. In the former, a single domain of
values for the substituting terms (e .g. my "b" and "d") is given. In the
latter, the entities referred to by the substitution-values have to be chosen
from different subdomains in different cases. For instance, in (I) b has
to be a living creature, whereas in (2) d has to be a location in space.
It lies close at hand for a logician to say that the only novelty here is
that natural languages employ many-sorted quantification theory (more
generally, many-sorted logic). And this need not by itself introduce any
complications (contrary to what is, e.g., implied in Moravcsik, 1976). In-
98
JAAKKO HINTIKKA
deed, many-sorted logics do not involve any serious new difficulties over
and above one-sorted ones.
Yet there is a new question present here. In many-sorted formal logics,
the sortal differences are indicated by notational conventions. How are
these differences marked in natural languages? How can one tell what
subdomain b or d must belong to?
Some clues are obvious, and the most obvious is the relative pronoun
which disappears in the process of instantiation. (These relative pronouns can be taken to be question words in a new role, except that
"what" is replaced by "that".) If the operative word is "who", the relevant subdomain consists of persons, if "where", of locations in space,
if "when", of moments (and/or periods) of time, etc. Further subdomains are introduced by prepositional phrases containing similar words,
for instance "like which" introducing a realm of qualities ("some color
like which you have never seen"). Clearly there is not a sharp one-to-one
correspondence between the ranges of natural-language quantifiers (my
"subdomains") and different relative pronouns (or other wh-words,
with or without prepositions or similar qualifiers), but a rough-andready correspondence certainly obtains.
The main discrepancy here is the fact that "what" covers several different subdomains. By asking, "what is X?", we can mean at least three
different things, to wit:
(i)
(ii)
(iii)
What is Sirius?
What is a gnu?
What is cordite?
Similar things can be said of Aristotle's Greek word esti: This ambiguity is made especially important in Aristotle's case by the absence of the
Frege- Russell distinction. For in terms of this distinction one could
distinguish the different what-questions (i) - (iii) from each other by the
different sense of "is" involved in them. But this distinction just is not
available to Aristotle.
In any case, the relative pronoun (or the corresponding question word)
cannot be the only clue to the choice of the subdomain. For one thing,
99
the whole relative clause can be missing from the quantifier phrase in
question, and hence be unavailable to supply any leads. Hence it is the
meaning of Yin (3) which must supply the main in formation as to which
subdomain (sort) we are dealing with. Presumably we must assume some
kind of semantical categorization of the terms (phrases) that can serve
as the Y in (3). In the case of simple terms these must be part of their lexical meaning. Since the Y's in (3) are basically predicate terms, we end
up in this way postulating a classification of all simple predicates of
English into certain equivalence classes. These classes wi\l be correlated
one-to-one with those subdomains of quantification, which we are dealing with, when using quantifiers in English, i.e., the largest classes of entities we can quantify over, and also correlated in a loose way with certain
wh-words and phrases.
The need of relying on Y for our choice of the subdomain is vividly
seen from the fact that if we try to eliminate Y (in the way in which we
could dispense with the relative clause), we would end up with an
ungrammatical expression. In order to preserve grammatically, we must
amplify the quantifier word itself so as to make it capable of conveying
the crucial information. For instance, some becomes someone,
something, somewhere, sometime, somehow, etc., where the added handle serves to betray the relevant sort (subdomain).
Furthermore, since each instantiation step (witness (3 introduces an
occurrence of "is", these correlated classifications are likewise correlated with a distinction between different uses of "is", viz., those that
could have originated from an application of the instantiation rules, plus
of course those that are logically on par with them.
Thus we are led to recognize four correlated multiple distinctions.
They distinguish from each other
(i)
(ii)
(iii)
(iv)
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
101
102
JAAKKO HINTIKKA
103
104
JAAKKO HINTIKKA
pertain to such things as the identity (in form) of indefinite relatives with
indirect interrogatives in Greek, and the close relationship of both with
quantifier words. These features of the Greek grammar serve to link the
different correlated distinctions explained above to each other especially
closely, and thereby to motivate Aristotle's theory. If I had to find
linguistic evidence for my interpretation of Aristotle, that is the direction
in which I could (and would) go. Even on the present superficial level,
it is not hard to see that my treatment of instantiation works mutatis
mutandis even better with Greek thari with English.
Likewise, we are now in a position to draw an interesting conclusion
from our observations. The different classes of questions with which
Aristotle correlated his other distinctions were primarily indirect questions. The correlation depends crucially on an analogy between relative
pronouns and question words, and this analogy (or near identity) can obviously be best argued for by comparing with each other the logical
behavior of relative clauses and indirect questions. (An especially useful
Mittelglied here is the class of relative clauses without antecedents. Their
logic is remarkably similar to that of indirect questions.) Aristotle's
distinction between different categories is less a distinction between different question types as between question words, and it pertains to these
words in so far as they are doing duty of their relative clause twins.
This observation reflects somewhat unfavorably on those scholars
who have made much of the classification of questions as the alleged cornerstone of Aristotelian categories. It seems to me that their thesis remains unproven. Admittedly, the importance of the dialectical questioning games practiced in the Academy for Aristotle can scarcely be exaggerated. However, there is little evidence in the Topics or elsewhere that
the theory of categories was developed for (or from) such games.
7. CONCLUSIONS FROM THE RECONSTRUCTED ARISTOTELIAN THEORY.
CATEGORIES VS LOGICAL TYPES
105
106
JAAKKO HINTIKKA
107
Aristotle's thought. In treating (at least in its first stage) all categories on
a par Aristotle (as well as my rational reconstruction of his theory) fails
to give a deeper account of the rationale of category distinctions. It is for
this reason especially important to realize the differences between
Aristotelian categories and logical types.
It is here that Aristotle's relative neglect of the Frege - Russell distinction (even as a difference in use and not just as a difference in meaning)
becomes a handicap for him. Admittedly, it was claimed by Maier that
Aristotle's theory of categories was calculated to accommodate certain
distinctions between different senses of "is" (see Vol. 2, p. 291 ff).
Maier's distinctions include most of the Frege - Russell ones. Indeed,
Maier's first two distinctions are identificatory being vs accidentally
predicatory being, p. 280, and existential vs copulative being, p. 282. No
major insights are forthcoming from Maier, however, into the way
Aristotle managed to combine the Frege - Russell distinction with his.
doctrine of categories. For he firmly believes that. according to Aristotle,
"immediate reflection on the concept of being [Maier's emphasis) ...
forms the principle of division for the table of categories" (pp.
298 - 299). Maier's immediacy claim notwithstanding, Aristotle himself
does not trust immediate intuition here, but discusses the relation of
other categories to that of a substance. However, these arguments are
either calculated to show the dependence of other categories on that of
substance, or (which may come down to the same thing) to point out the
role of focal meaning in relating the being of the other categories to that
of substance. They do not rely on the kinds of distinctions which Maier
mentions or which are likely to be made by a twentieth-century logician.
Instead, they mark a slightly different point of partial contact between
Aristotle's theory of categories and modern type distinctions. For the
primacy of substances over members of categories came to mean for
Aristotle something very much like the claim that only substances are individuals in a modern philosopher's sense. This is shown by Aristotle's
frequent reference to a substance as a "this" (TC>O) or as a "this
something" (TOOf n). Another indication is given by Aristotle iq Met. Z,
3, 1029 a 27 - 29, where he says that "it is taken to be chiefly true of a
substance that it is separable and a certain this". [}(~ )'ae TO xWeWTOP
}(at TO TOOf n lJ7rCxexHP oO}(fi p,CxALaTa Til ova[~.) More generally,
Aristotle's use of separability (TO xwew rop) as a characteristic of
substance (cf. e.g., Met. Z 1, 1028 a 24 - 25) points in the same direction.
In Met. Z I, 1028 a 17 - 20 he says of the members of the other categories
that they "are called beings because they are, some of them, quantities
108
JAAKKO HINTIKKA
of that which in that way is [sc. of substances], some qualities, some affections, some something else" .
Similar statements are found elsewhere, e.g., in Met. r 2, 1003 b
5 - 10. This passage is an especially clear indication of the fact that the
Aristotelian primacy of substance is not due to any related recognition
of the Frege - Russell distinction, for it is in the very same chapter that
Aristotle denies (as we saw in Section 1 above) in effect most emphatically this distinction. However, Aristotle provides little by way of closer
analysis of this mode of dependence of the other categories on
substances. Nevertheless it seems fairly clear that Aristotle's celebrated
manoeuvre of considering the differences between the uses of esti in different categories as not being homonymous but instances of focal meaning (7I'eos gv) is squarely based on the idea of treating substances as
something very much like individuals in the sense of a Frege - Russell ontology. (Concerning Aristotle's attempted Aufhebung of category
distinctions along these lines, see Owen (1960) and (1965).)
8. CATEGORIES, MATTER, AND FORM
We have found plenty of indication that Aristotle did not consider the
doctrine of categories as being completely satisfactory in itself. On
systematic grounds, too, it can be argued that the doctrine so far expounded is only an approximation to the real analysis of the relevant
parts of our Sprachlogik. As a matter of historical record, in the
Organon Aristotle seems to be satisfied with this approximation.
However, in the Metaphysics, especially in rand Z - H, he realizes that
he has to go beyond it. In order to understand how he does this, it is once
again advisable to turn again to topical considerations.
My discussion in Section 4 above was based on a simple analysis of
quantifier phrases. If we look away from the relative clause, the structure
we have presupposed is simply something like the following:
--
QUANT
some
every
each
NP
...............NP
109
This is too simple to be realistic, however. Indeed, Joan Bresnan has proposed the following more refined analysis:
NP
/~
/'"
QP
Det
/~
some
every
NUM
PP
/"'"NP
of
e:~h r;~
Here Q marks among other things quantity classifiers, such as the
following:
one (s)
number
part
herd
gallon
ton
110
JAAKKO HINTIKKA
the material of which the individuals over which the quantifier ranges as
being formed by imposing the individuation principle that goes together
with Bresnan's Q. The rightmost NP must in such cases involve a mass
term. The whole phrase thus expresses a kind of combination of matter
and a certain individuating principle. This principle is what Aristotle
calls the form, whereas the last NP specifies what Aristotle would call (he
matter. Thus the closer analysis of quantifier phrases just sketched is
closely related to the important Aristotelian contrast between matter and
form.
It is seen that this refinement is our analysis of quantifier phrases goes
beyond the conventional logical languages which can be traced back to
Frege and Russell. For in these languages, one starts from a given class
of basic individuals. The way they are constituted from more basic ingredients, such as matter and form, does not come up in them at all. Hence
the logic of Bresnan's refined analysis cannot be captured by means of
the usual logical languages, even when they are turned into many-sorted
languages capable of incorporating the reconstruction of Aristotle's
theory of categories outlined above.
From what has been said-it also follows that the systematic reconstruction of Aristotelian categories presented above can only be an approximation to the truth of the matter. It is based on an oversimplified
analysis of quanti fer phrases. This need not make my reconstruction of
Aristotelian categories any less interesting historically, however. On the
contrary, it seems that the pressures on the reconstruction due to its approximative character are very closely related to the reasons why the simple picture of categories so far adumbrated did not satisfy Aristotle,
either.
Thus we are beginning to see what light do these systematic observations throw on Aristotle's argumentation in the Metaphysics. First of all,
we can understand the role of one of the main concepts which Aristotle
did not use in the Categories but which he relies on heavily in
Metaphysics. That is the concept of matter. It is one of the main novelties
of the Metaphysics treatment of being and substance as compared with
the Categories.
Now the role of this concept is roughly the same in my systematic treatment as it is in Aristotle, for Aristotle, too, discusses how what in the
earlier approximation were unanalyzable values of quantifiers must now
be thought of as if they were combinations of matter and some individuality principle. This principle Aristotle labels form.
Several further similarities between Aristotle and my analysis can be
registered.
III
For one thing, Aristotle says that substances consist of matter and
form (cf. e.g., 1029 a 27 - 30). Moreover, apparently it is only substances
that do so, not members of other categories. For instance, in Met. H 4.
1044 b 8 - 9 Aristotle says that "things which exist by nature but are not
substances have no matter; their substrate is their substance". [ouo' Qaa
oh cj>vaH
J1.fV, J1.h
ouaLW Of, oux Ean TOVTOLS VA1], aAAO!
TO V7rOXfLJ1.fVOV ~ oua[a.] On the systematic side, too, it is at least questionable whether the full Bresnan analysis can be found among entities
other than individuals. Thus we can see how the introduction of the
matter-form contrast seriously upsets the symmetry between different
categories which they originally enjoyed in Aristotle's Categories.
Another relevant observation is that in many instantiations of the
Bresnan analysis we don't literally have to do with a clear matter-form
combination. In a large number of cases, we may have as the lexical instantiation of the rightmost NP, not a mass term but the plural of a count
noun. Then Q must be instantiated, not by a quantity word, but by a term
which indicates a structure the entities referred to by the count noun can
instantiate. Cases in point are the following:
some discrete set of moments of time
every large school of fish
many ordered groups of numbers.
We may describe these cases by saying that in them higher-order entities
are thought of as being from lower-order one. In contrast, earlier we
were dealing with the formation (construction) of individuals out of matter and form. It seems to me that Aristotle's notions of matter and form
are calculated to cover both formation processes. It is very dubious that
any unambiguous concept can bear such a burden. Hence we find here
some reasons to be suspicious of Aristotle's concept of matter. We have
to be very cautious, though. It might for instance seem that the more
refined analysis of quantifier phrases indicated above embodies a
mistaken admission to Aristotelian ways of thinking in that it in effect
disregards the modern contrast between count nouns and mass terms.
This distinction is linguistically much more dubious, however, than has
been realized in recent literature. Perhaps this is another direction in
which Aristotle is closer to the semantics of natural languages than
Fregean logic.
Be that as it may be, Aristotle's conception is far from unproblematic.
Aristotle's problems are increased by the fact that he assimilated the
form-matter contrast also to the traditional subject-predicate contrast
112
JAAKKO HINTIKKA
and to the actuality-potentiality contrast. It is doubtful that anyone concept can happily cover all these cases. Furthermore, the assimilation increases once again the distance between Aristotle's conceptual
framework and that of modern (Frege - Russell) logic. However, it
would require more space than I have here to follow Aristotle in these
ventures.
BIBLIOGRAPHY
Ackrill, l. L.: 1963, Aristotle's 'Categories' and 'De Interpretatione', Clarendon Press,
Oxford.
Albritton, Rogers: 1957, 'Forms of Particular Substances in Aristotle's Metaphysics',
Journal of Philosophy 54, 699 - 708.
Apelt, 0.: 1891, Beitriige zur Geschichte der griechischen Philosophie, Leipzig.
Balme, D. M.: 1980, 'Aristotle's Biology was not Essentialist', Archiv fur Geschichte der
Philosophie 62, 1 - 12.
Benardete, Seth: 1976 - 77, 'The Grammar of Being', Review of Metaphysics 30,
486-496.
Benveniste, Emile: 1966, 'Categories de pen see et categories de langue', in Problemes de
linguistique generale, Gallimard, Paris, pp. 63 - 74.
Bluck, Richard S.: 1975, Plato's 'Sophist': A Commentary, Gordon C. Neal (ed.), Manchester University Press, Manchester.
Bonitz, Hermann: 1853, 'Uber die Kategorien des Aristoteles', Sitzungsberichte der
Kaiserlichen Akademie der Wissenschaften, Phil.-Hist. Klasse, Vol. 10, pp. 591 - 645;
reprinted as a separate volume 1967 by Wissenschaftlicbe Buchgesellschaft, Darmstadt.
Bresnan, loan: unpublished seminar notes, 'On the Syntax of English Quantifiers', Stanford University.
Dancy, R. M.: 1975, Sense and Contradiction: A Study in Aristotle, D. Reidel, Dordrecht.
Dancy, R. M.: 1975, 'On Some of Aristotle's First Thoughts about Substances',
Philosophical Review 84, 338 - 373.
Dancy, R. M.: 'Aristotle and Existence', this volume, pp. 49-80.
Frede, Michael: 1967, Priidikation und Existenzaussage, Hypomnemata, Vol. 18,
Vandenhoeck & Ruprecht, Gottingen.
Furth, Montgomery: 1968, 'Elements of Eleatic Ontology', Journal of the History of
Philosophy 6, 111 - 132.
Gomez-Lobo, Alfonso: 1976 - 77, 'Aristotle's Hypotheses and Euclidean Postulates',
Review of Metaphysics 30,430-439.
Gomez-Lobo, Alfonso: 1980 - 81, 'The So-Called Question of Existence in Aristotle's An.
Post. 2. 1 - 2', Review of Metaphysics 34, 71 - 89.
Gomez-Lobo, Alfonso: 1981, 'Aristotle, Metaphysics H 2', Dialogos 38, 7 - 12.
Graeser, Andreas: 1978, 'Aristoteles und das Problem von Substanzialitat und Sein',
Freiburger Zeitschrift fur Philosophie und Theologie 25, 120 - 4 I.
Hintikka, laakko: 1959, 'Aristotle and the Ambiguity of Ambiguity', Inquiry 2,137 - 151:
reprinted with changes in Hintikka, Time and Necessity, pp. 1 - 26.
Hintikka, laakko: 1972a, 'On the Ingredients of an Aristotelian Science', Nous 6, 55 - 69.
Hintikka, laakko: 1972b, 'Leibniz on Plenitude, Relations, and the "Reign of Law" " in
Harry G. Frankfurt (ed.), Leibniz: A Collection of Critical Essays, Doubleday, Garden
113
City, pp. 152 - 190; reprinted (with an important change) in Simo Knuuttila (ed.), 1981,
ReJorging the Great Chain oj Being, D. Reidel, Dordrecht, pp. 259 - 286.
Hintikka, laakko: 1973, Time and Necessity, Clarendon Press, Oxford.
Hintikka, laakko: 1976, The Semantics oj Questions and the Questions oj Semantics,
North-Holland, Amsterdam.
Hintikka, laakko: 1979, '''Is'', Semaantical Games, and Semantical Relativity', Journal
oj Philosophical Logic 8, 433-468.
Hintikka, laakko: 1980, 'Aristotelian Axiomatics and Geometrical Axiomatics', in
laakko Hintikka et al. (eds.), Theory Change, Ancient Axiomatics, and Galileo's
Methodology, D. Reidel, Dordrecht, pp. 133 - 144.
Hintikka, laakko: 1982, 'Game-Theoretical Semantics: Insights and Prospect', Notre
Dame Journal oj Formal Logic 23, 219 - 241.
Hintikka, laakko: 1983, 'Semantical Games, the Alleged Ambiguity of "Is", and
Aristotelian Categories', Synthese 54, 443-468.
Kahn, Charles: 1966, 'The Greek Verb "To Be" and the Concept of Being', Foundations
oj Language 2, 245 - 265.
Kahn, Charles: 1973, The Verb 'Be' in Ancient Greek, D. Reidel, Dordrecht.
Kahn, Charles: 1973, 'On the Theory of the Verb "To Be" " in Milton K. Munitz (ed.),
Logic and Ontology, New York University Press, New York, pp. 1-20.
Kahn, Charles: 1976, 'Why Existence does not Emerge as a Distinct Concept in Greek
Philosophy', Archiv Jur Geschichte der Philosophie 58, 323 - 334.
Kirwan, Christopher: 1971, Aristotle's Metaphysics, Books r, ~, E, Clarendon Aristotle
Series, Clarendon Press, Oxford.
Loux, Michael 1.: 1974, translator and editor, Ockham's Theory oj Terms, Notre Dame
University Press.
Maier, Heinrich: 1896-1900, Die Syllogistik des Aristoteles, Vols. 1- 2, K. F. Koehler,
Stuttgart; reprinted 1969 by Georg Olms, Hildesheim.
Mates, Benson~ 1979, 'Identity and Predication in Plato', Phronesis 24, 211 - 229.
Moravcsik, 1. M. E. (ed.): 1967, Aristotle: A Collection oj Critical Essays, Doubleday,
Garden City.
Moravcsik, 1. M. E.: 1967, 'Aristotle's Theory of Categories', in Moravcsik (ed.), Aristotle: A Collection oj Critical Essays, pp. 125 -145.
Moravcsik, 1. M. E.: 1976, 'The Discernability of Identicals', Journal oj Philosophy 73,
587 - 598.
Owen, G. E. L.: 1%0, 'Logic and Metaphysics in Some Earlier Works of Aristotle', in 1.
During and G. E. L. Owen (eds.), Aristotle and Plato in Mid-Fourth Century (Studia
Graeca et Latina Gothoburgensia, Vol. ii), Goteborg, pp. 163 - 190.
Owen, G. E. L.: 1965, 'Aristotle in the Snares of Ontology', in R. Bambrough (ed.), New
Essays on Plato and Aristotle, Routledge and Kegan Paul, London, pp; 69 - 95.
Owen, G. E. L.: 1978-79, 'Particular and General', Proceedings oj the Aristotelian
Society 79, I - 21.
Ross, W. D.: 1924, Aristotle's Metaphysics, A Revised Text With Introduction and Commentary, Vols. 1-2, Clarendon Press, Oxford.
Ryle, Gilbert: 1937 - 38, 'Categories', Proceedings oJ the Aristotelian Society, Vol. 38, pp.
189-206; reprinted in 1971, Collected Papers, Vol. 2, Hutchinson, London, pp.
170-184.
Ryle, Gilbert: 1939, 'Plato's Parmenides', Mind 48, 129 - 151, 302 - 325.
Saarinen, Esa (ed.): 1979, Game-Theoretical Semantics, D. Reidel, Dordrecht.
Trendelenburg, Adolf: 1846, Geschichte der Kategorienlehre, Bethge, Berlin.
114
JAAKKO H!NT!KKA
Dept. of Philosophy,
Florida State University,
Tallahassee, FL 32306- /054, U.S. 4.
[Editor's note: This paper reproduces the manuscript left by the chimera, but I have added
references to books and manuscripts, plus a few notes which appear in square brackets.
The reader will notice that the chimera has wisely disregarded accidental changes of
philosophers' choices of example when they need a composite animal. The chimera takes
remarks about, e.g., the goat-stag as remarks aimed at itself. As a matter of fact, Aristotle
and the Greek Aristotelian commentators prefer the goat-stag (TQOI-YMOIIPOS) and the centaur ({1r1rOXEJlTaIlQOS). In the Hellenistic period, the centaur, the scylla and the chimera are
the standard examples. In Latin medieval texts the chimera (inherited from Manlius
Boethius) is vastly more popular than any of the other composite animals.)
"In front a lion, in the rear a serpent, in the middle a goat" - that's
what I am. Zeus bless Homer for his excellent description [Iliad 6.181].
2.400 B.C.
Philosophers exhibit an un savoury interest in the being and the nonbeing. Gorgias claims that I am not. He seems to say, though, that my
claim to being is as good as anyone's because I may be thought about
115
S. Knuuttila and J. Hintikka (eds.), The Logic of Being, 115-143.
1986 by D. Reidel Publishing Company.
116
STEN EBBESEN
('PQoPiat?m). But he does not intend to save me that way. Only to show
117
I once wished for help to understand Aristotle. During the last 500 years
or so there have been many Aristotelian commentators. By now they are
118
STEN EBBESEN
few. I therefore think it may be time to try to summarize what they have
said about us non-beings.
Some of the more recent writers have the rather crude notion that our
names are non-significative in the same way that 'blityri' and 'skindapsos' are; or that 'goat-stag', 'blityri' and 'skindapsos' share the property
of signifying non-existent things. [Stephanus, Int., CAG 18.3: 7.17 - 18;
Elias, Intr., CAG 18.1: 3.7 - 8; Ps.-Elias, Intr., 25.8, p. 52; David, Intr.,
CAG 18.2: 1.16-18; Anonymus Moraux, Intr., 111.107 -109 p. 80; cf.
Eustratius, APo., CAG 21.1: 95 - 96.] As if it made no difference that
you can describe a goat-stag or me, but not a blityri or a skindapsos since
no sort of meaning or notion is associated with those words - in fact,
they are "words" coined by philosophers precisely to show that it is
. possible to have an articulated string of sounds with n.9 meaning at all.
The better commentators - and even the naive ones, on occasion agree that 'chimera' and 'goat-stag' are significative and not to be confused with such nonsense "words" as 'blityri'. (After all, I have a
nominal definition [cf. Simplicius, Ph., CAG 9: 696.3sqq; Ps.Philoponus, APo. II, CAG 13.3: 359.26-360.9]). In their standard
theory, the semantic relation has three terms. The word signifies a concept which signifies a thing [cf. Ebbesen (l981a), 1: 141ff]. In my case,
they hold, the semantic relation is not satisfied as far as the third term
is concerned since there is no independently subsisting thing (or:
"nature") to be signified by the word via the concept; but at least there
is a concept of sorts. With 'blityri' there can be no semantic relation at
all, since both concept and thing are lacking. [Ammonius, Int., CAG 4.5:
29.8-9 (cf. 184-185); Elias, Cat., CAG 18.1: 129.15-17, Cf.
Boethius, Int. ed 2a p. 50; Ps.-Philoponus, APo. II, CAG 13.3:
362:32 - 363.2.]
I should like to know whether the commentators will allow me to be
in some Aristotelian category. The general attitude seems to be negative.
[See lamblichus apudPhiloponus, Cat., CAG 13.1: 9; Ammonius, Cat.,
CAG 4.4: 9 - 10; Dexippus, Cat., CAG 4.2: 7.20 - 24.] But perhaps Porphyry accepted me as a quasi-substance, in the sense that my name names
a concept of something which is not, as if it were a substance [Porphyrius
apud Simplicius, Cat., CAG 8: 11.6 - 12].
At any rate, if there is a concept corresponding to my name, it may
be argued that I have a sort of conceptual existence as long as people
think of me - though they may kill me by ceasing to do so - and that
it is possible to know and understand (E7rLamaiJw) what I am. This
knowledge and understanding (E7rtaT~/L7J, scfentia) will be as perishable
119
120
STEN EBBESEN
r,
121
122
STEN EBBESEN
6. A.O. 1260
After several dull centuries, the last one and a half have been exciting.
I have often wondered whether I am an individual or a universal. [The
same doubt has haunted the editors of CAG. Some print XLJLmQex and
give the references in the index nominum; others use lower-case and put
the word in their index verborum.] It is now clear that I am a universal.
An endangered species, in fact. Thus one introduction to logic says,
Est quoddam universale quod praedicatur de nullo actualiter, sed de pluribus secundum
intellectum, ut chimaera [Anon., Logica Cum sit nostra, LM 11.2: 432. 11 - 12; cf. Anon.,
Logica Ut dicit, LM 11.2: 387).
[There is one sort of universal which is predicated of nothing actually but of several
things in thought; thus (the) chimera.)
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
123
is' would be a move from 'is 2 ' (the "substantive 'is' ", as some say) to
'is l ' (which they call "adjective"); and so the move is illegitimate as it
hinges on the wrong assumption that 'is' has one sense only.
[An excellent collection of 13th-century texts bearing on this question
(as well as the others dealt with by the chimera in the present entry) is
found in Lewry (1978) and (1981a - b). The interpretation of est tertium
adiacens as a mere copula is at least as old as Abelard (see, e.g., his
Dialectica, pp. 135, 162, 164; cf. De Rijk (1981a- b)). It is earnestly
defended by Robert Bacon (early 13th c.) in his Syncategoremata; text
in Braakhuis (1979), I: 131 - 135). Robert's view is criticized in an
anonymous scholium on Sophistici Elenchi 166b37, in M. Osterr. Nat.bib!., lat. 166: 184v:
Item quaeri potest de paralogismis quos ponit Aristoteles. Sunt autem huiusmodi: 'quod
est opinabile, est; sed quod non est, est opinabile; ergo quod non est, est'; eodem modo:
'chimaera est opinabilis, ergo chimaera est'; alius paralogism us talis: 'hoc non est homo,
ergo non est'. Videtur quod in primo paralogismo sit aequivocatio, quoniam proceditur ab
hoc verbo 'est' secundum quod est substantivum ad idem secundum quod adiectivum.
Dicitur 'est' dupliciter sumi: quandoque scilicet est adiectivum, quando[que] simpliciter
praedicatur et non ponitur in numero, ut 'Socrates est'; quandoque est substantivum, ut
quando ponitur in numerum, ut 'Socrates est homo'. Fit ergo processus in praedicto
paralogismo secundum diversas acceptiones huius verbi 'est', scilicet in una est secundum
quod est substantivum, in altera secundum quod adiectivum; et ita erit ibi aequivocatio vel nulla est praedicta distinctio.
[Further, questions may be raised concerning the paralogisms in Aristotle's text. They
are as follows: 'what is opinable, is; but what is not, is opinable; therefore what is not, is';
in the same manner: 'althe chimera is opinable, therefore althe chimera is'; another
paralogism goes like this: 'this is not a man, therefore this is not'. It is arguable that in the
first paralogism we have to do with equivocation because there is a move from the verb
'is' in its substantival function to 'is' in its adjectival function. 'Is' is presumed to be
capable of two uses; sometimes it is adjectival (namely when it is precticatcd absolutely and
does not enter in a count, as in 'Socrates is'); sometimes it is substantival (namely when
it does enter in a count, as in 'Socrates is a man '). So, in the above paralogism the move
from premisses to conclusion is accompanied by a shift in the interprelation of the verb
'is', which in one proposition is laken in its substantival funclion, in another in the adjectival one; this, then, must be a case of equivocal ion - or else the said dislinclion is void.]
124
STEN EBBESEN
[As regards the criticism raised concerning the first paralogism, the answer must be that
there is no case of equivocation but a move from in-some-respect to absolutely, because
'is' in itself predicates being absolutely and actually, whereas the determination 'opinable'
makes it name being in some way - being in opinion, that is. Hence a move is performed
from being in some way - in opinion - to being absolutely and actually. Thus it may be
said that there is no foundation for claiming that the verb 'is' is equivocal, nor for claiming
that it is sometimes substantival, sometimes adjectival. For it is always substantival,
whether it enters into a count or is predicated absolutely. 1
The distinction between the predicative and the existential 'is' has met
with general disapproval, at least from the time of Robert Kilwardby (c.
1240). He flatly denies that 'is' is equivocal, and holds that the right way
to expound both 'is l ' and 'is 2 ' is "is being" [see text in Lewry (1978),
128]. This exposition brings out that a predicate has both matter and
form. The fo!'m is the means relating the matter to..the subject. When 'is'
is secundum adiacens, as in 'a/the man is', the matter of the predicate
is being simpliciter. If 'is' is tertium adiacens, the matter is a specified
sort of being - substantial in '(a/the) man is (an) animal', accidental in
'(a/the) man is just'. In his commentary on De interpretatione, Kilwardby says:
Dubitatur postea si praedicetur tertium, propter hoc quod in omni enuntiatione est medium
hoc verbum 'est' et subiectum et praedicatum extrema, cum praedicatur tertium adiacens;
et nihil unum et idem potest esse medium et extremum; ex quo sequitur quod non
praedicatur tertium. Sed intellege quod uno modo extremum, alio modo medium: ratione
compositionis medium, ratione substantiae sive rei verbi extremum. Est enim hoc ipsum
'est' praedicatum secundum materiam et formam; et dico formam praedicati medium per
quod comparatur praedicatum [medium) subiecto, et praedicatum secundum materiam
dir:o rem verbi; et ex hiis fieri unum praedicatum, ut cum dico 'homo est' id est 'homo est
ens', et sic praedicatur hoc ipsum 'est' secundum adiacens. Et quam vis sit copula aliquo
modo tertium (scilicet non ordine sed numero), quia iIIud cui adiacet (scilicet praedicatum
secundum materiam) non ponit in numero cum eo, non dicitur hoc ipsum 'est' esse tertium
adiacens, sed secundum, cum subiectum sit quod adiacet et praedicatum est quod adiacet;
et quia in hac 'homo est iustus' et 'homo est animal' praedicatur esse specificatum (scilicet
per substantiale et accidentale), quod quidem ponit in numerum cum esse simpliciter, ideo
dicitur hoc verbum 'est' in talibus praedicari tertium adiacens, ita scilicet quod sit tertium
numero, non ordine. Sic ergo aliquando praedicatur hoc verbum 'est' tertium adiacens.
[MS Cambridge, Peterhouse, 206: 7?rB.)
[Then doubt is raised whether it is predicated third. For in every statement the verb 'is'
is the mean and the subject and the predicate the extremes, when it is predicated as a third
supplement; and no one thing can be both mean and extreme. It follows that it is not
predicated third. But it should be realized that it is an extreme in one way, a mean in
another way: a mean as far as the composition is concerned, an extreme as far as the
substance or content of the verb is concerned. For this very 'is' is a predicate in respect of
matter and form. By a predicate's form I understand the mean through which the predicate
is related to the subject. By predicate in respect of matter I understand the verbal content.
Now, my claim is that these two together constitute one predicate, as when I say '(a/the)
man is' = '(a/the) man is being', and thus 'is' is predicated as a second supplem~nt. And
125
although the copula is in a way a third - not in order, that is, but in number -, as that
to which it is a supplement - viz. the predicate in respect of matter - does not add to the
count together with it, 'is' is not said to be a third supplement but a second one, since the
subject is one supplement and the predicate is another supplement. And since in the propositions '(a/the) man is just' and '(a/the) man is an animal' the being which is predicated
is specified (as substantial or accidental), and this adds to the count together with being
absolutely, for this reason the verb 'is' is said to be predicated as a third supplement in such
propositions, i.e., in such a way that it is third in number, not in order. Thus, then, 'is'
is sometimes predicated as a third supplement.)
There can be no doubt that for Kilwardby the matter of the predicate
in 'a/the chimera is opinable' is opinable being. In fact, there has been
quite a debate about the status of opinable being vis-a-vis simple being.
Is the one somehow included in the other? If opinable being were a subjective part (species) of being, it would be permissible to argue, 'this is
opinable, ergo this is'. The argument would be as good as 'Socrates is
a man, ergo Socrates is an animal'. But since the substitution of 'the
chimera' for 'this' would render the antecedent true and the consequent
false, people deny that opinable being is a species of simple being. But
then it might seem that opinable being is a wider term than being since
both the (actually) being and the non-being is opinable. [So already Alexander Aphrodisiensis, Topica, CAG 2.2: 359.17 -18 ad Top. 4.6
127a34sqq.). This view would imply that 'this is, ergo this is opinable'
is a sound inference. But then the converse inference, 'this is opinable,
ergo this is' would commit the fallacy of consequent rather than the
fallacy secundum quid et simpliciter which Aristotle says it commits.
Kilwardby, and many others, solve this difficulty by distinguishing
between a proper and an improper sense of 'opinable'. In the proper
sense, they hold, it means "in opinion only, not really". When
'opinable' is taken in this sense, the actually being does not fall under the
opinable, and so there is a fallacy secundum quid et simpliciter, but not
a fallacy of the consequent in 'this is opinable, ergo this is'. However,
in an improper and general sense 'opinable (being)' does comprise actual
being as well as non-being or imaginary being. When 'opinable' is taken
in this sense, there is a fallacy of the consequent in 'this is opinable, ergo
this is' - but then, did not Aristotle himself say that one and the same
argument may be fallacious for more than one reason?
[For an early (12th century) treatment of the above que~tion, see
Anonymus Cantabrigiensis, Commentarium in Sophisticos Elenchos,
MS Cambridge St John's College D.12: 90rB - vA (on Arist. SE 166b37).
The following texts are all doctrinally close to Kilwardby's: Anonymus
Monacensis, Commenlarium in Sophislicos Elenchos, MSS Admont
126
STEN EBBESEN
127
since everything that is can occur in opinion, but not vice versa. Exactly the same objection
is raised concerning the second paralogism. For there the fallacy of consequent from destruction of the antecedent is committed when he says 'it is not a man, therefore it is not,
as it follows vice versa, in this way: 'it is not, therefore it is not a man' .... As for the first
point, the response must be that, as Aristotle says in Book II of this work, nothing prevents
joint occurrence of several possible reasons of fallacy in the same utterance. Accordingly,
my position regarding the first utterance is that if 'opinable' is taken in a general sense,
there is the fallacy of consequent. For in one way the opinable is, in a general sense, all
that which can occur in opinion; and in that way 'it is opinable, therefore it is' is a non
sequitur, whereas it does hold vice versa. So when the opinable is taken in the general sense,
there is in this paralogism as well a fallacy of consequent as in-some-respect-andabsolutely. But when the opinable is taken in its proper sense, no fallacy of consequent occurs. For then it does not follow vice versa; thus 'it is, therefore it is in opinion only', as
that is properly speaking opinable which is in opinion only. Accordingly, I submit that
although there is in this paralogism as well a fallacy of consequent as in-some-respect-andabsolutely, more properly speaking there is here a fallacy in-some-respect-and-absolutely
- and I say "more properly" because the opinable is taken in the proper sense in somc
respect. But in the other paralogism, when it is said 'a/the donkey is not a man, therefore
it is not', there is a fallacy of consequent and likewise a fallacy in-some-respect-andabsolutely. But he introduces the paralogism for the sake of its vicious in-some-respectand-absolutely. Some, however, claim that no fallacy of consequent occurs there. For they
submit that the claims 'it is a man' and 'it is not' are compatible, and they say that Caesar
is a man even when he is not. And when they are confronted with the arg).lmcnt 'it is a man,
therefore it is', they say that this argument is not valid, and for the following reason: Every
essential and universal predicate expresses habitudinal being; but whenever the \crb 'is' is
predicated as a second supplement it expresses temporal being (or "being as of now"). Accordingly, they say that when it is argued as follows, 'Caesar is a man, therefore Caesar is'.
a fallacy of equivocation is committed because a move is performed from habitudinal being
to temporal being (or: "being as of now"). Whether they are right or wrong on this score
is scarcely a mattcr of doubt.]
128
STEN EBBESEN
not claim that I have some substantial being. Nor does 'althe chimera
does-not-run' contain a claim that I exercise any activity; all that is
posited or left is an unspecified being - in fact just as is the case when
the predicate is 'is opined' (opinatur) [see Lewry (1978), 62J or 'is
opinable' in the broad, improper sense.
I am but the most extreme example of that which is not actually. Less
extreme examples are extinct or not yet actualized species, and individuals of the past. Philosophers often perform an experiment in
thought, saying "suppose there were no men". Their standard example
of an individual of the past is Caesar (who has taken over the role Homer
used to have). Now, if people can admit that there is a sort of being which
is predicable of both me and actually existing things, they should have
little difficulty in admitting that there can be true affirmative 'is'propositions about non-actual natural species or a defunct individual.
Some do indeed claim that there is a way of being which renders this
possible. They call it "habitual being" (esse habituate) and explain that
no stronger being is needed for purposes of verification, and that all
essential predications may be interpreted as not positing any stronger being. Hence 'Caesar is a man' is true even when Caesar is not (because he
is dead), and 'man is an animal' is true whether the species has actual
representatives or not. [For the 13th-century debate about habitual being, see Ebbesen and Pin borg (1970), Braakhuis (1977, 1981), Lewry
(1981a, b), De Libera (1981), Fredborg (1981).J
There would seem to be several interpretations of esse habituate, viz. (I)
= esse habitu (as dinstinct from esse actu) , i.e. an incomplete way of being, being as a tendency to be actualized in a certain way. Thus 'man is
(habitually) an animal' means "the nature of man is such that man can
be actualized and an actual man will be an animal"; (2) = esse
habitudinis or esse consequentiae (distinguished from esse temporis or
129
esse ut nunc), i.e. a way of being which consists in entering into a relationship. Thus 'man is an animal' may mean "there is a relationship
(habitudo), viz. that of species to genus, between man and animal, such
that man entails animal and 'this is a man' entails 'this is an animal".
In other words, 'man is an animal' contains the true claim that there is
a relation of entailment (consequentia) between man and animal. [The
reader should be aware that the chimera follows general 13th-century
usage in talking about entailment as a relation between terms as well as
propositions. However, Roger Bacon (Compendium Studii Theologiae
p. 57) shows awareness that this usage is dangerous:
Adhuc cavillant de esse habitudinis, sed hoc in propositione (pro nomine ms & ed.) habet
locum, et ideo destruetur postea, cum de propositionibus fiet sermo.l
[They further use the trick of talking about habitudinal being. But this belongs in a proposition, and so it will be demolished below when I shall talk about propositions.l
130
STEN EBBESEN
[Answering the argument one may follow the view expressed by Boethius, viz . that an
infinite term does not posit anything, and say that an infinite noun extends indifferently
to the being and the not-being according to Boethius. Inasmuch as it is in just anything that
is or is not, it posits neither, but either may be truly said of it, and itself it may be truly
said of either, when the being in question is taken to be habitual.l
But habitual being has its enemies. Even Kilwardby dislikes it. For it
is clear that 'Caesar is a man' can have an esse actuate - interpretation,
not only an esse habituate - interpretation. But then, why should it be
different with 'Caesar is'? In short, 'is' becomes equivocal and 'Caesar
is a man, therefore Caesar is' will be a valid inference if 'is' is taken in
the same sense in the antecedent and the consequent, but invalid if it is
taken in the habitual sense in the antecedent and in the actual sense in
the consequent. Although he himself uses the habitus/actus distinction
in another, but not unrelated, context [De artu scientiarum
433 - 434, p. 150], Kilwardby will have none of such "an equivocal 'is'
[See the end of the extract from his Etenchi commentary, supra .]
Yet, doesn't he make being equivocal when he says that simple real being has a finite meaning and may be infinitized so as to leave unspecified
being, whereas unspecified being has no finite meaning and cannot be infinitized (or, if it can, the result will be "in-no-way-being")? [Kilwardby,
Cammentarium in Anatytica Priara, MS Cambridge Peterhouse 205:
116vB
Notandum autem quod non dicitur proprie infinitari ens nisi secundum (n.s. iecrio incerra)
quod dicit ens simpliciter secundum naturam et veri tat em et tunc privatur simpliciter ens
et derelinquit ens secundum opinionem sive secundum rem, illud tamen non ponitur per
ipsum. Si autem accipiatur ens commune secundum animam sive opinionem et additur
negatio, puto quod sola negatio erit, quia ilia privatio nihil derelinquit, cum privet tam ens
secundum rationem quam ens secundum rem. Sic enim omnino privatur ens, et sonat idem
quod nullo modo ens. Ens autem in tali communitate a<:<:eptum non l'st ali<:uius finitac
significationis, et ideo in tali acceptione non infinitatur . Et ho<: cst quod prius diximus talc
nomen non infinitari cuius non est aliqua finita significatio qualita!is.l
[But it should be noticed that being cannOl properly speaking be infinitized except in so
far as it means being absolutely, physically and truly. And in that case being. is eva<:uated
in the absolute sense, and it leaves being in opinion or in reality, but that is not posited by
it. If. however, we take it as general being in the mind or in opinion and add a negation ,
I think that the result will be a pure negation as this cvacuation Icaves nothing since it
evacuates as well being in notion as being in reality. For in this way being is totally
evacuated. and it means as much as "in no way being". Howevcr. being in this general
sense has no finite signifi<:ation, and so it cannot be infinitizcd when taken in this sense.
And that is what we expres sed earlier by saying that a name whIch is such as to have no
finite signification of quality cannot be infiniti/ed.l
131
as well as of things which are as of things which are not - the very property that makes habitual being an abomination to those who (like
Robert Bacon from Oxford) are equipped with a robust sense of reality.
By the way, Bacon is funnily old-fashioned in some respects. Thus it
reeks of the 12th century when he says that a proposition one of whose
terms has no referents is nonsensical (not: false) [Compendium Studii
Theologiae p. 62]. I think some 12th-century philosophers considered
sentences about me nonsensical. [On this aspect of 12th-century logic,
see Ebbesen (1981c).]
Something must happen. Bacon's semantics is bizarre and complex.
With others, senses of 'is' and modes of being tend to proliferate. This
may be to my advantage, but I do not think people will accept it in the
long run. Even Kilwardby's philosophy is not immune to the tendency
to multiply ways of being. And he lacks a good comprehensive theory to
back the claim that 'is' can mean either real or (in the strict sense)
opinable being, according to which environment it has in the
proposition.
7. A.D. 1300
For some thirty years a brand of philosophers whom I think one ought
to call "modists" have dominated the scene. They use the notion of
"analogy" to explain the semantics of 'man' in 'this is a dead man, ergo
this is a man' [cf. Ebbesen, 1979], and of 'is' in 'the chimera is opinable,
ergo it is'. They claim that 'is' is an analogical term. When it stands
alone, it means real being; when with an addition, the sort of being the
addition indicates. Thus the addition of 'opinable' makes 'is' stand for
being in the mind, which is a deficient sort of being ("esse deminutum").
Faced with the objection that an addition cannot change the significate
of 'is', they counter that if 'to be' had been a normal univocal term, this
would have been true. But as a matter of fact, it is "born" with the
liability to influence from additions, and so it is OK that it should change
its meaning according to the circumstances. They also say that my 'is'
and my 'opinable' mix in such a way as to form one predicate, so that
it is not possible to claim that one may take the 'is' out and predicate it
separately on the ground that 'opinable' may be detached and separately
predicated (however that was to happen). There is nothing revolutionary in most of this; but I think it is a step forward that they have
found a way to classify 'is' with other terms, so that they need fewer ad
hoc-rules for that verb. [For the above, see Incerti auctores, Quaestiones
super Sophis:icos Elenchos, CPhD vii: quu. 56, 89; Simon de
132
STEN EBBESEN
133
of them by which it is understood that they are (since this is false), but persisting in the mind
there is an understanding of them by which it is understood due to what causes, principles
and elements each of these things will be produced if it is produced, it being impossible for
any of them to be produced due to any other causes, principles and elements. And there
is an understanding of them by which it is understood which is the manner of generation
of each of these things from its principles, it being impossible for it to be produced from
its principles in any other manner.j
With Boethius the causal relationship has replaced being as the constant
factor that makes Aristotelian science possible. [Cf. Pinborg (1974,
1976).] - But of course some cannot resist the temptation to make this
a sort of being. Giles of Rome [Super Analytica Priara, quoted Pinborg
(1976), 247] speaks of an esse in causis which strongly resembles a
habitual being.
As far as I can see, what Boethius says amounts to saying that if you
can describe an object and state the precise conditions that will or would
actualize it, then that object is a legitimate citizen in the city of
understandable things. Now, if somebody could cogently argue that if a
she-goat were to conceive after an orgy during which she had had intercourse with both a snake and a lion, she would bear a chimera - 'would
the possibility of setting up such a causal explanation of my birth make
me a first-class potential entity that men could have scientific knowledge
of? On Boethius' theory it would. But he will not let me into the company
of understandable things, and so he stresses that I have no causes due to
which my being might be possible. I not only fail to be actual, I cannot
be actualized. In this I differ not only from things which must be actualized, but also from such generabilia as have causes due to which their
being is possible although they will not be actualized. In the sophism I
quoted above, he says:
Cum arguitur secundo "Quod non est, nemo potest scire quid est, ut tragelaphus et
sphinx" - verum est de illo quod non est penitus, ita quod non habet esse in re nec causas
ex qui bus esse suum est possibile; tale autem non est omne quod non est: eclipsis enim, cum
non sit, habet causas tamen in re ex quibus suum esse non solum est possibile sed etiam
necessarium, ut per easdem causas remaneat sua scientia apud animam .... Ad aliud dicendum, cum tu dicas quod omnis scientia est de ente: quia vel oportet iIIud esse ens actu vel
tale quod non habet esse, tamen causas et principia, ex qui bus suum esse est necessarium
vel possibile, habet; sicut apparet, cum eclipsis non sit, suum esse tamen necessario eveniet
ex suis causis (dico "possibile" propter generabilia: non enim omne genera bile generabitur,
habet tamen causas ex quibus ipsum esse est possibile).
[As for the second argument, that "Concerning that which is not, nobody can understand what it is (e.g., a goat-stag or a sphinx)", this is true of that which is not at all, so
that it does not have being in reality nor causes due to which its being is possible. But not
everything that is not is of this sort. An eclipse, for instance, even when it is not, does have
causes in reality due to which its being is not just possible but even necessary, so that in
134
STEN EBBESEN
virtue of these very causes an understanding of it may persist in the mind .... As for your
further claim that all understanding is about something that is, the response must be that
this will have to be either something actually being or such that without having being, it
does have causes and principles due to which its being is necessary or possible. A clear instance of the latter is an eclipse: even when it is not, its being will necessarily come about
due to its proper causes. (I say "possible" because of generable things, as not every
generable thing will be generated, yet does have causes due to which it is possible for it to
be).)
135
8. AD 1340
136
STEN EBBESEN
there never is, was or will be a time when 'this is a chimera' is actually
or possibly true. It is simply impossible that I should ever be.
I think it is due to a slip of attention that Ockham in one place [Expositio super libros Elenchorum, p. 264] accepts the traditional assignment of a fallacy secundum quid et simpliciter in the case of 'the chimera
is opinable, therefore it is' on the ground that 'if something is opinable,
it is' is false. Buridan is adamant on this point: 'the chimera is opinable'
is simply false. Here follows an extract from his quaestio 'Utrum
chimaera sit intellegibilis' [Quaestiones super Sophisticos Elenchos, MS
Krakow Blag. 736: 70r - v, minor scribal errors tacitly corrected.]
Breviter ad istam quaestionem pono primo ill am conclusionem: haec propositio est falsa
'chimaera est opinabilis' vel 'chimaera est intellegibilis' vel 'chimaera est significabilis'. Ad
quam probandam supponatur primo quod ad veritatem propositionis affirmativae requiritur subiectum et praedicatum supponere pro eodem. Secundo supponatur quod omnis
terminus supponens supponit vel pro eo quod est vel pro eo quod potest esse vel pro eo
quod fuit vel pro eo quod erit. Tertio supponatur quod sicut impossibile est chimaeram
esse, similiter impossibile est chimaeram posse esse, similiter impossibile est chimaeram
posse fuisse, similiter impossibile est chimaeram posse fieri. Hiis suppositis probatur haec
conclusio. Si haec propositio est vera 'chimaera est opinabilis', cum sit affirmativa, per
prim am suppositionem eius subiectum et praedicatum supponerent pro eodem. Sed hoc est
falsum. Falsitas probatur: subiectus terminus pro nullo supponit, ergo falsum est quod terminus subiectus et terminus praedicatus pro eodem supponunt. Consequentia nota est de
se. Antecedens probatur: omnis terminus supponens supponit pro eo quod est vel potest
esse vel fuit vel erit, sed hic terminus 'chimaera' non supponit pro eo quod est nec pro eo
quod potest esse nec pro eo quod fuit nec pro eo quod erit. Igitur hic terminus 'chimaera'
non supponit pro aliquo. Maior est supposito. Sed minor patet ex eo quod per tertiam suppositionem impossibile est chimaeram esse, similiter impossibile est chimaeram posse esse,
sim!1iter impossibile est chimaeram fuisse, similiter impossibile est chimaram fieri. Consimilem conclusionem poneremus de vacuo: supposito quod impossibile sit vacuum fuisse
vel fieri, posse esse vel esse, haec est falsa 'vacuum est opinabile', similiter 'vacuum est intellegibile' et huiusmodi.
Secunda conclusio: Non obstante quod haec sit falsa 'chimaera est significabilis',
'vacuum est significabile', tamen hic terminus 'chimaera' significat aliquid, similiter hic
terminus 'vacuum' significat aliquid. Patet hoc ex eo quod hic terminus 'chimaera'
significat caudam draconis, similiter vent rem virginis et collum ... is et caput bovis;
similiter hic terminus 'vacuum' significat corpus et significat locum. Modo quodlibet
istorum est aliquid.
Circa quod notandum quod hic terminus 'vacuum' et hic terminus 'chimaera' vocalis vel
scriptus subordinantur termino mentali in significando non incomplexo sed complexo,
propter quod talis terminus mentalis veri us debet did oratio, licet extenso nomine posset
dici terminus, non tamen complexus complexione distante quae fit mediante hoc verbo
'est', sed complexione indistante, ut si fieret talis complexio 'homo asinus' vel 'homo
equus' ...... et quia tales orationes non supponunt pro aliquo, videtur quod huiusmodi
termini 'chimaera' 'vacuum' etiam pro nullo supponunt, postquam tali bus orationibus
subordinantur in significando; nec etiam oportet terminum supponere pro omni illo quod
significat, nam hic terminus 'album' ,licet significet albedinem. sicut patet per definitioncm
137
eius exprimentem quid nomimis 'album est res habens albedinem', non supponit pro
albedine sed pro re subiecta albedini. Sic in proposito, licet hic terminus 'chimaera'
significet ventrem virginis, non tamen supponit pro eo. Similiter hic terminus 'vacuum'
licet significet corpus et similiter locum, pro nullo tamen eorum supponit.
[To deal briefly with this question, I first submit this conclusion: the proposition 'a/the
chimera is opinable' or 'a/the chimera is thinkable' or 'a/the chimera is signifiable' is false.
To prove this, let it be presupposed that for the truth of an affirmative proposition it is
required that the subject and the predicate suppone for the same. Secondly,let it be presupposed that every supponing term suppones either for that which is or for that which can
be or for that which has been or for that which will be. Thirdly, let it be presupposed that
just as it is impossible that a chimera is, so it is impossible that a chimera can be, and impossible that a chimera can have been, and impossible that a chimera can come to be. With
these presuppositions, this conclusion is provable as follows. If the proposition 'a/the
chimera is opinable' is true, then, since it is affirmative, according to the first presupposition its subject and its predicate would suppone for the same. But that is false. The falsity
of this is provable as follows. The subject term suppones for nothing, therefore it is false
that the subject term and the predicate term suppone for the same. The consequentia is selfevident. The antecedent is provable as follows. Every supponing term suppones for that
which is or can be or has been or will be; but the term 'chimera' does not suppone for that
which is, nor for that which can be, nor for that which has been, nor for that which will
be; therefore the term 'chimera' does not suppone for anything. The major is one of the
presuppositions, whereas the minor is obvious from the fact that according to the third
presupposition it is impossible that a chimera is, and similarly it is impossible that a
chimera can be, and impossible that a chimera has been, and i111possible that a chimera will
come to be. We could posit a similar conclusion with respect to the vacuum. On the presupposition that it is impossible that a vacuum has been or will come to be or can be or is, the
proposition 'a vacuum is opinable' is false, and likewise 'a vacuum is thinkable' and the
like.
Conclusion 2. Notwithstanding the fact that 'a/the chimera is signifiable' is false - and
likewise 'a vacuum is signifiable' - yet the term 'chimera' does signify something, and
similarly the term 'vacuum' does signify something. This may be seen from the fact that
the term 'chimera' signifies the tail of a dragon, and likewise the belly of a young woman
and the neck of a ... [illegible word) and the head of an ox. Similarly the term 'vacuum'
signifies body and signifies space. Now each of these things is something.
In this connection it should be noted that the term 'vacuum' and the term 'chimera' in
the sense of vocal or written terms are subordinated in signifying not to an incomplex but
to a complex mental term - for this reason such a mental term ought more truly to be
called a phrase, though through an extension of the name it might be called a term -, but
not one that is complex in virtue of the distant composition which is brought about by
means of the verb 'is', but in virtue of an indistant composition, such as woule! occur in
a combination of the type 'man ass' or 'man horse' .... And as such phrases do not suppone for anything, it appears that terms of the type represented by 'chimera' and 'vacuum'
also suppone for nothing, since in signifying they are subordinated to such phrases. Nor
is it required that a term suppone for all that it signifies. The term '(the) white', for instance, signifies whiteness, as is obvious from its nominal definition ("the white is a thing
having whiteness"), and yet it does not suppone for whiteness but for the thing that is the
subject of whiteness. So too in our case: although the term 'chimera' signifies the belly of
a young woman, it does not suppone for it. Similarly, although the term 'vacuum' signifies
body and space too, it does not suppone for either of these.)
138
STEN EBBESEN
139
The "some people" (aliqUl), mildly censured toward the end of this
text, may be Ockham. But, in general, Ockham and Buridan agree, and
all the other affirmative categorical propositions which used to be accepted as true receive the same treatment as 'the chimera is opinable'
does at Buridan's hands. 'The chimera is capable of being signified', the
chimera is not-being', even 'the chimera is a chimera'. Ockham has the
cheek to remark [Summa Logicae, p. 288]: "No proposition in which
something is predicated of the name 'chimera' taken significatively can
be truer than the one in which the name 'chimera' is predicated of itself.
But it is consistent with this that neither one nor the other should be
true. "
In Buridan's and Ockham's philosophy I have ceased to be the menace
to the theory of science that I was as long as people thought the objects
of scientific understanding were things of which something may be
predicated, and as long as they did not clearly distinguish between expressions, concepts and extramental referents. With the new semantics
and theory of scientific knowledge, in which propositions are the object
of understanding, the old difficulties disappear. If it is possible to know
that 'this is an animal composed of a goat, a snake and a lion' can never
be true, and if it is clear that the fact that there is a word 'chimera'
equivalent to 'animal composed of. .. ' does not mean that the proposition 'the chimera is nameable' is true or that 'a chimera is an animal composed of ... ' or 'the chimera is definable' are true, then there is no
dangerous similarity between the word 'chimera' and words associated
with concepts to which actual or possible individuals correspond.
So this is what is left of me: a word and a corresponding complex concept. But not a trace of old-fashioned being. The price Buridan and
Ockham must pay for their achievement is to recognize that 'is' is
equivocal. Buridan, at least, is not afraid of that consequence. Indeed,
he distinguishes between 'is' as (1) an atemporal copula, (2) a temporal
copula, (3) an abbreviation of 'is being', and (4) a catachrestic way of
saying 'means the same as'.
The only comfort I can find in this situation is that Ockham and
Buridan may not be very interested in me for my own sake (they do not
even bother to give me the same constituents parts on the different occasions on which they state my nominal definition). They assume for the
sake of argument that there cannot be such a compound animal as the
chimera. But it does not matter to them as logicians whether this assumption is true, as long as people will grant them that some name may be
subordinated to a complex concept that can never be associated with extramental referents.
140
STEN EBBESEN
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 Interpretatione (Codex Parisinus Graecus 2064), ed. by L. Tanin (Beitrage zur Klassischen
Philologie, Heft 95), Anton Hain, Meisenheim am Glan.
Aristoteles: 1831, Opera ex recensione Immanuelis Bekkeri, Vois. 1-2, Berlin.
Ashworth, E. 1.: 1977, 'Chimeras and Imaginary Objects: A Study in the Post-Medieval
Theory of Signification', Vivarium IS, 57 - 79.
Bacon, Roger: 191 I, Compendium Studii Theologiae, ed. by H. Rashdall, Aberdeen.
= Commentarii in librum
Boethius, Anicius Manlius Severinus: 1877, Int. ed.
.Aristotelis ITEPI EPMHNEIAE, pars prior, ed. by C. Meiser, Teubner, Leipzig.
Boethius, Anicius Manlius Severin us: 1880, Int. ed.2' = same edition, pars posterior,
Teubner, Leipzig.
Braakhuis, H. A. G.: 1977, 'The Views of William of Sherwood on Some Semantical
Topics and Their Relation to Those of Roger Bacon', Vivarium IS, 1 I 1- 142.
Braakhuis, H. A. G.: 1979, De J3de eeuwse tractaten over syncategorematische termen,
2 vols. (Diss., Leiden), Meppel.
Braakhuis, H. A. G.: 1981, 'English Tracts on Syncategorematic Terms from Robert
Bacon to Walter Burley', in Braakhuis et al. (1981), pp. 131 -165.
Braakhuis, H. A. G., C. H. Kneepkens, and L. M. De Rijk (eds.): 1981, English Logic
and Semantics from the End of the Twelfth Century to the Time of Ockham and
Burleigh, Artistarium, Supplementa I, Ingenium Publishers, Nijmegen.
Buridanus, Johannes: 1977, Sophismata, ed. by T. K. Scott (= Grdmmatica Speculativa
I), Frommann-Holzboog, Stuttgart - Bad-Cannstat.
Buridanus, Johannes: 1957, Summulae, tractatus 4, in M. E. Reina (ed.), 'Giovanni
Buridano, "Tractatus de Suppositionibus" " Rivista critica di storia della filosofla 12,
175 - 208 and 323 - 352.
Cassin, B.: 1980, Si Parmenide, Cahiers de philologie pub lies par Ie Centre de Recherche
Philologique de l'Universite de Lille III, Vol. 4, Editions de la Maison des Sciences de
I'Homme, Presses Universitaires de Lille.
De Libera. A.: 1981, 'Roger Bacon et Ie probleme de I'appellatio univoca', in Braakhuis
et al. (1981), pp. 193 - 234.
De Rijk, L. M.: 1981a, 'Die Wirkung der neuplatonischen Semantik auf das mittelalterliche Denken', Miscellanea Mediaevalia 13,19-35.
De Rijk, L. M.: 1981b, 'Abailard's Semantic Views in the Light of Later Developments'.
in Braakhuis et al. (1981), pp. I - 58.
Ebbesen, S.: 1979, 'The Dead Man is Alive', Synthese 40, 43 - 70.
Ebbesen, S.: 1981a, Commentators and Commentaries on Aristotle's Sophistici Elenchi,
3 vols. (Corpus Latinum Commentariorum in Aristotelem Graecorum VII, I - 3), E. J.
Brill, Leiden.
Ebbesen, S.: 1981b, 'Albert (the Great?)'s Companion to the Organon', Miscellanea
Mediaevalia 14, 89 - \03.
142
STEN EBBESEN
Ebbesen, S.: 1981c, 'The Present King of France Wears Hypothetical Shoes with
Categorical Laces. Twelfth-century Writers on Well-formedness', Medioevo 7,
91-113.
Ebbesen, S.: 1984, 'Proof a.nd its Limits According to Buridan, Summulae 8', in Z. Kaluza
and P. Vignaux (eds.), Preuve et raisons al'Universite de Paris. Logique, ontologie et
thiologie au XIV' siecle, Vrin, Paris, pp. 97 - 110.
Ebbesen, S. and J. Pinborg: 1970, 'Studies in the Logical Writings Attributed to Boethius
de Dacia', Cahiers de I'lnstitut du Moyen-Age grec et latin, Universite de Copenhague
3.
Ps.-Elias,lntr. = Pseudo-Elias (Pseudo-David), Lectures on Porphyry's lsagoge, ed. by
L. G. Westerink, North-Holland, Amsterdam, 1967.
Fredborg, K. M.: 1981, 'Roger Bacon on "Impositio vocis ad significandum" " in
Braakhuis et al. (1981), pp. 167 -191.
Harlfinger, D.: 1975, 'Edizione critica del testa del "De ia-eis" di Aristotele', in W. Leszl
(ed.), II 'De ideis' di Aristotele e la teoria Platonica delle idee (Accademia Toscana di
scienze e lettere 'La Colombaria', Studi 40, pp. 17fO, Olschki, Firenze.
Kilwardby, Robert: 1976, De Ortu Scientiarum, ed. by A. G. Judy (Auctores Britannici
Medii Aevi 4). The British Academy/The Pontifical Institute of Mediaeval Studies.
Lewry, 0.: 1978, Robert Kilwardby's Writings on the Logica Vetus (Diss., Oxford).
Lewry, 0.: 1981a, 'Robert Kilwardby on Meaning', Miscellanea Mediaevalia 13,
376- 384.
Lewry, 0.: 1981 b, 'The Oxford Condemnations of 1277 in Logic and Grammar', in
Braakhuis et al. (1981), pp. 235 - 278.
Lucretius: 1947, De rerum natura, ed. by C. Bailey, 3 vols., Oxford.
Ockham, Guillelmus de: 1979, Expositio super libros Elenchorum, ed. by F. del Punta
(Opera Philosophica et Theologica, Opera Philosophica 3), St. Bonaventure University,
St. Bonaventure, N. Y.
Ockham, Guillelmus de: 1978, Expositio in librum Perihermenias, ed. by A. Gambatese
and S. Brown (Opera Philosophica et Theologica, Opera Philosophica 2), St. Bonaventure University, St. Bonaventure, N.Y.
Ockham, Guillelmus de: 1977 -1979, Scriptum in librum primum Sen ten tiarum, Ordinatio, ed. by G. I. Etzkorn and F. E. Kelley (Opera Philosophica et Theologica,
Opera Theologica 3-4), St. Bonaventure University, St. Bonaventure, N.Y.
Ockham, Guillelmus de: 1974, Summa Logicae, ed. by Ph. Boehner, G. Gal, and S. Brown
(Opera Philosophica et Theoiogica, Opera Philosophica I), St. Bonaventure, N. Y.
Parmenides: 1966 12 , Fragmenta, in H. Diels and W. Kranz (eds.), Die Fragmente der Vorsokratiker, Vol. I, Weidmann, Dublin/Zurich.
Pinborg, Jan: 1974, 'Zur Philosophie des Boethius de Dacia. Ein Oberblick', Studia
Mediewistyczne IS, 165-185.
Pinborg, Jan: 1976a, 'Diskussionen urn die Wissenschaftstheorie an der Artistenfakultiit',
Miscellanea Mediaevalia 10, 240 - 268.
Pinborg, Jan: 1976b, 'The Summulae, Tractatus I De Introductionibus', in J. Pin borg
(ed.), The Logic of John Buridan (Opuscula Graecolatina 9), Museum Tusculanum,
Copenhagen pp. 71-90.
Rijk, L. M. de: see De Rijk.
Robert Kilwardby: see Kilwardby.
Roberts, L. N.: 1960, 'A Chimera is a Chimera: A Medieval Tautology', Journal of the
History of Ideas 21, 273 - 278.
Roger Bacon: see Bacon.
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.
KLAUS JACOBI
146
KLAUS JACOBI
147
and to this end, there must be a thing or the image of a thing for our conception (intellectus)
to be based upon. But what sort of a thing or construction of the mind do we conceive of
when we use 'if or 'and'? Does it belong to the category of substance or to that of quality
or to some other category? It must belong to one of them if we are to be able to have a
proper conception of it. 7
According to the thesis thus proposed, the copula 'est' is merely a connecting sign in affirmative statements or a sign of separation when used
with a negating word. The hearer is directed to join the conception of the
first term with that of the second or to separate the.two conceptions from
each other. The validity of the thesis is not demonstrated here, neither
148
KLAUS JACOBI
The essays in this book investigate whether the word 'to be' is indeed
ambiguous, as is assumed in most modern discussions of logic. The first
discussion sketched from Abelard's writings teaches that caution is required at the outset, in the formulation of the question itself. If asked
whether the word 'to be' always conveys the same thing or whether it
conveys different things according to context and use, Abelard would
surely have been unable to favor either alternative. For this formulation
of the question assumes that 'est' has a significative function wherever
it is employed, an assumption which Abelard does not share. In at least
one use, namely when' est' stands between two terms as the copula, it has
connective function but no content, and is to be reckoned among the
logical functors,13 the so-called syncategoremata. 14.
This thesis was obviously not uncontroversial in Abela:-d's time.
Abelard shares the following assumptions with his discussion partners:
(1) A word has significative function if and only if it conveys
something.
(2) Every noun or verb has a categorial content which it conveys.
(They are thus categorematic words.)
(3) The copula does not have a definite categorial content to convey.
How the negative thesis (3) could be reformulated in the affirmative
is a matter of dispute.
(3a) The copula can convey every categorial content in an indefinite
manner. Which definite content is meant can be determined in context
from the terms between which the copula is placed.
(3b) The copula has no content to convey. It indicates that that which
the predicate term conveys is to be joined with that which the subject
term conveys.
149
2. Theories on What Verbs Signify and the Question oj What the Word
'Esse' Signifies
A. Aristotle's Criterium jor Verbs: Consignificatio Temporis. 'Esse' as
a Linguistic Device jor the Verbalization ,oj Nouns
It is evident from the expositions of Abelard's Logica '/ngredientibus,ls
and Dialectica l6 that the theory of kinds of words was a central theme
in his time. In these works Abelard seeks to work out his own position
on a question on which many a debate had already taken place.
Following remarks by Priscian,17 grammarians q,ad attempted to
distinguish different kinds of words using the criterium of meaning.
They had manifestly taken the Aristotelian doctrine of categories as their
starting point. Each category was thought of as forming a distinct semantic range. In Abelard's writings the viability of this theory is debated only
as it applies to the verb. It can be inferred that the category of substance
encompassed only nouns, that of quantity numerals, that of quality adjectives, etc. and that there was contention as to whether the theory was
actually able to describe language adequately.
The theory teaches that the verb is a word which signifies an action or
a being-acted-upon. 18 In the discussion the objection is made that the
theory does not do justice to the full complexity of language. As counterevidence, verbs such as 'lies' (as in the sentence 'Cologne lies on the
Rhine'), 'live', 'have', and 'be' are cited. The reaction to this objection
is indicative of the kind of theory being sought. The validity of the thesis
is defended by drawing special distinctions. The object is obviously not
to make an empirical, descriptive statement as to the meaning of most
verbs but to-find a criterium which allows verbs to be distinguished from
other kinds of words. The defenders reply that the examples cited do indeed signify something other than an action or a being-acted-upon, they
signify something which must be ascribed to other categories. Since the
meaning of these words is heterogeneous, they contain a semantic ambiguity. But the different meanings, it is maintained, must be ascribed
to the different uses of the words. The words mentioned are used either
in the verbal or in the nominative function. In the verbal function they
signify either an action (active form) or a being-acted-upon (passive
form), and as such (called expressions of action) they have temporal
meaning (tense). In the nominative function, however, they signify
something belonging to another category (for example, that of being-in-
150
KLAUS JACOBI
In cases where no verb has been invented for a specific semantic content, one can form a substitute by combining a noun with a finite form
151
(4.2)
(4.3)
152
KLAUS JACOBI
is to be kept strictly separate from its use as a time word. In the meaning
of existence, 'esse' is a full and independent verb. Thus in the symbols
chosen, it does not merely correspond to the conjugational form' - t',
but to the full unit 'I/; - t'. Abelard emphasizes that one may assert the
existence not only of substances ('homo est'), but of substantivated accidents as well ('albedo est'). Differing from the grammarians with
whom he is disputing, Abelard does not consider the verb of existence
to contain a semantic ambiguity. For him, 'est' does not have the force
of 'exists as a substance' or 'exists as a quality' or 'exists as a quantity'.
It "has in all cases the same meaning", that is "is something from among
the multitude of things which exist". 32
Now the thesis:
(5)
'Esse' as a full and independent verb signifies 'existence'
does not harmonize with thesis (2) formulated above. 'Existence' is not
a categorial content. It seems that thesis (2) will have to be reformulated
for 'esse' also to be counted among the significative words. But then isn't
thesis (3a) more natural than (3b)? Or if (3a) is too unclear, would it not
be better to come to a thesis (3a '), according to which' esse' has existential import even in copulative use? Why does Abelard consider it
necessary to maintain such a sharp distinction between the use of 'esse'
as a full verb from its use as a copula, to the extent that all significative
function of the copula is denied?
153
systematic structure of Peri Hermeneias interests him less than the individual theses which Aristotle advocates. The correctness of the content
of Aristotle's remark is scarcely a subject of contention among the interpreters, however: "Verbs were above all invented to be predicated,
nouns that something be predicated of them". The "complete" proposition consists of at least a noun and a verb. 35
In Dialectica the function of .verbs to couple a predicate with a subject
is given even greater emphasis. In his commentary to Peri Hermeneias
Abelard said that this function is characteristic for verbs,36 but it was the
temporal co-signification which served as the criterium for the definition. Not so in Dialectica. Here the copulative function, on account of
which verbs were 'invented', is reckoned to be part of the 'concept' of
the verb,37 but temporal co-signification seems not even to be accorded
the recognition of being one of the verb's characteristic features. This
thoroughly new idea stands in sharp contrast to the long tradition
preceding it and must be investigated with great care.
In the section on nouns, Abelard first relates Aristotle's view on the
specific differences between the noun and the verb. 38 Then he starts to
argue against this view, the same he had defended in his commentary on
Peri Hermeneias. He asks if it is really so that only verbs, but not nouns,
bear a temporal connotation "so that they assign their main meaning to
the persons serving as subject in accordance with their tense". 39 Just as
one understands in 'curro' and 'currens' that running (cursus) is to be attributed to a person in the present, one understands in 'album' something
"which is determined by whiteness as an (accidental) form in the
present" ;40 and also when 'man' is used to denote something it is because
"a mortal animal substance endowed with reason is present", so that
"'man' amounts to 'is a mortal animal endowed with reason which exists
in the present' ".41 "Thus Aristotle seems to have been wrong in
distinguishing nouns from verbs by saying that nouns aie without temporal connotation. For nouns are shown to indicate tense, too, namely
that of the verb, that is, the present. ,,42
Does Abelard actually intend to contest all distinctions between nouns
and verbs? Does he take the extreme extensionalistic point of view that
nouns signify nothing more than their present denotata? The care which
Abelard devotes again and again to non-denotative words (that is, those
which have meaning, although they do not refer to anything which
exists)43 speaks against such a view. Then how are his remarks to be
interpreted?
'Album' does not mean 'white' here in the sense of a dictionary entry.
154
KLAUS JACOBI
155
156
KLAUS JACOBI
who seems to be good does evil. '). The point is that 'to be' cannot be
regarded simply as a predicate which is defined more precisely by a
predicate noun. The meaning of the other example sentences can be
made clearer by seeing the limitation on the predicate as a namelyrider: 'This is becoming something, namely good'; 'This seems to be
something, namely good'; 'He sees something, namely Socrates'. One
may not, however, interpret linkages using the copula 'est' as meaning
'This is something, namely
Can we thus safely conclude that 'est" in copulative function does not
retain the existential meaning it has as a verb of full value? Abelard exhorts to caution in his further discussions. One thing has been clearly
established: that the copulative use of 'est' should not be confused with
its use as a predicate. The copula is not "predicated in the essential sense
(proprie)". Its function is to link not its own semantic content but that
of the predicate term to the semantic content of the subject term.
Nonetheless, even if the copula is not predicated proprie, it is indeed
predicated per accidens. 55 And although its principal function is to link
~ubject and predicate terms, it may be that the meaning 'existence' plays
a certain role when 'est' is used as copula.
In speaking of 'esse' Abelard points out that "there is always an existential import in its linkage" and "it allows us to determine that
another thing exists", even when it is used as a copula. 56 But he says this
with regret. 57 It would be ideal if the copula had absolutely no semantic
content and functioned as a purely syncategorematic connecting
symbol. 58 'Est' remains the word which comes closest to this ideal. Any
other verb would convey a specific categorematic content. This content
would make the verb unsuited to serve as a link between the subject and
yet another such content as found in the predicate term. 59 'Est' is
suitable for use as a copula because its own meaning preoccupies least
of all and can be most easily kept in the background or, one could even
say, suppressed by the subject and predicate terms.
One can find passages in which recognition is granted to the existential
import of the copula near other passages which emphasize that the existential meaning by no means belongs to the copulative function. My
understanding of this pecularity derives from the role Abelard accorded
to the science of argumentative discourse, this being to reflect on and explain language and yet not to shy away from correcting it and bringing
to it a greater precision. The discussions carried on in these passages need
not be summarized individually here. Nonetheless I consider it important
at least to show briefly which semantic types of propositions play a
special role in them.
vI'.
157
In a first special case, the name of a person stands in the subject position of the proposition as an expression for a concrete substance. An expression for an accident stands in the predicate position. Take for exampIe 'Socrates est albus'. This predicate joins two items to the subject, first
the whiteness (albedo) as an accidental form (in adiacentia), and secondly something white (album) in existence (in essentia).60 Abelard is himself
unsure how the two should be weighted. In Logica '[ngredientibus' he
considers the speaker's real intention to be to predicate the attribution
or inherence of whiteness in the actual subject. The use of the concrete
term 'white' seems to be misleading here due to the existential import
which the copula bears. Because Socrates exists as something white, not
as whiteness, the abstract term cannot be used here, contrary to the
speaker's intention. In Dialectica. which was written later, Abelard
defended the form of the proposition which is naturally given in the
language, 'Socrates est albus' as being a sensible one. 61 The proposition
asserts that Socrates is one of the things which are white. In the first
chapter of Categories Aristotle makes a distinction between accidental
predicates, which "are in a substance" and substantial predicates, which
"are said of a substance". In reference to Socrates, 'white' belongs to
the former group. But the "inherence of attributed whiteness" is merely
"intimated" in this proposition. To make the meaning of this intimation
explicit is a task for a categorial analysis of predicates; it is not a task for
an analysis of the meaning and functions of predication.
A second special case, one which plays a considerable role ill
Abelard's investigations, is that of propositions whose subject terms
signify something which in principle or in fact does not exist. A few
classic examples: 'chimaera est opinabi/is', 'chimaera est non-existens'
and' Homerus est poeta' .!>l It is obvious that such propositions are not
to be understood as though the copula were asserting - even merely per
accidens - that such creatures as chimaeras ever existed or that Homer
were alive at the moment the proposition was uttered. Abelard and the
logicians whose writings he uses as groundwork tryout a long series of
transformations in search of a clear understanding of the meaning of the
sentences given as examples. Thus the sentence 'chimaera est opinabilis'
should be reworded as a sentence about someone who can imagine a
chimaera. 63 The true meaning of the sentence about Homer requires it
to be transformed into a sentence about a still existing poem written by
him or about a memory of him which has not died out. 64 The purpose
of the transformations is to return to sentences in which the verb 'to be',
in both in its full meaning and its copulative use, is dealing only with
158
KLAUS JACOBI
3. The Theory of the Proposition and the Question of the Exact Meaning
of the Assertive Formula <ita est in re'
In Abelard's time, the following terminological distinctions were
regarded by logicians and grammarians as being fundamental: Every
grammatically well-formed combination of several per se significative
words (dictiones), with or without the addition of co-signific~tive words,
is called a word string (oratio).69 A word string is called 'incomplete
(oratio imperfecta)' when the person hearing the sequence of words cannot help but expect to hear an additional word or words. When such an
addition is not absolutely necessary, that is, when the combination of
words is such as to enable a (relatively) full understanding, the word
string is called 'complete (perfecta). ,70 Some kinds of minimal complete
word strings are questions, wishes, orders, and propositions. 71 Their
distinguishing sign is normally the presence of an inflected verb form.
When this form is absent, as in many orders and questions, the listener
takes it as implied. The proposition (oratio enuntiativa, propositio) is
defined as a "sentence signifying something which is either true or
false". ~2
With great pertinacity Abelard uncovers the problems hiding behind
these common distinctions and definitions. His treatment of the theory
of the proposition is particularly painstaking. He begins with questions
arising out of very simple observations. In understanding any word
159
160
KLAUS JACOBI
When Abelard later distinguishes between "valid" and "empty concepts", he does not deviate from his basic assumption: "valid" concepts
are those in which the basic intention to mean a thing is fulfilled. Abelard
explains the meaning of 'valid (san us)' in relation to 'concept (intellectus)' as follows. "Every concept, whether simple or complex, through
which we heed how the thing behaves, is valid. ,,82 This explanation is illustrated with an example: the word 'man' "yields a valid concept as long
as (at least) one man exists". 'Chimaera' and 'hircocervus' are examples
of words the concepts of which do not denote any existing thing; if there
were no such thing as man, the concept of 'man', too, would be empty.
The same holds for complex expressions and that of which they are
concepts. 83
Simple words or incomplete word strings such as 'homo currens' fulfil
their naming or descriptive function when there is at least one thing
which is (in a specific respect) as the word or word string conveys. Now
it is of great importance to Abelard that a distinction can be made
between
(6.1.) the truth criterium, 'the thing is in reality so as the proposition
says', and
(6.2.) the validity criterium, 'there is some real thing which falls
under this naming or description'.
His arguments are apagogic for the most part. He finds a theory
among contemporary logicians according to which the content of a proposition is to be identified with the content of its components. 84 He
shows in a penetrating analysis that this theory is unsatisfactory: it cC'nnot explain what is asserted in propositions of necessary consequences
such as 'si est rosa, est jlos,85 or in tautologies such as 'Socrates est
Socrates' .86 Abelard's critical deliberations cannot be presented here
individually.87 The existence of the proof that (6.1.) must be differentiated from (6.2.) will have to be taken for granted. But what exactly is
the difference between the two?
In the first chapter of Peri Hermeneias, anticipating the more
thorough discussion to come, Aristotle touches on the difference between (simple and complex) nouns and propositions. The noun alone
does not signify anything which can be true or false, he says, unless "being or not-being, whether absolutely or with a definite tense, is added. ,,88
Boethius interprets this passage as a reference to the deep structure of
every proposition. He reformulates such subject-predicate-propositions
as 'homo vivit' and 'Socrates ambulat' into 'homo vivens est' and
'Socrates ambulans est'. 89 The word order is surely not arbitrary; it is re-
161
162
KLAUS JACOBI
position. And then they bring about that the things signified by the word string in the subject position are conceived of in a certain way, just as an interposed verb ('est') or a conjunction ('if. .. , then') whose coupling indicates necessity serve to do 50. 98
'. . . is the case' and '. . . is not the case' are syncategorematic expressions. Their meaning is nothing other than their function in a word
string. The content which Abelard mainly has in mind in his analyses are
events. The truth value of propositions about events is dependent upon
the point in time at which the proposition is asserted. It is the case in a
certain 'now' that Socrates is running, just as this is not the case in a different 'now'. The dictum as the content of a proposition does not receive
a specific truth value before it is actually asserted to be true or false.
Whoever understands the assertion comprehends the asserted content as
existing (or not existing, as the case may be) at the time the proposition
is asserted.
II. POSITIONS
The inquiry into the logic and semantics of 'est' was developed by
medieval authors as the inquiry into the nature of predication. It can be
seen from Abelard's discussions that the theory of propositions and
predication was a matter which provoked deep controversies. The discussion centers for the most part around the paradigm for all individual
analyses and does not occupy itself merely with puzzle solving within a
theory or the further development of an accepted one. The fact that all
parties appeal to Aristotle and Boethius only contradicts what I have said
at first glance. The deliberations of these authorities on semantics prove
on closer scrutiny to be less than firm, for they allow widely varying
interpretations.
(A) The common ground of all parties to the discussions consists of the
following assumptions:
(1) The search for a theory of the proposition is seen as a search for
a theory of predication. Following Aristotle, the basic structure of the
proposition is thus held to be that something is said about something else.
Relational propositions are put into predicative form. In 'A sees B', for
example, the object 'B' is considered to be a more precise determination
of the predicate. The first beginnings of a reorientation can be found
with William of Ockham: he investigates the supposition of both the
term preceding and following the relational word in relational propositions. The thought of constructing a theory of predicative propositions
on a necessarily more comprehensive theory of relations and thus of
163
164
KLAUS JACOBI
Accordingly, the explicit logical standard form of an affirmative proposition has three parts: a subject term (about which something is
predicated) is coupled to a predicate term (which is predicated) with an
interposed 'est'. 100 The theory of predication then presents itself as the
theory of the coupling capacity of this interposed 'est'. Now linguistically, 'est' is a form of the verb 'esse', which signifies 'to exist'. 'Est' has
this meaning in all cases when it is used as secundum adiacens in such
propositions as 'Socrates est'. It signifies here the existence of that of
which it is predicated. The question arises whether 'est' also has existential import when it is used as tertium adiacens, that is, as the copula. In
the answer to this question two contradictory stand-points part company: The first position (1.1.) maintains that the two uses of 'est' are not
related; 'est' is equivocal. The second (1.2.) maintains that the two uses
are interrelated.
(1.1.) According to the equivocation theory, the copula has no semantic content, it simply functions as a coupling device. Thought out to the
last consequence this would mean including the copula among such conjunctions as "si (if. .. , then)'. 101 The justification for this thesis is to be
found approximately as follows: 102 'est' as secundum adiacens, that is,
as a complete predicate, joins its own semantic content, namely existence, to that of the subject term. The copula has, however, a different
function. Its function is to join the semantic content of the predicate term
to that of the subject term. The copula can exercise this function only
when it has no semantic content of its own, for it is not possible for a
word to join itself and something else to the subject simultaneously. This
last sentence, which is crucial to this line of argument, is supported by
a comparison with other per se significative verbs: Just as one cannot add
a further predicate to 'Socrates currit' or 'Socrates est currens', say
'homo', one can add no further predicate to 'Socrates est' or 'Socrates
est aliquid existens'. Consequently, 'est' does not mean 'est aliquid existens' in the sentence' Socrates est homo'. By varying it a bit, one can
use this argument to attack the opposing thesis. 103 If 'est' always meant
'est ens', then the predicate term in 'Socrates est ens' would be
superfluous, seeing that it would already be contained in 'est'. To say
'Socrates est ens' would amount to saying' Socrates est ens ens', etc.
Another argument with critical intent refers to propositions in which the
subject term signifies something which does not exist. 104 The proposition
'chimaera est opinabilis' would be false if the copula had existential import. Further, the proposition 'chimaera est non-existens' would provide
ground for the paradoxical inference that the chimaera exists (by virtue
165
of the existential import of the copula) and does not exist (by virtue of
the semantic content of the predicate term). Consequently, 'Petrus est'
cannot be inferred from 'Petrus est homo', at least not by virtue of the
alleged existential import of the copula. \05
(1.2.) The point at which parties to the opposing position can assail the
equivocation theory is the following: the theory leaves one puzzle unsolved, namely why we use one and the same word, 'est', on the one hand
(as secundum adiacens), as both significative per se and copulative, and
on the other (as tertium adiacens), as merely copulative. This difficulty
weighs iheavily on the logicians of linguistic analysis of the time. If someone were to succeed in constructing a theory which explained the relationship of the two uses to each other convincingly instead of deriding
the double use of 'est' as a misleading accident of language, such a theory
would clearly be more comprehensive than the equivocation theory.
What must be demonstrated is that the word 'est' can join a term other
than itself to a subject and yet retain existential import. The opponents
of the equivocation theory want to do so by arguing for a stronger thesis:
the compatibility of existential meaning with the copulative function has
been established when one successfully derives the copulative function
from the existential meaning. And inasmuch as 'est' means everything in
being (essentia) indiscriminately, so it is said, it is capable of joining
together any and all content l06 and of being the sign of the "existence
of the thing" 107 in affirmative propositions. 'Socrates est' means
'Socrates is one of the things which exist'. When 'est' is used copulatively, so that a predicate term is added to it, what is thus determined is that
as which the subject exists, thus, for example, 'Socrates is (as) a man' or
'Socrates is (as) a person who is presently reading'. In the argument
brought to bear by the adherents of the equivocation theory that no word
can simultaneously join both itself and another term to the subject, they
overlook the special character of the semantic content of 'being': the
predicates added are not principally different from this semantic content
but are, rather, complementary predicative determinations of it. Thus,
in 'Socrates est ens' the speaker merely makes clear that he does not wish
to make a narrower determination of what Socrates is. Furthermore,
propositions whose subject terms stand for things which do not exist are
"non-literal" formulations. What they intend must be clarified by reformulating them. 'Homer' is the name of a person who no longer exists.
The proposition 'Homerus est poeta' is to be reformulated into 'the
renown of Homer is kept alive in human memory by his poetry'. 108 Propositions about the chimaera are to be interpreted as treating something
166
KLAUS JACOBI
167
168
KLAUS JACOBI
pelled to refute it, either. Abelard's theory was simply not to be found
in the universe of discourse. It was considered obvious that it is the
copula which forms the proposition inasmuch as it relates the subject and
predicate terms to each other. Now as long as the equivocation theory
(II. B. 1.1.) and the theory maintaining the semantic and functional interrelation of the verb of existence and the copula (II. B. 1.2.) competed
with each other, the thought of ending the dispute by taking up the position counter to both (II. B. 2) would not have been far-fetched. But the
actual resolution came from an unexpected source.
In the thirteenth century the reception of Aristotle's 'Metaphysics'
took place in the European schools. This is the theory of being as such.
One of the methodological instructions given by Aristotle in order to
open up a new area of research is to pay attention to the various aspects
of the meaning of words and to investigate the interrelations between
them. 116 The semantics of the word 'est' inevitably became the central
theme of discussion now. Since the equivocation theory did not satisfy
the second part of Aristotle's instruction, the theory quickly came to be
regarded as obsolete. The opposing theory, on the other hand, fitted the
theoretical requirements of the time perfectly. 117
Is "the logic of being" a theme which ontology brings to logic? Is this
theme of central importance to logic only when logic is understood as being bound up with a metaphysical "theory of being"? The readers of this
volume will be in a better position to reflect on the question than an
author who has presently but a single voice in the concert.
When surveyed from the distance of history, Abelard's contributions,
as reported on here, show a common tendency. From the discussions of
his contemporaries, Abelard was familiar with the problems which confront someone trying to answer the question whether 'est' is semantically
ambiguous. His strategy was first to remove the word 'est' from the
limelight in which it had stood for his predecessors and contemporaries.
The logic and semantics of the word 'est did not necessarily constitute
the central problem at the basis of a theory of the proposition and the
sermocinalis scientia. Abelard's question was no longer why 'to be' was
the verb used to form the proposition. The question he had to answer was
why 'to be' could be substituted for inflected verb forms. Unlike the
adherents of the equivocation theory, he sought to account for this
linguistic phenomenon. Unlike the opponents of the theory, he attempted to do so not by offering a broadly encompassing theory,IIS but by
offering one which bore a minimal burden of argumentation. 'To be'
meant that something is simply given, that it exists, has existed, or will
169
exist. And because it meant nothing more than this, it lent itself to functioning as an auxiliary verb to indicate the time for which the truth of
a proposition was to be asserted.
The theories of predication introduced in Part II of this essay are all
tenable and also flexible enough to come to terms with the material of
which such theories treat, namely with the proposition in normal
language. Abelard realized this. Although he is one who normally does
not let any opportunity for apagogic argumentation pass by, he makes
no attempt to demonstrate that the three-part analysis of the proposition
is inconsistent or not appropriate for the description of all types of propositions. Problems of the logic of tense and of propositions whose subject terms signify something which does not exist can also be resolved
within the framework of the copula theories. To this end it must be
shown that the meanings of the terms of a proposition are displaced or
amplified in a specific manner when the copula is put into the past or
future tense or when it is supplemented by a modal expression or an expression of knowing or being of an opinion.
Abelard is able to adopt the terminology of the copula theories without
reservation. When he refers to his theory as being "more elegant", he
is thinking of the problems just reviewed. But one must bear in mind here
that the logic of occasional sentences was accorded a much greater
relative weight in medieval systems of logic than in modern ones. For
Abelard, the occasional sentence is not a special case requiring the
development out of general logic of a special branch to be called the logic
of tenses. The occasional sentence is truly the normal case in his eyes, so
that logical theory must come to terms with it right from the beginning,
at its foundations. Thus a tense indicator is to be counted as a basic
feature of the predicate. It specifies the relationship between the point
in time at which the proposition is made and the content of the proposition. Is the two-part analysis indeed more elegant when seen in the context of logical theory as a whole and not merely of particular detail problems, assuming that the occasional sentence and a logic of tense corresponding to it playa central role?
What kind of a structure would a logic have to have which did not
from time to time merely bring in the two-part analysis to help clarify
specific problems but which took the two-part analysis as its foundation?
Up to now I have described the logical standard form of the proposition
rudimentarily as 'X Vt - t'. When, instead of a personal name, a general
term is placed in the subject position, distributors must be added specifying whether the predicate is to be asserted for all or only for some of the
170
KLAUS JACOBI
individuals which fall under the subject term. Now the denotative use of
a noun is based on a predication. 119 'A man' amounts to 'some thing that
is a man'. 'All men' amounts to 'every thing that is a man' .120 'Is' serves
here, as in all other cases, as a replacement for the tense indicator as contained in the verb ending. Abelard shows this to be the case in passages
on the proper conversion of temporally specific propositions. 121 He gives
the sentence, 'Every old man was once young (Omnis sen ex jui! puer)' ,
as an example. By interchanging subject and predicate, a universal affirmative proposition can be transformed into a particular affirmative proposition. Abelard emphasizes in his discussions that the tense indicator
of the predicate must also be transferred to the subject position: 'Some
of the .men who were once young ... '. Likewise, the original subject
must be given a tense indicator when it goes to the predicate position:
' ... are now old'. The original proposition can be reformulated in the
same fashion: 'All men who are now old were once young'. The form of
a proposition explained with the help of quantors and tense indicators
is then as follows:
'All/some which l/; - t, If' - t'.
It should be clear that a negating word can be added to the first predicate
as well as to the second. The tense indicator must also be given. Thus:
'All/some which [not] l/;-t(l/;-tovl/;-tm V l/;-t n ),
[not] If'-t(If'- (0 VIf'-tmVIf'-t n )'.
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.
171
Constable (eds.), Renaissance and Renewal in the Twelfth Century, Harvard University Press, Cambridge, Mass., 1982, pp. 488 - 501). Kretzmann's admirably clearly written contribution shares to a considerable
extent the themes which I have dealt with here. At the centre of his discussion is Abelard's theory of predication, more specifically the logic of 'is'.
Kretzmann distinguishes between three-piece predication and two-piece
predication, and discusses the various theories in whose terms Abelard
examines the role of the copula in three-piece predication and that of the
substantive verb in two-piece predication.
I cannot accept Kretzmann's attempt to prove that a development took
place in Abelard's thinking on these matters, one which started from an
"lngredientibus theory", then a "Dialectica revision" and a "Dialectica
theory", ending with a "Dialectica suggestion" . As Kretzmann indicates, the "lngredientibus theory" was not, even at the time when it
was written, Abelard's last word on the problems connected with the
theory of predication. The" Dialectica theory" , like the" Dialectica suggestion", can be documented by reference to the Logica 'lngredientibus'
(see above, 1.1.; 1.2.b., Note 52 and Note 55; 1.2.c. resp. 1.2.a.). The array of different theories in Abelard's work is not to be explained in
chronological terms. As the title of his article shows, Kretzmann sees
Abelard as an autumnal figure, a man reaping the harvest which had
been grown on the ground of the 'old logic'. I see Abelard as more flexible and more restless. In his discussions, he neglects no question which
has once been asked nor any suggested answer; but it is just through this
placing in juxtaposition of a variety of suggestions that he allows himself
to be led along quite new paths. And a theory which challenges the whole
of previous tradition as emphatically as does Abelard's two-piece theory
of predication must be placed more decisively in the foreground of his
work than appears in Kretzmann's paper.
The fact that it is possible precisely to describe the points of disagreement between us shows at the same time to what a great extent we are
in accord. In fact it is not only the manner of our approaches which is
very similar, nor do we merely both discover different theories of
predication in Abelard's thought; we also name, describe, and evaluate
them in ways which are largely similar. Such broad agreement encourages me to believe that in our attempts to understand this author,
we find ourselves moving in the right direction .
172
KLAUS JACOBI
NOTES
Parts of the following essay have been published in German; see Jacobi (1980) and Jacobi
(1981). The invitation to contribute to the present volume has given me a gratifying opportunity to reexamine my earlier research, to incorporate supplementary material, and to
strive toward greater precision and clarity. I wish to thank C. Sam Farler for preparing the
English translation.
ICf. Jolivet (1969), Chapter IV; de Rijk (1980). Also compare Haring (1975), who explains
the meager transmission of Abelard's works as at least partially attributable to Abelard's
style of thinking and writing. His philosophical "works" were not written as books intended to be recopied and handed down but as records of his own thinking to be used in
teaching. A thesis which he adheres to with conviction at one point in his writings may reappear later or even in a reworking of the first source as being subject to doubt or in need
of revision.
2pt' cus Abailardus, Logica 'Ingredicntibus', Super Peri ermenias, 336,27 - 340,18. Compare the notes of textual criticism in Jacobi (1981) pertaining to the edition of Abelard's
commentary on Peri Hermeneias from which I quote.
3 Cf. Petrus Abailardus, Dialectica, 118,1 - 120,20.
4 Super Peri erm., 337,11-32.
5 Loc. cit. 337,41-338,7.
6 Loc. cit. 336,24 - 37; 337,33 - 40. Cf. Nuchelmans (1973), pp. 140 - 141.
7 Super Peri erm., 338,21 - 339,4; cited 338,39 - 339,4: 'Si'vero vel 'et' sive aliae multae,
quid significant, non est promptum assignare vel cuius rei imaginem eorum intel/ectus
habeat subiectam. Si enim significant, utique intel/ectum constituunt, ad quod necesse est
esse vel rem vel imaginem rei, in qua nilatur intel/ectus. Sed cuiusmodi rem vel cuiusmodi
figmentum imaginis per 'si' sive per 'et' concipimuS, numquid ad modum substantiae vel
qualitatis vel ad quem modum, ut sanus consistat intel/ectus?
8 Loc. cit., 339,5 -7.
9 Loc. cit., 339,7- 11.
10 Loc. cit., 339,11-19.
11 Loc. cit., 339,20-32; cited 339,24-32: Licet ('est' et 'non est') intel/ectus non conslituant, quandam tamen coniunctionem vel disiunctionem intel/ectarum rerum in anima
haberijaciunt, quam tamen coniunctionem vel disiunctionem non significant, licet haberi
jaciant, quia intel/ectum non dant in se, sed intel/ectorum coadiunctionem vel separationem habere nos jaciunt. Sunt itaque tres actiones in intellectu proposilionis, intel/ectus
scilicet partium, coniunclio vel disiunctio intel/ectarum rerum. Nec est incongruum, si ea
actio, quae intel/ectus non est, sit pars intel/ectus totiusproposilionis. Cf. 339,32 - 340,2.
Cf. Nuchelmans (1973), pp. 141-142; Tweedale (1976), pp. 231-234.
12 Super Peri erm., 340,2 - 6: Sicut autem 'est'vel 'non est' coniunctiva vel disiunctiva
sunt, non significativa, ita 'si' vel 'non si' significativQS voces copulant vel separant, ut ipsa
tamen non significent, cum nullius rei in se conceptiones teneant sive verae sive fictae, sed
animum inclinant ad quendam concipiendi modum.
13 cr. Tweedale (1976), p. 234.
14 Cf. Kretzmann (1982).
15 Super Peri erm., 346,1-351,22.
16 Dial., 130,6-8; 130,26-131,3; 132,21-133,28.
17 Priscianus, Inst. gramm., 1118, t. I, p. 55,8-9; VIII I, t. I, p. 369,2-3; XVII 14, t.
II, p. 116,26 - 27. Cited by Abelard, Super Peri erm., 346, I - 2; Dial., 132,38 - 133,2.
173
174
KLAUS JACOBI
B, not In manuscript A, upon which B. Geyer's edition is based. Compare also 362, 20 - 23,
which is also to be found in Geyer's edition.
33 Aristoteles, De into c.3 16 b 7; 16 b 9 -10. Translatio Boethii, Arist. lat., 7,2 - 3; 7,5 - 6:
et est semper eorum quae de altero praedicantur nota . ... et semper eorum quae de altero
dicuntur nota est, ut eorum quae de subiecto vel in subiecto.
34 Super Peri erm., 351,23 - 354,5; cf. 357,16 - 363,24.
3S Loc. cit., 352,4 - 7: In quo innuit (Aristotelesj verba maxime propter praedicationem inventa, nomina vero propter subiectionem, ut ex his integram proposition is constitutionem
doceat.
36 Loc. cit., 351,23 - 31: "Etest semper'. Datadefinitione, qua omne verbum includit tam
copulativum praedicati quam non, tam rectum quam casuale, tamfinitum quam infinitum,
quandam proprietatem verbi supponit, ex qua vim maximam in propositione praedicativa,
de qua intendit, verbum habere monstrat. Unde in sequentibus dicet nul/am enuntiationem
absque verbo consistere. Haec est autem proprietas, quod verbum semper est nota, id est
copula, praedicatorum de altero, id est copulativum est praedicatorum, quae praedicata de
altero quam de ipsis verbis copulantibus necesse est praedicari.
37 Dial., 129, 21 - 26: Quod itaque dixit verbum semper esse notam eorum quae de altero
praedicantur, omne verbum monstravit habere officium copulandi praedicatum subiecto
nec iIIud "semper" ad temporum, immo ad verborum comprehensionem referendum est.
Potest enim verbum per se proferri nec aliquid copulare; semper tamen secundum inventionem suam copulativum est.
38 Loc. cit., 121,28-29; 122, 13-21.
39 Loc. cit., 122, 17 - 21: Nomina . .. non . .. sicut verba tempus consignijicant, ut
scilicet, quemadmodum dictum est, primam signijicationem subiectis person is secundum
tempus distribuere dicantur. Sed cur non?
40 Loc. cit., 122,22 - 29: Sicut enim 'curro' vel 'currens' cursum circa personam tanquam
ei praesentialiter inhaerentem demonstrat, ita 'album' circa substantiam albedinem
tam(quam) praesentialiter inhaerentem determinat; non enim album nisi ex praesenti
albedine dicitur. Unde et tan tum 'albi' nomen dicere videtur, quantum quidem praesentialiter albedine est informatum, sicut et 'currens' in quodam praesentialiter cursum parlicipat. Sicut enim substantivi verbi signijicatio, cui quoque tempus adiunctum est, verbis
adiungitur, sic et nominibus videtur. Cf. Twecdale (1982), p. 144.
41 Dial., 123,9-15: Quod itaque tempus verbis accidit, hoc etiam nominibus congruit,
praesens scilicet, sive ea sint substantiae sive adiacentiae vocabula. Sicut enim 'album' ex
praesenti albedine datum est, ita etiam 'homo' ex praesenti substantia animalis rationalis
mortalis, et quem hominem dicis, iam animal rationale mortale ipsum ostendis et tantumdem 'hominis' vocabulum sonat, quantum quidem praesentialiter 'est animal rationale
mortale'. Cf. de Rijk (1981 b), pp. 29 - 30.
42 Dial., 123,2 - 5: Male ergo per "sine tempore" nomina, quae etiam temporis
designativa monstrantur, Aristoteles verbis disiunxisse videtur; eiusdem, inquam, temporis
consignijicativa cuius et verba, idest praesentis.
43 Cf. Jacobi (1981), pp. 56 - 59; 67 - 68.
44 Tweedale's (1982) interpretation is different. He understands Abelard as associating a
tense with the isolated noun, generally the present (p. 146: cf. p. 144). In this case the noun
must change its meaning when it comes after a copula in past or future tense (p. 146). In
this interpretation the wording of the passage has not been attended to with sufficienti:are.
Nor is it clear how a noun in isolation is to convey a time connotation in addition to its
main meaning. Moreover, in Super Peri erm., 349,31- 33, Abelard speaks out against the
very theory which Tweedale attributes to him: Si qllis autem dicat 'homo' per adconiunc-
175
176
KLAUS JACOBI
6S
177
85
178
KLAUS JACOBI
sideramus. Unde aperte rem animalis cum re hominis copulat. Non tamen negamus idem
'est' consideratum in ipsa oratione vim verbi obtinere; sed aliud est agere de vocibus per
se consideratis, aliud de eisdem ad vim et officium quod habent in oratione posite relatis.
Nom quantum ad vim huius oration is 'homo est animal', 'est' non per se tan tum sed cum
aliis hoc solum significat, quod ilia res que est homo sit ilia res que est animal. Hoc autem
ex vi verbi habere non potest, immo ex vi substantivi. For Abelard's view compare Super
Perierm., 347,23-27; 360,13-18.
107 Super Peri erm., 357,36-358,3.
108 Dial., 135,23 - 35. In Dial., 168,11 -169, 28, these interpretations are introduced as being the opinions of Abelard's teacher V. (Ulger).
109 Super Peri erm., 361,19-25; compare above, p. 157.
110 Dial., 165,3-8; compare the further reference to the earlier discussion 170,21-31;
compare also the formulation at 167,7 - 8: secundum eos qui 'est' tertium adiacens
praedicato non componunt, sed dictionem per se sumunt.
III Super Peri erm., 359,22 - 28; 362,7 - \0; 362,20 - 23; Dial., 134,28 - 32; 161,28 - 32;
compare above, p. 155.
112 Super Peri erm., 361,33 - 362,7; 362,23 - 36; Dial., 134,32 - 135,6; 136,37 - 137,6;
138,11 - 17; compare above 1.2.c.
III Cf. Moody (1953), pp. 32- 38; de Rijk (1967), pp. \02-106; 183 -186; 203 -205;
561 - 562; the same (1956), pp. XXXVIII - XL; Pinborg (1972), pp. 46; 53 - 55; Maieru
(1972), pp. 199 - 206.
114 Aristotle's explication of the proposition based on the word 'hyparchei' surely forms
the background here.
115 See above, p. 157.
116 Cf. Aristoteles, Top. Icc. 13 and 15.
117 Cf. William of Sherwood, Syncategoremata, pp. 70-71; Thomas Aquinas, In Peri
herm., L.I, l.V, 70-73; see also Zimmermann (1971).
118 The countertheory to the equivocation theory is easily joined with metaphysical
theorems about transcendentals and the actus essendi in the thirteenth century.
119 See above p. 154.
120 Cf. Tweedale (1982), p. 149.
121 Dial., 139,12 - 140,14; cf. Super Peri erm., 350,35 - 39.
122 Cf. Dial., 160,17 - 19. At Dial., 279,8 - 282,33, Abelard distinguishes between the case
that no more men exist and the case that none have as yet existed. In the latter case, 'animaf
would have a different meaning, namely no impositio for men.
BIBLIOGRAPHY
Aristotle, De Interpretatione, ed. L. Minio-Paluello, Oxford, 1949; Translatio Boethii, ed.
L. Minio-Pa1uello, Aristoteles Latinus II 1 - 2 (Corpus Philosophorum p1edii aevi
Academiarum consociatarum auspiciis et consilio editum), Desclee de Brouwer,
Bruges - Paris, 1965.
Aristotle, Topica, ed. W. D. Ross, Oxford University Press, Oxford, 1958.
Boethius, Commentarii in lib rum Aristotelis nEPI EPMHNEIAE, ed. C. Meiser. Prima
editio [I], Teubner, Leipzig, 1877; Secunda editio [II], Teubner, Leipzig, 1880.
Boethius, De differentiis topicis, Patrologia Latina, t. 64.
Peter Abelard, Dialectica, ed. L. M. de Rijk, Van Gorcum, Assen. 1956,21970.
Peter Abelard, Glossae super Peri ermeneias, ed. B. Geyer, in Peter Abaelards
179
Philosophische Schriften, I. Die Logica 'Ingredientibus', (BGPhThMA 21,3), Aschendorff, Miinster, 1927.
Peter Abelard, Glossae super Periermenias B, cappo XII - XIV, ed. L. Minio-Paluello, in
Twelfth Century Logic. Texts and Studies II: Abaelardiana Inedita, edizioni di Storia
e Letteratura, Roma, 1958.
Peter Abelard, Glossae super Praedicamenta Aristotelis, ed. B. Geyer, in Peter Abaelards
Philosophische Schriften I. Die Logica 'Ingredientibus' (BGPhThMA 21,2), Aschendorff, Miinster, 1921.
Peter Abelard, Glossae super Topica, ed. M. dal Pra, in Pietro Abelardo, Scrittijilosojici,
Fratelli Bocca, Roma - Milano, 1954.
Priscian, Institutionum grammaticarum libri XVIII, ed. M. Hertz (Grammatici Latini, rec.
H. Keil, II, III), Teubner, Leipzig, 1855.
Thomas Aquinas, In libros Peri Hermeneias Expositio, ed. T. M. Spiazzi, Marietti, Rome,
1950.
William of Sherwood, Syncategoremata, ed. J. R. O'Donnell, Mediaeval Studies 3 (1941),
46-93.
Haring, N. M., 'Abelard Yesterday and Today', in Pierre Abilard - Pierre Ie Venerable,
Actes du Colloque international, Cluny, 2-9 juillet 1972, ed. R. Louis, J. Jolivet, J.
Chatillon, Editions du Centre National de la Recherche Scientifique, Paris, 1975, pp.
341-403.
Jacobi, K., 'Diskussionen iiber Priidikationstheorie in den logischen Schriften des Petrus
Abailardus. Versuch einer Ubersicht', in Petrus Abaelardus (1079-1142). Person,
Werk und Wirkung, ed. R. Thomas (Trierer Theol. Studien, Bd. 38), Paulin us, Trier,
1980, pp. 165 -179.
Jacobi, K., 'Die Semantik sprachlicher Ausdriicke, Ausdrucksfolgen und Aussagen in
Abailards Kommentar zu Peri hermeneias', Medioevo (1981), 41-89.
Jolivet, J., Arts du Langage et tMologie chez Abelard (Etudes de Philosophie Medievale
57), J. Vrin, Paris, 1969.
Kretzmann, N., 'Syncategoremata, exponibilia, sophismata', in The C-ambridge History of
Later Medieval Philosophy, ed. N. Kretzmann, A. Kenny, J. Pinborg, Cambridge
University Press, Cambridge, 1982, pp. 211-241.
Maieril, A., Terminologia logica della tarda scolastica, Edizioni dell' Ateneo, Roma, 1972.
Moody, E. A., Truth and Consequence in Mediaeval Logic (Studies in Logic and the Foundations of Mathematics), North-Holland Publ. Co., Amsterdam, 1953.
Nuchelmans, G., Theories of the Proposition. Ancient and medieval conceptions of the
bearers of truth and falsity, North-Holland Publ. Co., Amsterdam, 1973.
Pinborg, J., Logik und Semantik in Mittelalter. Ein Uberblick. Frommann-Holzboog,
Stuttgart - Bad Canstatt, 1972.
Rijk, L. M. de, Logica Modernorum. I1- J: The Origin and Early Development of the
Theory of Supposition, Van Gorcum, Assen, 1%7.
Rijk, L. M. de, 'La signification de la proposition (dictum propositionis) chez Abelard',
in Pierre Abilard - Pierre Ie venerable, Actes du Colloque International, Cluny, 2 - 9
Juillet 1972, ed. R. Louis, J. Jolivet, J. Chatillon, Editions du Centre National de la
Recherche Scientifique, Paris, 1975, pp. 547 - 555.
Rijk, L. M. de, 'Peter AbiiIard (1079-1142): Meister und Opfer des Scharfsinns', in
Petrus Abaelardus (1079- J142), Person, Werk und Wirkung, ed. R. Thomas (Trierer
Theol. Studien, Bd. 38), Paulinus, Trier, 1980, pp. 125 -138.
Rijk, L. M. de, 'Die Wirkung der neuplatonischen Semantik auf das mittelalterliche
Denken iiber das Sein', in Sprache und Erkenntnis im Mittelalter. Miscellanea
Mediaevaiia, Bd. 13/1, De Gruyter, Berlin-New York, 1981, pp. 19-35.
180
KLAUS JACOBI
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 existence, the "is" of predication, and the "is" of identity, which Hintikka
has labeled "the Frege trichotomy" (1979, pp. 433f), a modern student
of Thomas Aquinas's doctrine of being cannot fail to realize that this
distinction, though it seems not to have been ignored by Aquinas, is overshadowed in his writings by another distinction between two semantically different uses of the verb "be", which he borrows from Aristotle. My
aim in this paper is, first, to examine how the two distinctions are related
to one another; secondly, to show that Aquinas, though drawing these
distinctions, does not commit himself to the assumption that the verb
"be" is genuinely ambiguous! and, finally, to elucidate how Aquinas
avoids such a commitment.
Since it is an ontological and to a certain extent even a theological
rather than a logical point of view from which Aquinas approaches the
problem of the semantically different uses of the verb "be", what he has
to say concerning the logic of being is split up into a lot of scattered
remarks, mainly the by-product of metaphysical reflections, from the
larger context of which. they have to be gathered and put together like the
pieces of a jigsaw puzzle. In view of this, not a few of the moves I shall
make in my following attempt to trace a coherent picture of Aquinas's
logical treatment of the verb "be" will be little more than conjectural.
182
HERMANN WEIDEMANN
we make of the verb "be" to express the truth (i.e. the being the case)
of a true proposition,3 on the other hand:
Philosoph us, in V Metaphys. [... J, ostendit quod ens multipliciter dicitur. Uno enim
modo dicitur ens quod per decem genera dividitur: et sic ens significat aliquid in natura existens, sive sit substantia, ut homo, sive accidens, ut color. Alio modo dicitur ens, quod
significat veritatem propositionis ... (In II Sent. dis!. 34, q. I, a. I, C).4
Used in the first way, we are told by Aquinas, the verb "be" refers "to
the act of a being thing insofar as it is being, i.e. to that by which
something is called actually being in reality" (esse dicitur actus en tis in
quantum est ens, idest quo denominatur aliquid ens actu in rerum natura:
Quodl. IX, q. 2, a. 2 [3], c), e.g., to the act of living which is the being
of whatever is alive (In I Sent. dist. 33, q. 1, a. 1, ad 1). Understood in
this sense the term "being thing" (ens) applies to "something naturally
existing, be it a substance, like a man, or an accidental property, like a
colour" (In II Sent. dist. 34, q. 1, a. 1, c). Used in the second way,
however, the verb "be" serves to answer the question whether there is
(an est) such and such a thing (ibid., cf. S.th. I, q. 48,.a. 2, ad 2, De malo
q. 1, a. 1, ad 19). The distinction between the two uses of the verb "be"
so far considered thus amounts to distinguishing between two different
existential uses of this verb, which we may, following Peter Geach, call
its use in an actuality sense and its use in a there-is sense, respectively. 5
Whereas an entity A falling under one of the ten categories in virtue
of its having an essence, can be said to be in the sense that A actually exists as well as. in the sense that there is such a thing as A, an entity A
which lacks an essence, because it is rather a privation of some being than
a being in itself, can be said to be only in the latter sense. Sentences of
the form "A is" (i.e. "A exists"), where "A", to judge from the examples chosen by Aquinas, is a placeholder for concrete or abstract
general terms, like "(a) man" or "(a) colour", and for concrete or
abstract singular terms, like "Socrates" or "blindness", can thus be put
to two different uses. In analysing the second use Aquinas comes close
to the modern analysis of existence statements in terms of the existential
quantifier. What we convey by saying that blindness exists, to take one
of Aquinas's favourite examples, is, according to him, nothing but the
fact that it is true to say that something is blind (Dicitur enim, quod
caecitas est secundo modo, ex eo quod vera est propositio, qua dicitur
aliquid esse caecum: In V Metaph. lect. 9, no. 896; cf. De pot. q. 7, a.
2, ad 1). Being the privation of sight, blindness is not being in the sense
that it actually exists, however, because - unlike an animal that happens
to be blind - it has no essence which could be actualized in reality. Since
183
Socrates, e.g., has such an essence (or nature), to say that Socrates exists
(" Socrates est") is either to say, on the one hand, that he actually exists
(or is alive) as the human being that he essentially is, or to say, on the
other hand, that there is such a person as Socrates (cf. In V Metaph. lect.
9, no. 896).6
Of these two uses of the verb "be" the latter is regarded by Aquinas
as the more comprehensive; for whatever can be said to be In the sense
that it exists as a substantial thing or as an accidental property of such
a thing can also be said to be in the sense that its existence can be truly
affirmed, but not the other way round. The existence of blindness, for
example, can be truly affirmed by saying that there is blindness (caecitas
est); but none the less blindness is not to be found among the entities
which belong to the furniture of our actual world, because it is rather a
lack of actuality than something actually existing (cf. In II Sent. dist. 34,
q. 1, a. 1, c).
His distinction between these two existential uses of the verb "be"
aside, Aquinas is well aware of the difference between the "is" of
predication and the' 'is" of identity, as witness a couple of texts in which
he distinguishes between something's being predicated of something' 'in
the way of an identity" (per modum identitatis: S.th. I, q. 39, a. 5, ad
4) and something's being predicated of something "in the manner in
which a universal thing is predicated of a particular one" (sicut universale de particulari: ibid.). A predication of the latter kind, which in contradistinction to the so-called praedicatio per identitatem (S.th. I, q. 39,
a. 6, ad 2, In III Sent. dist. 7, q. 1, a. 1, c) he also calls (praedicatio) per
denominationem sive informationem (In III Sent. dist. 5, expos. textus),
is for Aquinas a predication "more properly" so called (magis propria
praedicatio: ibid.).7
Concerning predications properly so called, Aquinas draws a distinction between substantial and accidental predications, which embraces the
distinction between the two existential uses of the verb "be" already
mentioned, in that this verb, according to whether it is used in its actuality senseor in its there-is sense, either functions as a substantial predicate,
which corresponds to the question "What is ... ?" (quid est?), or as an
accidental one, corresponding to the question "Is there such a thing
as ... ?" (an est?) (cf. In II Sent. dist. 34, q. I, a. 1, c; In V Metaph. lect.
9, no. 896; De malo q. 1, a. 1, ad 19).
Aquinas's view that to use the verb "be" in its actuality sense and,
hence, existentially is to use it as a predicate which .corresponds to the
question of what a given thing is might seem rather odd. It is explained
184
HERMANN WEIDEMANN
185
is said to be true of what a subject-term stands for. 9 What Aquinas is obviously trying to say in this account of the copula is that, used in a simple
sentence of the form" Sis P" to make a statement, the copulative "is"
performs not only the predicative function of establishing a "propositional combination" 10 between the predicate-expression "P" and the
subject-expression" S" (where" S" stands proxy for general terms, like
"[a] man", II as well as for singular ones, like "Socrates"), but also the
assertive function of indicating the speaker's commitment to the truth
of what is expressed by the propositional combination of "S" and" P".
In other words, when used to make the statement that Sis P, the copula
serves not only as a propositional link between" S" and" P", but also
as an expression for the truth-claim which, whether it is warranted or
not, i.e., whether the sentence "s is P" is true or not, we are making to
the effect that" P" is true of what" S" refers to whenever we assert that
Sis P.
As for Aquinas's doctrine that it is by means of the copulative "is"
that the question whether there is such and such a thing is answered (cf.
S.th. I, q. 48, a. 2, ad 2), which seems to result from a confusion between
the predicative and the existential uses of the verb "be", it is accounted
for by the fact that existence statements of the form "There is F-ness"
or "F-ness exists" (e.g. "Blindness exists") are analysed by Aquinas in
terms of the corresponding predicative statements of the form
"Something is F', in which the "is" functions as copula. When we assert
the existence of a privation (e.g. the existence of blindness, which is the
privation of sight), "our intellect links that privation", we are told by
Aquinas, "to a subject as being something like a form (thereof)" (in-
186
HERMANN WEIDEMANN
This account seems to rest on the assumption that the predicative statement that a man is white is equivalent to the existence statement that
there is such a thing as a white man (or, in other words, such a thing as
a man who is white).
The relation between the so-called "Frege trichotomy" and what has
been called Aquinas's "fundamental ontological dichotomy" may be
depicted by the following diagram:
~"iS"Of
existence
predication
(properly so
called)
(predication of)
identity
in act
actuality sense
substantial
substantial (?)
true
there-is sense
accidental
accidental (?)
to be
II
187
scilicet ipsum ens, de quo dicit quod nihil est (ut Alexander exponit), quia
ens aequivoce dicitur de decem praedicamentis (In I Peri herm. lect. 5,
no. 70 [19]). "This explanation", Aquinas says, "does not seem to be
appropriate, for in the first place, 'being' is not, strictly speaking, said
equivocally, but according to the prior and posterior; whence, said absolutely, it is understood of that of which it is said primarily": t3
haec expositio non videtur conveniens, tum quia ens non dicitur proprie aequivoce, sed
secundum prius et posterius; unde simpliciter dictum intelligitur de eo, quod per prius
dicitur (ibid.) .
At first sight the passage just quoted might suggest that Aquinas'~ rejection of the view that the word "being" is used, strictly speaking, in an
equivocal or ambiguous way is meant by him to be confined to the use
we make of the verb "be" to express the actual existence of what falls
under one or the other of the ten categories. That this is by no means the
case, however, is shown by the fact, already pointed out, that on both
sides of his dichotomy Aquinas tends to assimilate the existential and the
predicative uses of the verb "be" by attributing to it, on the one hand,
the role of a substantial predicate when it is used in its actuality sense
and, on the other h.md, the role of an accidental predicate when it is used
in its there-is sense. In addition, Aquinas assimilates the predicative use
which, according to him, is made of the verb "be" even in a statement
of identity with the predicative use of this verb properly so called in the
following way:
In every true affirmative statement the predicate( -term) and the subject(-term) must in
some way signify what is really the same but conceptually different. That this is so is clear
in the case of statements whose llTedicate is an accidental one as well as in the case of those
whose predicate is a substantial one. For it is obvious that (the terms) 'man' and 'white'
(e.g.) are the same as regards the subject-thing (they signify), but different with regard to
the (respective) concept (under which they signify it); for the concept of (a) man is a different concept from that of (a) white (object). Similar considerations apply, when I utter
(the sentence) '(A) man is an animal'; for the very same thing which is a man is truly an
animal, because in one and the same subject-thing there is to be found both a sensitive
nature, on account of which it is called an animal, and a rational nature, on account of
which it is called a man. Hence in this case, too, predicate and subject are the same as
regards the subject-thing (they signify), but different with regard to the (respective) concept
(under which they signify it).
Even with statements in which the same thing is predicated of itself, this is in some way
the case. insofar as our intellect treats what it assigns to the subject(-position) as being on
the side of a subject-thing, whereas what it assigns to the predicate(-positilln) is treated bv
it as belonging to the nature of a form existing in a subject-thing, in 3(cordance "ith th~
saying that predicates are taken formally and subjects materially.
While it is the plurality of predicate and subject which answers to the conceptual difference, it is the (propositional) comhination (of subject and predicate) hy means of which
our intellect signifies the real identity (S.th. I. q. 13, a. 12. C).14
188
HERMANN WEIDEMANN
189
Of the two modes of being that the verb "be" can be used to signify,
namely being as being actually existent under one of the ten categories
and being as being true, "the second is comparable to the first", we are
told by Aquinas, "as an effect is to its cause; for it is from a thing's being
in reality that the propositional truth and falsity follows which our intellect signifies by means of the word 'is' insofar as it is (used as) a verbal
copula":
... iste secundus modus comparatur ad prim urn sicut effectus ad causam. Ex hoc enim
quod aliquid in rerum natura est, sequitur veritas et falsitas in propositione, quam intellectus significat per hoc verbum 'est' prout est verbal is copula (In V Metaph. lec!. 9, no. 896).
What prevents Aquinas's ontological dichotomy from reflecting a genuine ambiguity of the verb "be" is, according to the passage just
quoted, the fact that being as actually existing (or falling under one of
the ten categories), on the one hand, and being as'being true (or being
190
HERMANN WEIDEMANN
the case), on the other hand, are two modes of being of which the latter
depends on the former in such a way that the truth of what we say is effected by and founded upon the actual existence of what we talk about.
As an attempt to account for the non-ambiguity of the verb "be" the
passage to be considered next deserves especial attention. Commenting
on Aristotle's statement, made in Peri herm. (De int.) 3, 16 b 24, that the
word "being" (that is to say, the verb "is") "co-signifies some (propositional) combination" (or "composition"), Aquinas remarks that, according to Aristotle, it does so because
it does not signify such a composition principally but consequently. It primarily signifies
that which is perceived by our intellect in the mode of actuality absolutely; for 'is', said
simply, signifies to be in act, and therefore signifies in the mode of a verb. However, the
actuality which the verb 'is' principally signifies is the actuality of every form or act commonly, whether substantial or accidental. Hence, when we wish to signify that any form
or act actually is in some subject, we signify it by means of the verb 'is' [... 1; and for this
reason the verb 'is' consequently signifies a (propositional) composition (In I Peri herm.
lect. 5, no. 73(22)).23
Far from being an interpretation which could be said to reveal [he proper
sense in which Aristotle himself intended his saying that the word
"be(ing)" co-signifies (i.e. additionally signifies) some (propositional)
combination to be understood 26 the text just quoted is nevertheless
peculiarly enlightening with regard to Aquinas's own view of the relation
between being in the absolute sense of actually being (or being actually
existent) and being in the copulative sense of something's being.. true of
something else. In addition to what we have already been informed of,
namely the fact that being in the latter sense depends on being in the
former sense as an effect depends on its cause, the passage under consideration gives us the reason why it is by means of one and the same
word (namely the verb "be") that these two distinct senses can be
expressed.
For Aquinas, the actuality sense of the verb "be" seems to be, as it
were, its "focal meaning" (to borrow G. E. L. Owen's happy term),27
191
192
HERMANN WEIDEMANN
something is something, his view that the copulative being is, as it were,
only an intellectual being, "additionally invented by the (human) soul in
the act of linking a predicate to a subject" (S. tho I, q. 3, a. 4, ad 2), commits him to excluding what we would call facts from his ontology, reserving to them, in a sense which reminds one of Frege's dictum that "a fact
is a thought that is true",29 the peculiar status of true thoughts; or so it
seems.
That it is only substances and their substantial forms, on the one hand,
and accidental properties of substances on the other that Aquinas is willing to admit into his ontology is obvious from his assumption that only
things having an essence or a form, which makes them fall under one of
the ten categories, can be said actually to be (cf. Quodl. IX, q. 2, a. 2[3),
c; In II Sent. dist. 37, q. 1, a. 2, ad 3), because the actual existence of
anything is nothing but the actuality of its essence (cf. In I Sent. dist. 33,
q. 1, a. I, ad I) or its form (cf. In I Peri herm. lect. 5, no. 73[22]).
As for Aquinas's view that facts, as opposed to particular things and
their properties, are not real entities but mental ones to which something
in reality coaesponds, witness the account he gives of St Augustine's
definition of truth in terms of his ontological dichotomy in De ver., q.
I, a. I. The definition in question, according towhich "the true is that
which is" (verum est id quod est), we are told by Aquinas, can be accounted for either by saying that "it defines truth (only) insofar as it has
a foundation in reality, and not insofar as it belongs to the complete notion of truth that reality corresponds to (the way it is conceived of by our)
intellect", or by saying that, when the true is defined as that which is,
"the word 'is' is not (to be) taken in the sense in which it signifies the
act of being, but in the sense in which it indicates that our intellect is
establishing a (propositional) combination, i.e. in the sense in which it
signifies that a proposition(-al content) is being affirmed; the meaning
of (the words) 'the true is that which is' would then be that (truth obtains)
when something which is (the case) is said to be (the case)":
... dicendum quod diffinitio ilia Augustini datur de veri tate secundum quod habet fundamentum in re et non secundum id quod ratio veri completur in adaequatione rei ad intellectum. - Vel dicendum quod cum dicitur verum est id quod est, Ii est non accipitur ibi
secundum quod significat actum essendi sed secundum quod est nota intellect us componentis, prout scilicet affirmationem proposition is significat, ut sit sensus: verum est id quod
est, id est cum dicitur esse de aliquo quod est (ibid. ad I; cf. De ver. q. I, a. 10, ad I, In
I Sent. dis!. 19, q. 5, a. I, ad I).
What Aquinas seems to be contrasting here is, on the one hand, the
sense in which the expression "something which is" means "something
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.).
194
HERMANN WEIDEMANN
195
Sent. dist. 34, q. 1, a. 1, c), and, on the other hand, that it is "the actual
being in some subject(-thing) of a (certain) form or act" (formam vel actum actualiter inesse alicui subiecto: In I Peri herm. lect. 5, no. 73[22])
which the copulative "is" signifies, he seems to have in mind something
like a distinction between a word's signifying something to the effect that
it expresses a sense and its signifying something to the effect that it has
a reference. In the light of such a distinction, which is crucial in cases in
which, instead of a form actually existing in reality, a privation of some
actual being (e.g. blindness) is combined with a subject-thing by means
of the word "is" (cf. In II Sent. dist. 37, q. 1, a. 2, ad 3), it must be
noticed, as regards the passage quoted from Aquinas's Commentary on
Peri hermeneias, that, though it is the copulative sense or meaning of the
verb "be" which he intends to explain there by reference to the sense in
which "to be" means "to be in act", what he does in fact explain by
reference to this actuality sense of the verb' 'be" seems dot to be, strictly
speaking, what the copulative "is" means, but rather what it refers to,
namely a certain form's actually being in something subjected to it.
What may have helped Aquinas in this way to blur the distinction between meaning (or sense) and reference, which he is elsewhere, following
Aristotle, careful enough to observe (cf. In I Peri herm. lect. 2, no. 15[5]:
" ... necesse fuit Aristoteli dicere quod voces significant intellectus conceptiones immediate et eis mediantibus res"), is the fact that in drawing
his distinctions concerning the different uses of the verb "esse" ("to
be") he is making prominent use of the verb "significare" ("to
signify"), which is, as it were, neutral with respect to the sense/reference
distinction. 32
NOTES
1 By saying that a word is ambiguous, I mean that it has several distinct senses and can
thus be put to semantically different uses; by saying that a word is ambiguous but not genuinely so, I mean that its different senses are in some way or other systematically con
nected, for instance in such a way that "a number of secondary senses depend upon a single
primary one" (Hamlyn, 1977173, p. 6). Cf. B. Miller's distinction between "casually ambiguous" and "systematically ambiguous" expressions (1975, p. 346).
2 The question to what extent Aquinas's account of this dichotomy is faithful to the position held by Aristotle himself is not easy to answer. In the present paper it will simply be
disregarded. For Aquinas's Commentary on Aristotle's Metaphysics cf. Doig (1972),
together with the critical review by Georg Wieland, Archiv fur Geschichte der Philosophie
57 (1975), 73 - 78.
3 It must be noticed that it is, in general, a declarative sentence, and not the propositional
content thereof, that Aquinas uses the term "propositio" to refer to.
196
HERMANN WEIDEMANN
Cf. De ente et essentia, Ch. 1, De pot. q. 7, a. 2, ad 1, and the other passages listed by
Veres (1970, pp. 92 - 97).
5 Cf. Geach (1969): "Existence in the sense of actuality (Wirk/ichkeit) is several times over
emphatically distinguished in Frege's works from the existence expressed by 'there is a soand-so' (es gibt ein - ). Indeed, he says that neglect of this distinction is about the grossest
fallacy possible - a confusion between concepts of different level" (p. 65). Cf. ibid., Ch.
4 ('Form and Existence', pp. 42 - 64), where Geach is discussing "what Aquinas meant by
his term esse, or actus essendi, 'act of existing' " (p. 42). For an aHempt to defend Geach's
view on the different senses of "exists", which has been criticized by Dummett (1973, pp.
386f) , see Miller (1975).
6 Cf. Weidemann (1979).
7 Cf. McCabe (1970, p. 77).
8 Cf. Kahn (1973), Ch. VII: 'The Veridical Use' (pp. 331- 370).
9 Like Aristotle before him, Aquinas seems to have muddled together what Geach calls
"two sorts of truth", namely "the truth of propositions, and the truth of predications"
(1972, p. 15).
10 For this term cf. Strawson (1974, pp. 20 - 22). "A truth-or-falsity-yielding combination
we call a propositional combination" (p. 21).
11 If in a sentence of the form
is P" a general term functions as subject-term, the
sentence in question is traditionally called an "indefinite proposition" (cf. In I Peri herm.
lec!. 11, no. 150[8)). Since its general subject-term is not explicitly quantified, such a
sentence, e.g., "(A) man is white" (homo est a/bus: ibid.) can be treated as logically
equivalent either to a particular proposition ("Some Sis P') or to a universal one ("Every
S is P").
12 The "est" of the Marietti edition which I have enclosed in square brackets should be
deleted; cf. no. 894: " ... hoc totum, quod est homo albus, est ens secundum accidens, ut
dictum est".
13 Oesterle's translation (1962, p. 51), slightly modified.
14 My translation (for different ones cf. McCabe (1964, p. 95), Malcolm (1979, p. 394).
" Cf. S.th. I, q. 85, a. 5, ad 3: " ... compositio autem intellectu> est signum identitatis
eorum quae componuntur. Non enim intellect us sic componit, ut dicat quod homo est
albedo; sed dicit quod homo est albus, idest habens albedinem; idem autem est subiecto
quod est homo, et quod est habens albedinem". For the context of this quotation cf. Note
17 below.
16 Veatch (1974, p. 419). Similarly, Malcolm speaks of an "identity aspect of predication"
(1979, p. 394 and passim). - Referring to S.th. I, q. 13, a. 12, c, and I, q. 85, a. 5, ad
3, Veatch launches a heavy attack on Geach to the effect that "Geach in his reference to
both of these passages never gives his readers any intimation that they both contain unequivocal assertions as to the presence of an identity factor in affirmative predication"
(Veatch, 1974, p. 418). In the face of such criticism it must be acknowledged that in the
original version of his paper 'Subject and Predicate' (1950), apparently unknown to
Veatch, Geach had commented on the latter of the two passages in question as follows
(with an additional reference to S.th. I, q. 39, a. 5, ad 5): "As regards the truth-conditions
of an affirmative predication (compositio), he [Aquinas] rejects the view that subject and
predicate stand for two different objects, which we assert to be somehow combined; on the
contrary, the truth of the predication requires a certain identity of reference. Thus, if the
predicate 'white' is to be truly attached to the subject 'man' or 'Socrates', there must be
an identity of reference holding between 'man', or 'Socrate,', and 'thing that has
whiteness' ('quod est habens a/bedinem'); the two names must be idem slihiecto. Notice
"s
197
that what is here in question is the reference of a descriptive name, not of a predicate;
Aquinas does not hold. indeed he expressly denies, that predicates like 'white' stand for
objects (supponunt). His theory is that if the predicate 'white' is truly attached to a subject,
then the corresponding descriptive name 'thing that has whiteness' must somehow agree
in reference with the subject" (p. 478; cf. Malcolm, 1979, p. 3951).
Unfortunately this comment, which would have rendered Veatch's criticism almost
pointless, is absent from the rather different version, referred to by Veatch, in which 'Sub
ject and Predicate' has been incorporated into Geach's book Reference and Generality
(1968, pp. 22-46; 1980, pp. 49-72).
11 Leaving aside the controversy between Geach and Veatch, which it is not my present
concern to settle, I should like to point out the following problem, which seems to have
gone unnoticed by both authors: The text of S.th. I, q. 13, a. 12, c, suggests that to account
for the truth-conditions of a sentence of the form" Sis P" in terms of the identity of what
the subject-expression" S" stands for with what the predicate-expression" P" is true of
is to answer the question of what the sentence refers to in reality, and that to account for
the conditions of its truth in terms of the inherence of the form of P-ness in what" S"
stands for is to describe the sense expressed by the sentence, i.e. the mode of conceiving
the real identity it refers to; the text of S.th. I, q. 85, a. 5, ad 3, suggests, however, that
it is just the other way round. For the passage quoted from this text in Note 15 above is
embedded in the following context: "Invenitur autem duplex compositio in re materiali.
Prima quidem formae ad materiam; et huic respondet compositio intellectus qua totum
universale de sua parte praedicatur [... J. Secunda vero compositio est accidentis ad
subiectum; et huic reali compositioni respondet compositio intellect us secundum quam
praedicatur accidens de subiecto, ut cum dicitur 'homo est albus'. - Tamen differt compositio intellect us a compositione rei; nam ea quae componuntur in re sunt diversa; compositio autem intellectus est signum identitatis eorum quae componuntur" (ibid.). Accor
ding to this account, it is the combination of the accidental property of whiteness with a
man that corresponds to the (true) sentence "(A) man is white" in reality, whereas the iden
tity of a man with something that is white (or has whiteness) seems to be what the sentence
in question expresses as the sense in which it is to be understood: " ... et secundum hanc
identitatis rationem intellectus noster unum componit alteri praedicando" (ibid.). Other
texts relevant to the problem of reconciling the apparently different accounts given in S.th.
I, q. 13, a. 12, c, and in S.th. I, q. 85, a. 5, ad 3, are S.th. I, q. 16, a. 2, c, S.th. III, q.
16, a. 7, ad 4, In VI Metaph. lee!. 4, no. 1241, In IX Metaph. lec!. II, no. 1898.
18 If a predicate-term like "white" is combined not with a singular subject-term like
"Socrates" but with a general one like "man", a word which, like "some" or "every",
specifies "the kind of identity of reference" (Geach, 1950, p. 478; cf. p. 479) required for
the truth of the resulting sentence must be added or understood from the context (cf. Note
11 above).
19 Cf. Geach (1950, pp. 4761).
20 Cf. Hintikka (1973, p. 6), together with the critical review by Dorothea Frede,
Philosophische Rundschau 22 (1976), 237 - 242.
21 Cf. Veres (1970).
22 I have conjectured "veri" instead of the "'sui" of Mandonnet's edition, which does not
make good sense. My conjecture is based on Aquinas's remark, immediately following the
quoted passage, "sicut supra de veri tate dictum est", which refers back to dis!. 19, q. 5,
a. I, ad 1: "Vel potest dici, quod definitiones istae dantur de vero non secundum com
pletam sui rationem, sed secundum illud quod fundatur in re". Other pieces of evidence
are In II Sent. dis!. 37, q. I, a. 2, ad I ("verum dupliciter potest considcrari. Vel secundum
198
HERMANN WEIDEMANN
quod fundatur in re [... J. Vel secundum quod completur operatione animae compositionem formantis") and De ver. q. I, a. I, ad I (" ... secundum id quod ratio veri completur in adaequatione rei ad intellectum").
23 Oesterle's translation (1962, p. 53), slightly modified.
24 For a detailed analysis of this text cf. Zimmermann (1971).
25 I have deleted the comma after "formae" in the Leonine and Marietti editions, because
it is misleading. As the clause "cum volumus significare quamcumque formam vel actum
actualiter inesse alicui subiecto" shows, the words "actualitas omnis formae[,J vel actus
substantial is vel accidentalis" are not to be taken in the sense of "actuality of every form,
be it a substantial or an accidental act" ("Wirklichkeit jeder Form, sowohl eines substanzialen wie auch eines akzidentellen Aktes": Zimmermann, 1971, p. 292), but in the sense
of "actuality of every substantial or accidental form or act" (in Oesterle's translation "or
act" is missing).
26 Cf. Weidemann (1982).
21 For a critical account of the idea of "focal meaning" cf. Hamlyn (1977178).
28 Hamlyn (1977178, p. 5). Although Hamlyn states this condition as a necessary one ("It
will be possible to show the use or sense to be derivative only if ... ": ibid., my italics),
he obviously assumes that it is also sufficient; cf. p. 6: "Thus the example satisfies the condition that a sense is derivative from another when an explanation of its derivability is in
principle forthcoming. Without that it would have been a case of straight ambiguity."
29 "Eine Tatsache ist ein Gedanke, der wahr ist": Gottlob Frege, 'Der Gedanke: Eine
logische Untersuchung', Beitr. zur Phi/os. des deutschen Idealism us 1 (1918/19),58 -77;
reprinted in Gottlob Frege: Logische Untersuchungen, herausgegeben und eingeleitet von
G. Pat zig (Vandenhoeck & Ruprecht, G6ttingen, 1966), pp. 30 - 53 (quotation, p. 50). For
an English translation cf. Gottlob Frege: Logical Investigations, ed. with a preface by P. T.
Geach, transl. by P. T. Geach and R. H. Stoothoff (Basil Blackwell, Oxford, 1977);
quotation, p. 25.
30 Cf. Note 22 above.
31 Cf. Aquinas's statement that "this man committing a sin (i.e., the fact that this man
commits a sin) is a kind of mental entity insofar as it is called true": "hoc quod est istum
peccare est quoddam ens ration is prout verum dicitur" (In II Sent. dist. 37, q. I, a. 2, ad
I). Cf. also In VI Metaph.lect. 4, no. 1241:" ... compositio et divisio, in quibus est verum
et falsum, est in mente, et non in rebus [... J. Et ideo iIIud, quod est ita ens sicut verum
in tali compositione consist ens, est alterum ab his quae proprie sunt entia, quae sunt res
extra animam, quarum unaquaeque est aut quod quid est, idest substantia, aut quale, aut
quantum, aut aliquod incomplexum, quod mens copulat vel dividit."
32 For helpful comments I am grateful to Gregg Beasley.
REFERENCES
Aquinas, SI. Thomas, De ente et essentia (Leonine edition, Vol. 43, Rome, 1976).
Aquinas, SI. Thomas, Scriptum super libros Sententiarum/super Sententiis Magistri Petri
Lombardi, Vols. I -II, ed. P. Mandonnet (Lethielleux, Paris. 1929); Vol. Ill, cd. M. F.
Moos (Lethielleux. Paris, 1933); quoted: In I (II. III) Senl.
Aquinas, St. Thomas, In ArislOlelis libros Peri hermeneias exposilio (Leonine edition, Vol.
I, Rome, 1882; Marietti edition [Torino/Rome], 1964); quoted: In I (II) Peri herm.
Aquinas, SI. Thomas, In duodecim libros Melaphysicorum Aristolelis exposilio (Marietti
edition [Torino/Rome], 1964); quoted: In I (II, ... ) Melaph.
199
Aquinas, St. Thomas, Summa theologiae (Leonine edition, Vols. 4 - 12, Rome,
1888-1906); quoted: S.th.
Aquinas, SI. Thomas, Quaestiones disputatae de veritate (Leonine edition, Vol. 22, Rome,
1970 - 76); quoted: De ver.
Aquinas, SI. Thomas, Quaestiones disputatae de potentia, de malo (Marietti edition
[Torino/Rome], 1965); quoted: De pot., De malo.
Aquinas, SI. Thomas, Quaestiones quodlibetales (Marietti edition [Torino/Rome], 1956);
quoted: Quodl.
(If not otherwise indicated, translations from the Latin are my own.)
Doig, James C., Aquinas on Metaphysics: A Historico-Doctrinal Study of the Commentary on the Metaphysics (Martinus Nijhoff, The Hague, 1972).
Dummett, Michael, Frege: Philosophy of Language (Duckworth, London, 1973).
Geach, Peter Thomas, 'Subject and Predicate', Mind 59 (1950), 461 - 482.
Geach, Peter Thomas, Reference and Generality: An Examination of Some Medieval and
Modern Theories (Cornell University Press, Ithaca and London [1962], emended edition 1968; third edition, 1980).
Geach, Peter Thomas, God and the Soul (Routledge and Kegan Paul, London, 1969).
Geach, Peter Thomas, Logic Mailers (Basil Blackwell, Oxford, 1972).
Hamlyn, D. W., 'Focal Meaning', Proceedings of the Aristotelian Society 78 (1977178),
1-18.
Hintikka, Jaakko, 'Aristotle and the Ambiguity of Ambiguity', in J. Hintikka, Time and
Necessity (Oxford University Press, Oxford, 1973), pp. I - 26.
Hintikka, Jaakko, ' "Is", Semantical Games, and Semantical Relativity', Journal of
Philosophical Logic 8 (1979),433 - 468.
Kahn, Charles H., The Verb 'Be' in Ancient Greek (The Verb 'Be' and Its Synonyms:
Philosophical and Grammatical Studies, ed. by John W. M. Verhaar, Part 6; Reidel,
Dordrecht, 1973).
Malcolm. John, 'A Reconsideration of the Identity and Inherence Theories of the Copula',
Journal of the History of Philosophy 17 (1979), 383 - 400.
McCabe, Herbert, St Thomas Aquinas: Summa Theologiae, Vol. 3: Knowing and Naming
God (la. 12 - /3). Latin text. English translation, Introduction, Notes, Appendices &
Glossary (Blackfriars [in conjunction with Eyre & Spottiswoode, London, and
McGraw-Hili Book Company, New York], 1964).
McCabe, Herbert, 'Categories', in A. Kenny (ed.), Aquinas: A Collection of Critical
Essays (Macmillan, London, 1970), pp. 54 - 92.
Miller, Barry, 'In Defence of the Predicate "Exists" " Mind 84 (197),338 - 354.
Oesterle. Jean T., Aristotle: On Interpretation. Commentary by St. Thomas and Cajetan
(Peri Hermeneias) , transl. from the Latin with an Introduction (Marquette University
Press, Milwaukee, 1962).
Strawson, P. F., Subject and Predicate in Logic and Grammar (Methuen & Co., London,
1974).
Veatch, Henry, 'SI. Thomas' Do,trine of Subject and Predicate: A Possible Starting Point
for Logical Reform and Renewal', St. Thomas Aquinas 1274 -1974: Commemorative
Studies (Pontifical Institute of Mediaeval Studies, Toronto, 1974), Vol. II. pp.
401 -422.
Veres, Torno. 'Eine fundamentale ontologische Oichotomic im Den ken dcs Thoma' \on
Aquin', Philosophisches Jahrbuch 77 (1970), 81 - 98.
200
HERMANN WEIDEMANN
Weidemann, Hermann, '''Socrates est" /"There is no such thing as Pegasus": Zur Logik
singularer Existenzaussagen nach Thomas von Aquin und W. Van Orman Quine',
Philosophisches Jahrbuch 86 (1979), 42 - 59 (translated into English by David J. Marshall Jr., in Contemporary German Philosophy 3 {I 983), 159-178).
Weidemann, Hermann, 'Aristoteles tiber das isolierte Aussagewort: De into 3, 16 b
19 - 25', Archiv fur Geschichte der Philosophie 64 {I 982), 239 - 256.
Zimmermann, Albert, '''Ipsum enim ('est') nihil est" (Aristoteles, Periherm. I, C. 3).
Thomas von Aquin tiber die Bedeutung der Kopula', Miscellanea Mediaevalia 8 (Walter
de Gruyter, Berlin/New York, 1971), pp. 282-295.
Philosophisches Seminar,
Universitiit Munster,
Dompiatz 23,
D-4400 Munster, B.R.D.
SIMO KNUUTfILA
202
SIMa KNUUTTILA
substance, but rather the prime matter as the bare substratum. Aquinas
understood this to mean that the prime matter, which does not exist as
such, is a metaphysical principle having existence as the substrate of the
singular instantiations of essences. 3
Thomas Aquinas says in De ente et essentia that the nature or essence
of things, absolutely considered, is neither one nor many. Were it many,
the humanity could not be actual in Socrates. And were it one, Socrates
and Plato would be one and the same as men. Because essences as such
cannot be one or many, they do not exist as such. Although an essence,
absolutely considered, is abstracted from any being, it is said to have being in singular things and as a concept in the mind. 4 The essences exemplified in the sublunary world have being as many, because they are
instantiated in matter, which functions as a principle of individuation in
such a manner that it generates a plurality of instances of species. 5
These metaphysical ingredients of the sublunary world have being in
the extramental reality as constituents of singular things. 6 The objects
which they constitute belong to the domain of existent objects, although
they themselves do not exist. From the metaphysical point of view we can
characterize subjects of existence we are immediately acquainted with as
actualities of essences individuated by matter. Aquinas believed that on
higher levels of the great chain of being, essences can be actualized
without being individuated by the prime matter, 7 and so he refers in his
more general accounts of the quidditive being simply to the individual instances of actuality of essences. 8
The essences, which as such are neither one nor many, are intelligible.
An essential part of Plato's and Aristotle's worldview is the belief that
there is an intelligible framework of the cosmos consisting in a system of
invariant forms of being. For them the most attractive part of the human
perfectibility is the possibility of ascending into an immediate knowledge
of this intelligible structure of the sensible reality. 9 Thomas Aquinas includes this kind of jelicitas theoretica in his conception of the final end
of man. lo However, there are some remarkable reformulations in his
handling of the subject.
Prime matter has no properties of its own and it is unintelligible.
Therefore, when essences are individuated by prime matter they become
parts of compositions which, because of the unintelligible factor present
in them, are not intelligible as such. II However, the human intellect can
u'nderstand that there is an intelligible structure in the extra mental reality, because it can abstract the intelligible essence or form from the
matter. 12 Through sensations we apprehend particular objects. From
203
204
SIMO KNUUTTILA
205
thought that the area of being manifests itself to the human intellect in
different modes of being (modi essendl) and that there are fixed modes
of intellection (modi intelligendl) , through which a conceptual picture of
the extramental reality is formed in the intellect. Words as significant
units signify directly something which exists in the intellect. They refer
to extramental objects indirectly, through the intellect (sub modo
intelligendl).26 Bearing this triadic view of signification in mind, we can
say that in the first sentence of the above Guotation Aquinas claims that
a univocal name signifies in all its applications one and the same element of the conceptual correlate of the extramental reality. Correspondingly, an equivocal name signifies at least two different ingredients of
the intelligible order having no reference to each other in their
definitions.
In his characterization of an analogous name Aquinas says that in its
applications there is one basic use connected with a certain meaning. This
meaning is implied in the prima Jacie different meanings an analogous
word seems to have when predicated of various types of things. 27 This
corresponds roughly to what Aristotle says about the pros hen
equivocity.28 It is not quite clear how the new terminology was introduced in Scholasticism; at all events, it seems to have been commonly applied in the first half of the thirteenth century. 29 For example, in the
Summe Metenses written before 1220, the class of analogical terms is
delineated as follows:
Equivocity is understood in two ways, properly and commonly. The proper equivocity
belongs to signifying units (dictio) which are actually multiplex. The common equivocity
belongs to signifying units which are related to many, to one primarily and to others ex consequenti. And so analogical words are said to be equivocal, e.g., ens. unum, and aliquid.
They are primarily said of substance, and secondarily (per posterius) of quantity, quality
etc.; and similarly "healthy" is said primarily of animals and ex consequenti of urine and
drinks and food. 30
206
SIMO KNUUTTILA
207
There are some recent studies on this disjunctive approach to the concept
of being which is the direct object of Scotus's criticism. 41
In one of his arguments for the univocal metaphysical concept of being
Scot us says that it is possible to know of something that it is without
knowing what kind of being it is.42 In the disjunctive approach this kind
of knowledge could be interpreted so that we know that one or another
of the analogous meanings of being is applicable to something although
we don't know which one. Scotus's point is now that when something is
known to be in this way, that something is at the same time conceived
recognizable, although we cannot identify it. 43 Things considered identifiable can be univocally referred to by ens, whatever they are.
Scotus's position could be explained to the defenders of the disjunctive
approach as follows: Imagine that an omniscient identifier has made a
list in which a proper name is given to anything tqat is in the actual world
and can be called a being in some of the many analogous meanings of
the word. Whenever we hear a name taken from that list we know of that
208
SIMO KNUUTTILA
209
210
SIMO KNUUTIILA
211
212
SIMO KNUUTTILA
213
214
SIMO KNUUTTILA
perfectionem intellectus, et haec sunt species et genera rerum, et rationes eorum." S.th. I,
q. 12, a. 8, ad 4.
11 S.th. I, q. 79, a. 3c, q. 86, a. I, ad 3.
12 S.th. I, q. 54, a. 4c, q. 79, a. 4, q. 85, a. I.
13 Abstrahit autem intellectus agens species intelligibiles a phantasmatibus, inquantum per
virtutem intellectus agentis accipere possumus in nostra consideratione naturas specierum
sine individualibus conditionibus, secundum quarum similitudines intellectus possibilis informatur." S.th. I, q. 85, a. I, ad 4. "Species intelligibilis est similitudo ipsius essentiae
rei, et est quodammodo ipsa quidditas et natura rei secundum esse intelligibile, non secundum esse naturale, prout est in rebus." Quodl. VIII, q. 2, a. 2c. See also S.th. I, q. 14,
a. 12c.
14 "Unde necesse est quod species intelligibilis, quae est principium operation is intellectualis, differat a verbo cordis, quod est per operationem intellectus formatum; quamvis ipsum verbum possit dici forma vel species intelligibilis, sicut per intellectum constituta, prout forma artis quam intellect us adinvenit, dicitur quaedam species intelligibilis." Quodl.
V, q. 5, a. 2c.
IS "Nam primo quidem consideratur passio intellect us possibilis secundum quod informatur specie intelligibili. Qua quidem formatus, format secundo vel definitionem vel divisionem vel compositionem, quae per vocem significatur. Unde ratio quam significat
nomen, est definitio; et enuntiatio significat compositionem et division em intellectus. Non
ergo voces significant ipsas species intelligibiles, sed ea quae intellect us sibi format ad
iudicandum de rebus exterioribus." S.th. I, q. 85, a. 2, ad 3; cf. ad 2. For further texts and
literature pertaining to these questions see R. W. Schmidt, The Domain of Logic according
to Saint Thomas Aquinas, M. Nijhoff, The Hague, 1966, pp. 98-117, E. P. Mahoney,
'Sense, Intellect; and Imagination in Albert, Thomas, and Siger', in N. Kretzmann, A.
Kenny, J. Pinborg (eds.), The Cambridge History of Later Medieval Philosophy, Cambridge University Press, Cambridge, 1982, pp. 605 - 611.
16 S.th. I, q. 12, a. 4c, q. 55, a. 2c, q. 58, a. 3 - 4, q. 79, a. 2c, q. 84, a. 7c, q. 85, a. 5c.
17 S.th. I, q. 84, a. 7c and 8c, q. 85, a. Ic.
18 S.th. I, q. 13, a. 8, ad 2, III, q. 25, a. 5, ad 4.
19 S.th. I, q. 29, a. I, ad 3, q. 77, a. I, ad 7, q. 87, a. Ic, In II Post. an, Ie. 13, n. 533,
De spirit. creat., a. II, ad 3, ScG IV, c. I, De ver. q. 10, a. I. For further examples, see
W. H. Kane, 'The Extent of Natural Philosophy', New Scholasticism 31 (1957),90-92;
cf. A. Kenny, The Five Ways. St. Thomas Aquinas' Proofs of God's Existence (Studies
in Ethics and the Philosophy of Religion), Schock en Books, New York, 1969, p. 90.
20 S.th. I, q. 12, a. 8c and ad 4.
21 'Time and Modality in Scholasticism', in S. Knuuttila (ed.), Reforging the Great Chain
of Being (Synthese Historical Library 20), D. Reidel Publ. Co., Dordrecht, 1981, pp.
163 -207.
22 Ibid., pp. 208-217.
23 For some relevant texts, see Note 20, ScG III, 56, De un. verbi inc. a.l, In II Post. an.,
1.6.
24 See, for example, S.th. I, q. 13, a. I, a. 4, a. 9, ad 2, q. 85, a. 2, ad 3, De pot. q. 8, a. I.
2S See J. Pinborg, 'Speculative Grammar', in The Cambridge History of Later Medieval
Philosophy (Note 15 above), pp. 262 - 265, J. Pinborg, 'Die Logik der Modistae', Studia
Mediewistyczne 16 (1975), 39 - 97.
26 See, for example, S.th. I, q. 13, a. 1-4, q. 45, a. 2, ad 2, In VII Metaphys., Ie. I, n.
9; R. M. Mcinerny, The Logic of Analogy. An Interpretation of St. Thomas Aquinas, M.
Nijhoff, The Hague, 1961, pp. 49-66.
215
216
SIMO KNUUTTlLA
217
Ord. IV, d. 8, q. I, n. 2 (ed. Vives XVII), cf. Quod/. q. 3, n. 2 (ed. Vives XXV).
Ord. I, d. 43, q. un., n. 14 (ed. Vat VI): ... per ipsam potentiam "sub ratione qua est
omnipotentia" non habet obiectum quod sit primo possibile, sed per intellectum divinum,
producentem illud primo in esse intelligibili, et intellectus non est formaliter potentia activa
qua Deus dicitur omnipotens; et tunc res producta in tali esse ab intellectu divino - scilicet
intelligibili - in primo instanti naturae, habet se ipsa esse possibili in secundo instanti
naturae, quia formaliter non repugnat sibi esse et se ipso formal iter repugnat sibi habere
esse necessarium ex se." Whatever can be thought of is produced in intelligible esse by the
Divine Intellect. Possible objects are identifiable as members of possible worlds, of which
the Divine Will chooses one to be the actual world (cf. Lect. I, d. 39, q. 1 - 5, n. 62 - 63).
Possible individuals have a positive nature as identifiable candidates to existence, although
they as such have no kind of existence. Cf. Ord. I, d. 36, n. 61. For the connection between
Scotus' ideas of being and possibility, see also L. Honnefelder, op.cit. and 'Die Lehre von
der doppelten ratitudo entis und ihre Bedeutung fiir die Metaphysik des Johannes Duns
Scotus', Deus et Homo ad mentem I. Duns Scoli, Acta Tertii Congressus Scotistici Internationalis 1970 (Studia Scholastico-scotistica 5), Societas Internationalis Scotistica, Romae
1972, pp. 661-671.
48 See Wolter, op.cit.
49 See the discussions in the works mentioned in Note 43 above.
50 For the un~vocity of being in Ockham, see M. Matthew, The Concept of Univocily
Regarding the Predication of God and Creature According to William Ockham, The Franciscan Institute, St. Bonaventure, 1952; P. Boehner, 'Scotus' Teachings According to
Ockham I. On the Univocity of Being', Franciscan Studies VI (1946), 100 - 107; D. C.
Langston, 'Scotus and Ockham on the Univocal Concept of Being', Franciscan Studies 39
(1979), 105 -129. The influence of Scotus' doctrine of being on Suarez is discussed in W.
Hoeres, 'Francis Suarez and the Teaching of John Duns Scotus on Univocatio entis', John
Duns Scotus 1265 -1965 (see Note 43 above), pp. 263 - 290. Short remarks on later influence are to be found in L. Honnefelder, 'Duns Scotus/Scotismus II', The%gische
Rea/enzyk/opiidie 9, De Gruyter, Berlin, New York 1982, pp. 232 - 240.
51 Knuuttila, op.cit., pp. 217 - 234.
52 Basic texts are Lect. I, d. 39, q. 1 - 5 and Ord. I, d. 43, q. un. For a discussion of these
and some other texts, see Knuuttila, op.cit., pp. 217 - 234, 'Modal Logic', The Cambridge
History of Later Medieva/ Philosophy (see Note 15 above), pp. 353 - 355; 'Duns Scotus'
Criticism of the Statistical Interpretation of Modality', in Sprache und Erkenntnis im MiIte/alter (Miscellanea Mediaevalia 13/1), De Gruyter, Berlin - New York, pp. 441 - 450.
53 In addition to references mentioned in Note 47 above, see Ord. I, d. 35, q. un., n. 61;
d. 38, q. un., n. 10 (ed. Vat. VI).
54 For the notion of real concept in Scotus, see Wolter, op.cit. pp. 15 - 17, 65 - 66.
55 After the passage quoted in Note 45 above Duns Scotus states: " ... quia homini non
repugnat (sc. to be something), ideo est possibile potentia logica, ... et illam possibilitatem
consequitur possibilitas obiectiva, et hoc supposita omnipotentia Dei quae respicit omne
possibile (dummodo iIIud sit aliud a se), tamen ilia possibilitas logica, absolute - ratione
sui - posset stare, licet per impossibile nulla omnipotentia earn respiceret." Ord. 1, d. 36,
q. un., n. 61.
56 In Scotus's metaphysics actuality and potentiality (actus ct potentia) belong to the disjunctive transcendentals which are in disjunction proper to quidditative being. Actuality
as a transcendental attribute means existence. See Wolter, op.cit., pp. 145 -148. Wolter
quotes Ord. I. d. 7, q. 1, n. 72 where it is stated: " ... ens in communi non tantum dividitur
per actum et potentiam, sed etiam quodcumque genus entis, et quaecumque species et in46
47
218
SIMO KNUUlTILA
dividuum, quia sic albedo eadem primo est in potentia et postea in actu". Cf. Ord. II, d.
16, q. un., n. 5: "Illud enim individuum, quod nunc est in actu, iIlud idem fuit in potentia".
When an individual has become actual, that same individual was earlier a (merely) potential
individual, i.e. it was a not yet actualized member of that possible world which is the actual
one. It is a member of alternative possible worlds, too. See, for example, Ord. I, d. 41,
q. un., n. 7, d. 44, q. un., n. II (ed. Vat. VI).
S7 See, for example, Ord. II, d. 3, pars I, q. 5-6, n. 191.
S8 Cf. A. Wolter, 'The Formal Distinction', in John Duns Seotus 1265 -1965 (Note 43
above), pp. 54 - 60; 'Is Existence for Scotus a Perfection Predicate, or What?", De doetrina loannis Duns Seoti II (Note 40 above), pp. 175 -182.
S9 In S.th. I, q. 3, a. 5c Aquinas states: "omnia quae sunt in genere uno, communicant
in quidditative vel essentia generis, quod praedicatur de eis in eo quod quid est. Differunt
aut em secundum esse: non enim idem est esse hominis et esse equi, nec huius hominis et
iIlius hominis." For temporal necessity in Aquinas, see, e.g., In I Periherm., Ie. 15, n. 2,
Sent. I, d. 38, q. I, a. 5, ad 3, De ver. q. 2, a. 12, ad 4.
60 John Buridan, Traetatus de eonsequentiis, ed. by H. Hubien (Philosophes medievaux
16), Publications Universitaires, Louvain, 1976, pp. 27, 31 - 40; 58,4 - 60,56; 75, 196 -76,
204; William Ockham, Summa /ogieae, ed. by P. Boehner, G. Gal, and S. Brown
(Guillelmi de Ockham, Opera philosophica et theologica: Opera philosophica I), The Franciscan Institute, St. Bonaventure, N.Y. 1974, pars I, c. 72, pp. 215, 37-218,112. See also
S. Knuuttila, 'Modal Logic' (Note 52 above), pp. 355 - 357 and 'Topics in Late Medieval
Intensional Logic', in I. Niiniluoto and E. Saarinen (eds.), Intensional Logic: Theory and
Applications (Acta Philosophica Fennica 35), Societas Philosophica Fennica, Helsinki,
1982, pp. 32 - 38; Ockham's Theory of Terms. Part 1 oj the Summa logieae, transl. and
introd. by M. J. Loux, University of Notre Dame Press, Notre Dame, London, 1974, pp.
40-44.
61 W. Hoeres, 'Francis Suarez and the Teaching of John Duns Scotus on univocatio entis',
John Duns Seotus 1265 -1965 (Note 43 above), pp. 263 - 290.
62 Disputationes metaphysicae (reprinted from Opera omnia, Vives 1856 - 78), Olms,
Hildesheim, 1965, Vo!. II, pp. 176 - 177, 190- 203,207 - 223. There is a short discussion
on Suarez's views on possibility and reality in J. A. Trentman, 'Scholasticism in the Seventeenth Century', The Cambridge History of Later Medieval Philosophy (Note 15 above),
pp.826-827.
63 For Leibniz's modal theory, see H. Poser, Zur Theorie der ModalbegrifJe bei G. W.
Leibniz (Studia Leibnitiana Supplementa VI), Franz Steiner Verlag, Wiesbaden, 1969; H.
Schepers, 'Zum Problem der Kontingenz bei Leibniz. Die beste der moglichen Welten', in
Collegium Philosophieum. Studien Joachim Ritter zum 60. Geburtstag, Basel, Stuttgart,
1965, pp. 326-350; J. Hintikka, 'Leibniz on Plenitude, Relations, and the "Reign of
Law''', in S. Knuuttila (ed.) (Note 21 above), pp. 259- 286; B. Mates, 'Leibniz on Possible
Worlds' , in Logic, Methodology, and Philosophy of Science lIl, Proceedings of the Third
International Congress, ed. by B. van Rootselaar and J. F. Staal (Studies in Logic and the
Foundations of Mathematics), North-Holland Pub!. Co., Amsterdam, 1968, pp.
507-529.
64 See, for example, Discours de Metaphysique, Sections 9 - II, transl. in Gottfried
Wilhelm Leibniz, Philosophical Papers and Letters. A selection translated and edited with
an introduction by L. E. Loemker (Synthese Historical Library 2), 2 ed., D. Reidel Pub!.
Co., 1976, pp. 308 - 309.
65 Diseoursde Metaphysique, Sections 8 - 9, 13; the letter to Arnauld, 14 July 1686, transl.
in Loemker, op.cil., pp. 331 - 338.
219
BIBLIOGRAPHY
Aristotle, De anima, ed. with introd. and comm. by W. D. Ross, Oxford University Press,
Oxford, 1961.
Aristotle, Ethica Nicomachea, ed. by I. Bywater (Scriptorum Classicorum Bibliotheca Oxoniensis), Oxford University Press, Oxford, 1970 (first published 1894).
Aristotle, Generation of Animals, with an English Translation by A. L. Peck (Loeb
Classical Library), Heinemann & Harvard University Press, Cambridge, Mass. and
London, 1953.
Aristotle, Metaphysics. A Revised Text with Introduction and Commentary by W. D.
Ross, 2 Vols., Oxford University Press, Oxford, 1924.
Aristotle, Parts of Animals, with an English Translation by A. L. Peck (Loeb Classical
Library), Heinemann & Harvard University Press, Cambridge, Mass. and London,
1945.
Aristotle, Prior and Posterior Analytics. A Revised Text with Introduction and Commentary by W. D. Ross, Oxford University Press, Oxford, 1949.
Barth, T. A., 'Being, Univocity, and Analogy According to Duns Scotus', in J. K. Ryan
and B. M. Bonansea (eds.), 1965, pp. 210-262.
Boehner, Ph., 'Scotus Teachings According to Ockham I. On the Univocity of Being',
Franciscan Studies VI (1946), 100-107.
Brown, St. F., 'Avicenna and the Unity of the Concept of Being', Franciscan Studies 2S
(1965), 117 - 150.
Brown, St. F., 'Richard of Conington and the Analogy of the Concept of Being', Franziskanische Studien 48 (1966), 297 - 307.
Brown, St. F., 'Scotus's Univocity in the Early Fourteenth Century', in De doctrina loannis Duns Scoti. Acta Congressus Scotistici lnternationalis 1966, Vol. IV (Studia
Scholastico-Scotistica 4), Rome, 1968, pp. 35 - 41.
Dancy, R., 'On Some of Aristotle's Second Thoughts About Substances: Matter', The
Philosophical Review 87 (1978),372-413.
Doig, J. C., A quinas on Metaphysics. A Historico-doctrinal Study of the Commentary on
the Metaphysics, Nijhoff, The Hague, 1972.
Fabro, c., La nozione metafisica di partecipazione secondo S. Tommaso d'Aquino, 3 ed.,
Societa Editrice Internazionale, Turin, 1963.
Finance, J. de, Etre et Agir dans 10 Philosophie de Saint Thomas, 2 ed., Librairie Editrice
de l'Universite Gregorienne, Rome, 1960.
Geiger, L.-B., La Participation dans 10 philosophie de S. Thomas d'Aquin, 2 ed., Vrin,
Paris, 1953.
Gilson, E., Being and Some Philosophers, 2 ed., Pontifical Institute of Mediaeval Studies,
Toronto, 1952.
Happ, H., Hyle. Studien zum arislOtelischen Materie-Begriff, De Gruyter, Berlin, New
York, 1971.
Hintikka, J., 'Leibniz on Plentitude, Relations, and the "Reign of Law" , in S. Knuuttila
(ed.), 1981. pp. 259- 286.
Hintikka. J .. Time and Necessity. Studies in Aristotle's Theory of Modality, Oxford
University Press, Oxford. 1973.
Hoeres. W., 'Francis Suarez and the Teaching of John Duns Scotus on univocatio entis',
in J. K. Ryan and B. M. Bonansea (eds.). 1965, pp. 263 - 290.
Honnefelder, L.. Ens inquantum ens. Der Begriff des Seienden als solchen als Gegenstand
220
SIMO KNUUTTILA
der Metaphysik nach der Lehre des Johannes Duns Scotus (Beitrage zur Geschichtt! der
Philosophie und Theologie des Mittelalters, NF 16), Aschendorff, Miinster, 1979.
Honnefelder, L., 'Die Lehre von der doppelten Ratitudo entis und ihre Bedeutung fUr die
Metaphysik des Johannes Duns Scotus', in Deus et Homo ad mentem I. Duns Scoti. Acta Tertii Congressus Scotistici Internationalis 1970 (Studia Scholastico-scotistica 5),
Rome, 1972, pp. 661 - 671.
Honnefelder, L., 'Duns Scotus/Scotismus II', Theologische Realen~klopiidie 9, De
Gruyter, Berlin, New York, 1982, pp. 232-240.
John Buridan, Tractatus de consequentiis, ed. by H. Hubien (Philosophes medievaux 16),
Publications Universitaires, Louvain, 1976.
John Duns Scotus, Opera omnia studio et cura Commissionis Scotisticae, Civitas Vaticana,
1950 - .
John Duns Scotus, Opera Omnia, Vives, Paris, 1891 - 5.
Kane, W. H., 'The Extent of Natural Philosophy', New Scholasticism 31 (1957), 85 - 97.
Kenny, A., The Five Ways. St. Thomas Aquinas' ProoJs oj God's Existence (Studies in
. Ethics and the Philosophy of Religion), Schocken Books, New York, 1969.
Knuuttila, S., 'Duns Scotus's Criticism of the "Statistical" Interpretation of Modality',
in Sprache und Erkenntnis im Mittelalter (Miscellanea Mediaevalia 13/1), De Gruyter,
Berlin, New York, 1981, pp. 441-450.
Knuuttila, S., 'Modal Logic', in N. Kretzmann et al. (eds.), 1982, pp. 342 - 357.
Knuuttila, S., 'Time and Modality in Scholasticism', in S. Knuuttila (ed.), 1981, pp.
163 - 257.
Knuuttila, S., 'Topics in Late Medieval Intensional Logic', in I. Niiniluoto and E. Saarinen
(eds.), Intensional Logic: Theory and Applications (Acta Philosophica Fennica 35),
Societas Philosophica Feunica, Helsinki, 1982, pp. 26-41.,
Knuuttila, S. (ed.), ReJorging the Great Chain oj Being. Studies oj the History oj Modal
Theories (Synthese Historical Library 20), D. Reidel, Dordrecht, 1981.
Krllpiec, M. A., 'The Theory of Analogy of Being', in S. Kaminski, M. Kurdzialec, and
Z. J. Zdybicka (eds.), Theory oj Being, Towarzystwo Naukowe Katolickiego Uniwersytetu Lubelskiego, Lublin, 1980, pp. 31 - 106.
Kretzmann, N., A. Kenny, and J. Pinborg (eds.), The Cambridge History oj Later
Medieval Philosophy, Cambridge University Press, Cambridge, 1982.
Langston, D. c., 'Scotus and Ockham on the Univocal Concept of Being', Franciscan
Studies 39 (1979), 105 -129.
Leibniz, G. W., Philosophical Papers and Letters, trans!. and ed. with introduction and
notes by L. E. Loemker, 2 ed. (Synthese HistoricalLibrary 2), D. Reidel, Dordrecht,
1976.
Loux, M. J. (trans!. & introd.), Ockham's Theory oJ Terms. Part I oJthe Summa Logicae,
University of Notre Dame Press, Notre Dame, London, 1974.
Loux, M. J., Substance and Attribute. A Study in Ontology, D. Reidel Pub!. Co., Dordrecht, 1977.
Lyttkens, H., The Analogy between God and the World. An Investigation oj Its
Background and Interpretation oj Its Use by Thomas oj Aquino, Almqvist & Wiksell,
Uppsala, 1952.
Mahoney, E. P., 'Sense, Intellect and Imagination in Albert, Thomas, and Siger', in N.
Kretzmann et al. (eds.), 1982, pp. 602 - 622.
Maritain, J., Existence and the Existent (trans!. by L. Galantiere & G. Phelan), Pantheon,
New York, 1948.
Mates, B., 'Leibniz on Possible Worlds', in Logic, Methodology, and Philosophy oj
221
Science Ill; Proceedings oj the Third International Congress, ed. by B. van Rootselaar
and J. F. Staal (Studies in Logic and the Foundations of Mathematics), North-Holland
Publ. Co., Amsterdam, 1968, pp. 507 - 529.
Matthew, M., The Concept oj Univocity Regarding the Predication oj God and Creature
According to William Ockham, The Franciscan Institute, St. Bonaventure. 1952.
Mcinerny, R. M., The Logic oj Analogy. An Interpretation oj St. Thomas Aquinas, M.
Nijhoff. The Hague, 1961.
Mcinerny, R. M., 'Metaphor and Analogy'. Sciences Ecctesiastiques 16 (1964),273 - 289;
reprinted in J. F. Ross (cd.), 1971, pp. 75 - 96.
Mcinerny, R. M., The ratio communis of the Analogous Name', Laval Theologique et
Philosophique 19 (1962), 9 - 34.
Owen, G. E. L., 'Logic and Metaphysics in Some Earlier Works of Aristotle', in I. During
and G. E. L. Owen (eds.), Aristotle and Plato iti the Mid-Fourth Century (Studia
Graeca et Latina Gothoburgensia XI), Almqvist & Wiksell, Gothenburg, 1960, pp.
163 - 190.
Owens, J., 'The Accidental and Essential Character of Being in the Doctrine of St. Thomas
Aquinas', Mediaeval Studies 20 (1958), I - 40; reprinted in J. Owens, St. Thomas
Aquinas on the Existence oj God, pp. 52 - 96.
Owens, J., 'Common Nature: A Point of Comparison between Thomistic and Scotistic
Metaphysics', in J. F. Ross (ed.), 1971, pp. 185 -209.
Owens, J., St. Thomas Aquinas on the Existence oj God. Collected Papers oj Joseph
Owens, C. Ss. R. (ed. by J. Catan), State University of New York Press, Albany, 1980.
Pinborg, J., 'Die Logik der Modistae', Studia Mediewistyczne 16 (1975),39-97.
Pinborg, J., 'Speculative Grammar', in N. Kretzmann et al. (eds.), 1982, pp. 254-269.
Plato, The Republic with an English translation by P. Shorey, 2 vols. (Loeb Classical
Library), Heinemann & Harvard University Press, Cambridge, Mass., and London,
1930 - 35.
Poser, H., Zur Theorie der Modalbegrifje bei G. W. Leibniz (Studia Leibnitiana Supplementa VI), Franz Steiner Verlag, Wiesbaden, 1969.
Raeymaeker, L. de, The Philosophy oj Being (transl. by E. Ziegelmeyer), Herder, St.
Louis, 1957.
Rijk, L. M. de, Logica Modernorum II, I. The Origin and Early Development oj the
Theory oj Supposition (Wijsgerige teksten en studies 16), Van Gorcum, Assen, 1967.
Ross, J. F., 'Analogy as a Rule of Meaning for Religious Language', International
Philosophical Quarterly 1 (1961), 468 - 502; reprinted in J. F. Ross (ed.), 1971, pp.
35 - 74.
Ross, J. F. (ed.), Inquiries into Medieval Philosophy. A Collection in Honor oj Francis
P. Clarke (Contributions in Philosophy 4), Greenwood Publ. Co., Westport, 1971.
Ryan, J. K. and B. M. Bonansea (eds.), John Duns Scotus 1265 - 1965 (Studies in
Philosophy and the History of Philosophy 3), The Catholic University of America
Press, Washington, D.C., 1965.
Schepers, H., 'Zum Problem der Kontingenz bei Leibniz. Die beste der moglichen Welten',
in Collegium Philosophicum. Studien Joachim Riller zum 60. Gebur/stag, Schwabe &
Co., Basel, Stuttgart, 1965, pp. 326- 350.
Schmaus, M .. Zur Diskussion iiber das Problem der Univozitiit im Umkreis des Johannes
Duns Scotus (Sitzungsberichte der Bayerischen Akademie der Wissenschaften, Philos.hist. KI., 1957,4), Munchen, 1957.
Schmidt, R. W., The Domain oj Logic According to Saint Thomas Aquinas, M. Nijhoff,
The Hague. 1966.
222
SIMO KNUUTIILA
Shircel, C. L., The Univocity of the Concept of Being in the Philosophy of John Duns
Scotus, The Catholic University of America Press, Washington, D.C., 1942.
Suarez, Francisco, Disputationes metaphysicae (reprinted from Opera omnia, Vives, Paris
1856-78), G. Olms, Hildesheim, 1965.
Thomas Aquinas, In Aristotelis libros Peri Hermeneias et Posteriorum. Analyticorum expositio, ed. by R. Spiazzi, Marietti, Turin, 1964.
Thomas Aquinas, In duodecim libros Metaphysicorum Aristotelis expositio, ed. by R.
Spiazzi, Marietti, Turin, 1950.
Thomas Aquinas, Expositio super librum Boethii De Trinitate, ed. by B. Decker (Studien
und Texte zur Geistesgeschichte des Mittelalters 4), Brill, Leiden, 1955.
Thomas Aquinas, In Aristotelis libros Peri Hermeneias et Posteriorum Analyticorum expositio, ed. by R. Spiazzi, Marietti, Turin, 1964.
Thomas Aquinas, In Aristotelis librum De anima commentarium, ed. by A. Pirotta,
Marietti, Turin, 1959.
Thomas Aquinas, Opuscula philosophica, ed. by R. M. Spiazzi, Marietti, Turin, 1954.
Thomas Aquinas, Opuscula theologica, ed. by R. A. Verardo et al., Marietti, Turin, 1954.
Thomas Aquinas, Quaestiones disputatae I - II, ed. by R. M. Spiazziet al., Marietti,
Turin, 1953.
Thomas Al1uinas, Quaestiones quodlibetales, ed. by R. M. Spiazzi, Marietti, Turin, 1956.
Thomas Aquinas, Scriptum super libros sententiarum, Vols. I - II, ed. by P. Mandonnet,
Lethielleux, Paris, 1929.
Thomas Aquinas, Summa theologiae, in Opera omnia iussu Leonis XIII, P. M., edita,
Vols. 4-12, Vatican Polyglot Press, Rome, 1896-1906.
Trentman, 1. A., 'Scholasticism in the Seventeenth Century', in N. Kretzmann et al. (eds.),
1982, pp. 818-837.
Weidemann, H., 'The Logic of Being in Thomas Aquinas', this volume, pp. 181-200.
William Ockham, Summa logicqe, ed. by Ph. Boehner, G. Gal, and St. Brown (Guillelmi
de Ockham, Opera philosophica et theologica: Opera philosophica I), The Franciscan
Institue, St. Bonaventure, 1974.
Wolter, A. B., 'The Formal Distinction', in 1. K. Ryan and B. M. Bonansea (eds.), 1965,
pp. 45 -60.
Wolter, A. B., 'Is Existence for Scotus a Perfection, Predicate, or What?' in De doctrina
loannis Duns Scoti. Acta Congressus Scotistici lnternationalis 1966, Vol. II (Studia
Scholastico-Scotistica 2), Rome, 1968, pp. 175 - 182.
Wolter, A. B., The Transcendentals and Their Function in the Metaphysics of Duns
Scotus, The Franciscan Institute, St. Bonaventure, 1947.
LILLI ALANEN
1. INTRODUCTORY REMARKS
It is a well-documented and widely-known fact that Descartes, in defending his new philosophy, relied heavily on Scholastic theories and notions, and that he also adapted them skillfully for his own purposes. l For
instance, in arguing that the mind and the body are radically different
and mutually independent substances, Descartes used a theory of distinctions which he seems largely to have taken over from Suarez. 2 However,
although Descartes's argument for the mind-body distinction has often
been discussed and analyzed in recent literature, 3 the theory of distinctions on which it is based and to which Descartes also refers in developing
his argument has not been extensively studied. This is somewhat
surprising. 4 For the interpretation of Descartes's dualism, as well as the
assessment of the argument supporting it is, it seems to me, dependent
on the sense given to the notion of a real distinction as used by Descartes
and his predecessors. It is my contention that Descartes's argument,
when interpreted in the light of traditional uses of this notion, is both
more cogent and less unproblematic than is usually thought. As I will try
to show, Descartes can at least be said to have answered, in a satisfactory
manner, the criticism of Arnauld, who was the most perspicacious and
serious of Descartes's contemporary objectors. 5 I do not want to claim
that Descartes's dualism is an unproblematic doctrine. My aim, rather,
is to show that many of the difficulties of this doctrine are due to ambiguities in the terminology which Descartes took over from
Scholasticism, as well as to the application he made of them. 6
Descartes's own theory of distinction which is exposed In the Principles, seems to be relatively simple and clear. (Cf. below, Section 5.) It
is important, however, to bear in mind that Descartes's use of traditional
Scholastic notions and concepts is not always very consistent. For
although Descartes developed and exposed his views and arguments in
terms of common, traditional concepts used by the School, he was an innovator in many respects. Whenever he thought it arpropriate, he also
liked to stress the originality of his views. The problem is that it is not
always very clear when Descartes, in using the terminology of the
223
S. Knuuttila and J. Hinlikka (eds.), The Logic oj Being, 223 - 248.
1986 by D. Reidel Publishing Company.
224
LILLI ALANEN
Scholastics, aligns himself with the views that inform this terminology,
and when he is on the contrary innovating, giving new senses to traditional terms. 7 The interpretation of the key-concepts of Descartes's
theory of distinctions, such as substance, essence, attribute, and mode,
for instance, is therefore delicate and problematical. This also contributes to the difficulty of evaluating the originality of Descartes's
arguments and views. Descartes claimed to be the first to assert that the
mind consists in "thought alone" or the faculty of thinking, and he
seems also to have considered his proof of a real distinction between the
mind and the body as unprecedented in force and cogency. 8 However,
the mere assertion of a real distinction between the mind and the body,
or between form and matter, seems to have been something of a commonplace among Scholastic philosophers and could therefore hardly, in
itself, justify any claims to originality. (Cf. below, Section 4.)
In order to compare Descartes's theory of distinctions with different
accounts of distinctions given by some of his predecessors, I will begin
by outlining briefly the general problem-context in which the medieval
theories of distinctions were developed and discussed. I will also present,
in a general manner, the main distinctions and concepts discussed by the
Scholastics which are relevant to the understanding of Descartes's theory
and his application of it in the proof for mind-body dualism.
2. SUBSTANCES, REAL BEINGS AND CONCEPTUAL BEINGS
225
226
LILLI ALANEN
227
real distinction, i.e., of a distinction between thing and thing. This means
that although these entities are actually united (e.g., matter and form
which are always found together and the composition of which was supposed to constitute the individual substances or substances in Aristotle's
primary sense of the word) they are separable in the sense that at least
one of them can be conceived as existing independently of the other:
more precisely, their existence is not necessarily bound to that combination in which they are actually found. Thus, most of the Scholastics seem
to agree on the possibility of conceiving, for instance, this form without
this specificity or this particular matter, or the body without the soul, or
the soul without these particular thoughts, ot, more generally, a
substance without its accidental qualities. This may be contrasted with
the distinction between merely conceptual or mental entities (ens rationis) , for instance, between 'man' and 'rational animal', or with the
more controversial distinctions defined as intermediates between the real
and the mental distinction, e.g., the distinction between rationality and
animality in man. Another instance of this intermediate or "formal"
distinction is that between the Persons of the Trinity, or God's attributes,
as his goodness, mercy and justice. The entities belonging to this intermediate class were often defined as conceptual (or formal) but they
were not regarded as created by the intellect. Although they were supposed, contrary to the purely mental entities, to have some kind of basis in
the nature of (extra-mental) thin~s, they were regarded as inseparable
from these things. Not being capable of separate existence these entities
were often characterized as different aspects of real things, existing in
these things before and independently of the operation of the intellect. 16
A case discussed in this connection, besides those mentioned above, was
that of the soul and its faculties: the faculties of the soul, according to
many authors, can be conceived as distinct entities in spite of the fact that
the soul is one and undivided and its faculties, consequently, are inseparable from the soul and from each other.
As noted before, there was, however, much controversy on how these
various distinctions should be classified. The Thomists, for instance,
used the term real distinction in a somewhat different sense and extended
it to items which, according to other authors, are separable only in
thought, as the soul and its faculties, essence and existence. But later
Scholastics, such as Duns Scot us, Ockham and their followers, seem to
have restricted the real distinction to things considered as separable in the
extra-mental reality. Hence, as opposed to the other distinctions assumed by these authors, a real distinction, for the later Scholastics,
228
LILLI ALANEN
229
230
LILLI ALANEN
Before discussing this thesis and Descartes's argument for it, I will present, briefly, the theory of distinctions upon which Descartes's reasoning
is implicitly based, as it is developed in the Principles, I, 60 - 62. As
noted above it corresponds largely to the theory of Suarez. 2t Like
Suarez, Descartes admits three different kinds of distinctions, but he interprets them, characteristically, in his own way. They are: the real
distinction, the modal distinction and the distinction of reason.
The real distinction according to Descartes "is properly speaking
found between two or more substances" (AT VIII, 28; HR II, 243). By
a substance, according to a definition given earlier in the same text,
Descartes understands "nothing else than a thing which so exists that it
needs no other thing in order to exist". Absolutely speaking, there is only
one substance which can be said to fulfill this requirement, namely God:
other things can exist only by the help or concourse of God. Therefore,
Descartes adds, the Scholastics are right in saying that "the word
substance does not pertain univoce to God and other things" because
"no common signification for this appellation which will apply equally
to God and to them can be distinctly understood" (AT VIII, 24; HR
I, 239 - 249). However, the concept can be attributed to created
substances, in so far as these' 'need only the concurrence of God in order
to exist". And in this sense it can be attributed univocally, in Descartes's
view, to the soul and the body (AT VIII, 24 - 25; HR I, 240). As regards
the criteria or sign of a real distinction Descartes says that
we can conch;de that two substances are really distinct one from the other from the sole
fact that we can conceive the one ciearly and distinctly without the other. For in accordance
with the knowledge we have of God. we are certain that he can carry into effect all that
of which we have a distinct idea. (AT VIIl. 28; HR 1.243. Cf. also AT VII. 162 and
170-171; HR II. 53 and 59.)22
231
232
LILLI ALANEN
cidental or permanent properties. And although such entities can be conceived without the substance in which they inhere by way of abstraction,
they can never be conceived or known clearly and distinctly in this way.
In so far as they are conceived by abstraction they are not conceived or
known as "complete" things. What Descartes has to prove, therefore,
in order to show that the mind is really distinct from the body, is that
the knowledge of the mind is not acquired by way of abstraction and that
the mind is consequently neither an accidental nor a permanent property
of the body to which it is united, but that it can on the contrary be known
as a complete or self-subsisting thing in itself.
(iii)
(iv)
(v)
233
The structure of the argument, I take it, is roughly the following: (ii) can
be said to follow from (i), which, as we have seen, was a commonly accepted principle. That (i) is essential to Descartes's proof is easy to
understand against the background of the considerations above (cf. Section 4). This explains why Descartes insists that the distinctness of mind
and body cannot be fully demonstrated until the existence of a veracious
and omnipotent deity has been proved, although the argument for the
other relevant premises iii) and (iv seems to be given already in the Second Meditation (cf. AT VII, 219; HR II, 96). On the basis of the reasoning there presented, (iii) at least can be accepted as - relatively - unproblematic. But (iv) is more controversial. It is, presumably, supposed
to follow from (iii), since there seems to be no other place in the Meditations where the question of what really belongs to his essence (i.e., the
essence of his mind) is explicitly discussed besides the passages subsequent to the Cogito reasoning in the Second Meditation. 24 However, the
argument (the Cog ito reasoning) for (iii) can hardly be considered as sufficient for establishing (iv) as it is here formulated. And it is certainly not
sufficient for inferring (v). The problem, then, which has often been
raised, is to understand what further argument Descartes can give to
justify this claim that the essence of mind consists in thought alone, in
a way, as Descartes also seems to assume, which excludes the body and
other things from its essence. The proof of the mind-body distinction
presupposes, as Descartes himself recognizes, a move from the order of
clear and distinct perception to the order of things, or, in other words,
from concepts to reality. But how is it possible to infer, as he seems to
do, from what merely looks like a subjective state of certainty (expressed
in (iii, to the knowledge of the essential nature of the self or the mind
((iv) and (v? For, as I have argued elsewhere, the clarity and distinctness
of the knowledge of the self or the mind and its nature acquired in the
Second Meditation, seem to consist of nothing more than the certainty
of the facts expressed in the propositions" I think" and" I exist" . 25 And
how can this knowledge justify any further conclusion about the objective nature of self or the mind? (Cf. the objection of Arnauld in AT VII,
203; HR II, 84.)
However, instead of trying to look for a justification of these controversial premises, one could raise the question to what extent Descartes
really needs such strong claims for the conclusion he wants. to draw. (iv)
and (v), it seems to me, could be reformulated in a way that renders them
more plausible without altering the force of the proof:
234
(iv ')
(v ')
LILLI ALANEN
Now this, I want to argue, is all Descartes actually needs, besides the
assumption of an omnipotent Deity, in order to conclude that the mind
is really distinct from the body and can exist without it (vi). In this conclusion the same point seems to be repeated twice, since to say that one
thing is really distinct from another, is, as we have seen, merely to say
that it can be posited, i.e., exist, by the power of God, without the other.
But by the same means Descartes can also be said to have proved (iv) and
(v). Because, as the term is used by Descartes, essence is the attribute
through which a real thing or being, i.e., a substance, is known. 27
Thus, to Arnauld, who wonders where Descartes has shown "how it
follows, from the fact that one is unaware that anything else (except the
fact of being a thinking thing) belongs to one's essence, that nothing else
really belongs to one's essence" (AT VII, 199; HR II, 81), Descartes affirms that this is demonstrated in the place where he proves that God exists, "that God, to wit, who can accomplish whatever I clearly and
distinctly know to be possible" (qui potest omnia quae ego clare et
distincte ut possibilia cognosco, AT VII, 219; HR II, 96). And, he
continues:
For although much exists in me of which I am not yet conscious (for example in that
passage I did, as a fact, assume that I was not yet aware that my mind had the power of
235
moving the body, and that it was substantially united with it), yet since that which I do
perceive (adverto) is sufficient for me to subsist with alone, I am certain that God could
have created me without those other things of which I am unaware, and that these other
things do not pertain to the essence of mind ( .. .Quia tamen id quod adverto. mihi sUfficit
ut cum hoc solo subsistam. certus sum me a Deo potuisse creari absque a/iis quae non
adverto. atque ideo ista alia ad mentis essentia non pertinere. AT VII, 219; HR II, 97)
236
LILLI ALANEN
Descartes's proof of the real distinction between mind and body, as mentioned before, leaves his opponents unsatisfied. Let us now consider
some of the objections raised against it. The first to oppose Descartes is
Caterus who invokes Duns Scotus's formal distinction, understood by
Caterus as a distinction intermediate between a real distinction and a
distinction of reason. On Caterus's reading of Scotus two things which
can be conceived as distinct and separated from each other are not (by
that fact alone) really, but merely formally distinct, as is the case with
the Divine attributes justice and pity. They have, Caterus observes, concepts prior to any operation of the understanding, "yet it does not follow
that, because God's justice can be conceived apart from his pity (mercy),
they can also exist apart" (AT VII, 100; HR II, 8).
Descartes answers by identifying, mistakenly, as he later concedes,
Scotus's formal distinction with his own modal distinction, that "applies
only to incomplete entities, which", he says, "I have accurately demarcated from complete beings" (AT VII, 120; HR II, 22).30 He continues:
Thus, for example, between the motion and the figure of the same body the distinction is
formal, and I can quite well understand (inlelligere) the motion without the figure, and the
figure without the motion, and either when abstracting from the body: but I cannot
however completely understand the movement without the thing in which the motion is,
nor the figure without the thing in which the figure is, nor finally can I feign that the motion
can be in a thing lacking figure, nor the figure in a thing incapable of motion. Nor can I,
similarly, understand justice apart from a just being, or compassion from a compassionate,
nor may I imagine that the same being which is just cannot be compassionate. But yet I
understand, completely, what a body is, in thinking merely that it is extended, figured,
movable, etc., and by denying of it everyt!Jing which belongs to the nature of mind, and
conversely I understand that mind is a compl-:te thing which doubts. understands, wills,
etc., although I deny that there is anything in it of what is contained in the idea of body.
Which would not be possible if there were not a real distinction between mind and body.
(AT VII, 120- 121; HR II, 22-23)
Things which can be conceived apart from each other merely by abstraction of the intellect, are always conceived inadequately. Since they are inadequately conceived, they are not known as complete, self-subsisting
things or beings. As Arnauld understands him, Descartes hereby claims
to have proved not only that the mind can be conceived completely
237
without the body but also that it can be conceived adequately apart from
the body (AT VII, 200 - 201; HR II, 82). But, as we saw before,
Descartes has not proved that the knowledge of his mind is adequate in
the sense assumed by Arnauld: it is not a knowledge embracing all the
properties of the thing known (cf. above, Section 5). Such knowledge,
Descartes stresses, is unattainable for the human mind, which is created
and finite, and it is therefore not required. An adequate knowledge
presupposes that one knows not only all the properties which are adequate for a thing, but also that one knows that God has not given the
thing in question other properties than those of which one has
knowledge. In other words, it is necessary to know that the knowledge
one has is adequate, which would require an infinite capacity of
knowledge (AT VII, 220; HR II, 97). Such knowledge, i.e., a knowledge
which is entirely adequate, is to be distinguished from a knowledge which
"has sufficient adequacy to let us see that we have not rendered it inadequate by an intellectual abstraction" (ibid.). Similarly, in order to
understand that the mind is a complete thing, we need not have an (entirely) adequate knowledge of it, but we must be able to see that the
knowledge we have of it is not rendered inadequate by abstraction, i.e.,
that it is not incomplete.
How then, according to Descartes, are incomplete entities to be
distinguished from complete entities? A complete entity, in Descartes's
view, is recognized by the fact that it can be conceived as existing in itself,
i.e., it must be understood as a real thing or entity in itself, independently of any other entities. By a complete thing, Descartes explains to Arnauld, "I mean merely a substance endowed with those forms or attributes which suffice to let me recognize that it is a substance" (AT VII,
222; HR II, 98). Now certain substances, as Descartes recognizes, are
popularly called "incomplete substances" (e.g., the mind and the body,
or parts of the living body). But, he adds, if they are called incomplete
because they cannot exist of themselves (i.e., without inhering in some
subject); it is contradictory to call them substances. However, substances
can be called incomplete in another sense, namely, when referring to
some other substance together with which they form a single selfsubsisting thing:
Thus. the hand is an incomplete substance, which taken in relation with the body. of which
it is a part; but. regarded alone. it is a complete substance. Quite in the same way mind
and body are incomplete substances viewed in relation to the man who is the unity which
together they form; but. taken alone. they are complete. (AT VII, 222; HR 11.99)
238
LILLI ALANEN
I know that the mind or the thinking thing is a substance because I know
with certainty that I exist (i.e., that my mind exists) by the sole fact that
I think. This is sufficient to refute the contention of the Scholastics that
my mind is related to my body as species to genus. 32 " .it can nowise
be maintained", Descartes emphasizes, "that ... body is related to mind
as genus is to.species; for, although the genus can be apprehended apart
from this or that specific difference, the species can by no means be
thought apart from the genus" (AT VII, 223; HR II, 99). Since the
knowledge that the mind, conceived only as a thinking thing, exists, does
not imply the knowledge that the body exists, the mind cannot be an attribute or property of the body but is a complete thing, i.e., a thing
which, by the power of God, can exist independently of other things.
The same argument is also supposed to prove that the body, the existence of which is not fully established until the Sixth Meditation (and
to which the mind, according to Descartes, is in fact "very closely", intimately or "substantially' united, cf. AT VII, 222 and 228; HR I,
190 and II, 102), is a real, complete thing or substance in the proper sense
of the word. The possibility that the body is merely an attribute or a
species of the mind is hereby excluded. 33
The main problem with this argument, if it is accepted, and Arnauld
seems actually to have accepted it since he never returned to the charge,
is (as Arnauld also points out in his objections) that it proves too much.
It seems therefore to give support to the opinion of Platonists according
to which "nothing corporeal belongs to our essence, so that man is hence
only a soul (animus), while his body is merely the vehicle of the soul"
and according to which man is defined as "a soul that makes use of his
body" (AT VII, 203; HR II, 84). This is, however, a doctrine that
Descartes categorically rejects. He answers, characteristically, that "in
239
order to prove that one thing is really distinct from another, nothing less
can be said, than that the divine power is able to separate the one from
the other" (my emphasis). He also claims that he actually proved the
substantial union between mind and body in the Sixth Meditation while
dealing with the distinction between mind and body, by employing
arguments the efficacy of which he says he cannot remember to have ever
seen surpassed (!). He adds:
Likewise, just as one who said that a man's arm was a substance really distinct from the
rest of his body, would not therefore deny that it belonged to the nature of the complete
man, and as in saying that the arm belongs to the nature of the complete man no suspicion
is raised that it cannot subsist by itself, so I think that I have neither proved too much in
showing that mind can exist apart from body, nor yet too little in saying that it is substantially united to the body, because that substantial union does not prevent the formation of
a clear and distinct concept of the mind alone as a complete thing. (AT VII, 228; HR II,
102 - (03)3.
240
LILLI ALANEN
is "one of those things which we can only make obscure when we try to
explain them in terms of others" (AT V, 222; Philosophical Letters,
235).
The notion of the union of the mind and the body is therefore
characterized as "primitive": it is-given as such and cannot be rendered
more intelligible or clear by means of simpler or more primary notions. 37
Although Descartes accepts the notion of a "real" union of an immaterial form nd a corporeal body in his account of human nature, he
rejects the use of this same notion in the explanation of physical
phenomena. According to Descartes the familiar and daily experience
that we have as sentient and acting conscious subjects of being "closely
united to" and "intermingled" with our body is the only context where
the use of this notion can be considered as legitimate (AT VII, 81; HR
I, 192). For it is the only context where this notion is based on a clear
and immediate experience. Applied to external, physical bodies it is confused and unintelligible: we have no concrete or clear experience of any
immaterial forms attached to or operating in physical bodies, such as the
"real qualities", "substantial forms" or forces (e.g. heaviness),
postulated by the Scholastics. The assumption of such forms in the scientific explanation of nature is not only an illegitimate and "occult"
hypothesis, it is also superfluous in Descartes's view. 38
I will not discuss the difficulties of Descartes's dualistic doctrine here.
As the comparison of Descartes's argument for mind-body dualism and
the theory of distinctions on which it is based with the ontological distinctions discussed by the Scholastics shows, many of these difficulties are
connected to ambiguities inherent in the Scholastic terminology of
distinctions and also to Descartes's application of this terminology in a
conceptual and scientific framework which in many important respects
differs from that in which it was originally developed. Descartes's main
innovation can perhaps be said to consist in the definition be gives of the
traditional notions of the soul and the body in terms of the concepts of
thought and extension. 39 Relying on the view according to which a real
distinction requires a mutual separability of the distinguenda, and accepting, for instance, like Ockham and Suarez did, logical independence and
non-identity as a sufficient sign of mutual separability, Descartes has no
difficulty in proving the separability and hence the real distinction of the
soul and the body in his sense of these terms. At the same time Descartes
reduces all other properties and things, considered as "real" entities or
things by many Scholastics, Ockham and Suarez included, to modes in
Suarez's restricted sense of the term. 40 The only objects of knowledge
241
which according to Descartes can be considered as "real" things or beings in the proper sense of the word, are thinking and extended
substances. All individual as well as generic properties of things, other
than the attributes of thought and extension, are thus reduced to modes
of these two attributes. Far from removing the original ambiguity related
to the various uses and meanings of the term being and related concepts,
Descartes's restructuring of the traditional conceptual framework leaves
us with a host of unsolved and perhaps unanswerable problems. Among
these, besides those encountered above, as the problem of interpreting
the two apparently contradictory notions of a real distinction and a real
union between the mind and the body, is the difficulty of accounting for
individual things or substances in the framework of Descartes's theory
of distinctions. No wonder that Descartes's opponents and followers are
perplexed, or that the doctrine called Cartesian dualism is difficult for
modern readers to understand or interpret in a satisfactory way.
NOTES
I am greatly indebted to Erik Stenius, Georg Henrik von Wright, Simo Knuuttila, Norman Malcolm, and Karl Ginet for valuable remarks and useful criticism at different stages
of my work. I wish especially to thank Erik Stenius for helping me to structure the present
interpretation of Descartes's argument for the mind-body distinction by giving me in a
private communication a very clarifying analysis of the argument.
Following a current usage a double reference will be given to the works of Descartes: one
to the Adam & Tannery edition (1897 -1913), abbreviated AT, volume and page, and one
to the English translation of Haldane and Ross (1911, 1978), abbreviated HR, volume and
page. For Descartes's correspondence references will be given to the English translation by
A. Kenny (1970), cited as Philosophical Letters. I have endeavoured to give literal translations of the Latin text and have therefore, occasionally, departed from the Haldane & Ross
translation. For controversial or unclear passages, quotes are also given in Latin.
I This has been amply shown by the work of scholars like Etienne Gilson, Alexandre
Koyre, J. R. Weinberg and others. See, e.g., Gilson, Index Scolastico-Cartisien, 1912
(1964), and Gilson (1925), and (1930); Koyre (1922); Wei~berg (1977); Beck (1965), and
Wells (1965).
2 Cf. Gilson 1912 (1964), p. 87; Weinberg (1977), pp. 75 -77, and below Notes 16 and 22.
J See, e.g., Wilson (1978); Weinberg, J. R. 'Descartes on the Distinction of Mind and
Body', in Weinberg (1977), pp. 71 - 81; Williams (1978), pp.102 -129; and the articles by
M. Hooker, A. Donagan, F. Sommers in Hooker (ed.), Descartes, Critical and Interpretive
Essays (1978). As appears from these studies there is no general agreement on the premises
of Descartes' argument. It is not very clear either how the conclusion of the argument
242
LILLI ALANEN
should be interpreted or understood. Cf. Wilson (1978), pp. 180ff; Wilson 'Descartes: The
Epistemological Argument for Mind-Body Distinctness', in Noils (1976), pp . 3 - 15, and
Malcolm (1971), pp. 5ff. For other recent studies of Descartes' argument for dualism see
the references given in Alanen (1982), p. 92, Note 16.
4 Cf. Wilson (1978), p. 243, Note 14. It is true that Descartes does not give any detailed
exposition of his theory of distinctions before the Principles of Philosophy (AT VIII,
28 - 30; HR II, 244 - 245), written some years later than the Meditations on First
Philosophy where the proof of the mind-body distinction is found (AT VII, 78; HR I, 190).
It is, however, clear from the Objections and Replies, published with the Meditations (.e.g,
AT VII , 120, 169-170, and 220ff; HR II, 22, 59, and 97ff), that Descartes already relied
upon a definite theory of distinctions in developing his controversial argument, and that
the notion of a real distinction, discussed by the Scholastics, hence plays a central part in
it. The term distinctio realis appears also, e.g., in the title of the Sixth Meditation (AT VII,
71; HR I, 185). I have dealt, briefly, with Descartes's use of these terms in a previous study,
'Descartes on the Essence of His Mind and the Real Distinction Between Mind and Body',
in Acta Philosophica Fennica 33 (1982), 66 - 73.
S Descartes's argument for the mind-body distinction was severely criticized by Descartes's
contemporaries, and notably by Arnauld whose objections Descartes, according to many
later critics, was unable to meet. See Kenny (1967), p. 94, and Wilson (1978), p. 198f.
6 Thus, if Descartes's proof of a real distinction between mind and body is less problematic
than is usually thought, when it is considered in the light of the traditional uses of the notion of a real distinction, the interpretation of this notion remains an open question (cf.
below Note 29). As to the Scholastic theory of a substantial union between the mind and
the body that Descartes defends as the only correct view of the human nature it is neither
clear nor distinct in the framework of Cartesian dualism, as Descartes himself was forced
to admit. Also, in trying to explain the union of mind and body to Princess Elizabeth whose
question on this matter, as Descartes recognizes, is "the one which may most properly be
put to me in view of my published writings" (AT III, 663, Philosophical Letters, p. 137),
Descartes invokes a third "primitive" notion, and asks Elizabeth to forget about tne
arguments proving the distinction between mind and body "in order to represent to herself
the notion of the union which everyone has in himself without philosophizing" . AT III,
692ff; Philosophical Letters, 142. Cf. the Principles of Philosophy, I. 48, and Section
7, below, esp. Notes 35 and 37.
1 Cf. Emile Boutroux' remark, often quoted, concerning Descartes' ability of "pouring
new wine into old bottles". See, e.g., Wells (1965), p. 22.
8 See AT VIII 2, 347; HR I, 434; AT VII, 549; HR II, 335 . Cf. also AT VII, 3; HR I,
133-134; AT VII, 13 -14; HR 1,140-141; AT VII, 153-154; HR 11,47.
9 See Edwards (1974), pp. 1-2; Aristotle, Met. 1l7, 1017 6b33 and Topics, I, 7.
10 Whatever is, Aristotle says, is one thing, Met., G. 2 1003 b22. See also, e.g., F. Suarez,
Disputationes Metaphysicae (hereafter quoted as DM) VII, Opera Omnia, Vo!. XXV, p.
250, and Weinberg (1964), p. 245.
II C. Vollert (trans!.), Francis Suarez: On the Various Kinds of Distinctions (1947), Introduction, p. 12.
12 The comparison of Descartes's theory of distinctions to those of his predecessors is
therefore instructive in many ways. The account presented in this paper is preliminary and
tentative: I hope to examine the problems here discussed more thorougl1'ly in a larger study
on the same subject.
il See, e.g., Met. Il 1017blO - 25, and Categ., Ch. 5, 2a1lff. Cf. Lloyd (1968), pp.
114-115.
243
Edwards (1974), p. 10. Cf. also A. Wolters in Ryan and Bonansea (1%5), p. 45.
Cf. Wolter, op.cil., p. 45. See also Note 16 below.
:6 Suarez, for instance, divides the mental distinction into two kinds: (a) a distinction of
reasoning reason (distinctio rationis ratiocinantis) which "arises exclusively from the
reflection and activity of the intellect" and which is in this sense purely mental; and (b) a
distinction of reasoned reason (distinctio rationis ratiocinatae) which is defined as a mental
distinction preexisting in reality, and which, he says, requires "the intellect only to
recognize it, but not to constitute it". (F. Suarez, DM VII, Section I, 4; transl. by Vollert
(1947), p. 18.) Descartes makes a similar distinction but rejects the former kind: he does
not, he writes, admit "any distinction of reason rationis ratiocinantis - that is one which
has no foundation in things - because we cannot have any thought without a foundation ... " (AT IV, 349-350; Philosophical Letters, 188). The formal distinction assumed
by Scotus is reduced by Descartes to a distinction of reason of the latter kind, i.e., to a
distinction of "reasoned reason", the only kind of mental distinction that Descartes admits
and which he characterizes, like Suarez, as having some kind of foundation in reality. (See
the letter referred to above and Descartes's Principles, I, 62, AT VIII, 30; HR I, 245. Cf.
also Alanen (1982), pp. 68 -70 and below, Section 7.) According to Suarez's characterization of the objects of a distinction of "reasoned reason~', they are not to be considered as
merely mental or conceptual entities because they are not produced or created by the mind,
but "real entities, or rather, a single real entity conceived according to various aspects. ...
Hence it is not the objects distinguished but only the distinction itself that results from the
reasoning" (DM VII, Section 1, 6; Vollert (1947), p. 19, my emphasis).
17 See Wolter, op.cit., p. 46. Cf. also Wolter, A., in JP S9 (1%2), p. 726.
18 Cf. below, Section 6 and Note 19. Although Descartes seems to question this principle
in some contexts, and suggests that God's omnipotence cannot be subjected to any such
limitations (cf., e.g., AT IV, lI8; Philosophical Letters, 151, and AT V, 223f;
Philosophical Letters, 236) the consequences of such a radical view, destructive for all
knowledge, are not taken seriously by Descartes. Cf. Koyre, op.cit., pp. 20ff.
19 See, e.g., the Second Replies, Proposition III, Corollary and Demonstration:
" .. .Deus. . . potest efficere id omne quod clare percipimus, prout idipsum percipimus. . . Est autem in nobis idea tantae alicujus potentia, ut ab illo solo, in quo ipsa
est, coelum & terra & c. creata sint, ab eodem fieri possint" (AT VII, 169; HR II, 58- 59,
my emphasis). Note that nothing corresponds in this passage to the words doivent avoir
ete creees in the French translation (AT IX, 131). Haldane and Ross who follow the French
Translation render the latin fieri possint by: ... whatever is apprehended by me as possible
must be created by Him too" (HR II, 59). But this is not the view professed by Descartes.
What the text says is merely that whatever exists has been created by the power of God,
and whatever we conceive as possible can be created by him. Cf. the note of Alquie, F.
(ed.), Oeuvres phllosophiques de Descartes, Vol. II, Note 1 to p. 597. See also the letter
to Regius, June 1642, AT III, 567, and to Mersenne, March 1642, AT III, 544 - 545;
Philosophical Letters, 132.
20 Cf. Suarez's discussion ofthe criteria of a real distinction, DM VII, Section 2, 2-12;
Vollert (1947), pp. 40-49, and Weinberg (1977), pp. 75ff. For Ockham's application of
the principle of God's omnipotence, see e.g., Weinberg (1964), pp. 245ff.
The radical position of Ockham requires a special mention in this context. True to his
view that nothing other than individuals (ir.dividual things) exist in the extra-mental reality
(outside human consciousness) Ockham reduced all distinctions (at least in so far as created
things are concerned) having any basis in reality to real distinctions; and he consequently
held that anything that can be conceived as distinct is a real thing or unity which
14
IS
244
LILLI ALANEN
is really distinct from all other things to which it might be united. Ockham hence allows
for no other kind of real distinctions than numerical distinctions. Therefore, the only kind
of distinctions existing in nature, i.e., having any foundation in extra-mental reality, according to Ockham, are real, numerical distinctions. (Whether or not Ockham accepted any
forlT)al distinction is a subject of controversy which we need not go into here. If he admits
such a distinction he restricts it to supranatural things, e.g., the Trinity.) Whatever can be
conceived as .distinct is thereby also a singular thing, distinct from all other things, by the
principle of God's omnipotence. Any quality which can be distinctly conceived is a singular
thing and hence also really distinct from the substance in which it inheres. See Edwards
(1974), pp. 12-13 and pp. 180ff, and Weinberg (1964), pp. 248-249. See also Weinberg
(1965), p. 49.
It is interesting to note also how far Ockham carries this doctrine, which is based on purely logical arguments. Since matter and form, for instance, are distinctly conceivable parts
of a composite substance, according to the common view, this means, in Ockham's view,
that the particular matter of a given composite must be a singular thing which, logically,
can exist also by itself without any form. Matter, as such, is thus not mere potentiality, as
the Scholastics generally held, but has some actuality, being, nature, or whatever it is
called, in itself. The same holds, according to Ockham, for the substantial forms and the
qualities. See Weinberg (1965), pp. 51 - 52.
21 See Note 2 above and Alanen (1982) pp. 69ff. For Suarez's discussion of the distinctions
to be retained, see DM VII, Section I; Voller! (1947) pp. 16-39.
22 Cf. Suarez' discussion of the signs for discerning various grades of distinction, DM VII,
Section 2, and notably his discussion of the signs of a real distinction, ibid., 9 - 28;
Vollert (1947), pp. 46-6i. Cf. also Weinberg (1977), p. 75 and Alanen (1982), p. 78f.
2J Cf. Suarez, D., DM VII, Section I, 16 - 21, and Section 2, 6ff. - It is interesting
to note how Descartes, while retaining Suarez's criteria for the modal distinction, gives the
term mode (modus) a much wider application than Suarez. Suarez, in discussing different
uses of the term mode, restricts it to a particular aspect of a given attribute (e.g., quality
or quantity), namely the mode of inherence of the attribute in question. Thus, the inherence
of quantity, for instance, is called its mode by Suarez "because it is something affecting
quantity, and, as it were, ultimately determining its state and manner of existing, without
adding to it any new entity, but modifying a preexisting entity" (DM VII, Sc.;tion I, 17;
Vollert (1947), p. 28). The mode of being of the inherence is such "that it cannot exist
unless it is actually joined to the form of which it is the inherence" (ibid., 18; p. 29).
Suarez hence distinguished two aspects in quantity: the first, he says, is called the thing or
being of quantity "comprising whatever pertains to the essence of the individual quantity
as it is found in nature, and remains and is preserved even if quantity is separated from
its subject" (my emphasis); the second, "the inherence of quantity is called its mode
because it is something affecting quantity, and, as it were, ultimately determining its state
or manner of existing, without adding to it a new proper entity ... " (ibid., 17; p. 28).
A mode, hence, in Suarez's restricted sense of the word, is not a thing or entity in itself:
it has no being of its own. "Its imperfection is clearly brought out by the fact that it must
invariably be affixed to something else to which it is per se and directl)! joined without the
medium of another mode, as, for instance, sitting to the sitter, union to the things
united ... " (ibid., 19; p. 31). Having no being or essence of its own the mode "so
necessarily includes conjunction with the thing of which it is a mode that it is unable by
any power whatsoever to exist apart from that thing" (ibid., 20; p. 32, my emphasis).
Descartes, as we saw above, uses the term mode in a much wider and general sense: he
treats all the qualities and accidental attributes or properties of things as modes in Suarez's
245
restricted sense. Unlike Suarez Descartes makes no distinction between an individual quality, say this whiteness, and its actual mode of inherence, the whiteness of this paper. Thus,
what for Suarez holds for the mode of inherence of a particular quality or quantity, holds
according to Descartes for any particular quality or quantity: it has no being of its own and
can hence not be clearly conceived without the essential attribute of the thing, i.e., the
substance, in which it inheres. Since Descartes a<;lmits only two such essential attributes,
namely thought and extension, this means that all other properties of things are reduced
to mere modes of either thinking or extended substances.
24 AT VII, 24-25; HR 1,149-150. Cf. A1anen (1982), pp. 29ff. - In proving that he
exists as a thinking thing, Descartes takes it that he has proved that his mind or soul exists
(cf. ATVI, 32-33; HR 1,101; AT VII, 27; HR 1,152). In discussing his nature or essence,
in the Second Meditation, Descartes makes therefore no distinction between his essence
and the essence of (his) mind. This, however, does not mean that Descartes identifies his
essence (as a human being) with the essence of mind, which, strictly taken is a pure reason
or intellect. See AT VII, 78-79; HR I, 190-191; AT III, 479; Philosophical Letters,
125 - 126; AT III, 371; Philosophical Letters, 102, and below, Note 25 ...,
2S Alanen (1982), pp. 22 - 43, and pp. 63 - 64. Notice that the term to think (cogitare) is,
deliberately, used in a very wide sense by Descartes, and covers all kinds of acts or states
or consciousness, from the acts of the pure intellect to dreams and sensations. See, e.g.,
AT VII, 28; HR I, pp. 152ff; AT VII, 217; HR II, 52. Cf. Alanen (1982), pp. 115-116.
26 Cf. AT VII, 175; HR II, 63, and AT VII , 355; HR II, 209.
27 AT VIII, 24f, 30; HR I, 240, 245. Cf. AT VII, 219; HR II, 99. Descartes adheres to the
view that substances, as such, are not immediately known in themselves. See AT VII, 161;
HR II, 53; AT VIII, 8; HR I, 223; AT VIII, 25; HR I, 240.
28 As to the clear and distinct concept of matter, acquired partly through the analysis of
the wax in the Second Meditation, and partly through the considerations in the Fifth and
Sixth Meditations, it is not really indispensable for this conclusion, that the essence of mind
is thinking. Descartes's definition of matter as mere geometrical extension, and the conception of the body as a piece of a mechanically moved extended substance that this implies
(a conception that is, by the way, difficult to prove), can certainly - if it is granted - be
said to give it additional support. But it is not essential to the argument as it is here
understood. This is the reason for which I have omitted the second part of premise (v)
which states (v)/ I have a clear and distinct idea of the body inasmuch as it is only an extended and unthinking thing (AT VII, 78; HR I, 190). For a discussion of the part played by
this premise and Descartes's concept of body in the argument for a real distinction between
mind and body see, e.g., Williams (1978), pp. 213ff, Gueroult (1953) I, pp. 121ff, and II,
pp. 67ff, and Alanen (1982) pp. 45-65.
29 On this reading Descartes's argument is, it seems to me, both more simple and forceful
than I was earlier inclined to think. This, however, is not to say that the argument as here
interpreted is unproblematic, but I will leave the consideration of the probiems it raises
aside for the moment (cf. Alanen (1982), pp. 84ff). The conclusion of the argument,
notably, is perplexing. For what it states, as I want to stress, is the absence of a logical entailment between the concepts of mind and body, and the existence of the mind in separation from the body is therefore a mere logical possibility. What the implications of this
scholastic notion of a real distinction are on the ontological level is and remains an open
question. - For an interesting critical discussion of the import and consequences of
Descartes' distinction between the concepts of mind and body see Malcolm, 'Descartes'
Proof that He Is Essentially a Non-Material Thing, Thought and Knowledge (1977),
58 - 84.
246
LILLI ALANEN
In the Principles, Descartes corrects himself and assimilates Duns Scotus's formal
distinction with his own distinction of reason (cf. Note 16 above). This mistake or uncertainty concerning the classification of the formal distinction from Descartes' side is not important here. But it is quite interesting. For both the distinction of reason (i.e., reasoned
reason) and the modal distinction, as understood by Descartes, require in fact an abstraction of the mind and are therefore opposed to the real distinction, which, as will be seen,
is restricted by Descartes to entities or things which can be conceived in themselves as complete, i.e., self-subsisting things. Cf. Suarez's characterization of the distinction of reasoned reason: it "does not exist strictly by itself, but only dependently on the mind that conceives things in an imperfect, abstractive, and confused manner, or inadequately". DM
VII, Section I, 8; Vollert (1947), 20-21.
31 Cf. above, Notes 23 and 27.
32 Cf. Arnauld's objection to the effect that Descartes's argument that the body can be
completely understood merely by thinking that it is extended, figured, movable, etc., is of
little value: it does not exclude that the body might be related to the mind as genus is to
species, for, as is commonly agreed, the genus can be conceived without the species (AT
VII, 201; HR II, 82). See also the letter to (Mesland?), 2 May 1644, AT IV, 120;
Philosophical Letters, 152.
33 Note that the fact that he has a body with which he is "very intimately conjoined" is
fully established only in the Sixth Meditation, after the proof of the mind-body distinction
is given. What this proof can therefore be said to show is that if the self or the mind is united
to (or has) a body, then it is really distinct, i.e. can be separated from the body, and conversely. Cf. AT VIII, 28; HR I, 243-244.
34 According to Suarez and, I presume, most of the Scholastics, the soul and the body as
well as form and matter in general, although they are considered as separable by the divine
power and hence as really distinct in the sense given above, are regarded as incomplete and
partial beings in themselves, whether they are united or in a separate state. What Descartes
here says about the mind and the body would according to Suarez apply only to integral
parts, e.g., homogeneous parts of a continuum, which unlike form and matter are not of
themselves "ordained to the composition of another thing", i.e., to be parts of the union
or compound to which they actually belong. Cf. Suarez, DM VII, Section I, 23; Vollert
(1947), pp. 33 - 34.
35 See, e.g., the letter to Arnauld, 29 July 1648, AT V, 222f; Philosophical Letters, pp.
235 - 236. Cf. also the letter to Elizabeth, 21 May 1643, AT III, 667; Philosophical Letters,
p. 139, and references given in Note 6 above.
36 Cf. Spinoza, Ethica III, Proposition II, scholium, Vloten et Land (ed.), (1914), p. 124.
37 See, e.g., AT III, 665; Philosophical Letters, 138f. Cf. Alanen (1982), p. 87f.
38 See, e.g., Descartes's letter to Mersenne, 26 April 1643, AT III, 648; Philosophical Letters, pp. 135 - 136.
39 For Descartes's definition of matter, see, e.g., Meditations, V and VI and Principles,
I, 60, AT VIII, 28; HR I, 243 - 244; ibid., II, 22 and 64, AT VIII, 52 and 78 -79; HR
1,265.
40 See Note 23 above.
30
BIBLIOGRAPHY
Descartes, R.: Oeuvres de Descartes (AT), publies par Ch. Adam et P. Tannery, Leopold
Cerf, Paris, 1897 - 1913, 12 vols.
247
Descartes, R.: The Philosophical Works of Descartes (HR), transl. by E. S. Haldane and
G . T . Ross, London, 1911 (1978), 2 vols.
Descartes, R.: Philosophical Letters, transl. and ed. by A. Kenny, Clarendon Press, Oxford, 1970.
Alanen, L.: 'Studies in Cartesian Epistemology and Philosophy of Mind', in Acta
Philosophica Fennica, Vol. 33, Helsinki, 1982.
Alquie, F. (ed.): Oeuvres philosophiques de Descartes, Garnier Freres, Paris, 1%3 -1973,
3 vols.
Aristotle: The Works of Aristotle, Vols. I and VIII, ed. by W. D. Ross, Clarendon Press,
Oxford, 1928.
Beck, L. J.: The Metaphysics of Descartes, Oxford University Press, Oxford, 1965.
Donagan, A.: 'Descartes' "Synthetic" Treatment of the Real Distinction Between Mind
and Body', in M. Hooker (ed.), Descartes, Critical and Interpretive Essays, The Johns
Hopkins University Press, Baltimore, 1978.
Edwards, S.: Medieval Theories of Distinction, University of Pennsylvania, Ph .D., 1974.
University Microfilms International, Ann Arbor, Michigan and London, 1981.
Gilson, E.: Index Scolastico-Cartl!sien, Paris, 1912; repro Burt Franklin, New York, 1964.
Gilson, E.: Rene Descartes: Discoursde la methode. Texte et commentaire, J . Vrin, Paris,
1925 (1967).
Gilson, E.: Etudes sur Ie role de la pensee medievale dans la formation du systeme cartesien, J. Vrin, Paris, 1930 (1975).
Hooker, M. (ed.): Descartes: Critical and Interpretive Essays, The Johns Hopkins University Press, Baltimore and London, 1978.
Kenny, A.: Descartes, A Study of His Philosophy, Random House, New York, 1968.
Koyre, A.: Essais sur I'idee de Dieu et les preuves de son existence .chez Descartes, Ernest
Leroux, Paris, 1922.
Lloyd, G . E. R.: Aristotle: The Growth and Structure ofHis Thought, Cambridge University Press, Cambridge, 1968.
Malcolm, N.: 'Descartes' Proof That His Essence Is Thinking', in The Philosophical
Review (PR) LXXIV (1965), 315 - 338.
Malcolm, N.: Problems of Mind, Descartes to Wittgenstein, Harper Torchbooks, New
York, 1971.
Malcolm, N.: Thought and Knowledge, Cornell University Press, London, 1971.
Sommers, F.: Dualism in Descartes: The Logical Ground, in M. Hooker (ed.), 1987, pp.
223 - 233.
Spinoza, B. de: Opera, I - II, in l. van Vloten and J. P. N. Land, The Hague, MCMXIV.
Suarez, F.: Opera Omnia, Vol. XXV, Paris, 1866; repr. Georg Olms, Hildesheim, 1965.
Suarez, F.: Francis Suarez: On The Various Kinds of Distinctions, transl. by C. S. J.
Vollert, Marquette University Press, Wisconsin, 1947.
Weinberg, J. R.: A Short History of Medieval Philosophy, Princeton, New Jersey, 1964.
Weinberg, J. R.: Abstraction, Relation, and Induction. Three Essays in the History of
Thought, Madison & Milwaukee, 1965.
Weinberg, J. R.: Ockham, Descartes, and Hume, The University of Wisconsin Press,
Madison, Wisconsin, 1977.
Wells, N. J.: 'Descartes and the Modal Distinction', The Modern Schoolman XLII (1965),
1-22.
Williams, B.: Descartes, The Project ofa Pure Enquiry, Penguin Books, Hammonsworth,
1978.
Wilson, M. D.: 'Descartes: The Epistemological Argument for Mind-Body Distinctness',
Noils 10 (1976),3 -15.
248
LILLI ALANEN
Dept. of Philosophy,
University of Helsinki,
Unioninkatu 40B,
SF-00J70 Helsinki 17, Finland.
JAAKKO HINTIKKA
N(3x)[(y)(y exists
:::J
x exists)]
249
S. Knuullila and 1. Hintikka (eds.), Tne Logic oj Beini!. . 249--267.
<0 1981 by Dia/ecrica.
250
JAAKKO HINTIKKA
where "N" is the necessity operator. There are two things to be observed
about this sentence. First, it is trivially valid (logically true), and, second,
that it looks very much like a summary of the ontological argument. Let
me spell out these two points.
I have noted before that (1) is trivially valid completely independently
of what the "predicate of existence" used in it is. 5 I have also claimed
that it is the logical truth of (1) that makes the ontological argument so
perennially seductive. In fact, (1) might seem to express precisely the
desired conclusion, viz. the necessary existence of an existentially perfect
being. Indeed, it can readily be seen that (1) is closely related to the ontological argument in its actual historical versions. What (1) says is that,
necessarily, there is an individual such that if anything exists, it does,
which nearly says - or seems to say - that there is something which is
greatest with respect to existence ("pre-eminent in its mode of existence", to use Kant's words in A 586 = B 614). Thus the inside conditional in (1), viz. (y) (y exists => x exists), can be considered as a
characterization of god ( = x), conceived of as the most powerful being
with respect to existence. Kant asserts what is very similar to this inside
conditional of (1) when he says (A 588 = B 616) that "from any given existence ... we can correctly infer the existence of an unconditionally
necessary being", that is, a God. Thus, the whole of (1) seems to express
quite well the Anselmian idea that the most perfect being - a being
greater than which cannot be conceived of - must necessarily exist, with
the perfection in question restricted to perfection or maximal greatness
with respect to existence.
From (1) the defenders of the ontological argument in effect
fallaciously infer
(2)
This quantifier switch is the crucial mistake in the most interesting versions of the ontological argument. For, appearances notwithstanding, it
is (2) and not (1) that Anselm, Descartes & Co. really want to establish.
Thus the logical truth of (1) helps them only if they could take the further
step from (1) to (2). But this further step is illegitimate. I shall later return
to the question as to what further premises might serve to validate the
step.
It remains to spell out more fully what is involved here. First why is
it (2) and not (1) that the ontological argument is calculated to prove?
This is seen easily by means of the obvious possible-worlds semantics of
(1) - (2). What (1) says is that in each world there is something such that
251
if anything at all exists in that world, it does. The reason why this is not
enough is due to the fact that in different worlds such individuals can be
entirely different from each other. Indeed, the triviality of (1) is reflected
by the fact that any existing individual can be chosen as the value of the
existentially bound variable "x" in (1).6 We would be able to infer (2)
from (1) only if we could assume that all these individuals are (or can be
chosen to be) identical with each other.
In contrast, (2) attributes the status of existentially greatest being to
some one individual in all (nonempty) worlds. 7 This is obviously what the
argument is supposed to establish. In view of the validity of (1), the right
way of criticizing the ontological argument is hence to spell out the difference between (1) and (2) and to show how and wh~ the step from (1)
to (L) is fallacious.
A prolegomenon to such a criticism is to point out how natural
language tends to hide the differences between (1) and (2). Indeed, such
English sentences as
(3)
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)
252
JAAKKO HINTIKKA
of (2) from (I). Probably the most effective way of doing so is to appeal
253
N(3x)(g == x)
Now we know from modal logic that this term "g" can help to establish
(2) only if we have at our disposal the additional premise
(6)
(3x)N(g == x),
where no existential force is being assumed. What (6) expresses is precisely the identity (g == x) of g with some one individual (the x in "(3X)" in
all the possible worlds (introduced by "N"). As we saw, the failure of
this identity is precisely the fatal flaw in the usual versions of the ontological argument.
The auxiliary premise (6) is analogous to the proposition
(7)
which says that it is known who (or what) God is. In fact it is (7) and not
(6) that we need as an auxiliary premise if we want to establish that it is
known that God exists. It is no wonder, in view of these observations,
that the literature on the ontological argument is full of considerations
of whether God necessarily is who he (or she) is, and whether we can
"conceive of" or "understand" who God is. As we know, Anselm and
Gaunilo were already discussing the latter issue. As far as the former is
concerned, the missing premise (6) seemed to be conveniently supplied
by Exodus 3: 13 - 14, where God says: "I am who I am", presumably
meaning that He necessarily is who He is. II It would divert our purpose
to explore these historically important lines of thought here, however.
I think it is the time to lay to rest the myth that "existence is not a
predicate". It is embarrassingly clear what Kant's grounds for maintain-
254
JAAKKO HINTIKKA
ing this thesis were. They were largely due to the paucity of the logics and
languages he was contemplating. He envisaged only two types of judgment relevant here, viz. what I shall here call judgments of "essential"
predication and judgments of existence. (Where contingent predication
was supposed to find a niche was not explained by the good Immanuel. 12
In the former, exemplified by "God is omnipotent", a necessary connection is asserted to obtain between the subject and the predicate, without
prejudicing the existence of either. As Kant puts it, "the omnipotence
cannot be rejected if we posit a Deity, for the two concepts are identical". But this judgment carries no existential import. " ... if we say,
'There is no God', neither the omnipotence nor any other of its
predicates is given; they are one and all rejected together with the
subject" .
The other kind of judgment Kant mentions is the existential one, e.g.,
"God exists". In neither one is existence a predicate, Kant says in effect.
A judgment of essential predication has no existential force, whereas in
an existential one we take a subject as it were all ready-made with its
essential predicates and simply assert that this particular complex of
predicates is in fact instantiated in reality. Here existence is not one of
the configuration of predicates; it is what is asserted of the configuration.
Nothing is wrong here. A faithful Aristotelian would have worried
about the total absence of existential import in a judgment of essential
predication, for on certain conditions Aristotle seems to have maintained
such an import. He went so far as to worry lest this would lend other instances of copula a similar existential force, so that we could fallaciously
infer from "Homer is a poet" that "Homer is", i.e., exists. 13
However, our worries are not Aristotelian. Kant's mistake is not that
he says something false, but that his philosophical diet is one-sided: he
nourishes himself on too few kinds of examples. In reality, there is a
tremendous multitude of forms of proposition which go way beyond the
ones Kant envisages. Among them, I suggest, we can safely assume to be
included some in which' 'existence is a predicate" in whatever reasonable
sense we can give to this phrase.
The following argument may indicate why this assumption is eminently natural - and also why the use of existence as a proper predicate has
met with such resistance among philosophers. This line of thought
would of course have been rejected by Kant, but I think that it would
have been appreciated by Leibniz.
Obviously, we attribute to actual individuals all the time predicates
255
which turn on what they would be like in other possible worlds, for instance, what they could be or could do. Sometimes these predicates turn
on the existence or nonexistence of these individuals in those other circumstances. For example, speaking of the necessary conditions of life in
the case of some particular organism involves this kind of predication.
All that is needed to be able to use actual existence as a predicate (so as
to refute Kant) is then apparently a parity of cases. If we can take an individual in the actual world and assign to it a predicate which involves
existence or nonexistence in some other world, surely we ought to by the
same token be able to take a "merely possible individual", i.e., a denizen
of some other world, and attribute to it predicates definable in terms of
its actual existence, maybe the "predicate of (actual) existence" itself.
Basically, it seems to me that this argument is unanswerable. There certainly are concepts applied to actual individuals which can only . be
defined in general in terms of (merely) potential existence, e.g., the
biological concept of fertility. Even though we don't do it often, we surely can pick out one "merely possible" individual from others by specifying that it enjoys the dubious distinction of actual existence. Examples
are not very easy to come by, but speaking of the actual Hamlet seems
to be good enough. For many of us, Hamlet is first introduced as a merely fictional "possible individual", and we learn only subsequently that
the melancholy Dane has a real-life counterpart. (You didn't know that
Hamlet really existed? Yes, he did enjoy the predicate of existence!)
There are several different kinds of difficulties here which have led
some philosophers to deny the possibility of the sort of return of an individual from other possible worlds to the actual one which I am envisaging. Some philosophers have failed to see how we can individuate a merely possible individual. Doesn't the very possibility of considering some
one definite individual (to which predicates are to be ascribed) presuppose its actual existence? The fact that philosopher-logicians as eminent
as Montague and Kripke have maintained this presupposition shows that
we are not dealing with a mere idle worry. I cannot here discuss this complex of problems in its entirety. A good descriptive account of how merely possible individuals can enter into our discourse is given by David
Kaplan in 'Quantifying In' .14 In general, I believe that the denial of
merely possible individuals is based on an unrealistically narrow view of
how out language actually functions. IS
There is another reason why rec'.!nt logicians may have been wary of
the line of thought I just adumbrated. In it, we took an individual which
had been considered qua citizen of another world and began to consider
256
JAAKKO HINTIKKA
M[(x)(A(x)]
B(x
where "M" is the possibility operator. Of course (8) is ill-formed in traditionat modal logic. Yet it expresses a perfectly good semantical sense. It
says that there is some alternative possible world such that everything
that is there A is in fact (i.e., in the actual world) B. The remarkable thing
is that this sense cannot be expressed by any well-formed formula of conventional modal logic. 17
This is nevertheless merely a limitation of one particular kind of notational systems. One can, as Esa Saarinen has done, introduce special
"backwards looking" operators which effect just the kind of return
journey I was envisaging. 18 What is more interesting, Saarinen has
shown convincingly that the kind of anaphora which these operators are
calculated to facilitate does occur frequently and importantly in ordinary
discourse. He has thus removed one important obstacle from the way of
vindicating existence as a genuine predicate, and incidentally illustrated
how important metaphysical dogmas can be embodied in a perfectly
innocent-looking formalism. Although further arguments are still
needed, I hope to have persuaded you at least that there is no mistake
in considering existence a predicate. The mainstay of Kant's criticism of
the ontological argument is simply wrong.
But doesn't Kant deserve at least the honor of anticipating the Frege
257
In the Critique oj Pure Reason (A 598 = B 626), Kant says that" 'being'
is obviously [sic] not a real predicate .... It is merely the positing of a
thing, or of certain determinations, as existing in themselves".
The reason why Kant introduces the term "setzen" is probably a desire
to have a term which sits more happily with the cases in which "is" apparently has a merely predicative function. "God is omnipotent" could
according to Kant be true even if there were no God. It merely expresses
a necessary relation between the subject and the predicate. "God is omnipotent" does not logically imply for Kant that "God is", even though
the step might seem tempting. In order to avoid this temptation, it seems,
Kant uses his terminus quasi technicus "setzen" for positing something
as being - in any sense of being.
The explanation Kant gives of the difference between' 'God is omnipotent" and "God is" nevertheless shows that we are dealing with the same
"is" in both cases. In both cases, we are "positing" something. The only
difference is that in the former case the positing is relative but in the latter
case absolute. Otherwise, it is the same old positing.
lIn] the proposition, 'God is omnipotent' .... the small word 'is' adds no new predicate,
but only serves to posit !he predicate in its relation IKant's italics] to the subject.
258
JAAKKO HINTIKKA
If, now, we take the subject (God) with all its predicates ... and say 'God is', or 'There
is a God', we attach no new predicate to the concept of God, but only posit the subject
in itself with all its predicates ...
Here positing clearly means to assume existence. The relation of existence to predication is thus merely that of an absolute positing to a
relative one. This observation is confirmed by further passages; witness
e.g. the following:
[In an existential jUdgment) ... nothing can have been added to the concept, which expresses merely what is possible, by my thinking its object (through the expression 'it is')
as given absolutely [my italics).
Thus Kant clearly thinks of the "is" of predication (the copula) and the
"is" of existence as two uses of the same notion. Occasionally he even
seems to consider the copulative "is" (at least in necessary judgments)
as a variant of the "is" of identity. He thinks of a: necessary judgment
like "God is omnipotent" as expressing the identity of a God and an omnipotent God. "The omnipotence cannot be rejected if we posit a deity
... , for the two concepts are identical" (A 595 = B 623). Hence meaning differences between the first three elements of the Frege ambiguity
are rejected by Kant.
As to the fourth alleged sense of "is" apud Frege and Russell, Kant's
assimilation of it to other senses (especially to the "is" of predication)
is seen from his failure (or refusal) to distinguish the subsumption of one
concept to another from the application of a concept to a particular (in
other words, this particular's failing under the concept). This is particularly striking in the schematism chapter of the Critique oj Pure
Reason, as has been often remarked. 20
The insight that Kant did not assumm the Frege - Russell distinction
enables us to make further observations. Among other things, it follows
that Kant's main thesis is expressed somewhat inaccurately - and in
any case very narrowly - when it is said that according to him existence
is not a predicate. What he maintained, and frequent[y said, is that being
is not a real predicate. This applies to both existential and predicative
uses of "is"; predication is accordingly for Kant as little a predicate as
259
existence is. This parity of the two is of course just a corollary to Kant's
failure (or refusal) to distinguish the different Fregean senses of "is"
from each other.
This helps to put certain puzzling-looking statements of Kant's
perspective. For instance, one of Kant's main pronouncements on our
topic in the first Critique runs as follows:
'Being' [SeinJ is obviously not a real predicate .. . . In its logical use lim logischen
Gebrauchl it is merely the copula of a judgment. (A 598 = B 626; Kant's emphasis.)
Here Kant makes his claim about being in general and then goes on to
apply it to predication rather than existence. Indeed, this predicative use
is precisely what he means by the "logical use" of being. In other words,
Kant's distinction between the logical use and other relevant uses of "is"
is the same as his contrast between the relative and absolute positing
discussed above. (This is, among other items of evidence, shown by our
latest displayed quote from the Beweisgrund; see especially the words
respectus logicus.) Philosophers have been puzzled by Kant's remarks as
to what happens to "is" in its merely logical use, and declared it irrelevant to Kant's main thesis that existence is not a predicate. 21 Kant'~
remarks are indeed not directly relevant, but only because they pertain
to a different but parallel case of his more general claim that being is not
a predicate.
There is one superficial aspect of Frege's and Russell's formalism
which misleadingly encourages the idea that Kant's thesis "existence is
not a predicate" is an anticipation of Frege. In the most literal sense, existence is not a predicate for Frege, either, viz. in the sense of being an
explicit predicate of individuals. We cannot take a free singular term
(Frege's "proper name"), say "b", and go on to assert "b exists".
However, this is a merely contingent feature of Frege's notation. What
is more, it partially hides one of the most fundamental features of his
treatment of existence, viz. that existence is expressed only by the existential quantifier.
In fact, the reason why Frege can get along without a predicate of existence is that he assumes that all proper names (free singular terms) are
nonempty. This is reflected by the validity of existential generalization
in Frege's system: from any proposition F(b) containing "b" we can infer (3x)F(x). This obviously presupposes that b exists. If we do not make
this assumption, we have to amplify the rule of existential generalization
and formulate it as saying that from the two premises
(9)
260
JAAKKO HINTIKKA
we may infer
(10)
(3x)F(x).
From certain eminently natural assumption one can show (as I have
demonstrated)22 that the extra premise "b exists" must be equivalent
with
(11)
(3x)(b = x)
Indeed, all we need for this purpose is in effect that the logical constants
have their customary semantics and that the "predicate of existence" ,
whatever it is or may be, is subject to the same substitutivity principles
as other expressions of first-order logic .
This result shows that in a Frege - Russell logic it is in the last analysis
the existential quantifier alone that need carry existential assumption,
contrary to the misleading appearance created by Frege's notation. This
idea can be considered an integral part of Frege' s distinction between the
"is" of existence and other senses of "is". Indeed, this privileged position of the existential quantifier seems to me to be a much more important feature of the overall Frege - Russell approach to logic than the
alleged impermissibility of asserting the existence of an individual in
Frege's canonical notation. We can now see that Frege's distinction does
not presuppose that "existence is not a predicate". On the contrary, the
full import of Frege's approach cannot be spelled out without a
"predicate of existence". Hence Kant's thesis does not make him into a
precursor of Frege and Russell.
Thus we can likewise see that in the last analysis we could, and should,
have "a predicate of existence" also for the extremely simple languages
to which Frege (and mutatis mutandis also Kant) restricted his attention.
Consequently, the reasons for having such a predicate in one's language
are not applicable only to the rich languages envisaged above, but apply
also within the present-day Frege - Russell languages.
This observation nevertheless need not drive a wedge between Frege
and Kant. One way of expressing our result concerning Frege might be
to say that for Frege existence was a predicate, but not a normal or
"real" predicate. In its primary use, existence is a second-order
predicate, saying that a certain first-order predicate is instantiated. The
question whether this second-order predicate can be extended to the
trivial first-order predicates of the form "(b = x)" is of little interest to
Frege. But if so, there is after all a partial agreement between Frege and
Kant. For Kant frequently formulates his point by saying, no! that ex-
261
istence is not a predicate, but that It IS not a real predicate. 23 Indeed, Kant
must obviously allow us to express not only the nonemptyness of common nouns but also the nonemptyness of singular nouns. Then Kant's
injunction that existence is not a real predicate might per haps be interpreted as saying merely that it must not be used in the definition of
anything. 24 This is an interesting point, but on the reconstruction of the
ontological argument presupposed here it is neither necessary nor sufficient for a refutation of the ontological argument. However, it is far
from clear what the precise import of Kant's locution is when he speaks
of a "real predicate", and it is not obvious a priori that his exclusion
of existence from the definition of anyone thing cannot itself be turned
into a line of defense for his critical claims. We shall return to these
points later.
That Kant's criticism of the ontological argument is largely beside the
point can also be seen in terms of his own system. It is largely a Fremdkorper in the body of his own transcendental philosophy. Earlier, I
quoted Kant as saying (in A 599 = B 627) that the concept "expresses
merely what is possible [my italics]". Elsewhere, too, he clearly thinks
of what I have called essential judgments as expressing possibilities. 25 in
an existential judgment, this possibility is asserted to be actualized,
without adding anything to the concept itself. Now this is precisely what
Kant could not say as his definitive opinion in the case of God. For if
God were in the fullest sense of the word possible for Kant, in the sense
of being empirically possible (possible in experience), he would
presumably be sometimes actual, and hence (since we are dealing with a
putatively necessary being) always actual. This possibility of restoring
something like the ontological argument by means of the additional
premise that God is possible had been exploited by Leibniz. Even though
Kant presumably would have rejected Leibniz' argument for other
reasons, he could scarcely afford to admit God's possibility.26 (In order
to see this, we may for instance recall Bill, where Kant says that
"necessity is just the existence which is given through possibility itself".)
It is true that the step I envisaged a moment ago from God's experiential possibility to his actuality is not backed by any outright assertions in
the Critique of Pure Reason of what Lovejoy called the Principle of
Plenitude, that is to say, of the principle that each genuine possibility is
actualized in the long run. Howevec, a closer examination of Kant's position shows that he could not really countenance vi~lations of the Principle
of Plenitude among full fledged experiential possibilities. This examination I have attempted, jointly with Heikki Kannisto, in an earlier
262
JAAKKO HINTIKKA
Kant was likewise aware of the specific arguments that are lurking
here, ready to jump in at a mere admission of God's possibility:
And thus the celebrated Leibniz is far from having succeeded in what he prided ~imself on
achieving - the comprehension u priori of the possibility [my italics) of the sublime ideal
being. (A 602 = B 630)
263
Without taking up the details of Kant's line of thought in this direction, we can see the main idea clearly enough. God is a mere ideal because
He is an ens realissimum, the sum total of all reality. Few philosophers
or theologians would quarrel with this idea, and Kant is perhaps not so
very wrong in seeing in this definition a ground for assigning to God
merely ideal existence. The only remark I want to make here is that
Kant's move is beside the point as a criticism of the original ontological
argument, on the rational reconstruction offered above. For on this interpretation, nothing turned on God's being ens realissimum. God was
not assumed (or defined) to embody all perfections. For my
reconstructed argument, it sufficed merely to assume that God has one
particular perfection. It suffices to assume that He is the most powerful
being existentially, in precisely the sense asserted by Kant. (Cf. the quote
above from A 588 = B 616.) For this is what the inside conditional of (I)
expresses. Hence the move that Kant uses to undercut Leibniz-style
arguments for God's existence is completely beside the point as an objection to the original ontological argument. Once again, we see how Kant
misses the key fallacy in the argument.
Perhaps there nevertheless is a charitable interpretation of Kant's position which makes his transcendental vantage point relevant to a valid line
of criticism of the ontological argument. Very roughly, my initial
diagnosis of the ontological argument may be expressed by saying that
the trouble with the argument is not that existence is not a predicate, but
that we don't know who God is in the sense that the designation ("definition") of him as existentially greatest being provides us with no grounds
for concluding that the existentially greatest beings in different worlds
are identical. We may try to interpret Kant's position (and perhaps even
his slogan that existence is not a predicate) as emphasizing this aspect of
the situation. Kant's slogan is applicable in that Kant's saying that existence is not a "real" predicate may perhaps be taken to amount to saying that it does not help us to determine God's identity in the sense of
bringing us to know who God is. This idea is related to my earlier observation that according to Kant existence cannot be a part of the definition
of any entity. This would make Kant's discussion relevant to a valid line
of criticisms against the ontological argument. For it amounts to saying
that Kant was in the last analysis denying the indispensable auxiliary
premise illustrated by (6).
This highly interesting line of thought is the thesis of Hans Wagner's
paper (above, Note 5). If there is a kernel of truth in Kant's criticism of
the ontological argument, this undoubtedly is it. In spite of the evidence
264
JAAKKO HINTIKKA
In contrast,
if, now, we ... say "God is" or "There is a God", we ... only posit the subject in itself
with all its predicates.
Then I am
thinking its [my concept'sl object (through the expression 'it is') as given absolutely.
Here we can also see what the force of Kant's little word "real" (in
saying that existence is not a "real predicate") really is. Kant is merely
following his customary contrast between what is logical and what is real
and identifying the "real" use of "being" to its existential use. In other
words, what is "unreal" about the purely predicative use of "is" is the
absence of existential presuppositions. In fact, the quoted passage continues immediately:
Otherwise stated, the real contains no more than the merely possible.
We have already seen that Kant brackets together the merely possible,
265
the predicative (copulative) use of being, and logic (i.e., the world of
concepts).27 By contrast, the real or actual should go together with the
existential use of being. If I may turn Kant's point into a tautology, what
he is saying is that the predicative use of "is" is not its existential use.
If this suggestion is correct, then the burden that many philosophers
have tried to put on "real" in "real predicate" is largely misplaced.
Some philosophers have for instance tried to find links between Kant's
criticism of the ontological argument and his discussion of reality as one
of the modal categories in the Transcendental Analytic, and assumed
that they are what is highlighted by the word "real". Others have
thought that they could perceive in Kant a contrast between being as a
"real" predicate (as a predicate of individuals) and its "merely logical"
use (as a higher order predicate). There is no foundation in the text for
either view, and the passage we just examined suggests that what Kant
intended was something much simpler. 28
It might seem at this point that one part of the Frege - Russell distinction does after all playa major role in Kant, viz. the distinction between
the existential and the predicative uses of "is". It is true that Kant puts
a premium on this distinction, but we have already established beyond
all doubt that it is for him a difference in use and not a difference in
meaning.
NOTES
The writing of this paper was made possible by a fellowship from the John Simon Guggenheim Memorial Foundation. In writing it, I have profited greatly from discussions with
Merrill B. Hintikka, Russell Dancy, and Robert Beard, and esrecially from conversations
and correspondence with Robert Howell. I also profited greatly from the discussion of an
early version of this paper at the Fourth International Colloquium in Biel, May 1-4, 1980,
and I .would like to thank all the participants in that discussion.
1 Cf. e.g. The Many-Faced Argument, ed. by John Hick and Arthur C. McGill, the Macmillan Co., New York, 1967; The Ontological Argument, ed. by Alvin Plantinga and
Richard Taylor, Doubleday, Garden City, N.Y., 1965; Dieter Henrich, Der ontologische
Gottesbeweis, J. C. B. Mohr, Ttibingen, 1960 (second ed. 1967); W. L. Gombocz, Uber
1: Zur Semantik des xistenzpriidikates und des ontologischen Argumentes, Verband der
wissenschaftlichen Gesellschaften bsterreichs Verlag, Wien, 1974; Jonathan Barnes, The
Ontological Argument, Macmillan, London, 1972. All these give further references to the
literature. For a contemporary journalistic view, see Time, April 7, 1980, Pi>. 65 - 68.
2 Cf. my paper ' "Is", Seman tical Games, and Semantical Relativity', Journal of
Philosophical Logic 8 (1979), 433 - 468, which provides further references to the literature.
3 Cf. here my AP A presidential address 'Gaps in the Great Chain of Being: An Exercise
in the Methodology of the History of Ideas', Proceedings and Addresses of APA 49
(1975 - 76), reprinted in S. Knuuttila (ed.), Reforging the Great Chain of Being, D. Reidel,
Dordrecht, 1981, pp. 1 -17.
4 See e.g. Jonathan Bennett, Kant's Dialectic, Cambridge University Press, Cambridge,
266
JAAKKO HINTIKKA
1974, pp. 228 - 240. Bennett even speaks in the title of this 72 of "the Kant - Frege view".
See my essay, 'On the Logic of the Ontological Argument', in laakko Hintikka, Models
for Modalities, D. Reidel, Dordrecht, 1%9. This essay prompted a perceptive attempt to
show that my criticism of the ontological argument is related to Kant's; see Hans Wagner,
'Uber Kants Satz, das Dasein sei kein Priidikat', Archiv fur Geschichte der Philosophie 53
(1971), 183 - 186.
6 Cf. 'On the Logic of the Ontological Argument' (loc. cit.)
7 The difference between (I) and (2) is essentially a de dicto-de re contrast. Further discussion of the relation between the two constructions in Kant is found in Robert Howell's
paper in Dialectica 35 No. I (1981).
8 See my essays collected in Esa Saarinen (ed.), Game-Theoretical Semantics, D. Reidel,
Dordrecht, 1979.
9 See here my books Time and Necessity, Clarendon Press, Oxford, 1973; and (with Simo
Knuuttila and Unto Remes) Aristotle on Modality and Determinism (Acta Philosophica
Fennica 29, No. I), North-Holland, Amsterdam, 1977.
10 Kant's thesis has implications beyond what is being discussed here. It can be construed
as criticizing the medieval and neo-Platonic idea that existence qua existence carries with
itself interesting attributes of which we can profitably theorize. The problems goes back
to Aristotle's aporia concerning a science of being qua being. By and large, Aristotle was
l1luch more wary of such a science than were his followers. Kant's denial that existence is
a predicate may hence be viewed as the end of a long neo-Platonic and scholastic detour.
(Cf. Notes 13 and 24 below.)
II On the history of the interpretations of this pronouncement, cf. E. Gilson, History of
Christian Philosophy in the Middle Ages, Random House, New York, 1955, pp. 3, 69 -70,
92,149,216,253,293,368,371,438-439,579, and 591.
12 Some indications are nevertheless found in Kant, Logik, Academy Edition, Vol. 9, pp.
60-61.
13 Cf. Russell M. Dancy, Sense and Contradiction: A Study in Aristotle, D. Reidel, Dordrecht, 1975. (See esp. Appendix II, pp. 153-155.)
14 In Donald Davidson and laakko Hintikka (eds.), Words and Objections, D. Reidel,
Dordrecht, 1969, pp. 178-214.
IS Cf. also lerome Shaffer, 'Existence, Predication, and the Ontological Argument', Mind
71 (1962), 307 - 325. Shaffer maintains that "the most that the Ontological Argument
establishes is the intensional object, God ... ". Apparently Shaffer does not see any problem in the uniqueness of that "intensional object".
16 This observation helps in understanding other features of the literature on the ontological argument. For instance, why do most of the recent formal or semi-formal discussions of the ontological argument presuppose S5? Because the alternativeness relation is
symmetric in S5, and thus allows for an attenuated form of "return journeys", which
brings a merely possible individual back to the actual world.
17 Even some of the medievals seem to have been aware of the need of the kind of return
trip logic exemplified by (8); see Simo Knuuttila and Esa Saarinen, 'Backwards-Looking
Operators in Buridan', in I1kka Niiniluoto et al. (eds.), Studia Excel/entia: Essays in
Honour of Oiva Ketonen (Reports from the Department of Philosophy, University of
Helsinki, 1977, No.3.), pp. 11-17.
18 See Esa Saarinen's own contributions to Game-Theoretical Semantics, ed. by Esa
Saarinen, D. Reidel, Dordrecht, 1979.
19 See Note 2 above.
20 Cf. e.g. Henry Allison, 'Transcendental Schematism and the Problem of the Synthetic
A Priori', Dialectica 35 (1981); Gerold Prauss, Erscheinung bei Kant, Berlin, 1971, p. 103.
5
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 ambiguity. Jaakko Hintikka has recently argued that, except for John Stuart Mill
and Augustus De Morgan, the ambiguity claim did not play any major
role in philosophical thinking before Frege and Russell. l What Frege and
Russell accomplished was to make the ambiguity of is a cornerstone of
modern first-order logic. Therefore, as Hintikka has pointed out,
"anyone who uses this logic as his or her framework of semantical
representation is thus committed to the Frege - Russell ambiguity thesis"
(Hintikka, 1983, p. 449). Hintikka has shown that in an alternative
seman tical representation, namely, in game-theoretical semantics, no
ambiguity claim need be made, which, of course, is not to deny that there
are different uses of is. The operative question is whether the differences
between these different uses have to be accounted for by assuming that
one particular verb for being is ambiguous, i.e., has several altogether
different meanings.
Independently of Hintikka, Charles Kahn (1973) and Benson Mates
(1979), among others, have called attention to the ambiguity doctrine
and partly challenged the validity of what Frege considered as an "eternal truth" concerning the verb to be.
But how is the verb is ambiguous in Fregean logic? Frege distinguishes
from each other the following meanings of is:
(1)
(2)
(3)
269
S. Knuuttila and 1. Hintikka (eds.), The Logic of Being, 269-289.
1986 by D. Reidel Publishing Company.
270
(4)
LEILA HAAPARANTA
or
(ii) expressed by means of the existential quantifier and the
symbol for predication (e.g., There are human beings/There
is at least one human being; (3x)H(x,
and
the is of class-inclusion, i.e., generic implication (e.g., A horse
is a four-legged animal; (x)(p(x) ::::l Q(x))).
As is shown in the brackets, each putative meaning of is has its own formalization in first-order logic.
Since the Fregean view of is has held such an indisputable position in
modern first-order predicate calculus, few philosophers have tried to
find out what actually led Frege to the ambiguity thesis. This paper is an
attempt to give some hints of an interpretation of the Fregean distinction. I shall concentrate on the following questions: (1) How did Frege
arrive at the distinction between the is of existence and the is of predication? (2) What is the philosophical background which motivated Frege's
distinction between the is of identity and the is of predication?
2. FREGE ON 'IS'
271
272
LEILA HAAPARANTA
273
sions. Besides concept-words and relation-words, there are functionnames like "the capital of ( )" from which we do not derive sentences
by filling their gaps with complete expressions.
Proper names can refer to objects, e.g., particular things, numbers,
classes, etc.; and sentences, which are -included by Frege among proper
names, refer to truth-values, the True and the False. Function-names
refer to functions: concept-words to concepts, relation-words to relations, and other function-names to other functions. 8 The values of functions named by concept-words and relations-words are always truthvalues. A concept, when filled with an object, will turn into an object,
namely, into a truth-value, and the same applies to a relation (GGA I,
pp. 7 - 8). When a function which is neither a concept nor a relation is
filled with an object, it will also turn into an object, but not a truth-value.
Names stand for their references (Bedeutungen) and express their
senses (Sinne). 9 Frege tells us quite explicitly that the sense of a sentenceis a thought. It is, however, much more difficult to see what he actually
means by the sense of a proper name which is not a sentence. In 'Uber
Sinn und Bedeutung' he remarks that the sense of a proper name is a way
in which the object to which this expression refers is presented (die Art
des Gegebenseins), or a way of 'looking at' this object. Furthermore, he
states that the sense expressed by a proper name belongs to the reference.
In other words, for Frege, senses are not primarily senses of names but
senses of references. Hence, it is more advisable to speak about senses
expressed by names than senses of names. Frege also gives some
examples of Sinne, like the Evening Star and the Morning Star as senses
of Venus, and the teacher of Alexander the Great and the pupil of Plato
as senses of Aristotle ('Uber Sinn und Bedeutung', KS, p. 144).
On the basis of Frege's hints, it is obvious at least that his concept of
Sinn is throughout cognitive. I suggest that the Sinne are nothing but
complexes of individual properties of objects. Frege regards it as possible
for an object to be given to us in many different ways. For him, our
knowledge of an object determines what sense, or what senses, the name
of the object expresses to us. Furthermore, he argues that complete
knowledge of the reference would require us to be able to say whether
any given sense belongs to the reference but that such knowledge is
beyond our reach ('Uber Sinn und Bedeutung', KS, p. 144).
As far as the suggested interpretation of the concept of Sinn is correct,
it is Frege's view that we know an object completely only if we know all
its properties, which is not possible for a finite human being. It also
follows that, according to Frege, each object can in principle have an in-
274
LEILA HAAPARANTA
It is far from surprising that Frege's view of is, as not being a common
275
276
LEILA HAAPARANTA
tion actually shows that being can, after all, be used as a logical firstorder concept, although Frege himself does not propose anything of that
kind.
Frege's argument for the claim that sentences like "A is" or "A exists" are self-evident is based on the impossibility of negating these
sentences without contradiction. But it is not conclusive. Why does he
not suggest that the sentence "Something that has being is not" or
"Something that has being falls under the concept of not-being" means
that something for which it is possible to exist does not exist in the actual
world? Interpreted this way, the contradiction seems to disappear.
Precisely that kind of procedure can be used in possible worlds semantics. As laakko Hintikka argues, possible worlds semantics allows us "to
take a 'merely possible individual', i.e., a denizen of some other world,
and attribute to it predicates definable in terms of its actual existence,
maybe the 'predicate of actual existence' itself" (Hintikka, 1981 a, p.
134).
That kind of step is, however, impossible for Frege, due to his overall
conception of language and world. Frege points out on several occasions
that his aim in the Begriffsschrift was, in Leibniz's terms, not only to present a calculus ratiocinator but to create a lingua characteristica. 13 The
idea of logic as language seems to be incorporated in many of the doctrines maintained by Frege, as has been emphasized by lean van Heijcnoort (1967) and laakko Hintikka (1981b). For Frege, the conceptual
notation is a proper language, which must be learnt by means of suggestions and clues. Frege is committed to the doctrine that we cannot step
outside the limits of our language in order to consider the seman tical relations between language and reality. Moreover, in the Fregean framework
we cannot talk about changing universes. As van Heijenoort points out,
"Frege's universe consists of all that there is, and it is fixed" (van Heijenoort, 1967, p. 325). Frege cannot pick out an individual from some
possible world and attribute to it the property of actual existence for the
simple reason that, for him, there are no alternative worlds. That kind
of position prevents Frege from regarding the is of existence as an expression of a meaningful first-order concept.
Frege attaches the meaning of an existential judgement to the form of
a particular judgement. He states:
Every particular judgement is an existential judgement that can be converted into the 'there
is' form. E.g. 'Some bodies are light' is the same as 'There are light bodies'. (NS, p. 70;
Long and White, p. 63.)
277
The same point is repeated in 'Aufzeichnungen fur Ludwig Darmstaedter', which Frege wrote in 1919 (NS, pp. 274 - 275). A particular
judgement serves to express a connection between two concepts. Because
it always involves two concepts, it is, Frege notes, difficult to turn a
sentence like 'There are men' into the form of a particular judgement.
The problem can be solved, however, if the concept in question is
definable by means of two concepts, as the concept man is definable by
means of the concepts rational and living being.
Frege chooses the concept being identical with itse/f(sich selbst gleich
sein) as the most general concept of the hierarchy of concepts. It is a concept superordinate to all concepts and having no content since its extension is unlimited. Frege remarks that in natural language it is the copula
that is purported to express the most general concept without content.
He argues:
This makes it possible to say: inen = men that have being; 'There are men' is the same as
'Some men are' or 'Something that has being is a man'. Thus the real content of what is
predicated does not lie in 'has being' but in the form of a particular judgement. (NS, p.
71; Long and White, p. 64.)
Thus, Frege's view is that if we talk about an object and then state that
it exists, that statement does not add anything to what we have said so
far. We presuppose the existence of objects; we do not say that they exist. For Frege, every predication carries with itself the claim for existence.
None the less, Frege must admit that we talk about objects, for example,
fictional entities, which do not exist. In 'Uber Sinn und Bedeutung' he
states that a name may have a Sinn, although it lacks a Bedeutung and,
accordingly, a sentence may express a thought, although it lacks a truthvalue. In Frege's view, a sentence is deprived of a truth-value if it contains a name which has no bearer (KS, p. 148).
Even if Frege takes being, or existence, to be included in every predication, he regards it as a proper concept in itself, since he paraphrases it
278
LEILA HAAPARANTA
According to Frege's clue, the sentences "A is" and "A exists" can be
paraphrased as the metalinguistic sentence "the name 'A' has a
reference". However, if Frege is consistent in his view of language as the
universal medium, the limits of which we cannot exceed, he does not accept such statements as talk about the expressions of our language.
In 'Dialog mit Ptinjer tiber Existenz' Frege concentrates on discussing
the concept of being as an unanalysed univocal concept and insists on the
Kantian claim that being is not a real predicate. Only at the end of the
dialogue does he come to the idea which he later emphasizes and
develops, i.e., taking existence to be a property of a concept:
The existence expressed by 'there is' cannot be a characteristic mark of a concept whose
property it is, just because it is a property of it. In the sentence 'There are men' we seem
to be speaking of individuals that fall under the concept 'man', whereas it is only the concept 'man' we are talking about. (NS, p . 75; Long and White, p. 67.)
Frege maintains that we can use existence as a proper second-order concept which expresses that a concept is instantiated, that an object falls
under a concept. Therefore, it is, after all, possible for Frege to say
meaningfully that an object exists in the sense of saying that a certain
bundle of properties is instantiated. Thus, existence is here something
that is asserted of the bundle of properties.
Now we are able to see more clearly the ties between Kant's and Frege's
positions. In Der einzig mogliche Beweisgrund zu einer Demonstration
des Daseins Gottes Kant argues:
Consider any arbitrary subject, e.g., Julius Caesar. Combine all his conceivable predicates,
including those of time and space; then you will soon realize that it is possible for him to
exist or not to exist with all these determinations .... Existence, which occurs in everyday
speech as a predicate, is not a predicate of a thing itself but rather a predicate of the thought
which we have of the thing (Kant's Gesammelte Schriften, Band 1/, p. 72.)
Frege's view of existence implies that we cannot talk about objects and
their existence directly, i.e., independently of the properties of objects.
279
At the same time, that is also one of Frege's basic semantical tenets. Between the name and the Bedeutung there is always the Sinn. An object
is for us always an object that falls under some concept.
The above extract from Kant's text shows how close Frege's view of
existence as a property of a concept actually comes to Kant's statements.
For Kant, existence is a predicate of a thought concerning an object. The
same holds true of Frege since, in his view, objects exist for us only as
subsumed under concepts, which means that we can know an object only
by knowing some thought or some thoughts concerning the object. Thus,
it seems as if Kant's views on existence influenced Frege's ideas at least
in two ways: first, Kant argues that being is not a real predicate, which
is the starting-point of Frege's discussion concerning existence, and
secondly, the idea that being is a property of a thought is developed by
Frege in the view that existence is a property of a concept.
5. IDENTITY AND PREDICATION
280
LEILA HAAPARANTA
In the Grundlagen Frege not only suggests, but gives an explicit formulation of, Leibniz's law concerning the substitutivity of identicals (GLA,
65).15 Again, in 'Uber Sinn und Bedeutung' he repeats the same law
(KS, p. 150). In the Grundgesetze (Vol. I) he establishes his view of identity as the substitutivity of identicals, not only by referring to the
substitutivity of the names of identical objects, but by stating that if two
objects are identical, they fall under the same concepts, i.e., they have
exactly the same properties (GGA I, 20). Nevertheless, Frege does not
regard his account as a definition of identity, as he unquestionably indicates in 'Rezension von: E. G. Husserl, Philosophie der Arithmetik I'
(KS, p. 184). For Frege, identity is indefinable but Leibniz has offered
a valuable tool for understanding the identity relation. In this paper, I
shall not discuss the similarities and dissimilarities between Leibniz and
Frege in more detail.
We can, however, conclude that Frege gives four different accounts
of identity. First, he regards an identity statement as a rule for
substitutivity of names in different contexts. Secondly, he takes it to be
a metalinguistic statement concerning the number of senses and
references of two names. Thirdly, he considers identity to be a relation
between two objects. Fourthly, he takes it to be a relation of an object
to itself, which occurs in sentences like "a = a". As we saw above, the
fourth meaning is related to his view of existence.
As far as objects are concerned, the contexts where the names occur
express properties of the objects named. If two objects are said to be
identical, they are claimed to have the same properties. Considered this
way, identity can be regarded as a border-line case of predication. When
paraphrased in the manner Frege suggests, i.e., by means of the expression no other than, it predicates of the object under discussion all the
properties that it in fact has ('Uber Begriff und Gegenstand), (KS, p.
169). Nevertheless, I shall argue that Frege does not reduce objects to
their properties. 16 Nor does he reduce the identity of objects to the
sameness of their properties. What I seek to show is that those very tenets
281
Also, in the article 'Uber die Begriffsschrift des Herrn Peano und meine
eigene', Frege notes that subordination is alien to predication (KS, p.
244). What is reflected in the distinction between subordination and
predication is precisely the principle of always keeping in mind the difference between objects and concepts. Frege would also reject analyses
like that given by Aristotle in Analytica Posteriora, in which man is said
to be identical with animal, since man is identical with a species of animal
(An. Post. 83 a 23 - 35).
The innovation of Fregean logic that I am here aiming at is, however,
related to the epistemological import of the distinction between objects
and concepts. It amounts to an attack on the general philosophical doctrines implicit in the received logical analysis of Frege's day. Frege's view
is that we cannot reach the essence of an object by means of our concepts
and our cognitive capacities. To say, for instance, that Plato is a man,
282
LEILA HAAPARANTA
When the reference to be assigned is logically simple, we cannot, according to Frege, give a proper definition but we must comply with
elucidatory propositions, by means of which we give hints concerning the
wanted reference (BW, p. 63).
It is obvious that the logically simple elements of Frege's world can be
concepts and, more generally, functions. That is implicitly expressed by
Frege himself in 'Ober die Grundlagen der Geometrie I - III' (KS, pp.
288 - 289). Concepts can be decomposed into more primitive concepts,
which are their characteristics. Still, we are not able to decompose for
ever. The most primitive units that we reach are the Bedeutungen of the
concept-words that can no more be defined.
Eike-Henner W. Kluge argues that the concept of logisch einfach is important in Frege's ontology and that logisch einfach is, for Frege, always
a functjon and never an object (Schirn II, 1976, pp: 60 - 65). Kluge's
argument is as follows: For Frege, a thing without properties is an Unding. Now if an object were logisch einfach and had only one property,
it ought to coincide with that propt;rty. For if it did not, it would be composed of this property and something that has this property, which is in-
283
consistent with the assumption that the thing is logisch einfach. Consequently, if it coincides with the property, it is a property itself and, as
a property, incomplete, which is inconsistent with its nature as an object.
Since there are no other possibilities, it follows that objects cannot be
logisch einfach.
Kluge's argument is based to a considerable extent on the fact that
Frege regards an object having no properties as an Unding. It is true
Frege maintains that objects without properties are impossible or even
absurd, but his wordings do not warrant the conclusion that the concept
of Unding applies to ontological considerations. Frege formulates his
claim ironically as follows:
First, things are regarded as similar, so similar to each other, that they can no more be
distinguished from each other; that is, all properties by which they are distinguished would
be demolished ... .Is there still something left of those things? Certainly! There are still
left the natureless things I die naturlosen Dinge I .... Nonetheless, it will be difficult to find
the natureless pieces of wood, or, rather, not-pieces of wood, in the set of the natureless
things. But even the gre~test difficulties are overcome by good will. The best means is
always this: they are totally ignored. ('Uber die Zahlen des Herrn Schubert', KS, pp.
247 - 248.)
Frege continues:
I have been led astray because I know the names of fixed stars but I do not know the names
of natureless things. (KS, pp. 250 - 251.)
In these passages we only encounter a version of Frege's doctrine according to which it is via a'sense and only via a sense that a proper name
is related to an object (NS, p. 135). Frege does not think that it is possible
to talk meaningfully of an object without referring to the properties of
the object. This is precisely the view that makes Frege an antihaecceitist
(see, e.g., Kaplan, 1975, p. 725).20 For Frege, it is impossible to know
objects in themselves or to talk about them. But if we wish to understand
the structure of Frege's ontology, objects can be taken to exist without
having any properties to carry. They are not known to us as simple,
however, and strictly speaking we cannot even talk about their existing
or not existing without properties. That is exactly what I emphasized in
the preceding section when discussing existence. According to hege, we
can say what an object is like, but we cannot say what it is.
But it does not suffice to recognize Frege's concept of Unding as having an epistemological force, for his concept of logisch einfach is not applicable to ontological considerations, either. For Frege, a reference is
logically simple if it cannot have a proper definition. But we can con-
284
LEILA HAAPARANTA
sistently assume that a reference has a complex structure even if we cannot define the reference by means of our concepts. 21 Nothing can be
logically simple from an ontological point of view, for if one calls a
reference logically simple, one already considers it within the limits of
human knowledge.
Frege's epistemological view comes up in Die Grund/agen der
Arithmetik as follows:
It is this way that I understand objective to mean what is independent of our sensation,
intuition and imagination, and of all construction of mental pictures out of memories of
earlier sensations, but not what is independent of the reason - for what are things independent of the reason? To answer that would be as much as to judge withoUI judging, or to
wash the fur without wetting it. (GLA, 26.)
285
286
LEILA HAAPARANTA
des Herrn Peano und meine eigene', KS, p. 227. See also' Anmerkungen Freges zu: Philip
E. B. Jourdain, The Development of the Theories of Mathematical Logic and the Principles of Mathematics', KS, 341. Frege uses the expression lingua characterica instead of
lingua characteristica. For this, see G. Patzig's footnote 8 in Gottlob Frege, Logische
Untersuchungen, ed. by G. Patzig, p. 10.
14 See Schirn (1976).
IS Austin's translation of this passage is the following: "Now Leibniz's definition is as
follows: 'Things are the same as each other, of which one can be substituted for the other
without loss of truth'. This I propose to adopt as my own definition (Erkliirung) of identity .... Now, it is actually the case that in universal substitutability all the laws of identity
are contained." I think it is misleading to use the word definition as the translation of
Erkliirung, because Frege argues elsewhere that the concept of identity is indefinable.
10 This was originally suggested to me by Prof. Jaakko Hintikka.
17 For the sharp distinction between objects and concepts, see also 'Ober die Begriffsschrift des Herrn Peano und meine eigene', KS, p. 233, and 'Ober die Grundlagen der
Geometrie II', KS, p. 270.
18 I have changed P. Long's and R. White's translation to the effect that I have adopted
the word reference as the translation of the German word Bedeutung, while Long and
White use the word meaning in this connection.
19 A similar idea concerning Frege's central role in the history of logic has been put forward by Ignacio Angelelli. See Ange1elli (1967), pp. 253 - 254.
Aristotle's view is discussed by Jaakko Hintikka in the article 'The Varieties of Being
in Aristotle' (this volume).
20 Different kinds of haecceitism have been discussed by R. M. Adams (1979). He defines
haecceitas as the property of beillg identical with a certain particular individual (Adams,
1979, p. 6). For the details of Frege's doctrine, see Haaparanta (1985).
21 For this point I am indebted to Prof. Jaakko Hintikka.
22 Cf. Hintikka (1981 b) and Hintikka, 'The Paradox of Transcendental Knowledge'. See
also Haaparanta (1985).
BIBLIOGRAPHY
Frege, G., Begriffsschrift, eine der arithmetischen nachgebildete Formelsprache des reinen
Denkens, Verlag von L. Nebert, Halle a. S., 1879; repro in Frege (1964), pp. 1-88.
(Referred to as BS.)
Frege, G., 'Ober den Zweck der Begriffsschrift', in Sitzungsberichte der Jenaischen
Gesellschaft fiir Medizin und Naturwissenschaft fiir das Jahr 1882, Verlag von G.
Fischer, Jena, 1883, pp. 1-10; repr. in Frege (1964), pp. 97 -106.
Frege, G., Die Grundlagen der Arithmetik: eine logisch mathematische Untersuchung iiber
den Begriff der Zahl, Verlag von W. Koebner, Breslau, 1884; repr. and transl. by J. L.
Austin in The Foundations of Arithmetic/Die Grundlagen der Arithmetik,. Basil
Blackwell, Oxford. 1968. (Referred to as GLA.)
Frege, G., Funktion und Begri'f. Vortrag. geliaIten in der Sitzung vom 9. Januar 1891 der
Jenaischen Gesellschaft fiir Medizin und Naturwissenschaft, H. Pohle, Jena, 1891;
repro in KS. pp. 125 - 142.
Frege, G., 'Ober Sinn und Bedeutung', Zeitschrift fiir Philosophie und philosophische
Kritik 100 (1892), 25 - 50; repro in KS, pp. 143 - 162.
287
288
LEILA HAAPARANTA
Frege, G., Conceptual Notation and Related Articles, transl. and ed. by T. W. Bynum,
Clarendon Press, Oxford, 1972.
Frege, G., Wissenschaftliche Briefwechsel, ed. by G. Gabriel, H. Hermes, F. Kambartel,
C. Thiel, and A. Veraart, Felix Meiner Verlag, Hamburg, 1976. (Referred to as BW.)
Frege, G., Logical Investigations, ed. by P. T. Geach, Basil Blackwell, Oxford, 1977.
Frege, G., Posthumous Writings, transl. by P. Long and R. White, Basil Blackwell, Oxford, 1979.
Frege, G., Philosophical and Mathematical Correspondence, abridged for the English edition by B. McGuinness and transl. by H. Kaal, Basil Blackwell, Oxford, 1980.
Adams, R. M., 'Primitive Thisness and Primitive Identity', The Journal of Philosophy 76
(1979), 5 - 26.
Angelelli, I., Studies on Gottlob Frege and Traditional Philosophy, D. Reidel, Dordrecht,
1967.
Angelelli, I., 'Friends and Opponents of the Substitutivity of Identicals in the History of
Logic', in Schirn (1976), Band II, pp. 141-166.
Aristotle, Analytica Posteriora, in The Works of Aristotle, Vol. I, ed. by W. D. Ross, Oxford University Press, London, 1928.
Bell, D., Frege's Theory of Judgement, Clarendon Press, Oxford, 1979.
Dummett, M., Frege: Philosophy of Language, 2nd ed., Duckworth, London, 1981. (First
published in 1973.)
Dummett, M., The Interpretation of Frege's Philosophy, Duckworth, London, 1981.
Forgie, J. W., 'Frege's Objection to the Ontological Argument', NoUs 6 (1972),251- 265.
Grossmann, R., 'Structures, Functions and Forms', in Schirn (1976), Band II, pp. 11- 32.
Haaparanta, L., Frege's Doctrine of Being (Acta Philosophica Fennica, vol. 39), Helsinki,
1985.
van Heijenoort, J., 'Logic as Calculus and Logic as Language', Synthese 17 (1967),
324-330.
Hintikka, J., '''Is'', Semantical Games, and Semantical Relativity', Journal of
Philosophical Logic 8 (1979), 433-468.
Hintikka, J., 'Kant on Existence, Predication, and the Ontological Argument', Dialectica
3S (1981), 127-146. (Referred to as Hintikka, 1981a; in this volume.)
Hintikka, J., 'Wittgenstein's Semantical Kantianism', in E. Morscher and R. Stranzinger
(eds.), Ethics, Proceedings of the Fifth International Wittgenstein Symposium, HolderPichler-Tempsky, Vienna, 1981, pp. 375-390. (Referred to as Hintikka, 198Ib.)
Hintikka, J., 'Semantical Games, the Alleged Ambiguity of "Is", and Aristotelian
Categories', Synthese 54 (1983), 443 - 468.
Hintikka, J., 'The Varieties of Being in Aristotle', this volume, p. 81 -114.
Hintikka, J., 'The Paradox of Transcendental Knowledge', forthcoming in the Proceedings of the 1981 Cambridge Conference on Transcendental Argumentation.
Hugly, P., 'Ineffability in Frege's Logic', Philosophical Studies 24 (1973), 227 - 244.
lshiguro, H., Leibniz's Philosophy of Logic and Language, Duckworth, London, 1972.
Kahn, C., The Verb 'Be' in Ancient Greek, D. Reidel, Dordrecht, 1973.
Kahn, C., 'On the Theory of the Verb "To Be" " in M. K. Munitz (ed.), Logic and Ontology, New York University Press, New York, 1973, pp. 1 - 20.
Kant, I., Der einzig mogliche Beweisgrund zu einer Demonstration des Daseins Gottes, in
Kant's Gesammelte Schriften, Band II, Vorkritische Schriften II, 1757 - 1777, G.
Reimer, Berlin, 1905, pp. 63 - 163.
289
Kant, 1., Kritik der reinen Vernunft, 1781, 2nd ed. 1787, in Kant's Gesammelte Schriften,
Band III, O. Reimer, Berlin, 1904; trans\. by N. Kemp Smith, The Macmillan Press,
London and Basingstoke, 1929.
Kaplan, D., 'How t() Russell a Frege-Church', The Journal of Philosophy 72 (I975),
716-729.
Kauppi, R., 'Substitutivity salva veritate in Leibniz and in Modern Logic', Ratio 10 (1968),
141-149.
Klemke, E. D. (ed.), Essays on Frege, University of Illinois Press, Urbana, Chicago, and
London, 1968.
Kluge, E.-H. W., 'Freges Begriff des Logischeinfachen', in Schirn (I976), Band II, pp.
51-66.
Kneale, W. and Kneale, M., The Development of Logic, Clarendon Press, Oxford, 1962.
Mates, B., 'Identity and Predication in Plato', Phronesis 24 (I979), 211-229.
Schirn, M. (ed.), Studien zu FregelStudies on Frege /- III, Frommann-Holzboog, Stuttgart - Bad Cannstatt, 1976.
Sluga, H. D., 'Frege as a Rationalist', in Schirn (1976), Band I, pp. 27-47.
Sluga, H. D., Gottlob Frege, Routledge and Kegan Paul, London, Boston and Henley,
1980.
Stuhlmann-Laeisz, R., 'Freges Auseinandersetzung mit der Auffassung von "Existenz"
als ein Priidikat der ersten Stufe und Kants Argumentation gegen den ontologischen
Oot~esbeweis', in C. Thiel (ed.), Frege und die moderne Grundlagenforschung, Anton
Hain, Meisenheim am Olan, 1975, pp. 119-133.
Woods, M. J., 'Substance and Essence in Aristotle', Proceedings of the Aristotelian Society75 (I974-75)., 167-180.
Dept. oj Philosophy,
University oj Helsinki,
Unioninkatu 40B,
SF-OOI70 Helsinki 17, Finland
INDEX OF NAMES
Bach, E. 77
Bacon, Robert 123, 131, 141
Bacon, Roger 129, 141, 142
Bailey, C. 142
Balme, D. M. 86, 112
Bambrough, R. 25, 74, 113
Barnes, Jonathan 25, 26, 265
Barth, T. A. 216
Beck, A. L. 219,241
Beck, C. H. 75
Beck, L. J. 247
Bekker, Immanuel 116, 141
Bell, David 274, 288
Benardete, Seth 82, 112
Bennett, Jonathan 265
Benveniste, Emile 23, 24, 100, 103, 112
Black, M. 77
Blass, F. 74
Bluck, Richard S. 81, 112
Boehner, P. 142,218,219,222
Boethius, de Dacia 135
Boethius, A. Manlius S. 115, 117-121,
127,132-134,141,142,145,160-161,
176-178
Bonansea, B. M. 216,219,221,222,243,
248
Bonaventura 224
Bonitz, Hermann 77,79,100-101,112
Borgnet, A. 126, 141
Boutroux, Emile 242
Braakhuis, H. A. G. 123,128,141,142,
180
Brentano, F. 79
Bresnan, Joan 109-112
Brito, Radulphus 132
Brody, B. A. 77
Brown, S. F. 142,216,218, 219, 222
Brugmann, K. 23, 45, 47
Brunschwig, J. 79
Buchanan, E. 79
Buridan, John xv, 135, 136, 138-142,
211,218,.220
291
292
INDEX OF NAMES
INDEX OF NAMES
Guthrie, W. K. C. 81
Hager, F. P. 79
Haldane, E. S. 241-247
Hamlyn, D. W. 195,198,199
Happ, H. 213,219
Haring, N. M. 172, 179
Harlfinger, D. 121, 142
Harman, G. 76
Harms, R. T. 77
Heidegger, M. 63
Helias, Petrus 177
Henrich, Dieter 265
Hermann, G. 23
Hermes, H. 288
Herodotus 24, 25
Hertz, M. 179
Hick, John 265
Hilbert, D. 271
Hintikka, Jaakko x, xv, xvi, 23, 26, 78,
82,87,90, 101-102, 106, 112, 113,
181, 197, 199,215,218,265-267,269,
276, 285, 286, 288
Hoeres, Walter 212,217-219
Homer x, 11-12,66-67,85, 115-116,
128, 157, 165,254
Honnefelder, L. 216,217,219,220
Hooker, M. 241,247,248
Howell, Robert 266, 267
Hubien, H. 218,220
Hugly, P. 288
Hume, David 3, 247
Hunt, R. W. 177
Husserl, E. G. 280, 287
lamblichus 118
Irwin, T. 74, 78
Ishiguro, H. 289
Jacobi, K. 172, 174, 179
Jacobs, W. 74,78,79
Jolivet, 1. 172, 177, 179
Jourdain, Philip E. B. :86
Judy, A. G. 142
Kaal, H. 288
Kahn, Charles ix, x, 47, 74, 78, 79, 81,
96, 100, 103, 113, 196, 199, 269, 289
Kambartel, F. 288
293
Kaminski, S. 216,220
Kamiah, W. 75
Kamp, 1. A. W. 75
Kane, W. H. 214,220
Kannisto, Heikki 261-262, 267
Kant, l. xv, xvi, 249-267, 272, 275, 279,
284-285, 288, 289
Kaplan, David 255-256, 283, 289
Kaulbach, F. 288
Kauppi, R. 289
Keil, H. 179
Kelley, F. E. 142
Kemp, Smith 267, 289
Kennan, E. L. 75
Kenny, A. 179,180,199,214,220,241,
242,247
Kilwardby, Robert 124-128, 130-\31,
142
Kirwan, Christopher 74, 75, 80, 84, II?
Klowski, J. 24
Kluge, Eike-Henner W. 282, 288, 289
Kneale, M. 289
Kneale, W. 289
Kneepkens, C. H. 141, 180
Knuuttila, Simo 113,214,217-220,241,
266
Koehler, K. F. 1\3
Koyre, Alexandre 241, 243, 247
Krakow, Bjag 138
Kranz, W. 142
Krapiec, M. A. 216,220
Kretzmann, N. 170-172,179,180,214,
220-222
Kripke, S. 76
Kuhner, R. 23,45,47, 74
Kurdzialec, M. 216, 220
Laeisz, R. 267, 289
Land, J. P. N. 246, 247
Langston, D. C. 220
Larkin, T. 79,80
Laupp, H. 79
Lee, E. N. 77
Leibniz, G. W. 36, 70, 106, 112,
212-213, 218-221, 254, 261-263,
276, 280, 286, 289
LePore, Ernest 285
Leroux, Ernest 247
294
INDEX OF NAMES
Lewis, Frank 76
Lewry, O. 122-124,127-128,142
Lloyd, G. E. R. 242, 247
Lockwood, M. 76, 78
Loemker, L. E. 218, 220
Lombard, Peter 189
Long, P. 286, 288
Louis, R. 179
Loux, Michael J. 113,213,218,220
Lovejoy, A. O. 261
Lucretius 117, 120, 142
Luschei, E. C. 45, 47
Lyons, John 23, 24, 77
Lyttkens, H. 215,220
Magentinus, Leo 116
Mahoney, E. P. 214,220
Maier, Heinrich 79,81,83,100,107,113
Maieru, A. 178, 179
Malcolm, John 196, 199, 242, 245
Malcolm, Norman 241, 247
Mandonnet, P. 197,198,222
Manheim, R. 77
Mansion, S. 79
Maritain, J. 215
Marshal, David J. 200
Marshall, W. 286
Mates, Benson x, 23, 78, 82, 113, 220,
269,289
Matthen, Mohan 25, 27
Matthew, M. 217,221
Mazzarella, P. 143
McCabe, Herbert 196, 199
McGill, Arthur C. 265
McGuinness, B. 288
McInerny, R. M. 214,215, 220
Mei, Tsu-Lin 77
Meillet, A. 23
Meiser, C. 117,141,178
Melissus 11,13-14,17-18,25
Mersenne, M. 243, 246
Mesland, D. 246
Mill, John Stuart 4, 23, 60, 269
Miller, Barry 195, 196, 199
Minio-Paluello, L. 178, 179
Mohr, J . C. B. 265
Montague, Richard 75, 255
Moody, E. A. 178,179
Moos, M. F. 198
INDEX OF NAMES
Rashdall, H. 141
Reimer, G. 289
Reina, M. E. 141
Remes, Unto 266
Richard of Conington 219
Roberts, L. N. 140, 142
Rorty, R. 77,78
Ross, G. T. 247
Ross, J. F. 213,215,221,241-246
Ross, W. D. 25,44,47,75-77,79-81,
85,101,113,178,219,242,288
Russell, Bertrand 4, 23, 33-34, 45, 47,
106, 289
Ryan, J. K. 216,219,221,222,243,248
Ryle, Gilbert 81, 106, 113
Saarinen, Esa 102, 113, 218, 220, 256,
266
Schaffer, Jerome 266
Schepers, H. 218, 221
Schim, M. 76, 286, 288, 289
Schmaus, M. 216, 221
Schmidt, C. C. E. 45, 264
Schmidt, F. 47
Schmidt, R. W. 221
Schwyzer, E. 23,45, 47
Scott, T. K. 141
Sextus Empiricus 116
Shircel, C. L. 216, 222
Shorey, P. 28, 221
Siger 214, 220
Simplicius 118-120
Sluga, H. D. 289
Smyth, H. W. 45,47
Soames, S. 77
Socrates xiii, 7, 9,10,14,16,27,29-33,
38,40,43-46,50-56,58-61,64-65,
67, 70-72, 78-80, 83-84, 86, 94, 97,
106,123,125,156-157,159-166,177,
182-183, 188, 196, 197,200,202
Sommers, F. 241,247
Sorabji, Richard 25
Spiazzi, R. M. 80, 179,222
Spinoza, B. 246, 247
Stead, C. 80
Stenius, Erik 241
Stephanus 118
Stranzinger, R. 288
Strawson, P. F. 196, 199
295
296
INDEX OF NAMES
Williams, B. 247
Williams, C. J. F. 23,78,241,245
Wilson, M. D. 241,242,247,248
Wittgenstein, L. 45, 47, 247, 288
Wolff, Christian 23
Wolff, Robert Paul 266
Wolter, A. B. 216, 217, 218, 222, 243,
248
Wolterstorff, N. 74
Woods, M. J. 94. 114, 289
Zdybicka, Z. J. 216,220
Zeno 25
Zimmermann, Albert 178, 180, 198,200
INDEX OF SUBJECTS
barbara 87-88
Bedeutung 272-273, 277, 279-280, 282
Begriffe erster und zweiter Stufe 270, 285
being, philosophical concept of: Abelard
149-152, 155-162; Aristotle 17,
49-50, 55-59, 64-73, 81-96,
100-111,200-201; Boethius of Dacia
132-134; Descartes 224-234; Frege
269-285; John Buridan 138-139;
John Duns Scot us 207-213; Kant
249-265; Leibniz 212-213; Melissus
13-14,
17-18;
modists
131;
Parmenides 14-18, 25-27; Plato
16-21, 34-39, 42-44, 51-55; Protagoras 12-14; Robert Kilwardby
124-131; Roger Bacon 129-131;
de re 211-222
deep structure 155, 160, 163
definition 18-20, 76, 84, 102, 120, 129,
135, 139, 193,204,282
dictio 127, 158,205
dictum 159-162
297
298
INDEX OF SUBJECTS
INDEX OF SUBJECTS
'is': accidental 56, 64, 66, 68-71; actuality sense of 181-187, 191; copulative
(see copula), ellipsis hypothesis of
(86-92); of existence ix, 1-4,6,9-17,
27,53,59,61,65-67,71,73,81-82,
84-91,93,96,99,101,105-107, 124,
183-188,249,260,265,269; of generic
implication ix, 81-82, 249, 270; of
identity 35-36,38-39,63-64,81-84,
88-89,93-94,99,101,183,186,188,
249, 258, 269, 281, 284; of location
3-4,7-11, 15, 17; of predication ix,
1-10, 17, 20-22, 26-27, 35-36,
38-39,53,59,61,63-64,81-84,86,
89,93-94,99,124,152,183-188,195,
249, 265, 270, 281 (see also copula);
there-is sense of 182-187; veridical
3-4, 8-9, 12-13, 16-17, 20-22,
26- 27, 184. (See also being, einai, esse,
existence.)
299
300
INDEX OF SUBJECTS
suppositio 134-137,162,167,188,197
surface structure 155
syllogism 87-88, 90-92
syncategorematic words 148. 156, 162