NECESSITY

or
CONTINGENCY
 CSLI
Lecture Notes
No. 56

NECESSITY
or
CONTINGENCY
THE MASTER ARGUMENT

Jules Vuillemin

Publications
CENTER FOR THE STUDY OF
LANGUAGE AND INFORMATION
STANFORD, CALIFORNIA
 Copyright © 1996
Center for the Study of Language and Information
Leland Stanford Junior University
Printed in the United States
99 98 97 96 96 5 4 3 H

Library of Congress Cataloging-in-Publication Data

Vuillemin, Jules.
[Ne'cessite' ou contingence. English]
Necessity or contingency : the master argument / Jules Vuillemin.
p. cm. — (CSLI lecture notes ; no. 56)
Includes bibliographical references and index.

ISBN 1-881526-85-2 (pbk. : alk. paper)

ISBN 1-881526-86-0 (hardback : alk. paper)

i. Necessity (Philosophy) — History. 2. Contingency (Philosophy) — History.

3. Diodorus Cronus, 4th cent. B.C. 4. Philosophy, Ancient. I. Title. II. Series.

1996
— dc2O 95-50100
CIP

Cover design by Tony Gee

CSLI was founded early in 1983 by researchers from Stanford University, SRI
International, and Xerox PARC to further research and development of integrated
:heories of language, information, and computation. CSLI headquarters and CSLI
Publications are located on the campus of Stanford University.

CSLI Lecture Notes report new developments in the study of language,

information, and computation. In addition to lecture notes, the series includes
monographs, working papers, and conference proceedings. Our aim is to make
new results, ideas, and approaches available as quickly as possible.
Contents

Introduction xi
Acknowledgements xiii

Part I The Master Argument 1

1 The Master Argument. On the Shortcomings of Some
Past Interpretations. Conditions to be Fulfilled by
any Acceptable Interpretation. 3
1.1 The text of Epictetus. 3
1.2 Zeller's interpretation. Confusion of the logical and the
chronological. 4
1.3 Ambiguity in the first premise: Necessity and
irrevocability. Signification of the first premise. 7
1.4 Prior's interpretation: It supposes two supplementary
premises, one of which is explicitly rejected by Aristotle;
it supposes the first premise ambiguous. 8
2 Reconstruction of the Master Argument. 15
2.1 An Aristotelian paradigm: De Caelo, I, 28366-17; its
context. 15
2.2 The principle of the conservation of modal status. 16
2.3 The principle of the possible realization of the possible
interpreted as a principle of pure modal logic. 18
2.4 The principle of possible realization of the possible as
principle of synchronic contraction of the possible and
diachronic expansion of the necessary. 21
2.5 The principle of conditional necessity. 24
2.6 The irrevocability of the past or the principle of the
impossibility of realizing the possible in the past. 26
vi / CONTENTS

2.7 The principle of the subsistence of a possible that is not

to be realized. 31
2.8 Reconstruction of the De Caelo demonstration. 33
2.9 Reconstruction of the Master Argument. 35
2.10 Sketch of a formal reconstruction of the Aristotelian
reasoning at De Caelo, I, 283*6-17. 36
2.11 Sketch of a formal reconstruction of the Master
argument. 38

Part II Systems of Necessity: The Megarians

and the Stoics 41
3 A System of Logical Fatalism: Diodorus
Cronus. 43
3.1 Diodorus' Solution. 43
3.2 Two possible interpretations as regards the object of the
Diodorean modalities: nominalism and realism. 46
3.3 The meaning of Diodorean implication. 52
3.4 Diodorean nominalism. 56
3.5 Diodorus' necessitarianism. 62
4 Eternal Return and Cyclical Time: Cleanthes'
Solution. 69
4.1 First conjecture. Necessity of the past secundum vocem
and secundum rem: Ockham's conception on Prior's
hypothetical reconstruction. Modality de dicto and
modality de re. 70
4.2 Inadequacy of Ockham's solution. Incrimination of the
principle of conditional necessity: John Duns
Scotus. 79
4.3 Cleanthes again and the second conjecture: the
conditional character of the necessity of the past
according to Cleanthes; the interpretation of
Leibniz. 92
4.4 Third conjecture: cyclical time and the numerical
conception of the identity of beings in eternal
return. 97
5 Freedom as an Element of Fate: Chrysippus. 105
5.1 Were Chrysippus' doubts about the thesis of pure modal
logic according to which from the possible the impossible
does not follow, they would be about its negative form,
not about its positive form. 106
 CONTENTS / vii

5.2 Chrysippus' doubt about the interdefinability of the

modalities. From the non-possibility of an event's
occurrence it cannot be concluded that its opposite is
necessary. 110
5.3 The non-standard modal system according to
Chrysippus. 115
5.4 A system related to Prior's system Q; the double logical
square of Chrysippean modalities and the double
temporal index in the Master Argument's second
premise. 120
5.5 A Philonian doubt about the second premise? 126

Part III Systems of Contingency: The Lyceum,

The Garden, The Academy 131
6 Towards Rehabilitating Opinion as Probable
Knowledge of Contingent Things. Aristotle. 133
6.1 De Interpretatione, Chapter IX. 133
6.2 Outline of the passage: Introduction (18a28-34): The
problem raised. 137
6.3 Validity of the principle of non-contradiction and the law
of excluded middle (18a38 and 18fc17-25). 139
6.4 Critical examination of the Megarian theory (18a34-18617
and 18ft25-19a22). 141
6.5 Aristotle's solution (19a22-1964); conditional necessity
and exceptions to the principle of bivalence. 144
6.6 Aristotle's general conception confirms the De
Interpretatione solution; the difference between Aristotle
and Diodorus. 149
6.7 First interpretative hypothesis: More than two
truth-values. 154
6.8 Second interpretative hypothesis: propositions without a
determinate truth-value. 157
6.9 Third interpretative hypothesis: probability. 161
7 Epicurus and Intuitionism. 169
7.1 First logical interpretation of the Epicurean denial of the
excluded middle: the three-valued logic of Lukasiewicz;
reasons for rejecting this solution. 171
7.2 Second logical interpretation of the Epicurean negation of
the excluded middle: The Intuitionist System. 173
viii / CONTENTS

7.3 Are the Epicurean 'criteria' compatible with

intuitionism? 174
7.4 Consequences of the Epicurean criteria: Plurality of
hypotheses and rejection of the excluded middle. 182
7.5 Epicureanism and the Master Argument. 185
7.6 Other intuitionist conceptions of reality: Descartes and
Kant. 187
8 Carneades and the Skeptical Nominalism of the
Modalities. 207
8.1 What is the relation between the principle of the
excluded middle and the principle of causality (De Fato,
X-XII)? 207
8.2 Aristotle's dogmatic definition of truth called into
question (De Fato, XIV). 210
8.3 Carneades and the Master Argument (De Fato, IX). 212
8.4 From Carneades to the logics of "fictive" names:
Buridan's ampliation. 215
8.5 Carneades does not abandon the principle of conditional
necessity; he simply deprives it of the ontological
involvement conferred upon it by the dogmatic
interpretation of truth. 219
9 Platonism and Conditional Necessity. 225
9.1 Platonism and the principle of conditional necessity. 225
9.2 Consequences of the connection between conditional
necessity and the substantiality of the sensible for
modality, causality and freedom. 230
9.3 The consequences of abandoning the principle of
conditional necessity and the substantiality of the
sensible world for the Platonic and Platonistic theories of
modality, causality and freedom. The same abandonment
entails similar consequences for Duns Scotus. 233
10 Epilogue 243
10.1 The impasse of natural language. 244
10.2 A probabilistic reconstruction of the Master Argument:
Diodorus' solution. 246
10.3 The special status of premise (C): Chrysippus' solution
and the 'Unique Law of Chance'. 251
10.4 Contingency and ignorance: The statistical mix. 257
10.5 Contingency and nature: The state of
superposition. 261
 CONTENTS / ix
Bibliography 267
Index of quotations of ancient and mediaeval texts 279
Index of Proper Names 285
Introduction

The Master Argument, recorded by Epictetus, indicates that Diodorus

had deduced a contradiction from the conjoint assertion of three propo-
sitions. Epictetus adds that three solutions of the aporia had been
obtained by denying one or another of the propositions advanced while
maintaining the other two.
The argument, which has to do with necessity and contingency and
therefore with freedom, has attracted the attention of logicians above
all. In any case there have been many attempts at reconstructing it
in logical terms, without excessive worry about historical plausibility
and with the foregone conclusion that it was sophistic since it directly
imperiled our common sense notion of freedom.
On both of these counts I have taken exception to recent tradition.
The success of the argument with the Ancients, and with Ancients who
were no mean logicians, seemed reason for presuming that the Master
Argument is not sophistic and that the contradiction it produces is a
real one. On the other side, I looked for a classical text containing
the propositions stated by Epictetus and which could have furnished
Diodorus with the material for his argument. I believe to have found
such a text in Aristotle's De Caelo.
In order to demonstrate the contradiction in the propositions thus
restored, I had in my turn to translate them into logical terms. It is
unlikely that Diodorus proceeded in such a way. Although the trans-
lation I have proposed tries to remain faithful to its models as they
have been handed down to us, it inevitably gives them a precision they
did not have in themselves. This indulgence in precision amounts to
historical inexactitude but seemed necessary nevertheless as it had to
do with restoring a reduction to the absurd.
There is one distinctive feature of the translation that must be
noticed. The propositions figuring in the Master Argument are in-
xii / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

terpreted in terms of temporal modal logic where both the modalities

and the statements they govern have chronological indices. This means
that the force of the argument comes not from purely logical or modal
considerations, but from our experience of time.
To bring to its complete end the research undertaken here, I would
have had to assign to the Master Argument and to each of the solutions
it is susceptible of, an explicit axiomatic system formalized according
to a set of rules. For want of competence, of stamina, of time, I have
been content to formulate only what was needed to elucidate the ar-
gumentation.
Justification for the first four chapters is to be found in the Epicte-
tus passage. The remaining chapters extend the debate about the
Master Argument to Greek philosophy at large. In this way it is seen
that principles are challenged-even logical ones at that-which are not
mentioned in the Epictetus passage but which must have played their
role in the argument. The reader will judge whether that extension is
legitimate or not. As one well imagines, the debate on the issue was
continued by the philosophers of the Middle Ages and the Moderns.
This was quite generally done in ignorance of the Master Argument
itself; but I have had no hesitation in appealing to them where they
might be apt to help explain or specify the position of the Ancients.
 INTRODUCTION / xiii

Acknowledgements
This book builds on earlier studies (Vuillemin, 1979, 1983) and espe-
cially on Necessite ou contingence, I'aporie de Diodore et les systemes
philosophiques (Les Editions de Minuit-La Fondation Singer Polignac,
Paris, 1984). I considerably shortened this last work and eliminated
the systematic considerations not directly pertinent to the Master
Argument. I corrected it on the occasion of an objection raised by
M.H. Angstl (his letter and my reply to it were published in the All-
gemeine Zeitschrift fur Philosophic, XI 3, 1986, pp. 79-87). Finally I
added the present, unpublished Epilogue. M. Thomas F. Morran has
translated into English this new French version.
My thanks go to Professor J. Moravcsik for his criticism, Bill Gra-
ham, who revised the text, Mrs. L. von Kampen, who transformed a
manuscript into a book, and to Patrick Suppes, my good friend.
Part I

The Master Argument

The Master Argument. On the
Shortcomings of Some Past
Interpretations. Conditions to be
Fulfilled by any Acceptable
Interpretation.

1.1 The text of Epictetus.

Epictetus gives the following account of the argument.
Here, it seems to me, are the points upon which the Master Argu-
ment was posed: there is, for these three propositions, a conflict
between any two of them taken together and the third: 'Every true
proposition about the past is necessary. The impossible does not
logically follow from the possible. What neither is presently true
nor will be so is possible'. Having noticed this conflict, Diodorus
used the plausibility of the first two to prove the following: 'Noth-
ing is possible which is not presently true and is not to be so in the
future'. Another, for the two propositions to keep, will maintain
these two: 'There is a possible which neither is presently true nor
will be so; the impossible does not logically follow from the pos-
sible'; but then it is not exact to say that every true proposition
about the past is necessary; that is what the school of Cleanthes
seems to maintain (SVF I 489) with whom Antipater is generally
in agreement (SVF III Ant. 30). Others (namely Chrysippus, SVF
II 283) admit the two other propositions: There is a possible which
neither is presently true nor will be so; every true proposition about
the past is necessary'; but then the impossible follows logically from
the possible. But there is no way to maintain the three propositions
4 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

at once, since in every case there is a conflict between one and the
other two.1
To this account Epictetus adds an ironic commentary on the virtues
of erudition. To one who boasted of having read Antipater's treatise
on the Master Argument he exclaims: "What more do you have, you,
for having read it? What opinion have you formed on the question?
You might just as well speak to us of Helen, of Priam and of that isle
of Calypso that hasn't existed in the least and never will".2
The irony of Epictetus is aimed neither at the doctrine nor at the
moral consequences that could be drawn from the argument, but only
at the vanity of logical quibbles. We cannot speculate then on any
opposition of principle that Epictetus might have had to the supposedly
morally disastrous consequences of the argument.
Such is the only explicit text on the Master Argument.
Most interpretations and reconstructions of the argument that have
been given can be put into one of three categories. They all agree,
implicitly at least, on denouncing its disastrous moral consequences
and on flushing out some ambiguity or other in the premises, which
they in turn hold responsible for these consequences. In order to save
human freedom called into question by the Master Argument it has
been postulated that Diodorus was guilty of making one of the following
confusions. He took the word 'follow' in two different senses in the first
two premises. He played, in the first premise, either on two possible
senses of the way in which a proposition can be concerned with the past
or on two senses of the word 'necessary'. Or finally, he played on a more
general ambiguity hidden in the usage of indeterminate grammatical
tenses.

1.2 Zeller's interpretation. Confusion of the

logical and the chronological.
Zeller3 has given the following syllogistic interpretation of the argu-
ment. "If something was possible which neither is nor will be, an im-
possible would result from a possible. But an impossible cannot result
from a possible. Thus nothing is possible which neither is nor will be".
The minor premise illustrates Diodorus' second proposition. The con-
ditional major premise has Diodorus' third proposition as antecedent
Epictetus, 1916, II, 19 1-4; Doring, 1972, p. 131; the parentheses refer to the von
Arnim collection, 1905; I follow the Brehier translation, revised by Goldschmidt in
Brehier, 1962, pp. 932-933. Brehier translates axoXoutfetv correctly by 'logically
follows'.
2
Ibid., p. 933.
3
Zeller, 1910, p. 254.
 ZELLER'S INTERPRETATION / 5

and the negation of his second as consequent. The correlation of the

two must evidently express the argument's first proposition. "The hy-
pothetical major premise was in need of some basis; and it found it by
way of the proposition saying that every past fact is necessary. Indeed,
when of two mutually exclusive cases one has turned out to be realized,
the possibility of the other finds itself cancelled out, since that which
has happened can no longer be changed... This second case, then, is
now impossible; had it been possible beforehand, then, in the opinion
of Diodorus, an impossible would have resulted from a possible".
This reconstruction supposes a quaternio terminorum in the argu-
ment of Diodorus: one based on the ambiguity of the words OLnoXovdelv
and d&warof. In the major premise, the first of these words does in-
deed have the temporal sense of 'follow in succession', whereas, in the
minor premise, on the contrary, it has the logical sense of 'follow log-
ically'. In the major premise, from the realization of a possible can
and must result the de facto impossibility of another possible, since
the event realized excludes the contrary of what it produces. But the
minor premise signifies that nothing is possible of which the realization
would result in something impossible in itself.
According to P.M. Schuhl, it is possible to restore a purely logi-
cal meaning to the conditional major premise, thereby validating the
argument in rendering its elements homogeneous. It is not that the
impossibility of an event simply succeeds its possibility: it is rather
a strict consequence of it, given the realization of its contrary.4 To
illustrate with an example from Kurt von Fritz,5 the conditional major
premise would mean "If the Carthaginians win today at Cannes, it fol-
lows that the Romans don't win today at Cannes, which, beforehand,
had seemed possible". But as von Fritz points out, Zeller would have
answered that the impossibility referred to by Schuhl is not an impos-
sibility in itself, in the logical sense, but on the contrary a de facto
impossibility, which is not actually far removed from the impossibility
due to temporal succession, as understood by Zeller.
Mr. G.H. von Wright has given an elegant and logically impecca-
ble, formally modified version of the intuition contained in this type of
interpretation. 6 He begins by supposing p to be possible. As a con-
sequence of that supposition, for some future moment, t, it is possible
that p at t. But, in virtue of Diodorus' third premise, it is true that
not-p at t. As a consequence of his first premise, for every moment, i',
4
Schuhl, I960, pp. 74-75.
5
Von Fritz, review of Schuhl, 1962, pp. 138-152.
6
von Wright, 1979a. Compare, however, von Wright's interpretation of the second
premise with the interpretation that will be given in 2.3 and 2.4.
6 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

later than t, it will be impossible that p at t'. Thus the supposition

of the non-realization of the possible leads to an impossibility, if one
interprets the second Diodorean premise as follows: if the proposition
p at t is prospectively possible, that same proposition cannot become
impossible at some later moment of time.
In interpreting the second Diodorean premise as a prohibition of the
proposition p's changing modality over time, Mr. von Wright brings to
light the metaphysical, as opposed to the logical, sense that this first
type of interpretation is obliged to suppose. The difficulty is that the
retrospective impossibility of p, which, according to the first premise,
follows upon the non-realization of p at t begins at a moment, t', later
than t, no matter what hypothesis might be made as to the continuity of
time. The Master Argument would be incontestable only if t' coincided
with t, and the second premise is then supposed to be meant to assure
the validity of the synthetic judgment by which the conservation of the
modality of p from t to t' is guaranteed.
In all these cases, the commentators reject the validity of the Master
Argument in accusing Diodorus of confusing a de facto necessity with a
necessity of principle; and in so doing they are immediately in conflict
with all of the ancient commentators. The translation of Epictetus
accepted here, and against these modern commentators, disqualifies
these interpretations from the start, in taking v.K.oXovdelv as 'follow
logically'. In this way the tradition of the Ancients is respected, though
without, of course, an a priori refusal to examine whether the notions
of possible and of impossible here have a univocal meaning or not.7
As Mr. Boudot points out, "the term rendered by 'follow [logically]'
(aKoXovtieli1) is that which the Megaro-Stoics use for designating what
we call implication. Certainly the definition varies from author to au-
thor, but no one identifies implication and temporal succession.8 For
Diodorus, '<?' follows from 'p' if at no time do we have both 'p' true
and 'q' false. Moreover, and this is the most important point, the prin-
ciple 'the impossible does not follow from the possible' is false in the
Diodorean system if interpreted in the sense of temporal succession. A
statement of the type 'it is possible that p' can be true at present and
false in the future". 9 Though a temporal sense of the term "follow"
is excluded here, it is however not decided whether the logical connec-
tion between the antecedent and the consequent and the consequent
7
For example, Kneale, 1962, p. 121.
8
On this point, cf. ibid., pp. 128-138, especially, p. 132.
9
Boudot, 1973, p. 445. As the author points out, Rescher and Urquhart, 1971,
p. 192, agree with Zeller in taking 'follow' in the sense of 'succeed'.
 AMBIGUITY IN THE FIRST PREMISE / 7

themselves express temporal modalities connected by a simple material

implication.

1.3 Ambiguity in the first premise: Necessity and

irrevocability. Signification of the first
premise.
The first premise of the Master Argument has seemed ambiguous, and
even doubly so.
It says, literally, that all that is past and true (itav KapeXr)Au96<;
dXr)6e<;) is necessary. It seems, then, open to two different interpreta-
tions accordingly as it signifies that every statement grammatically in
the past is necessary if true, or that every statement about a past event
is necessary if true. Since we can give a grammatically past form to
every statement about a future event, in going, for instance, from the
form 'it will rain tomorrow' to the form 'it was true yesterday that it
would rain in two days', the immediate result of taking the first premise
in the grammatical sense would be universal necessity.
Is it believable that such a gross confusion could have escaped Aris-
totle, Epicurus and Chrysippus or that they would have preferred in-
curring the risk of logically onerous solutions to denouncing this simple
sophism? The translation retained here: 'about the past', eliminates
the incriminated ambiguity from the outset.
There is a second type of ambiguity resulting from the use of the
word 'necessary'. It might be said that a past event is simply irrevo-
cable. When a thrown dice has come up six, the throw doesn't cease
to have been aleatory. The proposition saying that the dice turned up
six could in no way be necessary, although it is about an irrevocable
event.
Such an objection, which is entirely justified, obliges us to recon-
sider the sense of the Master Argument's first premise and to ask our-
selves what could be the meaning of the necessity accorded to true
propositions about the past.
The modern logicians who have studied the relations between the
modalities and time have universally accepted the so-called axiom of
necessity, by virtue of which that which is necessary is a fortiori existent
(Ab oportere ad esse valet consequentia) and that which is existent is
a fortiori possible (Ab esse ad posse valet consequentia). This axiom
seems logically well-founded when the modalities are without relation
to time. If, in mathematics, a conclusion is necessary, it is a fortiori
possible. The logical sense of the modalities, as such, excludes time.
Naturally, if a logical necessity had a temporal statement as its object,
8 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

it would necessarily entail, ipso facto, the logical possibility having that
same temporal statement as its object. Supposing it necessary that a
certain thing should have happened, it is a fortiori possible that that
thing should have happened. But when the modalities are applied to
temporal events, they are generally understood, and rightly so, in a
different sense. They are taken in a real, rather than in a logical sense.
Irrevocability, which is a factual kind of necessity, applies to any event
whatsoever, even a contingent one, once it has come to pass. It follows
then that the real modality itself must be assigned a temporal index
distinct from the one affecting the event to which the modality applies.
At present it is irrevocable, or necessary in the factual sense, that the
battle of Salamis took place. Factual possibility, the modal counterpart
of this factual necessity, will likewise be assigned a temporal index of
its own. But it is notable that there is no way of getting from the
past conceived of as a factual necessity to the corresponding factual
possibility, where that factuality is taken to be that of a future or, at
most, a present event, to the exclusion of any event having taken place.
We shall see that for Aristotle this privileged temporal direction of the
possible constitutes the entire content of the Master Argument's first
premise.
The Master Argument then could not be accused of having an am-
biguous first premise, unless, in order to be demonstrative, it confused
either the grammatical form of the past with the factual past or tem-
poral necessity with irrevocability. Neither the one nor the other of
these confusions is required. Each axiom of Epictetus will, in the same
way, be given an interpretation in terms of temporal modalities. As to
the logical demonstration that these axioms are incompatible, it will be
shown to simply obey the laws of first order extensional logic without
having to resort to arguments borrowed from modal logic. Therefore
there will be no risk of subreption between the logical and the factual
meanings of modalities.

1.4 Prior's interpretation: It supposes two

supplementary premises, one of which is
explicitly rejected by Aristotle; it supposes
the first premise ambiguous.
A. N. Prior has proposed a formal reconstruction of the Master Argu-
ment giving a purely logical sense to the word 'follow', though without
deeming it useful to distinguish between logical necessity and the irre-
vocability of the past.
 PRIOR'S INTERPRETATION / 9

He shows that Diodorus' first two premises, A and B, joined to two

others, D and E, allow the derivation of the denial of the third premise,
C. What he proposes is roughly the following.10
A. If it has been true that something is the case then it is not possible
that it has never been true that it is the case.
B. If the consequence q necessarily follows from the premise p, then
if q is not possible, p is not possible.
D. Of whatever is the case it has never been true that it will never
be the case (i. e. it has never-been never-going-to-be the case).
E. Of whatever neither is nor ever will be the case, it has been true
(at some moment) that it will never be the case.
Consider the propositions B and D. Let the premise p in B be
the proposition 'something is the case' and the consequence q 'this
something has never been never going to be the case'. We are justified
in saying that if it is necessary that whatever is the case has never been
never going to be the case, then if it is not possible that it has never
been never going to be the case, then it is not possible that it is the
case. But since D gives us the antecedent of this proposition as true,
10
Prior, 1967, pp. 32-33 [Pp: it has been the case that p (it has been true that p);
Fp: it will be the case that p (it will be true that p)]:
A. Pp 3~ M ~ Pp
B. L(p 3 q) 3 (~ Mq 3~ Mp)
~C. (~ p • ~ Fp) 3~ Mp
D. p 3~ P ~ Fp
E. (~ p • ~ Fp) 0 P ~ Fp
Prior (1967, p. 34) starts from the system T of R. Feys, who takes L as undefined,
defines M as ~ L ~, and adds to propostional calculus only the one new rule
I- a —»h La and the two axioms: Lp 3 p and L(p 3 q) 3 (Lp 3 Lq). First let us
demonstrate B:
1. (p 3 q) 3 (~ q 3~ p) (Contraposition)
2. L[(p 3 q) 3 (~ q 3~ p)] (l.rule of necessitation)
3. L(p 3 q) 3 Z/(~ q 3~ p) (2, second added axiom, Mod.Pan.)
4. L(p 3 q) 3 (L ~ q 3 L ~ p) (3, second added axiom, Mod.Pan.)
5. [L(p 3 q) 3 (~ Mq 3~ Mp)] = B (4, definition of M)
Let us demonstrate that (A. B. D. E.) 3~ C
1. L(p 3~ P ~ Fp) (D, rule of necessitation)
2. L(p 3~ P ~ Fp) 3
(~ M ~ P ~ Fp 3~ Mp) (B with substitutions p/p, ~ P ~ Fp/q)
3. ~ M ~ P ~ Fp 3~ Mp (1,2, Mod.Pan.)
4. ~ p- ~ Fp 3 P ~ Fp (E)
5. P ~ Fp 3~ M ~ P ~ Fp (A)
6. ~ p- ~ Fp 3~ M ~ P ~ Fp (4,5 Mod.Pon.)
7. (~ p- ~ Fp 3~ Mp) =~ C (6,3, Syll., Mod.Pon.)
10 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

we can detach its consequent to get 'if it is not possible that something
has never been never going to be the case, then it is not possible that
it be the case'.
On the other hand, given E and substituting 'something will never
be the case' for 'something is the case' in A, the rule of syllogism will
allow us to write:
Of whatever neither is nor ever will be the case it is not possible that
it has never been never going to be the case. But we have demonstrated
(B and D) that if it is not possible that something has never been never
going to be the case then it is not possible for that something to be the
case. Hence, whatever neither is nor ever will be the case is impossible.
This is the denial of the Master Argument's premise C according to
which there is a possible that will never be realized.
Once the ambiguous interpretation of premise A is admitted, the
whole question comes down to one of justifying the introduction of D
andE.
In favor of D one might invoke11 Chapter IX of De Interpretatione,12
or again Cicero's De Fato.13 But in one case it is the Megarians, in the
other the Stoics, to which this principle of retrogradation is attributed.
Aristotle refuses it implicitly.
As for the premise E, it means that "... if a statement is false and
will always remain so in the future, then there has been a past moment
at which it was true that that statement would always subsequently
be false".14 What makes this premise seem plausible is that "... if
p is now and always will be false then it has already been true in
the past, at least at the moment just past, that p will never be true
anymore—it hasn't always been true, because at least in the moment
just past it wasn't true, that p would be true again".15 This thesis is
valid only if time is discrete and non-dense,16 in other words, only if,
as Diodorus holds,17 every instant, and in particular the present one,
"Becker, 1961, pp. 250-253.
12
See below, 6.4, pp. 140-141.
13
Potest factum quicquam igitur esse, quod non verum fuerit futurum esse? (Ci-
cero, De Fato, XII (27)).
14
Boudot, 1973, p. 447.
15
Prior, 1967, p. 49; but cf. also p. 8.
16
Boudot, 1973, pp. 447-448 (for an intuitive resume); Prior, 1967, pp. 49-50.
17
D6ring, 1972, frag. 116-120, p. 129 sq. According to Fraenkel (1960, pp. 204-
211) the arguments of Zeno treat of space represented as continuous just as well as
discontinuous, whereas Diodorus envisaged solely the case of its discontinuity. In
the third proof of Festus, the d^epi), as logical conceptions, are introduced only to
render motion impossible. According to Sedley, 1977 (pp. 88-89), basing himself on
Chalcidius, Diodorus' influence on some of the Stoics was so great that they would
have incorporated the notion of indivisible units of matter into their doctrine. For
 PRIOR'S INTERPRETATION / 11

has an immediate predecessor and successor. But suppose that time

is continuous or, what suffices for the argument, dense. Then between
the present moment, t0, time from which p is supposed by virtue of
the antecedent no longer to be the case, and the past moment, t\,
chosen arbitrarily near so as to verify the consequent 'p has been never
going to be the case', there will inevitably be an infinity of instants
and so at least one instant, £ 2 , different from t\ and t0. But thesis E,
which affords no condition on the status at time t% of the future case
in question, inevitably leaves the truth-value of the statement about
the future case indeterminate at t^. Taking the value of this statement
to be true we falsify the thesis. Thus E is valid only if it is excluded a
priori that that statement chosen as false at t\ and confirmed as false
at to be able to change its truth-value between ti and to. This exclusion
is satisfactory only if there are no intermediate instants between ti and
to, which implies that t\ is the immediate predecessor of to and that
time is thus discrete.
With the conclusion established, it can be shown that Diodorus'
propositions A and B are, in turn, demonstrable. Mr. Boudot con-
cludes; "If we admit four principles, three of which were in common
use in antiquity -viz. A, B and D- the fourth -viz. E- expressing
the discreteness of time, we conclude the Diodorean definition of the
possible.18 On the other hand, if we begin with that definition, the
premises used, linking the modalities with temporal determinations,
are demonstrable. Looked at in this way, the Master Argument reveals
such a perfection in its alliance of coherence and simplicity that it be-
comes easy to understand the reputation it established for its author
and the respect it inspired among the Ancients".19 What is more, con-
ceived of in this way, the Master Argument does not lead, not directly
at least, to determinism: a statement true sometimes, but not always,
is true without being necessary.
Two shortcomings, however, prevent us from accepting Prior's re-
construction. The first lies in the introduction of premises D and E,
the second in the hypothesis of an ambiguity in the interpretation of
A which would be imputed to both Diodorus and his opponents.
In the first place, premise E, of which an equivalent is supposed
Diodorus there is a relation between these unities and the definition of the possible:
since no body will be divided into an infinite number of parts, why call such a
division possible? Chrysippus holds that the division is indefinite, not reaching
infinity. It is to him that our explicit sources trace back the affirmation of infinite
divisibility (note 80, pp. 111-112).
18
That is to say: the possible is whatever is or will be and only that.
19
Boudot, 1973, p. 448 and p. 449.
12 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

to be attributable to Diodorus, is, like premise D, in direct opposition

to the Aristotelian theory of the continuity of time in book V of the
Physics. Aristotle distinguishes, according to an increasing degree of
'neighbourhood', the three notions of succession, contiguity and con-
tinuity. "The Aristotelian continuum is defined as a collection, in a
certain sense well-linked (Ephexes and Echomenon), of parts poten-
tially separated by limiting points" .20 The infinite divisibility of the
continuum thus involves its density. So Diodorus' argument, on Prior's
reconstruction, would fail to touch Aristotle in the least. Not only
would it not touch him in fact, but the introduction of clause E of the
discontinuity of time would be tantamount to a conscious admission of
defeat. If we think of Diodorus as having distinguished himself from
the "ancient" Megarians in saving the modal distinctions, it would have
to be added that as soon as confronted with a dense time-and we can
presume that on this point the Aristotelian representation had rallied
the assent of the scientific community-those distinctions collapse. The
Master Argument is thus demoted to the rank of a particular school's
argument.
In the second place, if we accept Prior's reconstruction, we have
the right to accuse Diodorus of having confused "a past-tense state-
ment" with a "statement about the past"21 and of having, in this way,
exploited an ambiguity of language. But this is to fly in the face of
the tradition which has it that, contradicted on this point by Chrysip-
pus, he conceived language to be free of any equivocation.22 Surely,
one might argue, this premise does not in itself produce necessity, for
Diodorus distinguishes the necessary and the possible. It does, how-
ever, produce it if we go from a tense-logic to a dated or 'pseudo-dated'
20
Granger, 1976, p. 306; for the philosophical analysis of this Aristotelian contin-
uum cf. Vuillemin, 1962, pp. 185-198.
21
Boudot, 1973, p. 470; Mrs. Kneale denounces the confusion, 1962, p. 121.
22
AuliusGellius, Nodes Atticae, Bk. XII, c.1-3 in Doring, 1977, 111, p. 31 and
p. 128.
 PRIOR'S INTERPRETATION / 13

event-logic.23 Prior and Boudot24 have shown that in constructing an

Ockhamist model in which the suspect confusion is avoided, necessi-
tarianism too can be avoided, even on a dated event-logic.
But should it turn out, as we hope to show, that there is no am-
biguity in Diodorus' first premise, it would no longer be possible to
avoid necessitarianism in a logic of dates by recourse of a rejection of
that supposed ambiguity.25 It would in fact suffice in this case to limit
ourselves to authentically past events, where there is no room for for-
mal confusion, to derive necessity. And if one wanted then to conclude
to necessitarianism, Ockham's distinction would no longer block the
move, for it would no longer be simply a grammatical mistake that
suggested necessitarianism.
Let us, on the contrary, follow historical likelihood; take the univer-
sal favor with which the Master Argument was met in Antiquity as a
presumption in favor of its solidity; dismiss the supposition that crude
ambiguities would have crept into its premises. Let us give, along with
Prior, a purely logical sense to 'follow'-though not necessarily the sense
of a thesis of modal logic-; and let us suppose that the past mentioned
in the first premise is that of events to show the incompatibility of the
three premises without having either to postulate the discreteness of
time, or to confuse the irrevocability of the past with logical neces-
sity, or even to invoke a retrogradation of the true, of which Epictetus
makes no mention and which the Stagirite expressly calls into ques-
tion. Above all, in order to avoid introducing principles inappropriate
to the interpretation, it is important to find a text contemporaneous
with the Master Argument as formulated by Diodorus and apt, by way
of comparison, to shed light on the objective sense of the latter.

23
Boudot, 1973, p. 451: "In tense-logic the conclusion from the truth to the ne-
cessity of 'Socrates will be sitting' does not follow. But in metric tense-logic, the
conclusion from the truth to the necessity of 'Socrates will be sitting tomorrow' does
follow, since that statement is equivalent to 'It was true yesterday that Socrates
will be sitting the day after tomorrow', which is itself necessary because past". This
variation of modal status with the formal expression of the tense of the statements
in question assures the originality of Diodorus with respect to the 'ancient' Megar-
ians (Blanche, 1965, pp. 133-149). It also limits the importance of that originality.
Sedley (1977, pp. 74-120) has contested Diodorus' adhesion to the Megarian 'school'
(pp. 74-78), making him rather a representative of the Dialectical 'school', which
had a separate existence. But even accepting this thesis, there remains sufficient
affinity between the two schools, from the point of view both of the theory of motion
and of that of the modalities, for us to still consider Diodorus to be a 'Megarian'
philosopher.
24
Prior, 1967, pp. 121 sq.
25
This constitutes the essentials of Boudot's elegant solution (1973).
14 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

There is such a text. It is that of an argument of Aristotle's con-

cerning demonstrations by the absurd in modal logic for propositions
involving time. It is a text that has not attracted as much attention
as it deserves on the part of modern commentators.26

26
With the exception of Cherniss (1962) and Hintikka (1973), though even here,
none of Hintikka's references to this passage (p. 94, p. 152, p. 164, p. 183) actually
analyzes Aristotle's manner of reasoning.
Reconstruction of the Master
Argument.

2.1 An Aristotelian paradigm: De Caelo, I,

28366-17; its context.
/6/"It is not true to say now /?/ that a thing exists last year, nor
last year that it exists now. It is therefore impossible /8/ that that
which does not exist at some moment should later be sempiternal. For
it will have later as well /9/ the capacity of not existing, though not
that of not existing while it exists (that which is /10/ in fact exists
in actuality), but of not existing last year /ll/ and in the past. Now
suppose this capacity it has to be realized in actuality. /12/ It will
then be true to say now that it does not exist last year. But /13/
that is impossible. For there is no potentiality of the past, but /14/
only of the present and the future. The same goes too if that which is
formerly sempiternal /15/ passes later to non-existence; for it will have
the capacity for what is not in actuality. /16/ Let us then actualize
the possible: it will be true to say now that this exists last year /17/
and generally in the past.".
In this passage Aristotle proposes a refutation of the Platonic thesis
set forth in the Timaeus, according to which the demiurge created the
world in imitating the eternal Ideas and in informing the receptacle
after their pattern. That creation accomplished, the world is then
supposed to exist sempiternally.1
1
That, in any case, is the "literal" interpretation Aristotle gives of Plato (Cherniss,
1962, pp. 414-417). This text is fundamental for insuring the separation of the
eternal, though sensible, substances which are not subject to generation and decay
from the properly material substances which are subject to generation and decay.
It is here that the separation of Heaven and Earth and astronomical theory find
their raison d'etre.

15
16 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

The passage here is composed of two parts. In the first, Aristotle

shows that nothing that has been generated can last sempiternally,
in the second, that nothing ungenerated can decay. He consequently
concludes the equivalence of the terms 'generated' and 'perishable'.
This difficult passage brings four principles to bear. The last three
of these, with the exception of changing a negation into an affirmation,
will serve to constitute the Master Argument as well. Since the first
of these last three principles has the form of a logical conjunction, it
will be analyzed, for clarity's sake, as two independent principles. The
complete list then will consist of 1) the principle of conservation of the
modalities, 2) the principle of the possible realization of the possible or
principle of the diachronic expansion of the modality together with the
principle of conditional necessity, 3) the principle of the impossibility
of realizing a contingent possible in the past, 4) the principle of the
subsistence of a possible that is never realized. All of these principles,
and these alone, are explicitly used in the De Caelo text. All but the
first, and these alone, are to be explicitly mentioned in Epictetus' text
on the Master Argument.

2.2 The principle of the conservation of modal

status.
This is the principle proper to the De Caelo to which the Master Argu-
ment will have no recourse. The modalities, according to the Stagirite,
are not simple operators governing statements or propositions. They
are the characteristic properties of substances. There are thus two sorts
of substances: a) necessary substances which are either eternal, that
is to say, have no relation to time, or sempiternal, that is to say, are
permanent in time, participating in topic matter alone and endowed
with perpetual motion, b) contingent substances, now existing, now
not existing, subject to full materiality, to generation and decay. It is
an axiom of Aristotle's philosophy that a substance cannot change its
modal status. Whether necessary or contingent, it is so by nature. It
will retain its modal status then and will never be seen to change it.
The consequence of this principle is used in lines 7-9 of the text.
Suppose that something that did not exist in the past should come, or
should have come, to be. It thus belongs to the category of contingent
things. By virtue of the principle of the conservation of modal status
it will then retain later the capacity of not existing.
This principle and its consequence would merit an investigation on
their own; but since they play no role in the Master Argument and
 THE PRINCIPLE OF THE CONSERVATION OF MODAL STATUS / 17

since its validity in no way depends on them, they will simply be taken
here as given.2
The three following principles, on the other hand, broken down into
four for the sake of clarity, are the very ones operative in the Master
Argument.
2
The principle of the conservation of modal status is a consequence of the Aris-
totelian distinction between three kinds of substance (Metaphysics A, 1069a30-
1069*2): a) the sensible corruptible substances, b) the sensible incorruptible sub-
stances, c) the immaterial and immobiles substances. The three kinds are such
that, if something belongs to one of them, it can never belong to another one. But
belonging to b) or to c) means being a necessary being, and belonging to a) means
being a contingent being. Themistius, in a subtle passage of his Commentary,
lays down the principle and shows the absurdities that its denial would entail. To
allow perishable substances to transgress their essence and accede to immortality
and indestructibility would be to destroy the very limits that define the nature of
things and to make contrary capacities persist indefinitely. "As the generable and
the perishable are not so by chance and by fortune, it is seen that they are so
by nature. Indeed, all that there is is either by nature or by chance-for we leave
aside here that which art produces; but all natural things maintain (custodiunt) the
capacities that are proper to them. If they don't keep them and they change into
other dispositions, their modification will be either the result of chance-in which
case there would be but one and the same disposition for both natural things and
for those resulting from chance and fortune-or the modification will come about,
even for chance things, according to nature. That is why, since there is conservation
of the modification even if the things themselves are to change, the modification in
turn will take place either by nature or as a result of chance. This is why natural
capacities must have limits. But if these capacities have limits, then that also,
which has a birth and is subject to death, will be so by nature before that death.
It can then be seen that what happens now to the nature and the matter thus
subject to these two dispositions, namely, on the one hand the generation of things
and their existence, on the other their privation and their inexistence, has a limit
beyond which it changes no more. That is why it is necessary that that which is
engendered should not be deprived of this power of being changed; that is why
death too will come to it, in its time. Likewise, we do not surprise that which is
always subject to death transgressing its proper nature. Otherwise, since it will
have persisted for some time by virtue of its nature, it will equally persist in the
disposition that renders it immortal and, keeping its nature, it will also retain the
capacity in virtue of which it was changed. That is why there must then be several
capacities at once for an infinite time, and since we have established that the action
of this capacity persists, then what we have established will be false, which can no
more be the case than the rest we have spoken of so often." (Themistius, De Caelo
A 12 [Arist. p. 28362-12] 1902, p. 86, lines 4-29).
18 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

2.3 The principle of the possible realization of the

possible interpreted as a principle of pure
modal logic.3
The reasoning, or rather the two parallel lines of reasoning that com-
pose our text are among those reductio ad absudum arguments applied
to singular propositions4 and obtained by showing that a supposed
possible cannot be realized.
In the text at hand Aristotle applies this manner of reasoning to
singular propositions of existence. What there is actually question of
is a determinate individual, the Heavens; but the reasoning has to do
with an individual in general, which happens at one time not to exist
(1.7-8). The reasoning is seen to be transcendental and universal, since
the only 'predicate' under which such indeterminate individuals are
subsumed is that of being and non-being which is not a genus. But
that very universality, transcending the categories, is the hallmark of
the syllogistic form of science and particularly of the modal syllogistic.
The directing principle of this mode of reasoning has generally been
expressed in terms of pure modal logic. It is this conception that is
to be set out first. But, on subsequently examining the Aristotelian
text, this first conception will be shown to be inadequate and it will
further be seen that the principle of the possible realization of the
possible comes down to none other than the diachronic and synchronic
movements of the modalities.
From the possible as such, nothing can be concluded, for a modal
conclusion is legitimate only if it proceeds a fortiori, that is to say,
from the necessary to the actual or from the actual to the possible.
Starting from the consideration of a possible then, the only thing
we can do is to analyze the consequences that would result from its
realization. If these consequences follow necessarily they would them-
selves have a necessary character, if the realization was necessary. By
contraposition, if these consequences which result necessarily from the
realization turn out to be impossible, that realization is impossible in
turn. By a second contraposition it can further be shown that, if the
consequences are supposed to follow necessarily from the realization of
the possible, then if that realization is possible the consequences will
be so as well.
3
Attested to, for example, in the De Caelo (281° 15): Su^paivei 5' dSuvcrcov It,
dSuvdtou ("The impossible comes from the impossible").
4
The principle applied to general statements and according to which one must
be able to realize a possible is called "ecthesis by demodalization" by G. Granger
(1976, p. 193, p. 214).
 THE PRINCIPLE OF THE POSSIBLE REALIZATION OF THE POSSIBLE / 19

Consider the possibility of a condition p. Let us posit p as existent

and examine the consequence q that follows necessarily from it. If q
is contradictory, i.e. not realizable in any possible world, we have the
right to conclude that p is also impossible. This is the import of the
Master Argument's second premise according to which the impossible
does not logically follow from the possible.5
As a typical application of the principle Aristotle gives that of the
incommensurability of the diagonal of the square with its side: "All who
establish an argument per impossibile syllogistically infer that which is
false and prove the hypothetical conclusion when something impossible
results from the supposition of its contradictory; for example that the
diagonal of the square has no common measure with the side, since if
one supposes it to have a common measure then the odd numbers are
equal to the even ones. One infers syllogistically that the odd are equal
to the even numbers and proves hypothetically the incommensurability
of the diagonal, since a falsehood results from its negation".6 Thus if p
signifies that there is a common measure between the side and diagonal
of the square, q that one and the same number is at once both even
and odd, then given that if p is true q is necessarily true as well,
from the impossibility of q the impossibility of p is inferred, i.e. the
incommensurability of the side and the diagonal of the square.
From this principle of the possible realization of the possible one
can draw yet another principle whose intuitive significance is perhaps
clearer. Namely, it is equivalent to say of a thing that it is possible and
that the conjunction of the possibility of that thing with its realization
is possible.7
At this point there is need for a note of caution. It happens, both
in the case of Aristotle's reasoning and in that of the Master Argu-
ment that the realization of the possible under consideration produces
a contradictory consequence only by virtue of an as yet non-explicit
hypothesis. It is thus the conjunction of the realization of the possible
with that hypothesis that produces the contradiction. That contradic-
tion justifies the conclusion that there is incompatibility between the
realization of the possible and the hypothesis but in no way does it jus-
5
This is Prior's premise B: L(p D q) D (~ Mq D~ Mp), deducible in the weak
system T (see p. 7, note 10).
6
An.Pr., I, 23, 41"23-30. This example is used by Aristotle in the De Caelo, I 11
281°7.
7
The principle in question here is: Mp = M(p.Mp). Since T contains L(p D
q) D (Lp 3 Lq) and the rule of necessitation, it is normal. Hence given p 3 Mp
(Hughes & Cresswell TH1), it follows that p = p • Mp, and since T is normal,
Mp = M(p • Mp) (Note of B. Graham).
20 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

tify the conclusion that there is incompatibility between the possible

in question itself (possibly non-realized) and the hypothesis.8
This latter illegitimate conclusion would be stronger than the pre-
mise. It would result in a sophism in the distribution of the modali-
ties. Consequently, whenever a reasoning by realization is employed to
demonstrate not the intrinsic impossibility of a thing, but rather the
logical incompatibility between the possibility of one thing and the re-
ality of another, simple application of the principle that the impossible
does not logically follow from the possible is not enough. What must in
fact be demonstrated here is that a contradictory consequence necessar-
ily follows, not from the conjunction of the realization of the possible
and the hypothesis, but from the conjunction of simply the possible
and the hypothesis. To this end a specific principle is required.
Such a principle indeed has been clearly stated by Leibniz. To An-
toine who in his Dialogue with Laurent Valla sets forth "the philoso-
pher's rule" that "whatever is possible may be considered as existent" ,9
Laurent replies: "Philosophers' rules are no oracles for me. This one
in particular is not at all exact. It often happens that two contradic-
tories are both possible; can they also both exist?" He thus denounces
the sophism of distribution which would have it that the real world,
in so far as it excludes the possible worlds, would also render them
impossible by the same token.10

8
Having recourse to the preceding thesis, let us posit p as existent. The condition
of validity of this position is that p be compatible with Mp. Suppose further that
the hypothesis of the argument imposes ~ p. Clearly ~ p excludes p but not Mp.
In other words:
~ p.M(p.Mp)
is compatible, for that conjunction is equivalent to:
~ p.Mp.MMp.
In order for the position that p is existent to enter into contradiction with the
hypothesis, this latter would have to impose ~ Mp. Then:
~ Mp.M(p.Mp)
is indeed incompatible, since that conjunction is equivalent to:
~ Mp.Mp.MMp.
9
Leibniz, Gerhardt, 1978, p. 359; Jalabert, 1962, p. 371. It is important to insist
here on the word 'may' which has as consequence the impossibility of applying
the rule in the case where the impossibility of the thing would have already been
posited.
10
Laurent warns Antoine that the philosophers' rule is false if one makes it say
that:
(1) (Mp.M ~ p ) 3 M(p. ~ p).
From (1), by contraposition, comes:
 THE PRINCIPLE OF THE POSSIBLE REALIZATION OF THE POSSIBLE / 21

That is why Aristotle inserted in the very expression of the princi-

ple of the possible realization of the possible the clause having to do
with the principle of conditional necessity (1. 10). It will be seen below,
however, that the modalities figuring in this latter principle, instead of
belonging to pure modal logic, are affected by temporal indices. If one
were then to continue interpreting the principle of the possible realiza-
tion of the possible in terms of pure modal logic an inconsistency would
creep into the Aristotelian passage, for the second principle of the De
Caelo text would then have to be expressed as the conjunction of non-
homogeneous terms, the first comporting "pure" modal expressions,
the second temporally indexed modal expressions.

2.4 The principle of possible realization of the

possible as principle of synchronic contraction
of the possible and diachronic expansion of
the necessary.
Coming back to Aristotle's text, just what is it that he enjoins us
to actualize the potentiality of in line 11? It is something which the
principle of conservation of modality (lines 8-9) assures us will have
retained the capacity of not existing. He makes it clear in lines 9 and
10 that this capacity cannot be actualized while the thing in question
is in actuality (namely, now, since we know from experience that the
world exists, and in the future when, by hypothesis, the world will exist
sempiternally) but, consequently, last year and in the past. The possi-
bility of not existing Aristotle speaks of then is not a simple logical or
atemporal possibility. It is a possibility subject to a double temporal
index, since as a possibility it will persist throughout all the future but
because of circumstances and admitted hypotheses its realization can
take place only last year and in the past. Such doubly indexed modali-
ties are homogeneous with the principle of conditional necessity, as we
shall see, and also with the principle of conservation of the modalities
(for what sense would it make to speak of the conservation of pure and
atemporal modalities?).
If we follow Aristotle's text to the letter, then the principle of pos-
sible realization of the possible signifies that if it is possible at a given
moment (that we shall fix as now for perspicuity's sake) that something
(2) ~ M(p. ~ p ) D~ (Mp.M ~ p )
and by virtue of non-contradiction and modus ponens:
(3) ~ ( M p . M ~ p )
which would have as consequence the strictest form of logical determinism.
22 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

should be realized or actualized at time t, then there is an instant, t\,

on the interval \N(i.e. 'now') -t\ < e situated either in the past or in
the future (possibly coincident with the now) such that it is possible at
ti that that thing should be realized at ti. The principle synchronically
contracts a possibility posited diachronically over the interval onto an
instant of the interval.
By contraposition, if at every instant, ti, of an interval \N —1\ < e,
where either t precedes N or N is identical with t or precedes t, it is
necessary that p at ti, then it is necessary at N that p at t.
It might be objected that this proposed formulation is entirely in-
adequate for expressing the possible realization of the possible. To
synchronically contract a possible by saying that it is possible at t\
that p at ti is surely not to realize that possible or to say that p at t\.
But we have forewarned that Aristotle did not separate the clause of
synchronic contraction from the clause of conditional necessity. Should
it turn out that this latter clause signifies or implies that a contracted
possibility is none other than a realized one, all the requisites of the
interpretation will have been satisfied. The aim of the following section
will be to show that this is so.11
Aristotle exemplifies his principle now by cases with no relation to
time, now by dated events. Thus in de Caelo 281b10-20 he argues about
the incommensurability of the diagonal, but also about the possibility
of standing and sitting. For a given agent, it is impossible at a given
moment to be at any time both sitting and standing. And it is clearly
having in mind such doubly indexed modalities that he says in the
chapter of Metaphysics where he refutes the Megarics (0,3,1047a23-
25): "A thing is capable of doing something if there will be nothing
impossible in its having the actuality of that of which it is said to have
the capacity".
Therefore, if it is now possible that p at t (M^pt)-, the contraction
of the possible at t (Mtpt) logically excludes the contracted impossible
(~ Mtpt)- Such an impossible would precisely obtain, were we to deny
B.
11
The principle of possible realization of the possible, when interpreted in confor-
mity with the De Caelo text in terms of doubly temporally indexed modalities, may
be expressed in the following manner:
(t){~ LN ~ p t D(3ti)[M t ,pt, • ( « < *i < NV N < ti < t ) ] } .
What is the relation between the law of modal logic
(a) L(p 3 q) D (Afp 3 Mq)
and the second axiom, B, of Epictetus?
 THE PRINCIPLE OF THE POSSIBLE REALIZATION OF THE POSSIBLE / 23

Then it would be true to say that:

LN ~ Pt • (t,)(N < tv < t D

When t = TV, a contradiction results. When N is indeterminate,

at all ti such that N < t\ < i,~ Mtlptl would obtain, destroying the
Let us show on what conditions B can be deduced from (a):
1. t-p, D ( 3 t i ) [ p t , • ( < < « ! < NVN <h <t)}
Existential generalization and conjunction of a truism)
2. h L{Pt D (Bii)[p ( l • (t < «i < 7VV TV < d < t)]},
(1, I- Q 3 h La).
3. h L{pt D (3ti)[pt, -(t<ti<NV N <h< t)]} D
{Mp< D A/(3ti)[p t l • (t < t! < N V N < tj. < t)]}
(a).
4. h Mp( D M(3«i)[pj 1 • (t < ti < Af V N < <i < t)]
(3,2, detachement).
5. I- Mjvpt D Mp ( .
6. h M;vpt D M(3(i )[p t j • (t < «i < TV V TV < ti < t)}
(5,4, Syll.).
7. h M(3t!)[p 4l • (t < «i < TV V N < ^ < 0] 3
(3«i)M[p (l • (t < ti < Af V N < tj < «)]
(7,BF).
8. I- MNpt D (3ti)M[p t l • (t < «i < N V TV < t j < t)]
(7,6, Syll.)
9. I- (StOfM^pt, • (t < ti < N V N < «i < t)] =D/
(3fi)M[p t l • t < ti < N V N < ti < t]
10. h MA,Pi D (3«,)[M (lPtl • (t < ti < N V N < ti < t)]
(8,9, Syll.)
11. H ~ L N ~ p ( D(3ti)[M ( l p ( - ( t < t, <NV N < ^ <t)}
(l-~ LN ^ pt = MNPI, 10)
There are four statements in this deduction that merit justification: 1, 5, 7 and 9.
The validity of the first statement stems from the fact that if p is the case at t then
there is a t\ falling between t inclusively and N or N inclusively and t inclusively.
The fifth statement is a characteristic axiom of the logic of the temporal modal-
ities. If it is possible now that p at t then it is a fortiori logically possible that p at
t, since the synchronic possibility puts added strictures on the logical possibility.
Statement 7 is a consequence of the Barcan formula. That formula is legiti-
mate when the variable is interpreted substitutionally. And in fact a substitutional
interpretation is sufficient when the values of the variable are the instants of time.
The definition 9 is that of synchronic possibility. To say that it is possible at t\
that p at (i , with ti a function of t and of N, is to say that it is logically possible
that p be the case at t\ with t\ a function of t and of ./V and with the logical
possibility bearing on the conjunction of the realization of p at the instant t\ and
the defined function of t\ and (.
The negative form given to the possible will be explained when we come to
Chrysippus' theory. (5.2, pp. 108-113)
24 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

potentiality of p at any such instant. Finally if we assign a possibility

with a fixed date, t — t\, at this date it will be impossible to contract
the possible.
An objection could be derived from this last case. "Suppose, it
would be said, that I can now do something at a given time tomorrow.
It may happen that this possibility will be maintained during some
time and us say from now on until t% < ^i< and that at (2 an event be
produced which makes impossible that patti. Therefore it may be now
possible that p at t\ and this possibility may persist until t^ without
that p be possible at t\, in contradiction with B". But Diodorus as well
as Aristotle exclude such a purely verbal interpretation of modality. In
the same book of Metaphysics (0, 5, 1048a16-21), Aristotle requires
that having a potency does not happen without given conditions that
distinguish a genuine potency from a rhetorical one: "To add, he says,
the qualification 'if nothing external prevents it' is no further necessary;
for it has the potency on the terms on which this is a potency of acting,
and it is not in all circumstances but on certain conditions, among
which will be the exclusion of external hindrances, for they are barred
by some of the qualifications of our definition".

2.5 The principle of conditional necessity.

In line 9 and the parenthesis of line 10 Aristotle uses, without formu-
lating it, the principle of conditional necessity. Which are the cases, he
asks, in which the consequence drawn from the principle of conserva-
tion of modal status (here applied to the negative possible) is invalid?
He answers that a thing which has not been but then comes to exist
will have later also the capacity of not existing, though not that of not
existing while it exists (that which is does indeed exist in actuality).
What is excluded by the actual existence of a contingent thing, then, is
not only that thing's inexistence, but its very capacity of not existing,
while it exists. The first exclusion is guaranteed by the principle of non
contradiction applied to temporal things, since it is impossible for one
and the same thing to be and not to be at the same time. The second
exclusion is stronger. It prevents a negative capacity's coexisting, in
so far as the time of the event governed by that capacity is concerned,
with an affirmative actuality; just as, symmetrically, it would prevent
an affirmative capacity's coexisting with a negative actuality. Thus the
actuality of p excludes the capacity of not-p while p and the actuality
of not-p excludes the capacity of p while not-p.
It would be futile to try to deduce such a principle from the prin-
ciples of the modal logics accepted today. We shall limit ourselves
 THE PRINCIPLE OF CONDITIONAL NECESSITY / 25

to pointing this out in the case of reasoning by possible realization.12

Suppose the negative possibility of an event and the hypothesis of that
event's existence actualize the possibility. By virtue of the principle of
contradiction an impossible consequence results necessarily from that
negative realization joined to the posited actuality of the positive re-
alization. But it would be giving in to the sophism of the distribution
of the modalities to conclude that the possibility of the non-realization
of an event is incompatible with its realization. The logical incompat-
ibility is in fact only between the non-realization and realization.
Since this consequence is unquestionably authorized by the passage
however, one of two things must be the case. Either Aristotle fell victim
to the illusion we have just censured, or he took this incompatibility
of the actuality and the contrary capacity to be a specific principle.
To decide the issue let us examine the consequences of this posited
incompatibility.
By the reciprocal definitions of the modalities, if p is necessary then
it is not possible that not-p. The necessity of a thing, then, excludes
the contrary potentiality of that thing. To maintain that the actuality
of a thing excludes the contrary potentiality of that thing is thus to
assimilate actuality and necessity in that respect. But it will be seen
that in Chapter IX of De Interpretations Aristotle does explicitly as-
similate actuality and necessity when he defines conditional necessity.
Simple necessity is the prerogative either of the insensible beings or
of the sempiternal substances, which always are in actuality. Contin-
gent substances sometimes are in actuality and sometimes are not. In
so far as, and while, they are in actuality they behave as necessary
substances. Only that necessity is temporally conditioned by the actu-
ality, and ceases with it. In short, if there are differences between the
eternal actuality of God, the sempiternal actuality of celestial things
and the temporal actuality of sublunar things, these differences are not
differences of degree and the actuality of a contingent thing has, for a
finite time, the same necessity as that belonging to the actuality of an
eternal or sempiternal thing.

12
The sophism of the distribution of the modalities may be written thus:
* ~ M(3t)(pt- ~ pt) 3~ M(3t)(pt • Mt ~ pt).
The principle of conditional necessity is written:
~ (pt • Mt ~ pt) or ~ (~ Pt •
that is to say:
Mt ~ pt D~ pt or Mtpt
26 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Does the principle of conditional necessity provoke a collapse of the

modalities? It does in that, while p, one can conclude the necessity
of p from p. But that conclusion can in no way be detached from its
condition,13 which limits the extent of the collapse.
The principle of conditional necessity seems to contradict numer-
ous passages where Aristotle speaks of the possibility for a thing to be
in a state opposed to that in which it is in actuality.14 For example,
Socrates, who is sitting, might be standing. The contradiction disap-
pears if, as we have proposed, a distinction is made between the logical
modalities and the real or temporal modalities, with the validity of the
principle of conditional necessity limited to these latter. A temporal
modality will be affected by a double temporal index: one situating
the modality itself in time, the other situating the event with which
the modality has to do. Suppose it possible at time ti (in the sense
of "real" possibility) that a certain event should take place at time
<2- Suppose further that at time ti the event in question is not tak-
ing place. What the principle of conditional necessity tells us under
such conditions is that time t\ cannot be identical with time t%. On
the other hand, there is nothing contradictory in maintaining that it is
possible at t\ that the event non-realized at t\ be realized at any time
£2 distinct from t\.15

2.6 The irrevocability of the past or the principle

of the impossibility of realizing the possible in
the past.
At the end of the second chapter of the sixth book of the Nicomachean
Ethics, Aristotle develops the following assertion (1139° 13-14): "No
one deliberates about things which are not susceptible of being other
than they are". And he winds up by saying: "concerning the past
there is no longer any choice. That is why no one chooses that Illium
should have been ravaged; for no one deliberates about that which has
been, but only about what will be and what is possible (ivdexonevov);
and indeed what has been cannot not have been. This is why Agathon
13
19°22-23.
"Especially Metaphysics Q, 3 (1047a20-29, for example).
15
The logical product of the principle of synchronic contraction of the possible and
the principle of conditional necessity has as consequence, in conformity with the
De Caelo passage, the realization of the contracted possible:
((t){MNPt D (3ti)[M tlPtl .(t = ti<NVN = tt= t)}} • (MtPt D P t ) ) D
(t){MNPt D (3ti)[ptl •(t = tl<NVN = tl= t)}}.
 THE IRREVOCABILITY OF THE PAST / 27

is right to say: 'God Himself is hindered in this one thing: to make

undone things that have once been done'." (113965-11). In short, there
is such a thing as deliberation. Deliberation has only to do with what
can be contingently. Nothing of what is past can be contingently. The
past is not among the objects of a possible deliberation. The same con-
sequence is drawn in the sentence of lines 13 and 14 of the De Caelo
passage: "There is no potentiality of the past but only of the present
and the future". For if the past is irrevocable it is not an object of
deliberation. Not being an object of deliberation, there is no (contin-
gent) potentiality with respect to it, all potentiality is regarding the
present or the future. From the fact that the contingent potentiality
of opposites regards the present or the future, the irrevocability of the
past follows by contraposition: this irrevocability residing in the dis-
junction of the 'necessity' of the affirmative member or the 'necessity'
of the negative member.
This disjunction of retrograde modalities could in no way be justi-
fied by arguing that if something has taken place then it is not logically
possible that it should not have taken place. The logical impossibil-
ity in question would be atemporal, even though having to do with
an event dated in the past. But once endowed with its two temporal
indices, this necessity, unlike real possibility, will be seen to take on
two distinct developments. Its natural movement will be in the same
direction as the natural movement of real possibility: i.e., the direction
going from past to future. But to this movement there will be added
another, retrograde and specific, no longer applying to things, but to
our relation to things. It is in virtue of this relation that the necessity
of the past will be spoken of. Let us then give each of the modalities a
double temporal index. Let us say that it is possible, in the sense of a
'real' possible, at time ti that a certain event should take place at time
£ 2 . What Aristotle asserts is that t2 can not be anterior to t\. What is
specific to irrevocability will be sufficiently distinguished from logical
necessity by the assignment of a temporal index proper to the modal-
ity; and it will be distinguished from real necessity by the exceptional
order of the indices relative to the modality and to the event.
Lines 6-7, 12 and 16 of the De Caelo passage add yet a further
determination to the principle of the irrevocability of the past. There it
is said that if one were to admit that there is a (contingent) potentiality
of the past one would be led to say of the now that it is last year, and of
the last year that it is now. Suppose in fact that we were to deliberate
about the fall of Troy. There would result the possibility both that Troy
should have fallen and that Troy should not have fallen, and since such
a possibility is open only to present action, the past time of the event
28 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

would have to come to be identified with the present and the present
with the past. In this way it would be denied that the past occupies a
different time from the present.
This text confirms the existence of a double temporal index of the
modalities, once they are used to characterize substances. The princi-
ple of the impossibility of realizing a possible in the past illustrates the
constraints affecting this double index just as the principle of condi-
tional necessity does. To say that a possible, which is forward looking,
has to do with the past, is to completely upset the order of time. It
comes down to postulating that the index of the possible at t\, which
must, by definition, be equal to or anterior to the index <2 of the event
whose possibility is in question, can, at the same time, be posterior to
it 16 .
16
This is attested to by the following Commentary of Simplicius: In L. De Caelo
I, 12, Ad 283i>6, 1894, pp. 355-356: "Having shown that the proposition saying
that a generated thing is indestructible implies that that thing has the capacity
of contraries' belonging to it at the same time, he goes on immediately to raise
the objection that can be brought against this line of reasoning. That which is
both generated and indestructible has the capacity of not being in the direction of
the past, since it was non-existent before being, and the capacity of being in the
direction of the future, since it is supposed indestructible. It is nevertheless not at
the same time that it will have the capacity of being and of not being, so that it
won't have the contraries in actuality either. But disposing of this objection, he
says that all potentiality is in the direction of present or future time. For above
all else we call possible the things which are not yet but are capable of becoming,
differing from existing things in that they will be but are not yet. If it is not true
then as regards anything to say now that it is last year or that it is not last year
(the two lessons exist = a./Kporepw; fap fpcuperai); for it is not true to say now
that the time of last year is, nor of any event that took place last year that it is
now, but neither was it true to say last year that now is a past of the time which
took place at the end of the year finished. In effect it is impossible to interchange
the times.
If this is then true, it is impossible that something which at some moment does
not exist should later be sempiternal, i.e., that that which was generated should
continue to be indestructible for the rest of time. Indeed, since that which after-
wards is was first inexistent it will also have, once having attained to being, the
possibility of not being, though not that of not being then when it has already
attained to being: indeed, at that moment it is supposed to be in actuality. It is
thus necessary that such an entity should have the potentiality of last year and in
the past. This is absurd, since there is no potentiality of that which has happened,
but only of that which is and of that which will be. Aristotle says even more clearly
moreover: 'let that of which it has the capacity exist in actuality: it will then be
true to say now regarding that which has the capacity now of not existing, not only
that it has last year the capacity of not existing, but even that it does not exist
last year'. What is even more absurd, the now itself is in the fact of not existing
last year. For last year will be now. Now is in effect supposed to have the capacity
of not existing last year. It is thus clear that the following reading is the more
consequent, namely, 'that it does not exist last year'. And it is in fact by these
 THE IRREVOCABILITY OF THE PAST / 29

Such a derangement of time leads to a physical impossibility and

it comes down to the same logically to say either that when true, the
words that he concludes: 'it will thus be true to say now that it does not exist last
year'. He nevertheless immediately applies the fact of existing last year to that
which is ungenerated though destructible. But it is right to say that there is no
potentiality of that which has happened: for the whole past is necessary (TO fa.p
irapeXri\vd6'; -nav 01/07notion) and is said to be neither possible nor contingent
(nal OUTE dwarbv OUTE ivbs.\6\iEvov \e~jeTca)".
The interpretation of premise A in the logic of the real modalities must fulfill the
following three conditions:
a) it must account for Epictetus' text saying that everything past and true is
necessary,
b) it must express the relation between deliberation and thus possibility ad
opposita, on the one hand, and the orientation of the possible towards the
future, on the other hand,
c) it must take into account the Aristotelian text saying that a possible oriented
towards the past would overturn the order of time.
Let us define what is now contingent as that which is such that it is possible now
that p take place at t and that p not take place at t:
CNPI —DS MNPI • MN ~ pt
(Definition of the contingent object of deliberation).
a) In leaving aside mention of the truth, which is implicit in the formulas, the
first condition can be written:
(1) p, • i < TV D LNpt
and, by symmetry:
(2) ~ pt • t < N D LN ~pt.
The logical product of the two has as consequence:
(3)t < N 3(LNPt V L j v ~ p t ) -
In effect,
h [(P.Q D R).(~ P.Q D 5)] D [Q D (R V S)],
with Sb: pt/P, t<N/Q, LNpt/R, LN~pt/S, 2 x Modus Ponens, (1), (2).
(4) (Lpjpt v LN ~ pt) =~ (~ Lffpf ~ LN ~ pt) =
~ (MN ~ pt • MNPI) s~ (Mftpt • MN ~ pt) =~ C^pt-
(5) t < N D~CTVP(.
b) The second condition is written:
(6) CNPt D TV = t.
The contrapositive of (6) is none other than (5). Consequently, to say that what
is past and true is necessary (Diodorus according to Epictetus) comes down to
the same as saying that there is no contingent but of the present or the future
(Aristotle).
c) The two expressions (5) and (6) are equivalent to:
(7) ~ (CNpt -t<N)
which is none other than the second term of the logical conjunction constituting the
principle of the irrevocability of the past in the De Caelo. This principle, henceforth
'A', contains however another term as well:
(8) (t)[(CNPt - t < N ) = (t<N-N = t)}.
To realize the contingent in the past is to completely upset the order of time in
saying that today is last year. In virtue of (7) the first member of equivalence (8)
is false, thus so is the second.
(9) (t)[t <ND~(N = t)}
30 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

past is necessary, as in the Master Argument's first premise, or that

it is impossible to realize a contingent in the past without upsetting
the order of time, as in the De Caelo, or again, that such a realiza-
tion leads to positing that time has two opposite directions, as in the
Nicomachean Ethics.
The following objection will come to mind. According to the princi-
ple of conservation of modal status every substance possesses an exis-
tence which is either immaterial and thus atemporal and absolutely
necessary, or sensible but sempiternal and thus necessary although
present in time, or, again, material, generable and destructible, and
thus contingent. But a contingent substance sometimes is and some-
times is not. In virtue of the principle of conditional necessity, when
it is, it is necessary while it is. When it is not, it is not with an equal
necessity and for the same reason. Then too, once it has been, it is no
longer possible that it be or it is no longer possible that it not be, in
virtue of the "necessity" of the past. The modality of the existence of
such a substance is thus subject to change, contrary to the principle of
conservation of modal status.
But what must be distinguished here is the modal status of a sub-
stance and the modality which expresses this status. The modal status
of a necessary sensible substance is expressed by the modality nec-
essary, but because that modality necessary expresses the essence, it
must be posited sempiternally. Due to the principle of conditional ne-
cessity, the modal status of a contingent sensible substance is expressed
at one time by conditional necessity, at another by conditional impos-
sibility; but these modal changes, relative as they are to a duration,
simply translate the immutability of the modal status characteristic of
In other words, the present is not eternal. Aristotle thus rejects, explicitly in (8),
and implicitly in (1) and (2) or in (6), the cyclical representation of time.
Intuitively, (8) expresses the case of the De Caelo: the negative possible or the
inexistence of the world resulting from creation is realized in the past, whereas the
positive possible resulting from the sempiternality of the world is realized in the
future; by symmetry, (9) expresses the case of the Master Argument. These two
statements are paradoxical because the two "pieces" of the contingent take opposite
temporal directions.
They become contradictory when the two pieces of the contingent are asserted
simultaneously. This is the case in the De Caelo, since, from the sempiternality
of the world results its positive possibility. It is the case in the Master Argument,
since, from the present and future non-realization of the possible results its negative
possibility.
(A) MNpt • (t < N • MN ~ pt) D (MNpt • MN ~ pt)
(simplification of the product)D~ (t < N) (by (4))
(A') MN ~ Pt • (t < N • MNpt) D~ (t < N).
This is the true sense of the 'necessity' of the past.
 THE SUBSISTENCE OF A POSSIBLE THAT IS NOT TO BE REALIZED / 31

contingency, a thing being contingent precisely if sometimes it is and

sometimes it is not.

2.7 The principle of the subsistence of a possible

that is not to be realized.
At lines 9 and 10 of the De Caelo passage Aristotle asserts that, be-
cause of the conservation of the modalities, this created world, which
was consequently non-existent in the past, conserves and will always
conserve the potentiality of not existing, but that since it does exist
and, by hypothesis, always will, this potentiality of not existing will,
because of conditional necessity, never be actualized. (That is why the
only means of actualization of this possibility to be found will be to
attempt to actualize it in the past).
The Master Argument's third premise will simply substitute for the
consideration of a negative possible, which by hypothesis will never be
actualized, that of a positive possible which, also by hypothesis, will
never be actualized either.
This third premise deserves further scrutiny, albeit because of only
a minor difficulty. In translating the Greek text as saying that the
possible is that which is not presently true and never will be so, one
might mean to say that no possible will ever be realized.
A possible that becomes actualized would falsify such a clause,
which is unjustified even as regards the possibles in utraque.
If one is tempted to accept this extreme interpretation, it is that
its negation is maintained by Diodorus when he says that that which
will never be actualized is impossible.
As regards the relations of the possible to actualization-inevitably
future, since there is no potentiality of the past-there are three concep-
tions that arise a priori. 1) One can, with Diodorus, treat the possibles
and the actualized futures (taking future here in the broad sense which
includes the present) as equivalent. This same thesis has been mistak-
enly ascribed to Aristotle because of a confusion between virtues which
verify and chances which falsify it.17 2) One can, with Leibniz, hold
17
Possibles ad unum are possibles which result from active irrational potencies (fire
can but heat). Possibles in opposita are possibles which result either from rational
potencies containing a deliberation (the doctor can give a remedy or a poison to
the patient) or from passive irrational potencies (air can be heated or cooled) (in
de Int., 22 6 36-23 a 5). Let me summarize Cajetan (in St. Thomas, Frette, 1875,
88-89; Oesterle, 1962, 217-221). These two sorts of possible are each defined with
respect to that which is susceptible of movement. The possible ad unum, however,
can, by equivocation, apply to the immobile beings of mathematics and of logic.
For that extended class of possibles (ad unum), the thing is said possible because
it is in act. Contrarywise, for the possibles in opposita, the thing is said possible
32 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

that the futures constitute a proper part of the possibles: they are the
events yet to take place of this world which was chosen by God, along
with the past and the present, as being the best. By possible, then, is
to be understood any idea in the mind of God. Those possibles that
are realized are realized by virtue of a "conditional necessity" which
is itself beyond the pale of chance and contingency and is thus fit to
retrograde, since it serves only to mark that which, in Creation, distin-
guishes the best from the non-contradictory. 3) One can again, with
Aristotle, distinguish two sorts of futures. Some have virtues that could
not not be realized. The others are contingent possibles. Among these
latter, some will be realized, but there remain others that will never
be realized. The range of possibles, then, is by no means exhausted
by the futures that will be realized. Diodorus wants to show that no
non-realized future is possible. But he need not for that suppose that
his adversary holds that all possibles are not and will not be; the fact
that Aristotle admits of the non-realization of some possibles is already
enough.
This is clearly the way Bayle understood the Master Argument's
third proposition. "The very famous dispute of possible things and
impossible things owed its inception to the doctrine of the Stoics re-
garding fate. The question was whether, among the things which never
have been and will never be, there are possibles, or if all that is not, all
that never has been, all that never will be, was impossible".18 And at
because it can be actualized (in de Int., 23a6-23°15). As for the natural (irrational)
possible ad unum, Cajetan (in St. Thomas, 1875, vol. 22, p. 89; Oesterle, 1962,
p. 219), in referring to the Metaphysics (Q, 2, 1046"36 sq.) remarks that, besides
their inapplicability to opposites, the active potency which is their raison d'etre
enters necessarily into operation as soon as the subject is present and impediments
are removed. Heat necessarily heats a material which is present, once all insulation
is removed.
To keep the original sense of the possible ad unum, that of a virtuality which
develops necessarily, once all impediments are dropped, let us call it a virtue. To
speak of the virtue of p is to say that it is possible that p ad unum. On the other
hand, let us call a possible in opposita chance. To speak of the chance of p is to
say that it is possible contingently that p.
Must we grant the axiom of necessity for virtue, while refusing it for chance
(23° 15-16)? If it is necessary that a material substance perish then it is possible ad
unum that it perish. This is in contrast to chance. If Socrates can be seated and
not be seated, in opposita, then there is no necessity of either one.
It must be added that a possible ad unum develops its actualization only on
the supposition of the existence of its subject. But that existence itself remains
contingent and has to do with possibles ad opposita. Once born, it is necessary
that Socrates die. That Socrates be born is contingent.
18
Leibniz, Gerhardt, 1978, p. 212; Jalabert, 1962, p. 223.
 RECONSTRUCTION OF THE DE CAELO DEMONSTRATION / 33

the end of the passage19 he goes on to add: "I think that the Stoics took
it upon themselves to give a greater extension to possible things than
to future things in order to soften the awful and odious consequences
drawn from their dogma of fatalism". Leibniz too gives reasons, and
different from Aristotle's, as to "why one is mistaken or at least speaks
very incorrectly when one says that there is no possible but that which
is or what God has chosen; this is the error of Diodorus the Stoic in
Cicero, and of Abelard, Wyclef and Hobbes among the Christians". 20
Cicero's text is indeed definite. In the De Fato (VII, 13) he inter-
prets this third premise as: "Some events which will not take place are
possible", when "Diodorus says 'That alone (solum) is possible which
is either true or will be true; whatsoever (quicquid) will not be is im-
possible' ".
The Master Argument's third premise will then be accorded the
same sense as the corresponding De Caelo premise—the sense of a
particular proposition: namely, there is a possible that will not be
realized. It is precisely because this premise is a particular that we are
permitted to reason on an illustrative case. In doing this, we will be
justified in construing the premise as a conjunction where it is affirmed
that a same event is possible which neither is nor will be realized.21

2.8 Reconstruction of the De Caelo demonstration.

What is meant to be shown in the De Caelo is an incompatibility
in the Platonic cosmological hypothesis according to which the world,
although engendered, is imperishable. And this is to be done by taking
into account only the principles of the conservation of modal status, the
possible realization of the possible coupled with conditional necessity,
and the impossibility of realizing the possible in the past.
In virtue of the conservation of modal status, what has been inex-
istent will conserve in the future the possibility of not existing. Such is
the case with the world. If the world has been engendered and is from
now on sempiternal, it is from now on sempiternal while conserving in
the future the possibility of not existing.
19
Leibniz, Gerhardt, 1978, p. 215; Jalabert, 1962, p. 225.
20
Leibniz, Gerhardt, 1978, p. 442, §22; Jalabert, 1962, p. 448, §22.
21
The De Caelo subsistence principle will thus be written:
(3p)[(3t)(~ LNpt -N = t ) - (t)(N = t D P t ) }
whereas the third premise of the Master Argument will result from the above by
simple substitution of ~ p for p:
(3p)[(3t)(~ LN~pt.N±t)- (t)(N i t D~ pt)].
34 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Examining the consistency of the Platonic hypothesis, then, comes

down to examining whether it is possible for the world to exist sempiter-
nally simultaneously with its conserving sempiternally the potentiality
of not existing.
According to the principle of the possible realization of the possible,
a thing's (here the world's) sempiternal capacity of not existing should
be able to be realized. It can and will be able not to be at every future
instant of time from "creation" on. Consider such an instant, say t\.
Given that a possible is situated by a double temporal index, ti and
<2, fix <i, leaving t% arbitrary. In other words,the possible realization
of the inexistence of the world signifies that at t\ it is possible that the
world should become inexistent at < 2 -
Instant t\ determines three cases of the universe, that is to say,
three possible worlds. Either (2, which situates the inexistence of the
world in time, is anterior to, identical with, or posterior to t\. This is
an exhaustive distribution of the universe.
Because of the impossibility of realizing a possible in the past, the
situation is physically contradictory if fa is anterior to t\. Again the
situation is contradictory on the hypothesis of the identity of fa and
ti, and this by virtue of the principle of non-contradiction, since it is
supposed that the world exists at ti. But the possibility Mtl ~ ptl
entails ~ ptl, by conditional necessity, and this is the contradictory of
Pt1. For there to be contradiction if fa is posterior to ti, it suffices to
consider the interval t\fa. We posit that it is possible at t\ that the
world not exist at fa. By the principle of contraction of the possible,
there is then a t on the interval tit? at which it is possible that the world
not exist at t. But because of conditional necessity this possibility is
realized. By hypothesis, however, the world exists at t. There is thus
a contradiction for we have simultaneously pt and ~ pt. If no case of
the universe representing the disjunction of all cases is possible, then
the universe itself is not possible.
Thus the realization of the possible, viz. of the possible inexistence
of the world, produces an impossible consequence. It is thus itself
impossible. In short, granted the principles of the possible realization
of the possible, conditional necessity, and the impossibility of realizing
the possible in the past, it is impossible to affirm at the same time that
the world exists and will always exist and that its inexistence should
become act.
In a second argument Aristotle demonstrates in symmetrical fashion
that an unengendered world that will perish is impossible. One gets
from the first to the second demonstration in changing the affirmations
into negations. Let us only point out that in this case the principle of
 RECONSTRUCTION OF THE MASTER ARGUMENT / 35

the impossibility of realizing the possible in the past takes on a positive

form: it is impossible to realize in the past, not now the inexistence
but rather the existence of the world.

2.9 Reconstruction of the Master Argument.

There are only two points of difference between the Master Argument
and the De Caelo passage considered. First, the Master Argument is
simpler, as it makes no use of the principle of conservation of modal
status. Secondly, the third premise of the Master Argument speaks of
a possible that will not be realized, whereas the Platonic hypothesis
specified that this possible was the destructibility of the world.
To fix our ideas, let us suppose the Master Argument's possible to
be situated at the present moment.
Then the event of which it is the possibility will be realized either
in the past, at present, or in the future.
In virtue of the principle of the impossibility of realizing a possible
in the past (premise A), the first case of the universe is excluded as
being physically contradictory.
Realize the possible then at present. According to the principle of
the possible realization of the possible, it is now possible that p now.
But, by virtue of conditional necessity, it is the case that p now. In
the third premise, however, it is supposed that not-p now. Thus the
second case of the universe is excluded by virtue of the principle of
non-contradiction.
Realize then the possible in the future. By the principle of contrac-
tion of the possible there is an instant t, on the interval between the
present and that future, at which it is possible at t that this possible be
realized at t. Thus, granted conditional necessity, there is an instant
t on the interval between the present and that future where p is the
case at t. But it had been supposed that the realization will never
take place. The third case of the universe is excluded by virtue of the
principle of non-contradiction.
With the three cases that exhaust the universe excluded as impossi-
ble, the universe, disjunction of these three cases, is itself contradictory.
Therefore, since the realization of the possible necessarily produces an
impossible consequence, there is an incompatibility between a possible
that will never be realized and that realization.
Thus a possible that will never be realized entails necessarily an
impossible consequence, if, along with the principle of conditional ne-
cessity and that of the possible realization of the possible, one admits
the principle of the irrevocability of the past. Diodorus turns against
36 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Aristotle the argument that Aristotle himself had developed against

Plato. Aristotle had combatted the creation and destructibility of the
universe. Diodorus combats the Aristotelian conception of a potency
that does not become actualized, that is to say, the representation of
future alternatives as required by the Stagirite to serve as a basis for
natural contingency and human freedom.
Against the validity of the Master Argument it has been alleged
that it uses simultaneously two different concepts of necessity: one
logical, the other real. But this objection turns out, in fact, to be
futile.
From the conjunction of the two premises A ("necessity" of the
past) and C (subsistence of a possible that does not become actual-
ized) together with the principle B of the possible realization of the
possible (that is to say, of the synchronic contraction of the diachronic
possible), it follows that for some instant t, idencical with or posterior
to now, it is possible at t that the event subject to this possibility be
realized at t, even though this event is not realized at t. But by condi-
tional necessity it follows that at t the event subject to this possibility
is realized at t, although C prohibits, by hypothesis, that that event
be realized at t. Thus we arrive at a conjunction of two contradictory
propositions without its mattering in the least that non-logical possi-
bles occur in each of them. The same remark can be made about the
impossibility of realizing the possible in the past. Even if A expresses
no more than a physical impossibility—that of reversing time—once it
is admitted, it does, for the past, preclude the possibility of contract-
ing the possible into the past and thus logically bars one of the three
alternatives constituting the universe of discourse. The conjunction of
C, B, A and of hypothetical necessity has thus a logical impossibility
as consequence. Therefore it is not logically possible itself. Q.E.D.

2.10 Sketch of a formal reconstruction of the

Aristotelian reasoning at De Caelo, I,
28366-17.
Axioms
The sign '<' is taken here as designating a relation which is simply
antisymmetric, reflexive, transitive and connected (a simple ordering
 FORMAL RECONSTRUCTION OF DE CAELO, I, 28366-17 / 37

relation).
D Ct0pt =D/~ Lto ~ pt- ~ LtoPt-
(a) (t0)[(3t')(t' < t0- ~ pt,) D (t)(t0 <t-~ Ltt
AA (t)(to){[(CtoPt •t<t0) = (t<t0-t0 = t)}-
(Ctopt • t < to)}
BA

CA (t0)(3p)[(3t')(t0 < t1- ~ LtoPt,) • (t}(t0 < t

* (t)(pt D~ Lto ~ pt)

2
Demonstration ^
= tDt0 = t V <

2. (O(0[(C4opt • « < < 0 ) = (t <t0 • t0 = t)}

(AA, Simplification, 1, Syll.).
3. (t0)(t) ~ (t < t0 • t0 ^ t)
(A A, Simplification, 2, h (~ P.P = Q) D~ Q.
4. ( t 0 ) ( t ) { ( C t o p t • t0 = t) D (~ Lto ~ pt • (3ti)[Mtl ~ P t l -
((t0 = t-t^tl<t0)V(t0^t-t0^t1< t))})}
(Add ~ L0 ~ p( • t0 = t in the premise and in the
consequent of BA).
5. («o)(0[(C to p t • t0 ^ t) D (3«i)(M 4l ~ ft, • «0 = ti = t)}
(4, Simplification, 3).
6. (t0)(t)(CtaPt -t0^t)D (3«! )(~ p tl .to = t!^t)
(5,NHA,\- [(P D Q.tf) • (Q D 5)] D [P D (5.^.)]).
22
I thank Mr. Jean Mosconi for his help.
38 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

7. (to)(t)[(CtoPt -t0^t)D (3*!)(~ ft, - 10 = h)}

(6, Simplification).
8. (t0)(3t)(CtoPt -t0^t)D (3«!)(~ p tl • < 0 = <i)
(7)
9. (*0)(3p){(3*')(*' < «o- ~ Pt>) D

((a) with simplification, CA,\- [(P 3 Q) • Q • R]

D [ P D ( Q - R)})
10. (t)(pt D~ Lt, ~ Pt)
(*)
11. (i 0 )(3p){(3«')(«' < t 0 - ~ Pt -) D

(9, 10, h {[P D {(3x)(Fa: • Gi) • (x)(Gx D

(x)(^x D J x ) } D {P D [(3x)(Fx • Jx • Gar)
(x}(Gx D
12.

(11, 8, h {(3p)[P D (Q - R ) ] - (Q D S)} D (3p)

13. ( t o ) ( 3 p ) ( 3 0 ( 3 * i ) ( * < t « , < « i - ~ P r ~ p t l - p t l )

(12)
14. ~((a)-AA.BA-CA-NHA).

2.11 Sketch of a formal reconstruction of the

Master argument.
Axioms
The sign '<' is to be taken here in the same sense as in the preceding
paragraph.

D Cwpt =D/~ LN ~ Pf ~ LNPI-

A (t) {[(CNPt -t<N = (t<N-N = t)}- ~ (CNPt -
B (*){~ LN ~ Pt D (3*i)[M tlPtl • (t < <i < N V JV < t <
C (3p)[(3<')(~ LN~Pt,-N< t') • (t)(N < t D
NH (t)(MtPtDpt)
* (t)(pt D~LN ~pt)
 FORMAL RECONSTRUCTION OF THE MASTER ARGUMENT / 39

Demonstration
tDN = tvN <tDN <t
(\~PDPVQ)
2- (t)[(CNpt • t< N) = (t < N • N < t)}
(A, Simplification, 1, Syll.).
3. (t)~(t<N-N<t)
(A, Simplification, 2, h (~ P • P = Q) D~ Q).

4. (t){(CNpt -N<t)D(~ LNPt • (3<! )[M tl p tl •

((N < t • t < ti < N) V (N < t • N < ti < t))})}
(Add ~ Lfjpt • N < t in the premise and
in the consequence of B).
5- (t)[(CNpt • N < t) D (3ti)(MtlPtl -N<t!<t)]
(4, Simplification, 3).
6. (t)[(Cf,pt -N<t)D (3ti )(ptl -N<ti<t)]
(5, NH, \- [(P D Q • R) • (Q D S)} D [ P D ( S - R)]).
7. (t)[(CNPt • N < t) D (3*,)( Pll • N < h)}
(6, Simplification).
8. (Bt)(CNPt -N<t)D (3*! )(Ptl -N<ti)
(7)
9. (t)(~ pt D~ LNpt)
(*,Sb~pt/pt).
10. (3p)[(3t')(CNPt. -N<t')- (t)(N < t D~ P t ) }
(C, 9, h {(3x)(Fx • Gx) • (x)(Gx D Hx) • (x)(Hx D
Jx)} D (3x)(Fx • Jx • Gx) • (x)(Gx D Hx)).
11. (3p)[(3*i)(p tl • ^V < «i) • (<)(^ < * D~ p t )l
(h [(3P)(P -Q)-(PD R)} D (3P)(72 • g), 10, 8).
12. ( S p K S i i K p t . - J V S t x . ^ p t J
(H)
13. ~(A-B-C-NH).
40 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

The following diagram will help the reader get a firmer grip of the
demonstration.

(B)
(Contraction)

-» v
N N

(A)
(Necessity
of the past)

(A,B)

(NH)
M P
(Conditional t t
necessity) P.t

(A, B, NH)
->t
N

(C)
(Existence of a
never realized T»t
possible)
Part II

Systems of Necessity: The

Megarians and the Stoics
A System of Logical Fatalism:
Diodorus Cronus.

This part of the book will be consecrated to an analysis of the different

responses which Epictetus tells us the Ancients made to the Master
Argument.
By means of a purely combinatory consideration Epictetus' text
deduces the possible responses. It amounts to the rejection of one
of the three explicit premises of the argument, the other two being
maintained. What we find thus laid out is an a priori system of the
history of philosophy.

Admitting one refuses

1. A and B C
2. B and C A
3. C and A B

The first response is that of Diodorus, the second that of Clean-

thes, the third that of Chrysippus. The history of the Megarian and
Stoic philosophies is to be found encompassed in this combinatorial
arrangement.

3.1 Diodorus' Solution.

If premise C of the Master Argument means that there exists a possible
such that it neither is now nor ever will be realized, Diodorus' solution,
which consists in denying that premise, must come down to affirming
that no possible is such that it is not now and never will be realized.

43
44 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Cicero's testimony confirms this interpretation. "For Diodorus then

the only possible is that which is true or will be true". 1
Premise C, a particular proposition, related the modality to tem-
poral determinations, without reducing it to them. It is this reduction
that Diodorus will undertake. In spelling out that the only possible is
that which is or will be true, Cicero furnishes the condition that per-
mits defining the modality in temporal terms. Indeed, that that which
is possible is or will be true stands as a necessary condition for some-
thing's being possible. Then again, since alone that is possible which
either is or will be true, being true now or in the future constitutes the
sufficient condition for something's being possible. When Boethius 2
writes that Diodorus defined the possible as that which either is or will
be, he explicitly draws out the consequence resulting from the negation
of C. For Diodorus the modalities have simply a derivative status. The
words expressing them are nothing but abbreviated ways of introduc-
ing complex temporal determinations; that is to say, they are logically
equivalent to conjunctions or disjunctions of such determinations.
Thanks to the interdefinability of the modalities, the definition in
terms of temporal determinations just laid down for the case of possi-
bility can be extended to them all. The impossible is that which neither
is true nor will be so. Since, according to Diodorus, every declaration is
true or false, the preceding definition could also be expressed by saying
that the impossible is that which is false and will never be true. The
necessary is that which is impossible not to be. Thus the necessary is
that which is true and never will be false. Finally, it is possible that
something not be when it is not necessary that it be. Consequently,
the possibly not is that which is false or will be so. These definitions
are attested to by Boethius.3 "Diodorus defines the possible as that
which is or will be, the impossible as that which, being false, will not
be true, the necessary as that which, being true, will not be false, and
the non-necessary as that which either is or will be false".
Diodorus' solution has two qualities in its favor.
First, it affords a means of defining the modalities and thus of
eliminating them as well.4 This renders modal sentences transparent
'Cicero, De Fato, IX (17); Brehier, 1962, p. 479; see above, 2.7, p. 31.
2
Boethius, 1880, p. 234.
3
Boethius, 1880, p. 234.
4
Mrs. Kneale points out that Diodorus' is the only non-circular definition of the
modalities in antiquity (1962, p. 125). There is nothing to prove however, that the
Stoics and Peripatetics thought of themselves as defining the modalities (Mignucci,
1978, p. 334, note 43). It is possible to interpret their circular definitions as simply
so many stipulations of the conditions that must be met by the modalities in their
relations to temporal determinations.
 DlODORUS' SOLUTION / 45

in translating them into general statements in which the bound variable

is time. The laws of modal logic are thus reduced to the syllogistic of
categorical propositions.
Secondly, the definition formally prevents any collapse of the modal
expressions that would reduce the possible to the actual and the actual
to the necessary. This in no way contradicts what has just been said.
The quantifiers serving to eliminate the modalities occur uneliminably
in the definiens. The modal declarations considered are singular. They
are translated into general temporal declarations, i.e. quantified dec-
larations; and these, in turn, are not reducible to singular non-modal
declarations.
If the possible is said to be that which is or will be, the actual
cannot be derived from the possible, for one member of a disjunction
(that which is) does not follow logically from the disjunction. Likewise,
if the necessary is said to be that which is true and will not be false,
the necessary cannot be derived from the actual, for a conjunction does
not follow logically from one of its conjuncts (that which is true). It
is right then to underline, as some have5, the formal capacity of the
Diodorean system to preserve that which is specific to the modalities.
The possible and the non-necessary, which are subcontraries, may both
be true at the same time. The predicate contingent can thus be formally
introduced in the Diodorean system. The contingent is that which is
possible and non-necessary, i.e. the logical conjunction of that which
is or will be and that which is not or will not be. As a result6 of this
definition the contingent is that which is not and will be or that which
is and will not be or that which will be and will not be.
The logical square of the modalities, according to Diodorus, sets
out an arrangement of two opposing categories. In the first place, the
necessary and the impossible contain two conjoined clauses. The first
of these fixes the present truth-value of the proposition about which
the modality is said to be. The second stipulates the exclusion, for
all future time, of the contradictory truth-value. Such a definition
could be compared to a complete induction in the mathematical sense
of the term. Assignment of a predicate to the first ordinal number
would correspond to the assignment of present truth-value, whereas the
'Blanche, 1965; Boudot, 1973, p. 440 and p. 444: to hold that possibility implies
necessity "would be to affirm that if a statement like 'Socrates is standing' is true
now or will be so at some moment of the future, then it is true and always will be
so
6 (p V Fp) • (~ p V F ~ p) =
(P- ~ P) V (p • F ~ p) V (~ p • Fp) V (Fp- ~ Fp) =
(p • F ~ p) V (~ p • Fp) V (Fp • F ~ p).
46 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

inductive clause would be extended from the non-continuous domain

of numbers to the contiguous domain of the instants of time. The
definitions in question lay it down that a declaration is necessary or
impossible when its truth-value does not change. The second modal
category, on the other hand, is compatible with a change of truth-value
though, given the sub-alternation relations between the modalities of
the two categories, it doesn't imply one. The contingent, defined as
a triple disjunction as stated above, specifies an inevitable truth-value
change for itself. For if the contingent is that which, being, will not be
or that which, not being, will be or that which will be and will not be
(this last excluding that something both will be and will not be at the
same moment), all three cases of the universe that realize contingency
imply a change of truth-value for the declarations expressing it.

3.2 Two possible interpretations as regards the

object of the Diodorean modalities:
nominalism and realism.
Up to this point our account has remained neutral. The term "declara-
tion" was used to designate that about which the Diodorean modalities
were supposed to be, without inquiring into the nature of that object.
But now such an inquiry is in order.
In scrutinizing the Diodorean definitions it will first be shown that
if the declarations do lend themselves to an interpretation as sentences,
they lend themselves equally well to an interpretation as propositions.
What will remain to be shown then is why the truth conditions set
down by Diodorus have given rise to yet another interpretation on
which the declarations would be construed as prepositional functions.
To begin with, Diodorus defines the modalities in assigning the
truth-value conditions. Since it is these values alone that are determi-
nant and since all the operations permitted in the linguistic construc-
tions are extensional, there is no need to retain distinctions that make
no difference whatsoever as regards these operations. An interpreta-
tion in terms of sentences here would come down to the same as one
in terms of propositions. We shall not insist on their distinguishing
characteristic, since they are in fact equivalent from the point of view
of the definitions at hand. We shall be further ahead remaining non-
committal in their regard, combining the two for the moment under
the heading of statement7.
7
A proposition is a class or sentences having the same meaning. The logical status
of this notion has fallen under fire because of the difficulties involved in identifying
propositions. This particular question does not have to be dealt with here.
 THE OBJECT OF THE DlODOREAN MODALITIES / 47

In the second place, once this subsidiary question is taken care of,
the Diodorean definition still leaves a doubt. In eliminating modality in
As a matter of fact, propositions have an essential occurrence only within opaque
contexts of discourse, i.e. in contexts liable to change their truth-value in the case
of substitution of expressions having the same denotation but not the same sense,
and particularly of sentences having the same truth-value but not belonging to the
same proposition. The modal operators are among these opaque contexts. For
example, granted the identity:
Scott = the author of Waverley
and the modal logical truth
necessarily Scott is identical with Scott
it is doubtful that the following sentence can be held true:
necessarily the author of Waverley is identical with Scott.
But Diodorus, in defining the modalities the way he does, does not overstep the
bounds or the means of expression characteristic of first order predicate logic. That
is a logic comprising only the connectives (and, or, not) and quantifiers over only
individuals and instants of time, which in turn guarantee its extensionality. All
opaque contexts are excluded.
Under these conditions it is only by virtue of their truth-values that sentences
enter into reasoning. Since sentences having the same meaning, have a fortiori
the same truth value, any proposition could be replaced by one or the sentences
belonging to it as to a synonomy class.
Why not then just settle for talk about sentences rather than insisting on the
neutral term statement"?
In the first place, Sextus Empiricus (Adversus Mathematicos, VIII, 11-13, and
132-139) speaks of the disagreement among the dogmatic schools on the subject of
truth. Some placed truth and falsity in the thing signified, others in the sound, and
others in the movement of the intellect. The first then distinguish the signifier or
sentence, the reference or existing thing, and the signified or "proposition"; and it
is to the proposition that they attribute a truth-value. This is the stance taken by
the Stoics. The Epicureans do not admit or meanings (the signified) between the
signifier and the thing. It is to the sentence that they attribute truth-value. Sextus
does not specifically mention the adherents of the third doctrine, which we would
call "psychologist" today. It can be seen that Aristotle doesn't hesitate to insist on
the importance of beliefs. In any case, accordingly as one considers an individual
movement of the intellect or a class of such movements, one falls back upon the
distinction between sentence and proposition, psychologically interpreted. We do
not know to which of these parties Diodorus belonged. It is possible that it should
have been to the Stoic one.
In the second place, in its original formulation, the Master Argument does use
modal operators; and it can be taken as refuting Aristotle only on the supposition
that such operators have propositions as argument. Even if the Diodorean defini-
tions free us from making that supposition, it is still a good idea to carry over a
trace of it in the polemical means of expression Diodorus used.
It is important for the sequel (especially 3.5) that propositions, in the sense here
adopted, are temporally indexed. For example, the proposition that I shall be in
Paris on January 14, 1995, as the class of sentences like 'Ich werde am 14. Januar
1995 in Paris sein', is at the future as are all the sentences of the equivalence class.
Therefore the truth-value shifts are the same for the proposition and for these
48 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

favor of a quantification over time, is the intention to turn the modality

into a quantification that would fix the satisfaction conditions for a
propositional function? 8 In this case every modal proposition would
be a particular or a universal general proposition. But is it also possible
to construe quantification as being over the truth-values of complete
statements, whether general or singular, and whether they be taken to
be sentences or propositions?
In constructing models for the two sorts of interpretation it will be
noticed that different consequences arise.
Consider the strong modalities first.
Suppose that necessity applies to a propositional function of which
the argument is time. For such a function to be true now and never
false in the future it is necessary and sufficient that it be true at the
present moment and at all future moments. The Diodorean predicate
'true now and never false in the future', translated in terms of quan-
tification over time, will have as consequence the temporal determi-
nation of the propositional function and its ineluctable transformation
into a proposition about a permanent state of affairs, at least from
the present moment on. The existence of the world on the Platonic
hypothesis would be necessary, for instance. On the other hand, ne-
cessity would have to be refused to past and finished occurrences not
only of such events as naval battles but even of such events as eclipses,
the very model of necessary events for the Ancients. Interpreted in
terms of propositional functions, the necessary signifies either a formal
implication ('snow is white (at all times)'), or a singular proposition
sempiternally true from an initial moment (now) on.
sentences. -A completely new situation would arise, were we to admit 'eternal
dated propositions' such as 'My being in Paris on January 14, 1995 is a fact', where
the word 'is' is given an intemporal meaning. Such an eternal proposition would
not admit of truth-value changes. In the following there will never be question of
such propositions.
8
A propositional function is derived from a singular proposition by the transforma-
tion of a constant into a variable. For example, from 'Peter is running' the function
'x is running' can be derived; or again, from 'It is raining now' the function 'It
is raining at time t' can be derived. A propositional function is different from a
proposition in that it has no truth-value. The corresponding concept for the func-
tion is that of satisfaction. For example, 'x is running' is satisfied by individuals
such as Peter and 'it is raining at time t' is satisfied by moments such as 'now'.
One more thing must be made clear. The propositional functions that could be
the objects of the Diodorean definitions are of the form /(TV, t), where TV standing
for Nunc designates a fixed term 'now', and £ is variable with respect to it. It would
be possible to construct propositional functions of two variables, with the position
of TV not fixed on the time axis. But since the texts of Diodorus eliminate the
interpretation in terms of propositional functions, we shall have no recourse to such
constructions.
 THE OBJECT OF THE DlODOREAN MODALITIES / 49

But now apply necessity to a statement. Necessity signifies the

stability of the truth of that statement from now on. The necessity
of the prepositional function entails that of the statement. But the
statement can be necessary without it's being the same for the cor-
responding prepositional function. Take the statement 'There was a
naval battle at Salamis', which is of the temporally indefinite kind with
a determinate subject. It was false before and during the battle, for
then the battle had not yet taken place. But as soon as the battle was
over it became necessary, for then it was true and has never since nor
will ever become false that the battle has taken place. On this inter-
pretation the movement of truth has nothing to do, properly speaking,
with the time of the event about which the statement is made.
This difference might be diagrammed roughly as follows. Let hor-
izontal lines directed from left to right represent the temporal and
'parallel' truth-axes respectively, with the point on them representing
the designated moment now. With a perpendicular raised on the time-
axis representing the occurrence of a verifying event, on the truth-axis
the truth of a statement, the necessary interpreted in terms of prepo-
sitional functions, gives rise to just one state of affairs:
TNP.(t)(t>N =>Ttp)

axis of truth

Now Now
Interpreted in terms of statement, the necessary gives rise to two states
of affairs. The first is the same as that given above, to which is added:
T N P.(t)(t>N z>Ttp)

II
Now Now

where the 'T' in the formulas is to be taken as 'it is true that'. Analo-
gous schemata, symmetrical for the false, would model the sense of the
two interpretations for the impossible.
The situation is similar with respect to the weak modalities. In-
terpreted in terms of prepositional functions, the possible signifies that
it is true or that it will be true that a certain event should occur: in
50 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

other words, that a certain event is occurring or will occur. Two states
of affairs correspond to this first interpretation:
P| T N pv(3t)(t>N.T tP )
I 4 > !
Now „ ! !

P|
ii -^ L
Now t Now
Any model in which the realization of the verifying event would pre-
cede the now is excluded. The interpretation in terms of statement
however is not subject to this limitation. Whatever the statement is,
thus whatever the real time at which its verifying event is seen to be
realized, that statement is possible if it is true now or at some ulterior
moment of time. But any genuine past proposition fits in precisely
with this condition. There are consequently three states of affairs cor-
responding to the interpretation of the possible in terms of statement.
The first two are identical with those given above, to which is added:
T N pv(3t)(t>N.T tP )

III
Now

Analogous schemata, symmetric for the false, would model the sense
of the two interpretations for the possibly-not.
It is seen then, in the case of both the strong and the weak modali-
ties, that if a declaration is necessary, impossible, possible or possibly—
not on the interpretation in terms of prepositional function then it is
a fortiori necessary, impossible, possible or possibly—not on the in-
terpretation in terms of statement. The inverse implication does not
hold.
In the two interpretations can be found the trace, so to speak,
that the two modal usages have left in the very definition that elim-
inates them. The interpretation in terms of prepositional function
corresponds to the modalities de re, which, since they attach to the
properties of things, fall within the quantifier's scope. The interpre-
tation in terms of statement corresponds to the modalities de dicto,
 THE OBJECT OF THE DlODORBAN MODALITIES / 51

where the object is a complete dictum. The first will be referred to

here as realist, the second nominalist, where these terms here shall
simply be taken as designating the two modal usages mentioned.9
Before examining the respective claims to validity made by the two
interpretations in the face of the Diodorean definitions of modality,
we shall make a detour through the Diodorean notion of implication
because of the deceptive light it could throw on the issue.
9
Take the following definition:
MX=Df(3t)(t>N-Xt)
If M is a quantifier then X is a prepositional function f ( t ) . But it is in nowise nec-
essary that this be the case: X may be taken as designating a complete statement.
The first interpretation is that of Mates (1961) who regards the objects of the
Diodorean modalities as propositional functions (p. 37). Independently of any his-
torical considerations, Russell appears to have expressly adopted such a conception
of the modalities (von Wright, 1979c, pp. 232-233).
The second interpretation is that of Boudot (1973, p. 442).
The two may be written as follows:
1. Modalities defined in terms of propositional functions:
MNPt = (3t)(t > N • pt)
MN ~ p ( =(3t)(t > N-~pt)
LNPt = ( t ) ( t > N Opt)
~ MNpt = (t)(t > N D~p t )
2. In terms of statement:
MNP = TNP v (3t)(t > N • Ttp)
MN ~ p = TN ~ p V (3t)(t > N • Tt ~ p)
LNP = TNP • (t)(t > TV D Ttp)
~ MNp = TN ~ p • (t)(t > N D Tt ~ p)
Commentators often juxtapose the two interpretations without distinguishing them.
Sedley, for instance (note 139, p. 116) has the following to say: "necessity belongs
only to propositions which can be stated truly both now and at all future times.
These must include, above all, eternal and analytic truths, but also certain state-
ments about the past (as in premise (1) of the Master Argument). By contrast, a
proposition is possible if it will express a truth now or at some future time". (It is
difficult to see how only certain statements about the past would be necessary).
Boudot reproaches Mates for conceiving the propositions (of the object language
itself) as if they contained a temporal variable. "For Diodorus", he adds, "the
statement is 'It is day' and not 'It is day at time t'. The date, if it must figure
in the expression, determines not the statement but the moment of its assertion;
and the temporal variables belong to the metalanguage, not to the object lan-
guage". Once granted that it is sentences that the modalities govern, the temporal
determinations of truth and falsity will be taken as belonging to the metalanguage
which will be carefully distinguished from the object language. The neutrality of
the statement, grouping both sentences and propositions, doesn't commit us to the
same extent. Excepting this last nuance, our presentation here owes much to the
clear and penetrating paper of Boudot.
52 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

3.3 The meaning of Diodorean implication.

It would be possible for partisans of the realist interpretation to invoke
the Diodorean theory of implication, arguing that, since it is a theory
of formal implication, it does indeed govern prepositional functions and
not propositions.
We shall examine the nature of this implication then and the false
appearances that have beguiled the partisans of prepositional func-
tions.
Philo10 admitted that a conditional is true if and only if the an-
tecedent is not true when the consequent is false. That is precisely
the conception of material implication taken up by the moderns. As
for Diodorus, according to Sextus' account , "he says that a condi-
tional is true whenever it neither was nor is nor will be possible for the
antecedent to be true and the consequent to be false, which is incom-
patible with the thesis of Philo. For according to Philo, a conditional
such as 'if it is day then I am conversing' is true when it is day and I
am conversing for, in that case, its antecedent 'it is day' is true and its
consequent 'I am conversing' is true. But according to Diodorus it is
false. For it is possible for its antecedent 'it is day' to be true and for
its consequent 'I am conversing' to be at the same time false, namely
when I have become silent. And it was possible for the antecedent to
be true and for the consequent to be false, for, before conversing, the
antecedent 'it is day' was true, but the consequent 'I am conversing'
was false".11
Unless the criterion of Diodorean implication is to be taken as cir-
cular, the modalities figuring in it must be spelled out in terms of
temporal determinations. We can say that a proposition p implies a
proposition q a la Diodorus if and only if it never has been the case
and it is not and never will be the case that p should be true and q
false.
But just what kind of formal implication is this? The two interpre-
tations demand a hearing once again.
Sextus gives us two cases of conditionals, valid according to Diodo-
rus, of which the first belongs to an inconclusive syllogism. Consider
first the Epicurean argument: 'If motion exists, the void exists'. As it
always begins with a falsehood and ends with a falsehood—Diodorus
denies both motion and the void—, it is in itself true; but as the
minor premise 'But motion exists' is false in turn, no conclusion can be
10
Sextus Empiricus, M, VIII, 113.
n
Sextus Empiricus, M, VIII, 115; Mates, 1961, p. 98.
 THE MEANING OF DlODOREAN IMPLICATION / 53

drawn.12 The second example is: 'If there exist no indivisible elements,
then indivisible elements exist'. It begins with the false and ends with
the true, and that disposition always holds.13
We could, to begin with, construe the expressions 'motion exists',
'the void exists', 'indivisible elements exist' not as statements in them-
selves but as prepositional functions. The Epicurean conditional would
then signify that for any time t whatsoever, if motion exists at that
time then the void exists at that time. The complete and exact form of
the expressions 'motion exists', 'the void exists', 'indivisible elements
exist' would be rendered by 'motion exists at time t\ 'the void exists at
time t' and 'indivisible elements exist at time t\ Diodorean implication
then would not be different from a universal proposition in terms of
times.14 As the binding of the variable would be over two prepositional
functions, the implication would express nothing other than a law of
physics in the modern sense of the term. Its model would be a propo-
sition of the type 'Snow is white' that would be rendered in temporal
terms as 'if something is snow at time t, it is white at the same time V.
Consequently, since the truth-values are fixed simultaneously for the
antecedent and consequent of the conditional, it would be excluded
that, at one and the same time, the antecedent could be true when the
consequent is false. And this is in conformity with Diodorus' requisite.
The specious character of this interpretation becomes obvious as
soon as we consider what Diodorean implication was meant to do.
Sextus has conserved two of Diodorus' arguments intended to block
objections made to him when he said that there is nothing which is in
motion but only something that has been in motion.15 It was objected
that "if preterites are true, it is impossible that their presents should
be false and they must be true; and similarly, the preterites must be
false when the presents are false".16 Among Diodorus' replies, two
have to do with the question of the role played by the movement of
12
Sextus Empiricus, M, VIII, 330-333 (Bury, II, pp. 412-413; Doring, fr. 143, p. 44).
13
Sextus Empiricus, PH, II, 110-111 (Bury, I, pp. 222-223; Doring, fr. 141, p. 43).
14
Mates, 1961, pp. 45-46.
15
Sextus Empiricus M.X.86 (Bury, III, pp. 252-253; Doring, 1972, fr. 123, p. 35).
As pointed out by Sedley (1977, pp. 85-86), the Diodorean passage here echoes
that of Aristotle showing that if the theory of indivisibles applied to time is true,
then there will have been motion without there being motion (Physics, Z, 231*
21-232a17, 24068-24106). If we go along with Sedley in thinking that Diodorus
is replying to Aristotle here, this correspondence suggests that Diodorus takes the
Aristotelian implication as holding and that, since he does admit the antecedent
(temporal indivisibles) as well, he is contesting the absurdity of the consequent,
which Aristotle had simply rejected.
16
Sextus, M, X, 91 (Bury, III, pp. 255-256; D6ring, 1972, p. 35).
54 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

truth. To begin with,17 he says that preterites may be true when their
presents are false. The past-tense proposition 'these men married' may
be true at one and the same time when the corresponding present-tense
proposition 'these men are marrying' has never been true. It is suffi-
cient for this to be so that the men in question should have married
at different moments of the past. The correspondence would require
that they should have married simultaneously. The second argument18
is of the same sort. The past-tense proposition 'Helen had three hus-
bands' may be true, because Helen was married three times, without
it's ever having been the case that the present-tense proposition 'Helen
has three husbands' is true, since the three marriages in question were
successive.
Independently of whether these arguments are sophistical, as Sex-
tus maintains,19 they prove that Diodorus does systematically use com-
plete temporal statements as antecedents and consequents of the con-
ditional. Formal implication over time has nothing to do with the time
proper of the propositions but only with that of their truth values. The
following schema illustrates why the conditional of the second example
is invalid. It could only be valid if the antecedent was never true when
the consequent is false.
,t 0
first marriage second marriage
The antecedent is the past-tense proposition about the conjunction
of the marriage at time t\ and the marriage at time fa. It is true
immediately after the accomplishment of time t^ and will never be false.
The consequent, on the other hand, is never true; it says that there was
a moment of the past at which both marriages were simultaneous.20
To say that there has been a first marriage at time t\ and a second
at time t? is to utter authentic statements, and not to speak in terms
of prepositional functions as the first interpretation would have had
it. Formal implication has to do with the truth, which varies with the
time of these statements. Formal implication here is nominalist.
What is more, Diodorean implication, as opposed to a formal im-
plication that would directly involve propositional functions, must be
17
Sextus, M, X, 97 (Bury, III, pp. 258-259; Doring, 1972, pp. 35-36).
18
Sextus Empiricus, M, X, 98 (Bury, III, pp. 258-259; Doring, 1972, p. 36).
19
Sextus Empiricus, M, X, 99-100 (Bury, III, pp. 260-261; Doring, 1972, p. 36).
20
Formally, the falsified implication is the following:
 THE MEANING OF DlODOREAN IMPLICATION / 55

able to connect an antecedent and a consequent situated at two dif-

ferent moments of time. This is the case when one tries to bring into
correspondence the present truth of a past motion with the past truth
of that motion as present.
It is owing to this temporal peculiarity of formal implication that
Diodorus distinguishes himself from the ancient Eleatics in leaving
some room for motion,21 quite like he left room for the possible de-
nned in terms of time, thereby distinguishing himself from the ancient
Megarians. To do this, he had to admit a succession of states while
rejecting a change of state. But to admit the first and reject the second
comes down to denying that the implication of the second by the first is
true, since to the truth of a past motion there is no corresponding truth
of a present motion that one could have apprehended in its passage.
The Diodorean definition of implication is closely tied to the concep-
tion of indivisible instants of time—a conception mentioned by Sextus
ad nauseam.22 The clearest instance of this conception is probably
that of the wall, collapsed without ever collapsing. "If the wall col-
lapses it collapses either while the stones are touching and fitted to-
gether, or when they are disjoined. But it is neither while they are
touching and fitted together, nor when they are disjoined that the wall
collapses; therefore the wall doesn't collapse". Sextus' commentary is
clear. "Two times are conceived of, that in which the stones touch
and are fitted together and that in which they are disjoined, and, aside
from these no middle time can be conceived. If the wall collapses then,
it must be in one or the other of these times", but that cannot be in
any case.23 The present, conceived of by Aristotle as a breadthless
mobile limit, disappears. If it is different states of the world that are
observed, and it is, then it is impossible to apprehend or to posit an
instantaneous change of state in being. Time is made up of indivisible
parts. One can then say that there has been motion without being
committed to admit of present motion, i.e. instantaneous motion.
In defining implication in the way he does, what Diodorus is really
doing is attacking the Aristotelian definition of motion as the fulfille-
ment of what is potential as potential. According to Aristotle, po-
tentiality is present as privation in the moving thing. In denying the
infinite divisibility of time, Diodorus is rejecting this kind of potential-
21
Sextus, M, X, 48 (Bury, III, pp. 234-235; Doring, 1972, fr. 122, p. 34).
22
Sextus, PH, II, 242 (Bury, I, pp. 312-313, and the retort of doctor Herophilus);
PH, II, 245 (Bury, I, pp. 314-315; Doring, 1972, fr. 127, p. 37); PH, III, 71 (Bury,
I, pp. 376-377; Doring, fr. 124, p. 36); M, X, 87-88 (Bury, III, pp. 252-253; Doring,
fr. 123, p. 35); M, X, 142-143 (Bury, III, pp. 282-283; Doring, fr. 125, p. 36).
23
M, X, 347-349 (Bury, III, pp. 378-379; Doring, fr. 126, p. 37).
56 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

ity. The future does not haunt the limit-present in the guise of a need.
It will be, or else it already is. There is time, but no becoming. There
is the generated and the destroyed, but no generation or destruction.
From the truth of the fact that there has been motion, it cannot be
concluded that it has been true that there was motion then—or, again,
that it is true that there was a passage. The whole demonstration
hinges on an implication of which the antecedent says that there has
been motion and the consequent that there has been passage. Those
are both complete statements: the first true, the second false. The
formal implication that fails to establish a connection between them
has to do with their truth-values and not with their content.

3.4 Diodorean nominalism.

In his definitions of the modalities, the conditions that Diodorus im-
poses on the truth-values have to do with their behavior over time.
Given what they are presently, what they will be is assigned. This
behavior is thus relative to the 'egocentric particular' now. But the
temporal reference of a declaration is itself susceptible of three differ-
ent degrees of precision. The time may be indefinite: Til undertake
this task one day'. Or the time may be linked to the present moment
of the declaration by the indication of a distance fixed by recourse to a
calendar: Til undertake this task in the second week of next month'.
Or, finally: the date of the event is fixed in relation to a chronology
within which is situated the experience of the speaker himself: Til un-
dertake this task on January 1, 1995'. The reference to time gives rise
in the first case to a grammatical tense, in the second to what is called
a pseudo-date,24 in the third to a date. It is only the dated declaration
that doesn't depend, apparently at least, on the assignment of now.
That is what has made it possible to think of it as 'eternal'.25

24
To use Rescher's expression, 1971.
25
Accordingly as dated declarations are made before, during or after the dated
event described, some languages maintain a reference to the time of speaking in
the verb of what is said. In such languages, this recourse to conjugation is seen as
encompassing the variations necessary for establishing the synonymy of statements
in virtue of the rule: accordingly as a statement about a dated event is made
before, during or after that event, the statement must be in the future, present or
past tense. The logical status of the declaration though is not affected by these
necessary grammatical modifications.
A date can only be determined with respect to a point of origin. The question
will still arise whether there isn't ultimately need of an "egocentric particular" for
fixing the point of origin. We won't go into this question here, as its solution would
have no bearing on the interpretation of the Diodorean definitions.
 DlODOREAN NOMINALISM / 57

The plan the inquiry must follow lies traced before us. We must
determine the respective merits of the two interpretations distinguished
in section 3.2, first for the case of indefinite declarations, then for the
case of dates and pseudo-dates. It is the first case that presents the
major difficulties.
In cases where the modality governs an apparent present or a sub-
junctive, as in English and French, one will be tempted to think that
the realist interpretation is imperative: that the modal expression is
derived from a prepositional function and that the movement of truth
will have then determined the time of the event. Instead of indicating
the simultaneity of the event with the time of utterance, this present
and this subjunctive will be taken as indicating the simultaneity of
the event with the time induced on the variable of the prepositional
function by the truth predicates or directly by a grammatical tense.26
Quoting the Diodorean definition of the possible, Alexander of Aphro-
disias does away even with the predicates of truth and falsity, assign-
ing the time directly from a non-temporal verbal form. "According
to Diodorus it is possible for me to be in Corinth in the case where
I am in Corinth or in the case where I will go to Corinth. But if I
were never to be in Corinth, that would not have been possible. And
to become a grammarian is possible for a child, if it becomes one one
day".27 There is no room here for ambiguity. There is a possible when
the thing presently is or will be realized.28 Boethius also reports the
Diodorean definition of possibility in purely objective terms on two dif-
ferent occasions,29 without any mention of the truth-predicate. That
26
To designate the object of the modalities the definitions employ a neuter pronoun
(Cicero and Boethius use quod, Plutarch onep Alexander o). When any precision
is furnished it is by way of translation into a substantified infinitive (TO -fevecrdat.,
says Alexander). In short, one is tempted to think that where the French and
English employ the subjunctive ('il est possible que j'aille a Paris', 'it is possible
that I should go to Paris'), the Ancients use expressions that exclude a determinate
tense, consequently appealing to propositional functions determined by the truth
predicates, for want of a direct translation in terms of grammatical tense.
27
1883, p. 184.
28
Plutarch (De Stoic, rep. 46, 1055 D-E; Doring, 1972, fr. 134, p. 40; in Plutarch's
Moralia, 1976, XIII, 2, p. 589), Philoponus (1905, 169, 17-21; Doring, 1972, fr. 136,
p. 41), Simplicius (1907, 195, 31-196, 24. Doring, 1972, fr. 137, p. 41) all record
the Diodorean definition of the possible in objective terms of that which is or will
be. Cicero, in the De Fato (6, 12, 7; Doring, 1972, fr. 132 A, p. 39) links the two
definitions. Diodorus, he says, "says that that only can happen which is true or
will be true, and all that is future, he says that it is necessary that it happen, and
all that is not future, he denies that it may happen". The time that counts for
characterizing the modality is that of the event, not that of truth.
29
Boethius, 1880, 234, 10-235, 9 and 412, 8-21 (Doring, 1971; fr. 138-139, pp. 42-
43). The commentary of Boethius leaves no doubt as to his interpretation of the
58 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

is possible which is or will be, says the definition. The possible then
assigns the present or the future to an indefinite declaration in the
apparent present, which is nothing but a propositional function.
Those who would settle for this semblance, however, would be mak-
ing a double mistake. Historically, they would be taking over arguments
of the commentators close to Aristotelian realism to interpret Diodorus,
whom we saw staunchly refusing to admit the notion of real potential-
ity. Logically, they would be forgetting that a realist possible justifies
a nominalist possible a fortiori. Take the statement 'I am now or will
be in Corinth'. To say that that statement is possible is to say that it
is true now or that it will be true that I am now or that I will be in
Corinth. The truth conditions on the statement are the same as the
satisfaction conditions that were on the propositional function.
To decide between these interpretations two conditions must be
met.
1. Cases must be found in which the realist interpretation errs by
defect or, what comes down to the same thing, cases where the nom-
inalist interpretation errs by excess. The discovery of such cases is a
necessary condition for distinguishing between the two interpretations.
It would not be a sufficient condition however if one could impute an
equivocal usage to Diodorus.30
2. This supposed ambiguity, which moreover contradicts the rep-
utation Diodorus enjoyed,31 loses all verisimilitude if we show that
it upsets the relations of the 'logical square of opposition' that the
Diodorean definitions of the modalities lead us to expect.32
matter. "While there are two principal parts of the possible: one which is said of
that which, though non-existent, can nonetheless exist, the other which is predi-
cated of that which already is something in act and not just potentially, this sort
of possible that already is in act gives rise in turn to two further sorts of possi-
ble: one which, though existent, is not necessary, the other which, existing, further
renders this possible necessary. And it was not only the subtlety of an Aristotle
that recognized this fact. Diodorus too defined the possible as that which is or will
be. Thus Aristotle holds possible that which Diodorus calls 'future', which, while
it doesn't exist, can nonetheless come to be. As for what Diodorus called 'present',
it is what Aristotle interprets as being possible, which is said possible precisely
because it already is in act". Whatever the validity of the comparison Boethius
makes between Diodorus and Aristotle—for in speaking of a possible that is not
but can be, Aristotle is not automatically talking about what will be—there is one
thing that is sure: Boethius does take Diodorus to hold that the possible is that
which is or will be in act.
30
According to Mates (1961, p. 39), the attribution of necessity in the Master
Argument's first premise requires taking the word 'necessary' in another sense than
that used in the definitions of the modalities, or of the weak modalities at least.
31
Sedley, p. 103. See above ch. I, note 22, p. 10.
 DlODOREAN NOMINALISM / 59

It is easy to show that the realist theory is in the wrong. If we

begin with the function 'there is a naval battle at Salamis at time t',
the Diodorean definitions fail to assign its modal status. Since such a
function is satisfied neither at present nor in the future, it would be
impossible at most. The nominalist interpretation, on the other hand,
is legitimate, since the statement 'there was a naval battle at Salamis'
is true now and will never be false and is thus necessary.
The realist interpreter is thus reduced to accusing Diodorus of am-
biguity. But once interpreted in terms of statement, the necessity of
the past takes its place alongside the so-called definitions in terms of
propositional functions, with the result that the possible no longer fol-
lows logically from the necessary. 'There was a naval battle at Salamis'
is surely a statement that is true and will always be true: but it cannot
be said to express the present or future act of an event.33
At this point we should examine how declarations about the future
fare with respect to the two interpretations. Let us first derive 'Fabius
will perish at sea' from the propositional function 'Fabius will perish
at sea at t\ If the statement derived from such a function is true
then, since the death of Fabius represents simply a unique event, the
statement will not be necessary. It will therefore be contingent. If
the statement is false it will always be false, and is thus impossible.
From the nominalist point of view on the other hand, where what we
begin with is the complete statement 'Fabius will perish at sea', if the

Necessary Impossible

[is true and will not be false) < tries (is false and will not be true)
SB SB
%>. -^P «
•Jx ,fN C
<X rv°
•y^'JO |-OJ

3
CO to
Possible Non-necessary
(is true or will be true) subcontraries (is false or will be false)
Square of Modal Opposition according to Diodorus.

(Kneale, 1962, p. 125). Of two contradictories one is true, the other false. Two
contraries cannot both be true, but can both be false at the same time. Two sub-
contraries cannot both be false, but can both be true at the same" time. There is a
relation of descending implication between subalternates: the necessary (impossi-
ble) is possible (non-necessary). A realist expositor like Mates agrees that the laws
of the logical square must be respected (p. 37).
33
See next page.
60 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

statement turns out to be true it will become false once the event is
accomplished, since from that moment on the death of Fabius at sea
will take place no more. If the statement is false it will always be false
and is therefore impossible. Both interpretations are seen to agree
on the two cases: impossibility in case of falsehood, contingency in
case of truth. Nevertheless, the contingency for the realist is based on
the unicity of the event, whereas for the nominalist it is based on the
changing truth-value. Both interpretations will confuse the case of the
eclipse with that of the death of Fabius in the same treatment. But
the nominalist confusion, based simply on the movement of truth, is
due to a more superficial reason and is therefore more acceptable than
the realist confusion that puts the two events on a par.
From the realist point of view 'There was a naval battle at Salamis'
is an assertion which stricto sensu is verified neither now nor in the
future:
^ [ I 1 i i I I ' 1 [ I I ' I I i ! ! I [ I i * t

p N
It is therefore impossible and entails a negative possibility. From the
nominalist point of view it is necessary and entails a positive possibility,
given that its truth-value doesn't change any more after the event.34
34
The nominalist interpretation may be written as follows:
(3t)(t<N-Pt) DTN[(3t)(t<N.pt)].
(t')(t'>NoTt,[(3t)(t<N.pt)])
= Lnom(3t)(t <N-pt)
3 Mnom(3t)(t <N-pt)
= TN[(3t)(t<N.pt)]
V(3t')(t'>N-Tt,[(3t)(t<N-pt)]).
(The symbols 'Ln0m' and 'M nom ' are here taken as designating nominalist necessity
and possibility respectively, as 'Lreai' and 'A/ reo /' will be taken as designating the
corresponding notions on the realist interpretation).
If, on the realist interpretation:
Mr€alP5PNV(3t')(t' >N-pt>),
the formula expressing that p is past is substituted for p:
Mreal P = [(3t)(t < N •Pt)]N V (3t')(t' > N • [(3t)(t < N •Pt)]t, )

the resulting formula will illustrate the Aristotelian De Caelo passage: it will be
true to say now that next year is last year. All semblance of paradox vanishes when
the invalidity of the conditional
*Lnom 3 A/reaf

is brought to light.
The realist can, with Aristotle, demonstrate the 'necessity' of the past in his
language. He will posit:
t < N D~ MptV ~ M ~ PJ,
 DlODOREAN NOMINALISM / 61

The realist interpretation seems to fare better once we come to

consider dates or pseudo-dates. Even though the prepositional func-
tion is satisfied without any indetermination as soon as the Now is
fixed, it remains contingent because of the fact that it is not satisfied
either always or never. There is a price to be paid for this univer-
sal contingency though, since the complete determination of the dated
or pseudo-dated occurrence of an event—an eclipse, for instance—will
not prevent it's being contingent. On the nominalist interpretation all
dated and pseudo-dated statements are, on the contrary, necessary or
impossible. If I say that it is day now or that it will rain tomorrow and
my statements are true, they cannot become false, and if they are false,
they cannot become true.35 Here, even the statements that common
sense would take as contingent ('It will rain tomorrow', 'It will rain in
Paris on May 20, 2000') are transformed by nominalism into necessary
or impossible ones.
Should we then, in the case of dates and pseudo-dates, follow the
middle-of-the-road lead of common sense, giving equal weight to the
two interpretations that entail complementary paradoxes? To do so
would be to forget that Diodorus challenges the position of common
sense regarding the paradoxes, and that there is no balance to be had
and if the prepositional function was satisfied at t, then
~ M ~ pt-
The fact will remain however that this necessity is not in conformity with realism,
for it does not entail realist possibility. On the contrary, it logically excludes it.
35
The real present is a pseudo-date since it asserts the simultaneity of the utterance
with the event. To say that the statement 'It is now day' will be false tonight would
be to render variable both the now and the distance separating the event from the
now. The synonymy class of the statement 'It is now day' includes all sentences of
the same form—in English or in another language translation—and uttered during
the lapse of time fixed by the chosen unit of measure of the present (the hour, for
instance). The sentences that will belong to the same class five hours hence are:
'Five hours ago it was day'. These sentences are true and will never become false
(in so far as they incorporate the appropriate compensation with respect to the
unit chosen). To overstep the bounds of the chosen unit, though, in saying that
the statement 'It is now day' will be false in five hours, would be to employ the
statement ambiguously and to violate its strict identity conditions which require
that the sentences be appropriately modified.
Those who speak in terms of sentences will have to be content with saying that a
sentence is tacitly bound to its moment of utterance. To demonstrate the necessary
or impossible character of dated sentences they will use a formal detour. With 'm'
designating the number of units in question, they will posit:
=
Pin the present tense ™™Fmp.
Admitting the principle
PmP D LPmp,
the conclusion Lp immediately follows (Boudot, 1973, pp. 450-451).
62 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

between realism's false contingencies and nominalism's shocking ne-

cessities. That necessity should attach to present and future-tense
propositions, once unambiguously dated or pseudo-dated, is something
that will appear philosophically legitimate if, along with Diodorus, we
reduce contingency to the indetermination of the temporal occurrence
of events. Far from striking a balance, as common sense would have
it, the nominalist and realist paradoxes in fact add up to a corrobora-
tion of the nominalist interpretation, since Diodorus' principles oblige
acceptance of the former and rejection of the latter.
Finally, there are all those passages of Epictetus and Cicero in which
Diodorus explicitly applies truth-values to closed statements.36 There
is no way of avoiding the nominalist interpretation here. The same is so
when truth-values are attributed to particular ('There are indivisible
parts') and universal ('There is no motion') general propositions, of
which we know nothing, moreover, beyond the mere examples recorded
by Sextus.37 There can be no doubt. Only nominalism conforms to
the letter and the spirit of Diodorus.
It must still be explained however why the realist interpretation is
so attractive wherever there is question of a genuine 'possible' and why
it is only the indefinite character of propositions that blocks necessi-
tarianism.
In the absence of textual evidence there will be something arbitrary
about all our conjectures. But the dispute over Diodorus' necessitari-
anism will amply show that we could not do without such conjectures.

3.5 Diodorus' necessitarianism.

The Ancients, like the classical philosophers, thought that in denying
the Master Argument's third premise Diodorus was implicitly commit-
ting himself to necessitarianism.
The nominalist interpretation, on the other hand, which has the
merit of showing that Diodorus stood out against the early Megarians,
the enemies of all possibility,38 maintains that as long as the propo-
sitions involved are neither dated nor pseudo-dated Diodorus is in no
36
Cicero, De Fato, VI, 12; IX, 17.
37
Sextus Empiricus, PH, II, 242 (Bury, I, pp. 312-313); Sextus Empiricus, PH, II,
110-111 (Bury, I, pp. 222-223; Doring, 1972, fr. 141, p. 43).
38
Aristotle, Metaphysics, 0, 3, 1046629-32: "There are those philosophers, the
Megarians, for example, who hold that there is no potency except in act, and that
where there is no act there is no potency. Thus he who is not building has not the
power to build; but only he who builds has, at the moment when he is building
(OTIXV oixoSonf))."
 DlODORUS 1 NECESSITARIANISM / 63

way committed to necessitarianism. He must admit to it however once

there is question of these sorts of proposition.
Let us examine the arguments of both parties.
That of the Ancients would be given the following form. 39 The
Diodorean definition of the possible is to be taken to mean that "only
that is possible which either already is the case or will be the case.
The consequence of this definition is that all future events are neces-
sary. For, since their contrary is not already present and will not be,
their contrary is not possible. But that of which the contrary is not
possible must itself be necessary. Therefore, all future events would be
necessary".
It is only the detour by way of the absurd that renders the reasoning
acceptable here. Supposing a unique possible true now or at some
moment of the future, its negation is false now and at all moments of
the future. Remaining false, it is therefore impossible; and in virtue
of the interdefinability of the modalities, the possible with which we
began is necessary.
This is surely the way me must take Cicero's comments on Diodo-
rus. "If, he says, addressing himself to Chrysippus, you sanction the
prophecies of the deviners, you will reckon among impossible things
events falsely predicted as to the future, and as to veracious predictions
of future events that are to occur, you will maintain that these are
necessary. But this is precisely Diodorus' view, completely opposed to
your own". And in giving the example of a necessary causal connection
he is careful to put the antecedent in the past ('You were born at the
rising of the dogstar') in order to render it necessary in conformity
with the Master Argument's premise A, and careful too to give the
consequent a negative form ('You will not die at sea') to avoid the
possibility of a truth-value shift. 40 The negative of the consequent
here connects the objective necessity of the event (not dying at sea)
with the grammatical necessity of the statement (you will not die at
sea), since this statement is and will remain true.41 But Cicero takes
great care in commenting on Diodorus' chapter, "Possibles", to say
that "nothing happens that was not necessary; whatever can happen
either is now or will be; propositions about the future can no more,
from true, become false than can propositions about the past; but as
regards the latter, the impossibility of changing them is apparent, and
39
M. Frede, 1974, p. 115.
40
Cicero, De Fato, VII, 13, 14; Brehier, 1962, p. 478.
41
The same holds true for the example discussed in De Fato, VI, 12; Brehier, 1962,
p. 477.
64 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

since it is sometimes not apparent as regards propositions about the

future it just doesn't seem to be there".42
Putting it this way, Cicero takes no pains to show that a propo-
sition about the future has a constant truth-value even though it be
future, whereas a proposition about the past has a constant truth-
value because it is past. He simply shows that a difference of degree
in the specification of the temporal determination of the future makes
no difference of degree as to the necessity of the proposition. He thus
takes no account of the grammatical form of the future which is the
cause of truth-value instability. When a Bayle or a Leibniz takes up
the question again they spontaneously adopt a point of view that ex-
cludes from the outset both ambiguity of designation and grammatical
cavils. They don't even discuss the matter and identify Diodorus' phi-
losophy with the system of most absolute necessity. There is only a
nuance, evanescent at that, to distinguish the immobilism of the An-
cient Megarians from the logical fatalism of Diodorus, for they both
agree in recognizing in each essence "in some sense, a universe complete
and closed upon itself, which contains from all eternity all that will be
in the course of its duration: it is fated, said Diodorus, either that you
reap or that you not reap; it cannot be but the one or the other, be-
tween the two contradictories there is no middle; therefore, necessarily,
whatever happens, either you will reap or you will not reap".43
Nominalism's response to this accusation is to argue that it is based
on a sophism. If the possible is interpreted in terms of statement,
every future, true or false, will undergo a truth-value change and the
realist's detour by way of the false is illusory. Granted, every possible
will be necessary, since it will be transformed into a past. But it is not
necessary. A possible at moment t(t > N) has always been possible,
since at every moment of the past it has always been true that it will
be. It can be concluded from this that if something is possible it has
been necessary that it be possible,44 but not that it is necessary that
42
Cicero, De Fato, IX, 17; Brehier, 1962, p. 479.
43
Robin, 1944, p. 120, who points out the connection between the Reaping Ar-
gument and the Idle Argument. The Reaping Argument is set out by Ammonius
(1961, pp. 251-252) as follows: "If you are going to reap, it is not that you are
going perhaps to reap, perhaps not, but that you are going to reap in any case
(modis omnibus); consequently, it is of necessity that you are either going to reap
or not going to reap". And Ammonius comments: "What is eliminated then is the
'perhaps' itself, to the extent that it obtains neither according to the opposition
between what it is to reap in the future and what it is not to reap in the future,
since necessarily one of the two will occur, nor according to the consequent referred
to either one of the two suppositions".
44
The Diodorean system belongs to what the moderns call the £4 modal system.
 DlODORUS' NECESSITARIANISM / 65

it be possible, nor a fortiori that it is necessary. Retrogradation of the

truth does not, therefore, change the possible into the necessary. As to
the discreteness of time, if something is necessary at time t, it allows
us to say at time t + 1 that it always will be true. But as that is a
future truth no conclusion can be drawn from it as regards time t - 1
and it may be that that thing was not the case at t — I . Necessity,
therefore, does not retrograde.
We're left with an argument however that proves too much and not
enough.
It proves too much, for all possibles that will be realized must be
seen as contingent.
It proves too little, for once a statement is temporally determinate
it becomes necessary.
These two flaws seem moreover to be related to a self-same de-
fect of the Diodorean doctrine regarding contingency. For according
to Diodorus, a statement that changes truth-value is contingent. But
there are two possible interpretations of such a change. When it is said
that the closed proposition 'Sometimes it rains' is contingent, what
is meant is that simultaneously (it is true now or it will be true in
the future) and (it is false now or it will be false in the future) that
it is raining now. The contingency then is due to the succession of
occurrences of showers and occurrences of dry spells. The realist inter-
pretation is also legitimate in this case since the satisfaction conditions
of the prepositional function 'It's raining at V will have the same con-
sequences as the truth conditions of the statement 'It sometimes rains'.
But consider the statement 'Fabius will perish at sea'. Here again real-
ists and nominalists agree in holding it to be contingent, but this time
for entirely different reasons. The realist will invoke the unicity of the
verifying event, but will deny that the proposition will have gone from
being true to being false. The nominalist will keep track only of the
grammatical tense shift to attribute a truth-value change to the propo-
sition. The nominalist interpretation thus resorts to a verbal quibble
to preserve by excess the contingency of all indefinite futures, and it is
also this verbal recourse that explains why all determinate futures are
turned into necessaries. The prestiges of the realist interpretation are
understandable. The truth-value shift of statements seems too feeble
a ground upon which to found freedom.
There would be a way out however if, in examining what is positive
in the Diodorean conception and what it is that distinguishes it from
that of the Ancient Megarians, we could make sense of how and why it
is that the nominalist interpretation, the only one historically and log-
66 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

ically admissible, ends up with arguments that are at once insufficient

and excessive.
It will be said that the only difference that allows our distinguish-
ing Diodorus from the Ancient Megarians is that he deploys over the
succession of time the necessary events that the others concentrate in
a timeless eternity, thereby allowing human ignorance to ludicrously
break the dike of fatalism. It will be added that grammatical contin-
gency has to do simply with the uncertainty of the time at which an
event will occur, which if dated or pseudo-dated would become neces-
sary. Real contingency, on the other hand, should reflect a hesitation
due to Nature itself. The Diodorean system thus makes modality a
subjective property of our knowledge. If this contingency sometimes
passes for real contingency it is because the realist interpretation, also
illegitimate moreover, in part parallels the nominalist one and so gives
rise to the illusion of a modality that would partially simulate the con-
tingency of things.
But let us take this described difference to its term. To elucidate
the meaning of a contingency reduced to temporal indetermination let
us consider a series of 'sophisms' to which Diodorus had recourse. They
are 'the Horned Argument', 'The Veiled Argument' and 'the Sorites'.
It will be seen that all these sophisms have to do with the problem of
decision.
How is one to answer the question 'Have you lost your horns'? The
answer is neither yes nor no, although Diodorus accepted the principles
of bivalence and excluded middle.45 What can be said to the following
argument? 'You say that it is impossible to know and not to know the
same thing. Well, you know your father. But if I show you a man with
his head veiled you will say that you do not know him. So if it turns out
that the man with his head veiled is your father, you will both know
and not know your father'.46 How is one to answer the Sorites?47 A
property P will be called 'predominant' if, given a body composed of n
indivisible particles, P(m) and m > n/2. Every predominant property
must be considered as applying to the body as a whole. Let a body
consisting of three indivisible particles have two of these moving and
one stationary. According to the principle of predominance then, the
body as a whole is counted as moving. Adding a new stationary particle
to the body of three already moving, the new composite of particles
45
Sextus, M.X., 112-117, Doring, fr. 129; Sedley, 1977, p. 102.
46
Sextus, M, X, 112-117, Doring, fr. 129; Sedley, 1977, p. 102.
47
Sedley, 1977, p. 95.
 DIODORUS' NECESSITARIANISM / 67

will also be moving; and continuing in like manner any number of

stationary particles whatsoever will be moved by two moving ones.
In these three cases Diodorus is exposing a typical nominalist
sophism. 'To have lost one's horns', 'to know someone', 'to be pre-
dominant' are not predicates that give rise to well-formed sentences,
that is, sentences capable of verifying the laws of logic. 'To have lost'
should be analyzed into the conjunction 'to have had and have ceased
to possess', of which each conjunct may be independently falsified. 'To
know x' as it is used here should be expanded into 'to know what x
looks like', which may be falsified without falsifying 'to know x\ The
Sorites arises from our treating any composite body, including the one
we are reasoning about, in terms of 'predominance', as a 'black box'
and forgetting to account for the process of attribution. But let us
penetrate into the composite. Let us postulate that for any number n
of particles we shall attribute P to the whole only if m > n/2. The
Sorites vanishes for the predicate 'predominant' must be analyzed into
'predominant given the internal composition', and this predicate may
be falsified (as in the case of the Sorites once m = 2 and n > 4) without
falsifying the absolute predicate 'predominant'.
These syntactical precautions are obviously imperative when the
infinite comes into play. There are two complementary arguments of
Euclid and Diodorus that attest to the fact. The first is 'the Elusive
Argument', important for its critical consequences regarding the per-
fect wise man of the Stoics.48 This is in a sense the antique sketch,
though negative here, of the notion of a chain used by Dedekind to
demonstrate the axiom of infinity starting from the knowing Subject. 49
Since an act of reflection can always apply to any number whatso-
ever of cognitive acts, self-knowledge yields a given infinite. From this
line of reasoning the Megarians conclude that one reflection always es-
capes self-knowledge. Self-knowledge therefore is ever-elusive. A state-
ment, therefore, must be finite. On the other hand, it is possible that
Diodorus proved atomism in applying his definition of the possible to
division. Since no division is nor will be carried out to infinity, how
could such a division be considered possible and what sense does it
make to speak of infinite divisibility?50
In warning us against the false appearances of incomplete state-
ments, Diodorus shows us by way of contrast what a genuine statement
is. It is a finite statement whose truth-value we are able,- in principle,
48
Sedley, 1977, p. 84.
49
Vuillemin, 1962, p. 297, note 2.
50
Sedley, 1977, p. 88.
68 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

to decide. But let us go back to the Master Argument's final premise,

to that possible that neither is nor ever will be realized. The state-
ment positing it is not a finite and decidable one and is therefore to be
rejected as syntactically deficient.
It was thus wrong of Cicero to ignore the degree of precision of the
temporal determination of statements. The whole question of freedom
resides in that degree. There is what is decided, what is decidable and
what is undecidable. We cannot change what is decided. Dates and
pseudo-dates will entail necessity then, and they will be its domain.
Nor can we on the other hand be misled by what is undecidable and
chimerical. And when in order to avoid these we define the possible as
that which is or will be true, we must not press the existential character
of the statement any more than taking it against Diodorus' intention
in supposing that the occurrence of the event could be deferred in-
definitely, which would only be a back-handed way of reinstating the
rejected premise. The indetermination required for freedom is not the
indefinite and that is why the truth-value change of indefinite futures
is guaranteed.
Is this indetermination a result of our ignorance or of irresolution
proper to Nature? We do not know. Whenever the Master Argument's
third premise has been challenged it has been so in the name of sub-
jective requirements characteristic of the supposed limitations of our
cognitive faculty. But does Diodorus rest the weight of indetermina-
tion on our faculty for knowing, as others, probably inspired by him,
will do? He loosens the vice-grip of fatalism through indetermination.
If he refuses to open it further as the third premise seems to do, it is
surely that he rejects speculation and accepts only finite statements. It
is not for all that that he makes indetermination a subjective property
of our faculty of knowing. His critical nominalism brings him close to
a philosophy of inquiry. He remains compatible nevertheless with a
dogmatic conception of indetermination.
Eternal Return and Cyclical Time:
Cleanthes' Solution.

Cleanthes, in rejecting the Master Argument's first premise, denied

that the past was necessary. Beyond this one certitude all is conjecture.
We shall examine, and subsequently reject, two of these conjectures,
each ascribing to Cleanthes a logical solution of the Master Argument.
The first, of nominalist inspiration, sees in Cleanthes a precursor
of Ockham. On the one hand, he would have contested, not the neces-
sity of the past in general, but the necessity of propositions having the
grammatical form of the past without being about the past. Calling
this first into question has moreover the added advantage of extending
the benefit of contingency to dated and pseudo-dated propositions—
something Diodorus had managed to assure only for the case of certain
indefinite propositions. Fatalism is avoided by eliminating modality
de re in favor of modality de dicto. Should one wish to rid this first
conjecture, however, of its imputation of ambiguity to Diodorus' Mas-
ter Argument, it would be necessary to extend the principle, throwing
doubt not on the necessity of the past, i.e. the Master Argument's
first premise but rather on the principle of conditional necessity, i.e.
the premise the argument borrows from the paradigm of the De Caelo.
But in doing this one deliberately parts ways with Cleanthes.
The second conjecture, exemplified by Leibniz, reverses the first. If
Cleanthes doubts of the necessity of the past, it is that he would refuse
to accord it any other than a conditional necessity. The first premise
loses all force but what would accrue to it from that additional premise.
The second conjecture, although historically more plausible than
the first, will have to be rejected as well. It is acceptable neither his-
torically, as we shall see, nor logically, in that it deprives itself of the
very possibility of distinguishing that which specifically characterizes

69
70 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

the 'necessity' of the past. The conjecture to be advanced and espoused

here is founded rather on physics than on logic. In so far as the modal-
ity of the past is concerned it results in the notions of eternal return
and cyclical time.

4.1 First conjecture. Necessity of the past

secundum vocem and secundum rem: Ockham's
conception on Prior's hypothetical
reconstruction. Modality de dicto and
modality de re.
The first interpretation of Cleanthe's thesis to be examined may be
summarized in the following three arguments.
I. The validity of the Master Argument does not depend, as it does
in the interpretation advanced in this book, on conditional necessity
but on the adjunction to premises A, B and C (as interpreted by Prior)
of two additional premises. The one of these additional premises posits
the retrogradation of truth, the other the discontinuity of time. What
leads to the incompatibility of the premises, once granted the discon-
tinuity of time, is the interplay of the modalities and time which is
governed solely by the grammatical form of the sentences involved.
That is why this interpretation can be called 'nominalist'. In this per-
spective,it is possible to get out of the fix by denying premise C, as
Diodorus Cronus did. Necessitarianism is then avoided, but only in
so far as the sentences reasoned about are indefinite ones. As soon
as pseudo-dates (yesterday, today, tomorrow) or dates are brought in,
necessitarianism is back to claim the field. To challenge necessitarian-
ism at its very root then, it is that axiom which is its veritable reason
for being that must be denied. The axiom in question is axiom A and
what warrants casting doubt upon it is precisely its ambiguous gram-
matical form. Denying A then is surely in conformity with the spirit
of nominalism. It is precisely the step that Ockham took.
II. The Ockhamist modal system allows for the explication and
formal expression of the nominalist intuitions regarding the logic of
indefinite, pseudo-dated and dated sentences. Possible ambiguities that
creep into modal and temporal sentences are brought to light.
III. The Master Argument again. In distinguishing modality de
dicto from modality de re, the confusion of which had led to the illusion
of necessitarianism, Diodorus' challenge will have been met.
 THE FIRST CONJECTURE / 71

I. The nominalist reconstruction of the Master Argument 1 will con-

clude that, limited to the temporal logic of indefinite propositions,
the Diodorean definition of the modalities does preserve the distinc-
tion between possible and necessary. It is only the step to a logic of
pseudo-dates or, a fortiori, to one of dates, that causes the distinc-
tion to collapse and renders necessitarianism inevitable. If this can be
the case it must be that there is something only tentative about the
Diodorean modal distinctions. And they are indeed bound to a meager
and very particular system of temporal expression. But "the sense of
a system of temporal logic may remain hidden as long as that system
incorporates only a part of the temporal determinations of discourse" .2
The introduction of pseudo-dates, and then of dates, will lift the veil
and bring out into the open that which seemed dubious in Diodorus'
principles only because of a lack of thoroughness due to the system's
incomplete formulation.
What renders Diodorus' principles dubious is the confusion that
crops up between the temporal form and the temporal scope of the for-
mulas. No one has done better at exposing that confusion than Ockham
has. The third postulate that he posits in order to solve the problem
of predestination, God's foreknowledge and future contingents, does in
effect distinguish the word from the thing.
"Some propositions are about the present as regards both their
wording and their subject matter (secundum vocem et secundum rem).
Where such [propositions] are concerned, it is universally true that
every true proposition about the present has [corresponding to it] a
necessary one about the past -e.g.,'Socrates is seated', 'Socrates is walk-
ing', 'Socrates is just', and the like. Other propositions are about the
present as regards their wording only and are equivalently about the
future, since their truth depends on the truth of propositions about
the future. Where such [propositions] are concerned, the rule that
every true proposition about the present has [corresponding to it] a
necessary proposition about the past is not true. And this is not re-
markable, since there are true propositions about the past and about
the future that have no true [proposition] about the present [corre-
sponding to them]. For example, 'what is white was black' and 'what
is white will be black' are true while their [corresponding proposition]
about the present—'what is white is black'—is false".3 In treating of
'Prior, 1967; Boudot, 1973.
2
Boudot, 1973, p. 463
3
Ockham, 1945, pp. 12-13; Ockham, 1969, pp.46-47. As the translators point
out this last example does not justify the conclusion that there is no necessary
proposition corresponding to such a proposition as 'Peter is predestinate' which,
72 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

the question as to whether 'A was predestinate' is a necessary proposi-

tion Ockham makes a distinction between propositions about the past
with respect both to form and to subject matter (secundum vocem et
secundum rem), which are necessary, and those which, though having a
past form, are really about the future. The proposition 'A was predes-
tinate' is equivalent to the proposition 'God will grant A eternal life',
"...that proposition which is in the present tense in such a way that it
is nonetheless equivalent to a future-tense proposition, and the truth of
which depends on the truth of a future-tense proposition, has nothing
of the necessity of the past; what is more, even though put in the past
tense, it has the same contingency as its corresponding present-tense
proposition" .4
Ockham's solution, then, comes down to flushing out ambiguities in
the determination of real time which are concealed beneath the appar-
ent determination of grammatical time. In the Treatise on Predestina-
tion Ockham presents the following objection: "To every proposition
about the present that is true at some time corresponds a necessary
proposition about the past. For example, if 'Socrates is seated' is true,
'Socrates was seated' will be necessary forever after. But suppose 'Pe-
ter is predestinate' is now true. In that case 'Peter was predestinate'
will always be necessary. Then I ask whether or not he can be damned.
If he can be, suppose that he is. Then 'Peter is reprobate' is true of
the present. Thus 'Peter was reprobate' will always be necessary of
the past. In this way 'Peter was predestinate' and 'Peter was repro-
bate' would be true at one and the same time." Ockham replies to
the objection: "I maintain that the major premise is false, for to that
proposition which is about the present in such a way that it is neverthe-
less equivalent to a proposition about the future, and the truth of which
depends on the truth of a proposition about the future, there does not
correspond a necessary proposition about the past. On the contrary,
the past-tense proposition is contingent, just as is the present-tense
one corresponding to it. All propositions having to do with predesti-
although in the present tense, is really about the future. Ockham's point is that
the truth of this last proposition is something that has to be established by what
will have come to pass in the future, which makes it like the proposition 'Socrates
will be seated' and unlike 'Socrates is seated' whose truth is already established at
present.
Due to the necessities of quotation the distinction between proposition and sen-
tence has largely been set aside in section 3.1. For most occurrences of the word
proposition it would be more suitable to read the word sentence as is natural for a
nominalist system.
"Ockham, 1945, p. 13; Boudot, 1973, pp. 452-453; Prior, 1967, p. 121; Ockham,
1969, p. 38.
 THE FIRST CONJECTURE / 73

nation and reprobation are of this sort, for they are all equivalently
about the future, even if grammatically they are in the present or the
past. Therefore 'Peter was predestinate' is contingent just as is 'Peter
is predestinate'."5
II. The entire difficulty lies in giving formal expression to that in-
tuition. Consider first a logic of pseudo-dates. To this effect we shall
construct an 'Ockhamist structure' representing a plurality of possi-
ble evolutions of a possible world. An 'Ockhamist model' is obtained
by the assignment of truth-values to the prepositional variables of a
structure. Since a structure is made up of a unique past, but of several
branches for the future, the evaluation of a variable will be relative to
a particular route. In such a model,8 a) the logical pre-determination
of the future is satisfied, just as it is in the metaphysics of Leibniz;
b) the distinction between propositions about the past and those sim-
ply formally in the past can be read off the formulas themselves; c)
propositions about the past are necessary; and d) the entailment of ne-
cessitarianism is blocked. It will be possible to distinguish between a
proposition about the past, even if couched in a grammatical future ('In
two days I shall have arrived four days ago'), and a proposition about
the future ('There will be a conflict in ten years') even if it occurs in a
grammatical past ('Three years ago there was bound to be a conflict in
thirteen years').7 In 'Ockham's system' thus conceived it is legitimate,
on the basis of p, to conclude to the necessity n units of time ago that
it will be true, n units of time later, that p; the necessity here being
de dicto. By contrast, it is not allowed to infer from p that n units of
time ago it was the case that necessarily n units of time later it will be
true that p. This would amount to a de re predication of necessity.8
Predetermination then is seen to be purely logical. From among all
future propositions, equally determined from the logical point of view,
we could distinguish the proper subset of propositions which are, in ad-
dition, causally true and determined in that the present or past causes
of their truth already exist. In any case, there is nothing that obliges
taking this subset as coextensive with the set of all futures. 9
5
Ockham, 1945, pp. 5-6; 1969, p. 38.
6
Boudot, 1973, p. 456.
Propositions about predestination and reprobation are of this last sort (Ockham,
1969, p. 38).
8
The formula p D LPnFnp is thus legitimate, but not the formula p D PnLFnp.
9
Boudot, 1973, pp. 460-462. It is seen here that predetermination in the sense
of the Ockhamist system is weaker than Leibnizian predetermination which in-
cludes, rather than excludes, causal necessity (Leibniz, Gerhardt, VI, p. 27 and
pp. 123-124; Jalabert, 1962, p. 38 and p. 130). What is called predetermination
in Ockham's sense corresponds simply to foreknowledge and futurition, whereas
74 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

If we turn now to a system of dates properly speaking, expressed

according to fairly natural conditions, and adjoin it to an Ockhamist
model, we run the risk of necessitarianism again. The necessitarian
thesis depends on two principles: the one maintaining that the truth
of a dated sentence is omnitemporal; the other defining the necessary
as that which is sempiternally true. The first, which does no more
than express the purely logical predetermination of futures, is plausible.
But the second, in going from the logical predetermination to a real
necessitation, conceals just that confusion denounced by Ockham of
past-tense sentence with sentence about the past. It will suffice to
escape determinism, then, to admit the validity only of the implication
from necessity to sempiternality, refusing the converse implication.10
III. What of the Master Argument then?
As will have been noted, in order to explain Ockham's position we
have tentatively accepted Prior's reconstruction of the Master Argu-
ment: reconstruction already rejected for the reasons given in Chap-
ter 1. We shall thus begin by pointing out the virtues of Ockham's
solution on the hypothesis of that reconstruction. But it is only after
having abandoned that hypothesis that it will be possible to make a
truly critical examination of Ockham's principles in their capacity to
resolve the Master Argument. That examination, moreover, will lead
us to abandon the philosophy of Ockham for that of Duns Scotus.

Leibnizian predetermination requires more, since it nullifies the choice of a middle

science.
-It is easily seen in comparing the Aristotelian and Ockhamist notions of the
determinateness of a thing, which serves as basis for the truth-value of the corre-
sponding proposition, that the sempiternality of the true does not entail necessity.
According to Aristotle (Adams and Kretzmann, in Ockham, 1969, p. 14), "z's being
A at tm is determinate at tn just in case it is settled by something actual in the past
or present relative to tn that x is (was, will be) A at tm". Determinateness of the
truth-value then implies the existence of a causal chain going back as far as it does.
For Ockham, on the other hand "x's being A at tm is determinate at tn just in case
it is settled by something actual at some time or times that x is (was, will be) A at
t m ". The truth-value of the proposition 'x will be A at tm' can thus be fixed at tn
chosen arbitrarily far back: thus from all eternity. The determinateness of the fact
'x will be A at tm' does not entail the existence of a corresponding eternal causal
chain. It could be, for instance, that the determinateness of x's being A at tm goes
back only to tm itself or to an instant close to tm. Predetermination then turns out
to be compatible with contingency. In other words, Ockham's predetermination is
logical, having to do only with the truth-value of the proposition and without in any
way prejudging of the determination of the thing. Leibnizian predetermination on
the other hand—adhering, on this point, to the Aristotelian realism of truth—links
the logical assignation of truth-value to the physical assignation of cause.
10
Boudot, 1973, pp. 468-472.
 THE FIRST CONJECTURE / 75

Ockham's solution consists in accusing the ambiguity of the premise

attaching necessity to indefinite or pseudo-dated past-tense proposi-
tions. (In the system of corresponding propositions in the logic of
dates it is possible to formulate the Master Argument but easy to re-
fute it.) For the first premise, asserting the necessity of the past, is
not universally applicable. To suppose it so is to suppose that to the
truth-value of a past-tense proposition there corresponds the already
given actuality of a verifying or falsifying event, even if, in spite of its
grammatical form, that proposition is really about the future. This
comes down to substituting, right from the start, a linear and unique
series of causes for the junctions of contrary possibles. It amounts to
settling an ontological question by grammar.11
Of the nominalist solutions that can be given to the problem of
future contingents, Ockham's is the most elegant. It preserves all the
laws of the logic of sentences and quantification, of modality and tense
(with the one exception of that dubious 'law': 'that which is always
true is necessary'). The amendment it makes of Diodorus, in order to
avoid his conclusion, concerns only the formulation of the first premise,
for no doubt is cast on the necessity of sentences which truly are about
the past. The first amendment has to do with a metaphysical subrep-
11
"It is easily verified in Ockhamist semantics that Diodorus' conclusion is invalid.
From the truth of '~ a and F ~ a' we cannot conclude '~ Ma' (or 'L ~ a' which
is equivalent to it). To prove this suppose a to be 'Fp' and that two routes branch
out from a point x. At every point of the first route 'p' is false, whereas it is
true at certain points of the second. The prima-facie assignment at x according
to the first route gives the true as value to '~ p and ~ Fp', while 'L ~ Fp' gets
the false as value. Is this to say that the Diodorean argumentation was somehow
defective? Not at all; it is the reasoning's premises that must be called into question.
The formula expressing that from the possible the impossible cannot follow is a
law of Ockhamist semantics. Likewise, among the additional premises we have
introduced, 'a 3~ P ~ Fa' is valid and, subject to the supposition that the
structures are discrete, that is to say that every point in these structures has an
immediate predecessor, '(~ a and ~ Fa) D P ~ Fa' is valid as well. It is naturally
the formula schema 'Pa 3 LPct', by which the necessity of the past was expressed
in Diodorean logic, that is now invalidated. /( could be justified only if we restricted
ourselves to the consideration of linear Ockhamist structures. But in that type of
structure the schema 'a D La' that expresses necessitarianism is equally valid".
(Boudot, 1973, pp. 462-463).
It is seen through this quotation that for Ockham the Master Argument's first
premise, understood as signifying 'Pp D LPp\ is not universally admissible. Neces-
sity, taken de dicto, applies to a grammatical form in itself ambiguous, since it is not
known a priori whether the event about which the sentence is has already taken
place or not. -The realist interpretation we have proposed of the first premise
as signifying the irrevocability of the past excludes, on the contrary, any formal
ambiguity in the sentences. It consequently renders such an Ockhamist solution
inoperative.
76 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

tion, the second with a grammatical one. They are in conformity with
the maxim of economy proper to Ockham but are characteristic of all
nominalism as well.
Nominalism applied to the modalities posits as principle that they
are properties of discourse, not of things. This principle, moreover, is
general, as can be seen from its application to the problem of universals.
Either language is separable from ontology and the laws governing the
entities of language commit us to nothing with respect to being: any
liberties taken then with the laws of language will be without danger,
since these laws govern only the use of signs. Or, if there is a suspicion
that such-and-such linguistic usage does commit us with respect to
being, all the laws having to do with those usages, and whose mixed
character would allow the surreptitious extension of simple linguistic
conventions to ontological commitments, will be eliminated as dubious.
As for the modalities then, either it will be posited that the modal laws
have to do only with the manipulation of certain symbols, or those that
seem to have a mixed—viz. a metaphysically dangerous—character will
be proscribed.
But just what is a modality that has to do only with discourse? It
is a modality whose scope contains a complete sentence , an expres-
sion of discourse. When the expression contains a quantifier, whether
explicitly or implicitly—as, for instance, when 'p was' is taken as equiv-
alent to 'it has not always been the case that not-p'—the modality is
de dicto if the modal operator precedes the quantifier. Consider on the
other hand an expression in which one of these quantifiers has been
'exported' and placed in front of the modal sign. This latter sign will
then govern an expression containing a free individual variable. This
is the case that arises, for instance, when it is said that there is an
x such that it is possible (necessary) that x has a certain property.
Two interpretations seem to be open here.12 Either exportation of the
quantifier will be given a weak sense in which modality and predicate
are still taken separately. The modal operator will govern a preposi-
tional function then, rather than a proposition. Or the predicate could
be taken as argument of the modal operator, in which case the two to-
gether would henceforth be treated as a single predicate. The reading
then would no longer be 'it is possible (necessary) that x has a certain
property' but 'a; has a certain possible-(necessary-) property'.
The Scholastics had a clear vision of the distinction between modal-
ity de re and modality de dicto. Buridan, for example, in his commen-
12
Hughes and Cresswell, 1972, pp. 183-184, footnote 131, which suggests distin-
guishing between L(ifx) and [Ltp}x.
 THE FIRST CONJECTURE / 77

tary on the De Caelo,13 exposes the ambiguity of the proposition: 'All

that is destructible will be destroyed necessarily', which can be taken in
a 'composite' or in a 'divided' sense. In the composite sense it means
that for any x whatsoever, if x is subject to destruction, then it is
necessary that x will be destroyed (Omne corruptibile de necessitate
corrumpetur). In the divided sense it means that it is necessary that,
for any x whatsoever, if x is subject to destruction then x will be de-
stroyed (De necessitate omne corruptibile corrumpetur). The first sense
is de re, the second de dicto.
This last distinction appears most clearly when the modal calculus
is extended to include identity. Accordingly as the modality is or is not
treated as a possible property of objects, this adjunction leads to two
sorts of entirely different systems: necessary identity systems and con-
tingent identity systems. In the first, every identity is necessary, as is
every difference. This necessity stems from the fact that, if the identity
of two things entails that a property which appertains to the one apper-
tains to the other, then, identity with itself, appertaining necessarily
to the one, is communicated necessarily to the other. 14 In this case,
necessity is treated just as any property whatsoever. Of course some
well known paradoxes follow from this conception. If that man who is
sitting is Socrates, it seems difficult to admit that it is necessary that
that man who is sitting be Socrates.15 Contingent identity systems
avoid these paradoxes in restricting the substitution of prepositional
functions containing identities to those not falling within the scope of a
modal operator.16 In these systems, possible-property and necessary-
property do not count as genuine properties.17 Identity is defined only
within a given world, and may not obtain in another. Contingent iden-
tities are thus preserved, but only at the expense of evacuating the
truly logical content from the very notion of identity. These last sys-
tems are extremely nominalistic with respect to the modalities which
play a harmless role in the realm of discourse. At the same time, it
becomes no longer possible to identify an individual, and the entire
edifice of quantification theory is threatened.
13
Buridan, 1942, p. 118.
14
Hughes and Cresswell, 1972, p. 189 sq.
I\ : x = x, /2 : x = y 3 [<p(x) 3 <f(y)] and L(x = x).
(1) x = y 3 [L(x = x) 3 L(x = y)], substitution in /2 : x = z/<f(z).
(2) L(x = z) 3 [(x = y) 3 L(x = y)\, permutation in (1).
(3) x = y 3 L(x = y), modus ponens applied to (2).

16
Id., ibid, p. 191.
16
Id., ibid, p. 195: cf. I' 2 .
I7
ld., ibid, pp. 199-200, ft.n. 151.
78 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

When modality and predicate together form a new predicate sui

generis, the modality is said to be de re. It introduces distinctions in
the universe between those properties that happen to belong to things
by chance, those that are inherent in them and essential, and those
that, as possibilities, are attached to their development. In this usage
modality goes from the sign to the signified. It is no longer linguistic,
but ontological. That is why nominalism must banish it, in showing
that such usage can be eliminated from discourse (as is supposed when
in a de re context the modality is interpreted as operating from the out-
side on a prepositional function which is its argument) either by means
of some general and regular procedure, or by adopting particular mea-
sures. We would dispose of a general and regular procedure if we had a
modal logic in which all modalities de re were automatically eliminable
in favor of modalities de dicto. To this end von Wright has suggested
the adoption of a 'principle of predication' according to which all prop-
erties could be exhaustively partitioned into two exclusive categories:
formal properties, whose belonging to an object is always necessary or
impossible, and material properties, whose belonging to an object is
always contingent.18 But, on the one hand, even the adjunction of the
'principle of predication' to the most classical modal logic19 falls short
of providing for the elimination of the modalities de re and it is an
open question as to whether there is an effective method allowing for
the construction of a formula of elimination for all the modalities de
re.20 On the other hand, just what would the principle of predication
come down to? Its point is evidently to separate the truths of reason
from the truths of fact; but this is accomplished by reducing the former
to simple reports of linguistic transformations, the latter alone being
susceptible of expressing any empirical content. The question might be
raised however as to whether such a transparent language, thus freed
from the opacities favoring the confusion of ontology and grammar,
would still have a real power of communicating ideas.
To be effective the nominalist must formulate particular rules of
exclusion. Ockham's method consists precisely in doing this. Implica-
tions in which the consequent involves a modality of necessity de dicto
are accepted;21 those same implications where the modality would fig-
ure de re are refused.22 Necessity is escaped, then, either through
18
von Wright, 1951, pp. 26-28; Hughes and Cresswell, op. cit, pp. 184-188:
(x)(Ltpx V L ~ <fx) V (x)(M<fx V M ~ ipx).
19
S5.
20
Hughes and Cresswell conjecture that there is no such method (op.cit., p. 187).
21
For instance: p D LPnFnp.
22
p D PnLFnp.
 INADEQUACY OF OCKHAM'S SOLUTION / 79

excluding Diodorus' first premise or through systematizing, within the

context of a richer and thus more distinct system than that of Diodorus,
the distinction between the two modal usages. In the one case it is the
universality of the necessity of the past that is denied; in the other,
with respect to what has taken place, it is accepted that it is necessary
that it has taken place, without its being accepted for all that that it
was necessary that it should take place. What is called into question
then is not the first premise properly speaking; it is only its formal
expression in which a distinction is drawn either between grammatical
and real necessity or between necessity of the thing and necessity of
the sentence.

4.2 Inadequacy of Ockham's solution.

Incrimination of the principle of conditional
necessity: John Duns Scotus.
Ockham's nominalism does succeed in avoiding necessitarianism. The
question remains as to whether this is so for the reason alleged.
That reason is said to be the ambiguity of the past tense in a sen-
tence. But Diodorus would surely have responded by purging his first
premise of any ambiguity himself. Could he, in fact, have done this?
Yes, he could have, on the condition of considering propositions hav-
ing a determinate subject instead of sentences. For such propositions
the distinction between past and future is fixed absolutely by the now.
From there on the necessity of the past becomes an ineluctable truth,
even on an Ockhamist interpretation. Ockham himself, in commenting
on Aristotle's Ethics (Eth. Nic. vi. 2, 113965-11) where it is said that
"in this alone is God deprived: to make undone things that have been
done", says that "if some assertoric [non-modal] proposition merely
about the present that is not equivalent to one about the future is true
now, so that it is true of the present, then it will always be true of the
past. For if the proposition 'this thing is'—some thing or other having
been indicated—is true, then 'this thing was' will be true forever after,
nor can God in His absolute power bring it about that this proposition
be false".23
If one accepts Prior's interpretation, the reason why the Master Ar-
gument doesn't go through has to do with the ambiguity of the sentence
about which one reasons. A sentence about the future (third premise)
is transformed (by means of Prior's second additional premise) into a
past-tense sentence and then, thanks to the first premise, its modality
is fixed. The remainder of the argument serves only to deduce the im-
23
Ockham, 1945, p. 4; 1969, p. 36.
80 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

possibility of what had been proposed. Necessitarianism is thus seen

to result from the fact that the first premise is applied to a proposition
about the future although expressed grammatically in the past.24
There is a twofold line of objection however that does away with
this argument.
The first is inherent in Prior's conception itself. Just what is the
meaning of saying of something that it neither is the case nor ever
will be? The statement is at once both present and future and its
future part does no more than continue and repeat its present part. If
we represent the model of this future, it excludes, by hypothesis, any
branching and is thus linear. In this way the thing is determined now
and causally so, not simply logically so. This is also the reason why
it can retrograde. The Master Argument's third premise, then, is one
of those propositions about the future that Ockham would take to be
necessary:25 it is not a future contingent. Granted that, the argument
from the formal ambiguity of past-tense propositions no longer applies
since the future spoken of here is in reality already given.
Should one reject this first line of objection in denying that the pre-
determination of truth entails a causal predetermination here, there
will remain an even greater difficulty if one abandons Prior's recon-
struction of the Master Argument which was in fact meant to reveal
an ambiguity in the argument's first premise. Suppose that the first
premise does simply signify the irrevocability of the past due to the
fact that a possible cannot be realized in the past. Thus construed it
can have to do only with true pasts, and Ockham's maneuver loses its
point.
On what condition could there possibly be another interpretation
that would invalidate the Master Argument while applying Diodorus'
first premise only to genuine pasts? The proposition saying that p is
not true now nor ever will be true in the future is true now. On Ock-
ham's principles then its truth can retrograde without ambiguity.26 Is
the Master Argument irresistible? It will be unless one holds that the
realization of the possible in the past leads to a proposition which is
certainly false but not impossible. The non-realization of the possible

{[(~ p • ~ Fp) D P ~ Fp] D~ M ~ £ ~ Fp} D~ Mp

3 5 1 4 and 2
The letters accused of ambiguity have been underlined; the numbers refer to Prior's
premises (1967, pp. 32-33).
25
Ockham, 1945, note 6.
26
Si haec propositio sit modo vera: Haec res est, quacumque re demonstrata, sem-
per postea erit haec vera: haec res fuit (Ockham, 1945, p. 4; quoted by Prior, 1967,
p. 35).
 INADEQUACY OF OCKHAM'S SOLUTION / 81

would not then entail its non-realizability. That apodeictic incompat-

ibility that is required before the second Diodorean premise can be
applied would hence disappear. The Ockhamist model of branching
time is excluded in reality because it has been posited that the pos-
sible will not be realized. Different possible worlds are not dismissed
for all that. In short, what would be challenged here is no longer the
validity of the first premise, nor that of any of the Master Argument's
explicit premises for that matter, but the validity of conditional neces-
sity, which is required for transforming non-realization in the future
into non-realizability.
Ockham's interpreters that we have referred to assert that, to avoid
determinism in a system of dates, Ockham refused to admit that what
is always true is necessary. They justify this refusal, which allows go-
ing from the necessary to the sempiternally true but blocks the inverse
route, by limiting the validity of the axiom: if p is true at t, then it
is true at f that it is necessary that p be true at t. The restriction
they adjoin is: t' > t. Abandoning this last condition would result
in necessitarianism. Compelling predetermination would then be the
rule. But this is precisely because one would have taken the formal
past, £', of the retrograded statement of p, for its real past, t."21 Deny-
ing the passage from the sempiternally true to the always true in the
logic of dates comes down once again to eradicating the ambiguity of
grammar and reality in the first premise. Yet these same interpreters
pride themselves on having saved the principle of conditional necessity
for Ockham. And indeed, positing t' = t, the preceding axiom reads 'if
p is true at t, then it is true at t that it is necessary that p at t', which
is simply another expression of the principle of conditional necessity
recognized as legitimate.
But does this argument, designed to render the logical predetermi-
nation of a dated event compatible with the real contingency of that
event, apply in effect to the Master Argument? Not really, because
the third premise precludes the realization of a certain possible for all
time going from now to the most distant future. What is at stake then
is not the logical predetermination of a future event, but the contin-
uous determination of a sempiternal event—as was required by both
the Timaeus model and the De Caelo interpretation. But the sense of
the contested passage from the sempiternal to the necessary changes
radically when it is no longer applied to a transitory event but to a
sempiternal duration (in so far as the future is concerned).

27
Boudot, 1973, pp. 471-472.
82 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Suppose that conditional necessity holds. Since the possible men-

tioned in the third premise does not exist and continues not to exist for
all future time, the posited axiom compels us to say that it is necessary
that it should not exist for all this time. And since it is necessary that
it not exist during all this time, it is impossible that it exist during all
this time. Q.E.D. We do not see then how Ockham can avoid having
the Master Argument's third premise reduced to the impossible, if he
does admit the principle of conditional necessity.
His interpreters were not mistaken in constructing an Ockhamist
model that verifies the principle of conditional necessity. The question
is how such a model can still meet Diodorus' challenge.
In the third question of his Treatise on Predestination Ockham dis-
cusses, and subsequently rejects, the calling into question of the princi-
ple of conditional necessity. "How can the contingency of the will, both
created and uncreated, be preserved [in the case of its] causing some-
thing external? That is, can the will, as naturally prior to the caused
act, cause the opposite act at the same instant at which it causes that
act, or can it at another, subsequent instant, cause the opposite act or
cease from that caused act?"28
Duns Scotus' doctrine maintains that there are two possibilities. It
recognizes in the will, alongside an evident capacity for opposite acts
or opposite objects in succession, a nonevident capacity for opposites
without succession. Ockham himself admits the first capacity. He
contests the second.29
28
Ockham, 1945, p. 32; 1969, p. 71.
29
Ockham, 1945, pp. 32-33; 1969, p. 83. The fundamental text is that of Duns
Scotus, Vives, t. 10, 1843, In librum primum sententiarum, D. XXXIX, Q. 1, 628-
629. "This freedom is accompanied by an evident capacity for opposites. It is not
allowed that there be in it a capacity to will and to not-will at the same time, for
this does not exist; but there is in it the capacity to will and then afterwards to
not-will, i.e. the capacity for a succession of opposite acts, and this capacity is
evident in all changing beings in which opposites succeed one another. There is [in
it] too, nevertheless, another capacity that is not evident so and that is outside all
succession. For if we posit a created will existing only at one instant and having
a certain volition in that instant, then it will have that volition in a non-necessary
manner. Proof: if in that instant it did indeed have it in a necessary manner,
since there is no cause outside that instant where it would cause, the will, when it
would cause, would then simply cause necessarily. Indeed, it is not only the cause
that preexisted that instant in which it causes and which then, as preexistent, was
able to cause or not cause, that is contingent, for just as that being when it is
necessary or contingent, so the cause, when it causes, causes then in a necessary or
contingent manner. From which it follows that in that instant it causes this willing
in a non-necessary, and thus contingent manner.
There is thus a capacity of the cause, that causes without succession, for its
opposite, and that real capacity is the capacity of that which is naturally prior, as
first acts, to opposites, which are naturally posterior, as second acts. For the first
 INADEQUACY OF OCKHAM'S SOLUTION / 83

Consider first created will. Scotus, criticizing the naturalism of the

philosophers, refuses to conceive of the created will as a passive capac-
ity to be moved by a representation. His argument is based essentially
on the existence of a volition and the question as to how the negation
of an object would have the force to determine the will.30 In short,
nothing extraneous to the will is the total cause of volition in the will.
Regarding the nature of the things with respect to which our will is
free, three different freedoms must be distinguished. In the first place
the will is free ad oppositos actus, i.e. to will or to refuse a same
object. Next, it is free ad opposita objecta, since it has the power of
tending towards opposite objects. Finally, it is free ad oppositos ef-
act, considered in that instant where it is naturally prior with respect to the second
act, posits this latter in being as its effect in a contingent manner, so that in so
much as naturally prior it could equally posit its opposite in being."
30
Duns Scotus, Vives, t. 13, 1843, In secundum librum Sententiarum, D. XXV. Q.
1, 199-201.
"A modern doctor (Godfrey) says here that something other than the will is the
effective cause of the volition in it and he posits that this other thing is a phantasm.
His principal reason is the following: mover and moved must be distinct as to the
subject, but in the intellective part of the soul there is nothing that can be distinct
as to subject from the will; hence there is no motive for it there; so there must be
something else outside the intellective part, and that is the phantasm. He proves
the first proposition thus: to posit that motor and moved are not distinct as to
subject is the same thing as to posit that a thing moves itself, which is impossible.
First, because one and the same thing would then be in act and in potency at the
same time with regard to itself, because it is said (Physics, 3) that as the motor is
in act, so the moved is in potency. He says also that that, viz. that there cannot
be a self-mover, is the principle of metaphysics; to do away with it therefore is to
deprive oneself of the veritable principle of research in metaphysics.
The second reason is that matter and efficient cause do not coincide in a numer-
ically single thing (Physics, 2); thus neither do agent and patient.
Likewise (Metaphysics, 5), the motor is really in relation with the moved, but the
relation of the motor with itself is not a real relation. The proof of this is that what
are really relatives are really opposites, but opposites cannot be simultaneously
proper to a same thing".
"Another opinion is that of a more ancient Doctor (St. Thomas) who posits the
same conclusion, viz. that the will is moved by something else, but posits that that
other thing is the object cognized or comprehended".
"Against the conclusion in itself there are reasons of principle, and I advance
these in reasoning as follows. A natural agent cannot by itself be the cause of
contraries concerning a same passive being..., but it is in the power of our will to
have not willing and willing, which are contraries with regard to a same object.
These two things, therefore, cannot result from an agent naturally, thus not from
an object either, which is a natural agent. If it is posited then that an object would
be the cause of willing, there must nevertheless be another that woufd be the cause
of not willing. But that object, other than the will, could not be but a bad object;
but since evil is only a privation, it could not be the cause of the posited acts of
the not willing type. Therefore it must be effectively by the will". Commentary in
Gilson, 1952, p. 583.
84 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

fectus, i.e. with respect to the opposite effects it can produce.31 As

regards the manner in which contingency or possibility results from
freedom, Scotus distinguishes two capacities. The first results from
the first freedom, ad oppositos actus. Our will can engage in opposite
acts successively, i.e. it can will after having refused or refuse after
having willed. On the other hand, it is impossible to will and not will
at the same instant: this would be nothing.32 The second capacity is
not manifest and excludes succession. "For if we suppose that there is
a created will that exists at only one instant, and that at that instant
it has this or that volition, then it does not necessarily have it at first.
For if at that instant it had the volition necessarily (since it is a cause
only at that instant when it caused the volition) then, absolutely, the
will, when it caused the volition, would cause it necessarily. For it is
not now a contingent cause because it pre-existed before the instant at
which it causes and pre-existing then could either cause or not cause.
For just as this or that being, when it is, is then either necessary or
contingent, so a cause, when it causes, causes then either necessarily
or contingently. Therefore whatever this willing causes in that instant,
and causes not necessarily, it causes contingently. Therefore this ca-
pacity to cause the opposite of that which it does cause is without
succession. And this real capacity is a naturally prior capacity (as of
first actuality) for opposites-[opposites] that are naturally posterior (as
of second actuality). For first actuality, considered at that instant at
which it is, is naturally prior to second actuality. Thus [first actuality]
contingently posits [second actuality] in reality as its effect, so that as
naturally prior it could equally posit the opposite in reality".33
31
Ockham, 1945, p. 33; 1969, p. 82; Gilson, 1952, pp. 586-587.
32
Ockham, 1945, p. 33; 1969, p. 82.
33
Quoted in Ockham, 1969, pp. 82-83. Gilson (1952, p. 587) comments as follows:
"Consider a created will whose existence would last but a single instant. Imagine
moreover that in that unique instant it exercises a determinate volition (hanc vo-
litionem). Obviously its freedom of choice would be exhausted in that unique act
that could be said to fill out the totality of its existence. Yet that unique volition
would not be necessary, ant here is the proof: if the will exercised that unique
volition necessarily, since it would be a cause only in that one instant, it would be
a necessary cause, thus a 'nature', that is to say the contrary of a will. It makes no
difference here whether it existed in a preceding instant where it would have willed
the contrary. It is in the present instant, the only one in which it exists, that it must
be considered. But just as in each present instant of whatever duration a being is
either contingent or necessary, so also, in the present instant in which it exercises
its causality, the cause exercises it either in a contingent or in a necessary manner.
Since by hypothesis it is a will that is spoken of, it does not cause in a necessary,
but in a contingent, manner. A will is thus capable of willing the contrary of what
it wills and of causing the contrary of what it causes, in the same time in which it
wills and causes. Assuredly, it cannot will or cause contraries simultaneously, but,
 INADEQUACY OF OCKHAM'S SOLUTION / 85

Duns Scotus reasons on the false supposition of the instantaneous

existence of a created will. But this false supposition implies no con-
tradiction: it constitutes a possible world. In such a world there would
be a single act that would exclude its contradictory. If we were reduced
simply to the manifest contingency in the succession of opposites, such
a world would be necessary because unique. But it was supposed that
it was a free, even though instantaneous, will that acted in this world.
It must be then that this freedom is not absorbed by its act and that
there is a capacity for contraries which is not deployed and which is not
deployable in time. This is to affirm logically that an act is compatible
with the potentiality of its contrary; and such is the radical contin-
gency ad extra presupposed by the freedom of a created will reduced
to an instant of life.
Scotus next takes up the case of the divine will. Regarding the
things with respect to which it is free, it differs from human will es-
pecially in that it is not free ad oppositos actus, as God is immutable.
As for that kind of contingency that relates it ad extra—and it is this
radical contingency that characterizes the Christian God, distinguish-
ing Him from the God of the philosophers-,34 we find again what was
said of the non-successive contingency for man (successive contingency
being excluded by the divine immutability): "...our will, as naturally
prior to its act, elicits that act in such a way that it could at one and
the same instant elicit its opposite. In the same way the divine will,
insofar as volition itself is naturally prior to such an intention (tenden-
tia), intends the object contingently in such a way that at the same
instant it could intend the opposite object. And this [is] as much by
virtue of a logical capacity—i.e. the compatibility (non repugnantia)
of the terms—as by virtue of the real capacity—i.e. [the will's being]
naturally prior to its act".35
at the same time as it wills and causes the one, it retains its essential aptitude to
will and cause the other: est ergo potentia hujus causae ad oppositum ejus quod
causal, sine successione".
34
Gilson, 1952, p. 34, p. 262, p. 363; Duns Scotus, Vives, t. 4, 1841, De rerum
principio (Against the necessary production of creatures by God, according to Avi-
cenna), Q. IV, art. I, sect. I, p. 305. "One cannot posit then that the creature,
which has the capacity for not-being, issues necessarily from the necessary being,
for, even though formally it would not be the necessary being, it would nevertheless
have the necessity of being, in virtue of which one opposes to it, in a contradictory
manner, the fact that it can not-be".
Q. IV, art. I, sect. Ill, p. 307. "It is therefore not due to any necessity that God
wills any extrinsic thing whatsoever; He does only what He wills and how He wills
it; thus He neither makes nor produces anything extrinsic necessarily".
35
Ockham, 1969, p. 83; Duns Scotus, Vives, t. 4, 1841, De rerum principio, Q.
Ill, art. Ill, sect. I, p. 297. "Thus, in the creature, the will is equally called will-
86 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

The psychology of volition supposes a logic of propositions in con-

tingent matter.36 Consider the proposition: the will that wills A may
ing, because the same will, by its nature, insofar as it is a free capacity, involves
opposites, like to read and not to read, and is free as regards any time difference
whatever, because it can will or not will in the divided manner (divisim) as regards
any now. It is nevertheless constricted by the act of willing, for its act is such that
it is one thing to will that it be, it is another thing to will that it not be, and it
is one thing to will to do a certain thing tomorrow and not before, another to will
the opposite or to do nothing. But it is the contrary with God, as it is the same
volitional capacity by which He can will that this be and will that it not be, and
that this be tomorrow, and by which He can will that the same thing be before or
after. And, in short, it is the same capacity by which He wills or does not will all
that He wills or does not will; what is more, it is the same capacity by which He
wills that this not be and wills that this be, and before or after or once again: so
it is with the same act of willing neither changing nor varying nor repeating itself
that He wills that this be and can will the opposite".
P. 299 (same section). "Thus by the same act of willing God wills contradictories,
not that they exist together, for this is impossible, but he wills them together; in
the same way, it is by the same intuition, or by the same science that He knows
that contradictories are not together, but they are known together by the same act
of science, which is one and the same act."
Leibniz is on Ockham's side here, rejecting Scotus' subtleties as purely verbal.
In his Essays on the Goodness of God, the Freedom of Man and the Origin of Evil,
second part, §132-XVII (Leibniz, Gerhardt, VI, p. 184; Jalabert, p. 93), he quotes
the principle of conditional necessity as it is formulated in Chapter IX of Aristotle's
De interpretatione and adds: "The nominalists adopted this Aristotelian maxim.
Duns Scotus and many other scholastics seem to reject it, but at bottom their
distinctions come down to the same thing".
36
Gilson, 1952, p. 588; Duns Scotus, Vives, t. 10, 1843, In librum primum senten-
tiarum, pp. 629-630. "...The will willing A can not-will, that proposition is false
in the sense of composition in that it signifies the possibility of the composition:
the will willing A does not will A. But it is true in the sense of division in that it
signifies the possibility for opposites successively, because the will willing at A can
not will at B. But if we also accept a proposition about the possible uniting the
extremes in a same instant, like this one: the will not willing something at A can
will it at A, this last must also be distinguished as to composition and division.
And in the composed sense it is false, for it says that it is possible that it [the will]
itself be at the same time willing at A and not willing at A. But the divided sense
is true; it signifies then that to that will, to which willing at A belongs, there can
belong not willing at A, but without this latter having to exist at the same time,
because then the willing would not belong to it.
"And to make this second distinction which is more obscure understandable, I
say that in the composed sense there is one single categorical proposition, whose
subject is: the will not willing at A, and the predicate: willing at A; and so this
predicate is simply attributed possibly to this subject to which it is repugnant, and
therefore it is impossible that that which is noted as belonging possibly to itself
should belong to itself. In the divided sense there are two categorical propositions,
attributing two predicates to the will. In a proposition of class-membership the
predicate not willing A is attributed to the will and the categorical falls implicitly
under composition. But in the other categorical [proposition] on the subject of the
possible, it is stated possibly willing A. And these two propositions are verified at
 INADEQUACY OF OCKHAM'S SOLUTION / 87

not will A. In the composed sense: the will that wills A does not will
A is an absurd and impossible proposition. In the divided sense and
taken in the order of succession the proposition is acceptable, for the
will37 can will a certain object at a moment A and not will it at mo-
ment B. But even in a given instant, the proposition remains true in
the divided sense; for even though the will that wills A cannot not will
it that will that wills A is such that it may not will it. Even willing it at
moment A, it is of itself capable of not willing it at moment A. Scotus
recognizes that this is a difficult distinction to grasp: obscurior, but it
is well-founded, for, in the divided sense, it justifies the possibility of
two distinct affirmative possibilities: the one saying that the will wills
at the moment A, the other that it is possible that the will not will at
the moment A. Being about the same time, but not about the same
object, they can be simultaneously true; and indeed they are, for it is
true that at the very moment at which the will is not willing it can
will, and that, willing one object, it could will another.
Ockham refuses Scotus' capacity without succession in arguing that
such a possible could not be realized. "This nonevident capacity can
be actualized by no capacity, since if it were actualized the will would
will something at t\ and not will it at ti, and so contradictories would
evidently be true at one and the same time".38 Therefore, "it is in-
consistent to say that the divine will as naturally prior posits its effect
in reality at t\ in such a way that it can not posit it in reality at the
same instant. For there are no such instants of nature as he [Scotus]
imagines, nor is there in the first instant of nature such an indiffer-
ence as regards positing and not positing. Rather, if at some instant
it posits its effect in reality, it is impossible by means of any capacity
whatever that both the instant occurs and the effect does not occur at
A because they signify that their predicates are attributed to the subject during
the same instant. But this is true. For proper to that will in the same instant is
not willing A together with the possibility of the opposite at A as is signified by
the proposition about the possible.
"There is an example of this distinction in the proposition: all men who are white
run. Being posited that all white men run and no blacks or dark-skinned men, it is
true in the composed, and false in the divided sense. In the sense of composition it
is a proposition having a unique subject determined by: who are white. In the sense
of division there are two propositions stating two predicates of the same subject.
Likewise for the proposition: all men who are white are necessarily animals. In the
sense of composition it is false, because the predicate is not proper necessarily to
all this subject. But in the sense of division it is true, because two predicates are
affirmed of a same subject, one necessarily, the other absolutely without necessity,
and both are appropriate, and the two categorical propositions are true".
37
Created.
38
Ockham, 1945, p. 33; 1969, p. 85.
88 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

that instant, just as it is impossible by means of any capacity whatever

that contradictories are true at one and the same time".39
Let us now bring the discussion back to its logical significance. Sco-
tus maintains that one can at one and the same instant act and possibly
act otherwise than one does.40 This means that the act is compatible
39
Ockham, 1969, p. 87. Concerning the question of instants of nature cf. Baudry,
1958, p. 128. "Instans naturae. To explain the knowledge God has of future
contingents Scotus imagines what he calls instants of nature. In his treatise on
future contingents (p. 35) Ockham rejects the Scotist theory in a word. 'Non teneo
Scotum quia in omnibus illis instantibus naturae erravif. The reasons for which
he does so are found in E.A., fol. 85 a. When it is said that one thing is prior to
another by a priority of nature, it must not be imagined that it exists in an instant
in which the second does not exist. For that instant would be either in the soul or
outside the soul. It cannot be said that it exists in the soul, for on that hypothesis,
if the soul didn't exist that priority of nature wouldn't exist either, and moreover,
that instant of nature could be neither a habitus, nor a capacity, nor a species, nor
the act of intellection nor its object. It cannot be said to exist outside the soul,
because it would then be either a substance or an accident. Which is impossible.
For, consider the priority of nature Socrates has with regard to his whiteness. If
one admits that Socrates exists in an instant of nature in which his whiteness does
not exist, either that instant of nature is none other than Socrates and it will have
to be said that Socrates is in Socrates, or that instant is a part of Socrates and
then Socrates will be in a part of Socrates, or that instant of nature is an accident
of Socrates and Socrates will be in one of his accidents. On the other hand, it
cannot be said that that instant is something exterior to Socrates. It must be held
as indubitable, therefore, that a priority of nature no more entails the existence of
an instant of nature than the highest honorability or highest perfection entail the
existence of instants of perfection and honorability. The instants of nature admitted
by Duns Scotus do not exist': 'Indubitanter est tenendum quod non magis sunt
instantia naturae quando aliquid est prius alio natura quam sunt instantia honoris
et perfectionis quando aliquid est honorabilius et perfectius alio... Universaliter
tenendum est quod non sunt aliqua instantia naturae nee originis praeter instantia
tempo ris".
40
Duns Scotus, Vives, t. 4, 1841, De rerum principio, Q. IV, art. II, sect. V,
p. 315. "But willing and refusing, which are one in God, are names expressing
the connection of this same divine willing to different effects in such a way that
this connection and relation is really in the creature, but according to reason and
our concept in God, just as creator and God express for an eternal thing the new
relations which, according to reason, are in God, but, according to the thing, in
the creature. Whence it is seen that, from one side, it is true that God wills and
at the same time does not will a thing, in so far as that act is identical with God,
and connotes no thing.
From another side, they are neither a same thing nor in the same time, but
successively, as long as they imply an exterior connotation, in such a way that
there is a succession according to the thing in what is connoted, but according to
reason and our concept in God".
(Same section, p. 316) "Just as divine intuition sees that the one and the other
part of a contradiction in the same time do not exist, but can happen (fore), not
in conjunction (conjunctim) but in division (divisim), so does the divine will too,
by a single immutable willing, bear on the one and the other extremes of the
 INADEQUACY OF OCKHAM'S SOLUTION / 89

with the potentiality of its opposite. Ockham replies that there is in-
compatibility here, since to realize the contrary possible would be to
enter into contradiction with the act. The divine will itself cannot do
that. What is clear is that Ockham admits the principle of conditional
necessity. The act, while it is, excludes the capacity for its opposite
whose realization can take place only after it itself will have disap-
peared. But that successive representation of potentiality, as legitimate
as it may be in suggesting something positive in the potentiality and
going beyond the simple logical reality or non-contradiction, appears
to be ill-founded in Ockham. When he claims to reduce Duns Scotus
to the absurd, he is, in reality, simply giving in himself to the sophism
of the distribution of the modalities, confusing the real contradiction
between act and contrary act with the apparent contradiction between
act and contrary possible.41

contradiction. It wills in willing them simultaneously, not however that they should
exist together, but it wills simultaneously that they should be able to happen [here
I propose correcting forte to fore] not in conjunction but in division. And just
as divine intuition sees the contingency of causes and the effects that will result
contingently from them, so does the divine will will that created causes act in a
contingent manner, or, on the contrary, a necessary one, according to the order of
causes" (see the end of note 36).
41
Duns Scotus explicitly contests the principle of conditional necessity in a text
that highlights his virtuosity as a dialectician: Vives, t. 10, 1843, In librum primum
sententiarum, pp. 630-631.
"Aristotle's proposition: it is necessary for whatever is that it be while it is can
be categorical or hypothetical, as is also the proposition: it is necessary that an
animal run, if a man runs. As a conditional, this latter gives rise to a distinction
accordingly as necessary can mean to say the necessity of consequence or of the
consequent. In the first sense it is true, in the second false. But the second which
is a categorical, signifies that the whole: should run if a man runs is predicated of
animal with the mode of necessity, and the categorical is true, because the predicate
thus determined belongs necessarily to the subject. But it is not a question of the
predicate absolutely and hence to conclude from the predicate thus determined to
the predicate taken absolutely is the error secundum quid ad simpliciter. Likewise
I say here that if Aristotle's proposition is accepted as a temporal hypothetical,
then the necessity notes the necessity of the consequence or that of the consequent.
If of the consequence, it is true; if of the consequent it is false. But if taken as a
categorical, then what is, while it is, does not determine an implicit composition
in what is, but a principal composition signified by the fact that what is should be,
and then the predicate that that be when it is is affirmed of the subject which is
with the mode of necessity, and in this way the proposition is true. Thus it doesn't
follow that it is necessary that it be, but there is a sophism of secundum quid ad
simpliciter, as appears in the other case. Hence no true sense of this proposition
denotes that being something in the instant where it is be simply necessary, but
only be necessary secundum quid, i.e. when it is, and as this is in harmony with the
fact that in the instant where it is the thing is simply contingent, it is consequently
in harmony also with the fact that in that instant its opposite can belong to it."
90 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

In all likelihood neither Scotus nor Ockham knew the Master Argu-
ment. It just happened that similar difficulties gave rise to responses
that can be compared systematically with those of the Ancients.
Scotus clearly expresses the consequence for freedom of calling conditional neces-
sity into question (ibid., p. 637).
"Just as our will, in so far as naturally prior with respect to its act, produces
that act in such a way that it should be able at the same instant to produce the
opposite act, so the divine will, in so far as it is itself under a volition, is naturally
prior with respect to such a tendency and tends to that object contingently, in such
a way that in the same instant it can tend to the opposite object. And this is so
from the point of view both of the logically possible (which is the non-contradiction
of terms, as was said regarding our will) and of the real capacity, which is prior
with respect to its natural act."
In his Commentary in the Vives edition, Lychet sets down Ockham's objections:
1. There is not such a nonevident capacity for opposites in the same instant, for
no capacity, not even an infinite one, could reduce such a capacity to act except by
a contradiction in the facts (p. 641). This argument is none other than the sophism
of distribution.
Gregory of Rimini answered this first argument in positing first (p. 643) "that
whatever is in a given instant can not-be in that instant, so that the proposition
stating that it is not can be true in that instant. This can be understood in three
ways, a) That this is possible in the composed sense, so that in that instant that
thing, at the same time as it is, is not in that instant. Then both the proposition
stating that it is and the proposition stating that it is not conjunctively and si-
multaneously in the same instant will be true. But this sense is truly impossible
because it implies a contradiction, b) It can be conceived or understood to be true
in the divided sense; and this in two ways. According to the first, the thing that
is in a given instant, can already, being posited in being, cease to be, and pass
from being to non-being in the same given instant. The proposition stating that
it is ceases to be true and from true becomes false, so that in that instant it is
false; and, conversely, its opposite ceases to be false and becomes true. But this
sense is impossible because such a passage and such a succession are not possible
in the instant. According to the second way, it is absolutely and simply possible
that this thing not be then and that it not be posited in being by its cause. The
proposition stating that it is not in that instant is true. This sense is true and in
no way improper."
2. Ockham's second argument against Scotus is based on the nature of the past
(p. 641).
"It is generally admitted by philosophers and theologians that God cannot make
a non-past of the past without its afterwards being true to say that the past was.
Therefore, since by hypothesis 'the will wills at A' is determinately true and, con-
sequently, always will be true afterwards, whereas 'the will does not will this at A'
never was true, 'the will did not will this at A' always was impossible."
Gregory of Rimini answers (p. 641). "Suppose that God can not make it happen
that the past not be the past (though many theologians contest that supposition).
I say that if the will wills something in the instant B, afterwards the proposition
'the will willed this in the instant B' will always be true, so that the consequence
'the will wills this at B' is necessary. Therefore, after B, the proposition 'the will
willed this in B' will always be true if it is formed. But I say that the antecedent
is contingent and can not-be-true, even in the instant B. And if one posits that it
is not true at B, as is possible to do in virtue of the third sense permitted, 'that
 INADEQUACY OF OCKHAM'S SOLUTION / 91

Ockham hasn't really called the first premise into question. He has
simply removed an ambiguity attaching to one of its modal formula-
tions. He is not then comparable to Cleanthes. His entire thrust is
concerned with the distinction between the logical and the real, be-
tween predetermination of the true and predetermination of the cause.
If we measure the system we can attribute to him by its effectiveness
in resolving the Master Argument, Ockham's system remains weak
even when taken within the confines of Prior's reconstruction. Ockham
adopts the principle of conditional necessity. Applying to a sempiternal
event, it confers necessity on its whole duration. The one thing more
that could be said in Ockham's favor has to do with the negative char-
acter of the sempiternal event in question in the third premise: the
exact symmetry postulated between affirmative and negative propo-
sitions is dubious for the nominalist. What is more likely is that if
Ockham had known the Master Argument he would have taken it after
the manner of a Carneades: on the one hand, he would not have been
frightened by Diodorean necessitarianism since it concerns discourse
only but not reality; on the other, the formal truth of a proposition
does not entail a causal consequence, the predetermination of the true
is without ontological implication.42
will willed this at B', will not be true after B, nor, consequently, necessary either,
no more than its opposite will be impossible."
As for Ockham, it has been maintained that he too contested the Aristotelian
principle of conditional necessity. He does indeed write in his Commentary on
Chapter IX of the De interpretatione: "Sciendum est, quod ista propositio: Omne
quod est quando est necesse est esse, de virtute sermonis est simpliciter falsa."
(Quoted by Boehner in Ockham, 1945, p. 71). But by that Ockham means simply
that the absolute or simple necessity, i.e. not qualified by the temporal condition of
the quando of the thing or the event, must not be accepted. That false interpretation
of Aristotle rejected, Ockham accepts the principle in its authentic Aristotelian
form: "It is necessary, if a certain thing exists, that it exist then." (ibid.). He
elaborates: "...But the Philosopher says that this proposition is necessary: whatever
is, is, when it is; for this proposition cannot be false. The following proposition is
likewise necessary. Whatever was, was when it was. Likewise: whatever will be,
will be, when it will be." (ibid., p. 72).
What misled the commentator (ibid., pp. 70-72) is Lukasiewicz' interpretation
of Aristotle from which the temporal condition is omitted. The discussion against
Scotus proves the point. What is at stake, in the Aristotelian principle, is the
possibility of the coexistence at the same instant of an act and of the contrary
capacity. Ockham, along with Aristotle, denies that possibility that Scotus asserts
(cf. below, 9.1, pp. 227-228).
42
What Ockham refuses is not the principle of necessity but the inference that goes
from truth to necessary truth. His position is reminiscent of that of a Carneades
(cf. below, 8.2, pp. 209-210). Ockham does indeed refuse to pass from truth to nec-
essary truth. Boehner interprets this refusal (Ockham, 1945, p. 68) as a rejection
of the principle of conditional necessity, in view of Qu. Ha, B, ad sec. of the Trac-
92 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Duns Scotus, taking up Plato again unawares, casts his doubts on

the principle of conditional necessity. The refusal of that necessity,
altogether consonant with modern thought, entails no contradiction.
Scotus' theory is thus consistent. By this refusal it escapes the Mas-
ter Argument without having to challenge any of the three explicit
premises of Epictetus' text. Therefore it has no relation to Cleanthes
solution either.

4.3 Cleanthes again and the second conjecture:

the conditional character of the necessity of
the past according to Cleanthes; the
interpretation of Leibniz.
Those who have interpreted Cleanthes through Ockham have attribu-
ted to the former a nominalistic attitude that is rendered highly im-
probable by what is known of the rest of his doctrine. Not that the
erstwhile boxer Cleanthes, whom Diogenes Laertius presents as a lout,
would have disdained the dodge. It is just that this dodge supposes
an ambiguity in the Master Argument's first premise. Had it been
possible, no doubt everyone would have adopted that simple solution
instead of affronting the detours and the obscurities of non canonical
systems of logic or physics. It must therefore be supposed that in deny-
ing the validity of the first premise Cleanthes meant to say not that
past-tense propositions not about the past are not necessary, but that
propositions about the past tout court are not necessary. It is in this
strong sense that Leibniz understood his position.
He writes in a passage that could be clearer: "There is a question
whether the past is more necessary than the future. Cleanthes was of
this persuasion. It is objected that it is necessary ex hypothesi that the
future happen, like it is necessary ex hypothesi that the past should
have happened. But there is this difference: it is not at all possible
to act on the past state, that's a contradiction; but it is possible to
bring about some effect on the future. Nevertheless, the hypothetical
necessity of the one and the other is the same; the one cannot be
tatus and of the fourth article (L et seq.), whereas it implies only the abandonment,
in the sense of Carneades, of the dogmatic definition of truth. The texts alleged
say: 1) that from necessary science to the necessity of the known the consequence
does not hold (ibid., p. 18); 2) that from the immutability of the truth-value of a
statement about a future contingent to its necessity, the consequence does not hold
(ibid., p. 29). But this implies only that a proposition can be determinately true
without its causal chain being already present for all that.
 THE SECOND CONJECTURE / 93

changed, the other will not be and, with that posited, it cannot be
changed either."43
Leibniz seems to say that, after having identified the hypothetical
necessity of the past with that of the future, Cleanthes added something
to the hypothetical necessity of the past that explained why delibera-
tion could not apply to it. But what is that something else? Isn't it
precisely what for Aristotle and Diodorus and nearby all the Ancients
made the hypothetical necessity degenerate into a sort of absolute ne-
cessity, insisting that it is only a sort of absolute necessity because it is
solely retrospective and, insofar as the event is concerned, it has only
to do with its time, not its mode of production? And what was Leibniz'
own position? "Voluntary actions and their consequences", he writes
"will not take place whatever we do, whether we will then or not, but
because we will do and because we will will to do what leads to them.
And this is contained in the forecast and in the predetermination and is
even their reason for being. And the necessity of such an event is called
conditional, hypothetical, or again, necessity of consequence, because
it supposes the will and the other requisites; whereas the necessity that
destroys morals and that renders punishment unjust and reward su-
perfluous is in the things that will be no matter what we do and what
we will to do, and, in a word, in what is essential. And this is what we
call an absolute necessity. Nor will it serve at all, as regards what is
absolutely necessary, to make prohibitions or commandments, to pro-
pose penalties or rewards, to blame or to praise; it will not be any the
more or any the less for all that."44 Let us compare this passage with
that already quoted from the Nicomachean Ethics (Z, 2, 113965-11),
keeping in mind that voluntary choice and project do not apply to the
past.
Aristotle says that the past makes the conditional necessity of the
event degenerate into a sort of absolute necessity. Leibniz seems to
agree with him on this point, though he himself makes the conditions
on conditional necessity more stringent. It no longer suffices to apply
it to momentary events, because now there is the additional demand
for the requisites of voluntary deliberation. Leibniz even gives as his
reason for agreeing that it is a contradiction to suppose that we have
any effect on the past; but what is possible is that which is not the
antecedent of a contradictory consequence. If something is possible we
can have an effect on it; but we cannot have an effect on the past;
therefore the past is not possible. This is clear. Leibniz concedes that
43
Leibniz, Gerhardt, VI, p. 214; Jalabert, p. 226.
44
Leibniz, Gerhardt, VI, p. 380; Jalabert, p. 383.
94 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

the necessity of the past signifies nothing other than the impossibility
of realizing the possible in the past. But it is just this conclusion that
contradicts Cleanthes and puts Leibniz in agreement with most of his
adversaries, among whom Aristotle must be numbered as well.
To situate these Leibnizian intentions let us return to the Master
Argument. Leibniz praises Aristotle for having distinguished between
conditional and absolute necessity. He even claims to borrow this dis-
tinction from him. But he blames Aristotle for his doubts relating to
the principle of the excluded middle. But in Chapter IX of the De
Interpretatione Aristotle combines these doubts with that distinction
in order to escape the Master Argument whose three premises, which
Diodorus borrowed from him moreover, he admits. In crediting Clean-
thes with the distinction between the two sorts of necessity, Leibniz is
interpreting it as tantamount to a denial of the first premise, i.e. of
the absolute necessity of the past. Since he himself, given his theory
of possible worlds, admits the existence in the divine understanding of
possibles that will never pass over to an empirical existence and since
he does not challenge Diodorus' second premise, we can suppose that
what he attributes to Cleanthes is a doctrine quite like his own. To
think that the past is absolutely necessary is to blot out the distinc-
tion between conditional and absolute necessity. Every event, past or
future, is determined. But what is determined is in nowise necessary
if its existence-conditions, which caused or will cause it to be chosen,
depend not solely on the logical principle of non-contradiction, but on
the principle of the best, as well. But any event, in so far as it is part of
this world, exists on the condition of what Leibniz calls the free choice
of the best and what Cleanthes called Divine Providence.
This distinction of two necessities, which for Leibniz and for Clean-
thes as interpreted by Leibniz, though not for Aristotle, suffices for
avoiding fatalism, is itself amenable to two different forms of expres-
sion. On the one hand, in asserting that the past is hypothetically
necessary, one might be saying that it is necessary that an event oc-
cur if it has occurred; whence it is not possible to conclude that it is
simply and absolutely necessary that that event should have occurred.
On the other hand, one might posit that if an event has occurred, it
was necessary, during the time of its occurrence, that it should occur;
the necessity is thus conditional in that it depends on the duration of
a past event, whence again the impossibility of concluding that it was
simply and absolutely necessary that that event should have occurred.
On either of the two different forms of expression, the first of which
is more Leibnizian, the second more Aristotelian, although Aristotle
himself does not apply it to the past,conditional necessity entails a
 THE SECOND CONJECTURE / 95

complete symmetry of past and future. For it can be said of a future

just as of a past, on the one hand, that it is necessary that it will
be if it is, as the past is necessary from the fact that it was, on the
other, that it is necessary, while it will be, that it be, as it is neces-
sary that the past should have been while it was. From the point of
view of Divine Providence moreover, the principle of the best and the
principle of determination apply equally to the one and the other. If
Leibniz distinguishes them it is because he appeals to the impossibility
of realizing the possible in the past. He consequently superposes on
the conditional necessity applying to all existences, past and future, an
irrevocability characteristic of the past and appropriate to rendering it
necessary, hypothetically too, though in a sense other than the future.
It is probable that he attributed the same distinction to Cleanthes.
According to Aristotle, conditional necessity degenerates into a sort
of absolute necessity whenever either the limits of an event belong to
the past or, although they belong to the future, there is already present
a causal chain determining them and allowing us to consider them as
already given then. 45 The separation of the ways comes in between
that about which one may still deliberate and the rest, including that
about which one could but no longer can deliberate—the irrevocable—
and what arises from external necessity and upon which deliberation
can have no effect. For the Leibnizian Cleanthes, on the contrary,
the past in itself, although removing the event from the province of
deliberation, does not plunge it into necessity that is somehow stronger
than the conditional variety common to all existents. The irrevocability
blocking realization of the possible in the past does not confer brute
necessity on that past.46
The authority of Leibniz is not to be taken lightly. There are two
arguments however that prevent us from accepting his opinion. The
first has to do with the internal economy of the Master Argument, the
second with the historical likelihood of a Leibnizian Cleanthes.
In the first place, thinking that he is interpreting Cleanthes, Leibniz
adopts a peculiar stance regarding the argument's first premise. On
the one hand he weakens its modal force in saying that the necessity of
the past is conditional, not absolute. On the other hand, he enhances
this in refusing to reduce it to the general hypothetical necessity such
as he attributes to futures. In doing this he refuses the complete sym-
metry of past and future. At no point does he clearly spell out just in
45
This is the case for "tomorrow's eclipse". (Metaphysics, K, 8, 1065" 16; cf. below,
6.5, p. 144).
46
This distinction corresponds to Schuhl's solution of the Master Argument. As
von Fritz has pointed out (1962, p. 145) this turns out to be Cleanthes' solution.
96 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

what this modal surplus consists that enables distinguishing the neces-
sity of the past from the necessity of the existent in general, without
assimilating it to metaphysical necessity for all that. For Leibniz, a
complete symmetry of past and future would probably have been seen
as entailing the nonsense of a Pierre Damien saying that God could
make what was not have been. If this is nonsense, it is because the
necessity of the past is specific in just that it is impossible to realize
the possible in the past. Time must then be taken to be asymmetrical
and linear. But if this is the case what has been gained for solving the
Master Argument? The interpretation of the first premise requires no
more than the irrevocability of the past. On the other hand, Leibniz
does not seem to contest the argument's other premises. He does have
the merit of showing the real difference there is between the irrevoca-
bility of the past and the necessity of a mathematical deduction. But
that is a difference Diodorus can accommodate.
There remains the question of the historical likelihood of the Leib-
nizian reconstruction. Let us first point out that the distinction be-
tween two sorts of necessity was known to the Ancients as an Aris-
totelian one. Had Cleanthes, the Stoic, taken it over from Aristotle,
the antique sources would have noted it. Above all, however, in in-
terpreting the first premise, Leibniz no more denies its validity than
had Ockham before him. The premise asserts the necessity of the past:
Ockham had clarified the sense of the word past, Leibniz clarifies the
sense of the word necessity. These clarifications purify, they do not
contest the premise itself.
The Epictetus passage reports that Cleanthes meant to reject the
premise, not to amend it. The conjectures advanced thus far are there-
fore irrelevant to his project. In particular, we have seen Leibniz de-
fend, without justification, the linearity of time, in order to safeguard
the asymmetry of past and future. It is in this way that it remained
impossible to realize the possible in the past. But if one is to inval-
idate the first premise, isn't it just that impossibility that must be
questioned?
There remains open then one last possible conjecture that would
allow for an interpretation in terms of physics and of the properly
Stoic conception of time and would negate the Master Argument's
first premise. First the condition of possibility of that interpretation
will be examined. It will then be shown that it seems to correspond to
Cleanthes' theory.
 THE THIRD CONJECTURE / 97

4.4 Third conjecture: cyclical time and the

numerical conception of the identity of beings
in eternal return.
There is a passage in the Problems of Aristotle that helps us fix the
issue. The impossibility of realizing the possible in the past was based
on the linearity of time. Realizing the possibility in the past would
make last year coincide with now. But there are philosophers who
have maintained that absurdity.
"After what manner", asks the pseudo-Aristotle,47 "are we to un-
derstand the words before and after! Are they to be taken in the
following manner: Those who lived in the days of the Trojan War are
before us, and before them are those who lived earlier still, and so on
ad infinitum, those who are found to be in the past being always taken
to be before the others? Or, if it is true that the Universe has a be-
ginning, a middle and an end; that what by aging has reached its end,
should, in virtue of that have come back again to its beginning; if it is
true moreover that things which are before are those that are closest to
the beginning, what is to prevent our being nearer the beginning [than
those who lived in the days of the Trojan War]? If that were so, then
we should be before them. Since by its local motion the firmament and
every star goes in a circle, why shouldn't it be the same for the coming
into being and decay of every perishable thing, such that that same
thing might also come into being and decay again? So they say too
that human affairs go in a circle. To think that the men who are being
born are numerically the same is absurd; but a better opinion would be
expressed in saying that they are preserved as to their form. It could
be then that we are before even [the contemporaries of Troy]. The
series of events will then be taken to be such that there is an inevitable
return to the state which served as point of departure and the taking
up again, without discontinuity, of a course which passes again by the
same things. Alcmeon said that men are perishable because they are
unable to join their end to their beginning. That is well put, if he is
taken to have spoken figuratively and one doesn't insist on taking what
he said literally. If the series of events is a circle, and since a circle has
neither beginning nor end, we cannot, by virtue of a greater proximity
to the beginning, be before those others, nor they before us."
Suppose then that the Master Argument's first premise signifies
simply the impossibility of realizing the possible in the past. On what
condition would that impossibility disappear? It is necessary and suffi-
47
Problems, XVII, 3, 916" 18-39.
98 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

cient 1) that the universe be subject to comings-into-being and decays

in such a way as to have a middle, beginning and end, 2) that these
generations and decays form a reproductive cycle, 3) that that repro-
duction be numerical, not simply specific. The first two conditions are
easy enough to see. The third is necessary. For if the reproduction
of the universe were only specific, we could number each successive
universe or, if that succession is from all eternity, we could at least dis-
tinguish it. There would be the universe in which Socrates lived and
the universe in which Socrates didn't exist. The clock, so to speak,
would be extraneous to the universe, since its phases would be dis-
cernible. But if it is numerically that the universe repeats itself, its
phases become indiscernible and it is the same time that repeats itself.
Then, and only then, is there a complete symmetry between past and
future. Then, and only then, can one contest the Master Argument's
first premise taken in its Aristotelian sense.
Just which schools of the followers of Heraclitus and Pythagoras fall
under this Aristotelian opposition of specific and numerical systems is
of no interest here. The issue is to know whether this same conflict
existed within ancient Stoicism. Zeno, in his treatise on the Eternity of
the World, which is directed against the Peripatetics, maintains that
the eternity of the universe is cyclical. After 365 times 10,800 years,
according to the count of Diogenes the Babylonian, the final confla-
gration purifies the universe and restores time.48 Neither Cleanthes
nor Chrysippus seem to have departed from Zeno on this part of the
doctrine. There is nothing that justifies doubting that ancient Sto-
icism was in harmony with the periodic palingeneses49 of the universe.
Panaetius is the first in the Stoic school to question the renewal of the
universe at each universal conflagration.50
Eusebius, in his Praeparatio Evangelical1 has the following to
say. "Here is what the Stoics have to teach about the conflagration
(lKTrvpuo-i<;) of the World: the oldest adherents of that sect thought
that all things, at the end of certain extremely long periods of time,
48
Brehier, 1962, pp. XVIII-XIX.
49
Goldschmidt, op.cit., 1977, p. 42.
50
Cicero, De Natura Deorum, XLVI, 118, in Brehier, p. 450.
51
Quoted by Duhem, 1951, t.I, p. 277. Following the suggestion of Father M.
Regnier we have corrected Duhem's translation to read had some reserve about in
place of insisted on. Duhem adds the following passage from Aetius (Arnim, S.V.F.,
II, p. 184) "They say that the organization [of the Universe] subsists eternally,
that there are certain periodic times at the end of which the same things are all
engendered again and in the same way, at the end of which the same disposition
and same organization of the World turn out to be safe and sound".
 THE THIRD CONJECTURE / 99

were etherized, that all things disintegrated into a fire resembling

ether...
"From this it is evident that Chrysippus did not think of that dis-
persion as affecting the very existence [of the World], for that is im-
possible, but as playing the role of a transformation (n£Ta(3o\ij). For
those who teach that disintegration of the Universe in a state of fire,
that they call conflagration (inTrvpuaK;), do not admit that that de-
struction of the World that recurs after very long periods of time is
a destruction strictly speaking. They use the expression 'destruction'
(ipflopa) in the sense of a natural transformation. Indeed, it was in
accordance with the Stoic philosophers that the Universe should be
transformed into fire, as into its seed (cnrepfia), and then that from
that fire should be produced again a disposition exactly like the one
that existed before. This dogma was admitted by the principal and
most ancient philosophers of the sect: Zeno, Cleanthes and Chrysip-
pus. It is said that Zeno [of Tarsus], who was a disciple of this last and
his successor at the head of the Stoic school, had some reserve about
the conflagration of the Universe...
"Then the common reason reappears to take up again the selfsame
course (iitl TCKTOVTOV). The common nature, become more ample and
full, finally drying out all things and taking them back into itself,
is engendered to full existence. It starts its course again according
to the rule it followed a first time. It recommences that restoration
(di/doTQcrtc) that accomplishes the Great Year. For, following that
Great Year, there comes about this renewal (<XTroKaTaoTacn<;) [of the
World] which starts from a certain state and returns again to the same
state. Nature begins again, in the order in which it deployed itself a
first time in like manner, to accomplish again, according to the same
law, the same series of events. And from an eternity the same periodic
cycles ceaselessly recur".
In such a world view one is led to say of the now, as opposed to
Aristotle, that it is last year, and of last year that it is now. The
upheaval among physical notions naturally entails an upheaval among
modal ones. The possibility of realizing the possible changes radically
its sense, once it involves the past as well as the future.
Simplicius reports that the Stoics were divided on the question of
whether the Eternal Return reproduced numerically or only specifically
identical beings. He also distinguishes two sorts of numerical identity:
the first seemingly absolute, the second having to do only with the
100 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

essence of the individual. "This question52 arises in connection with

the palingenesis (KaXi^eveaia) of the Stoics. For they say that, by
palingenesis, there will be reborn a man who is the same as me. And
they rightly ask if I will then be numerically the same as now, if I will
be the same by essential identity (6t,a TO if) ouoiqt cfvca 6 auioc;), or if
I will be different due to my insertion into another Universe from this
one." So it all boils down to knowing whether there was agreement in
the ancient Stoa to maintain the return of numerical identity and just
what kind of identity that was taken to be.
The testimony of Alexander of Aphrodisias allows us to attribute
to Chrysippus the thesis of numerical identity understood as having to
do only with the individual's essence. "According to the Stoics,53 there
must really be a time after the death of Dion at which the separation of
body and soul of the one designated by the name Dion takes place. For
they think that after the conflagration all things will be regenerated in
the world, and numerically the same, so that such-and-such a man in
particular (6 Wt'u>c TTOIO?) will be once again the same as before, and
will so be born in this new world. That is what Chrysippus says in his
books On the Cosmos...54
"They say further, that the particular men born afterwards are
affected only by differences, with respect to those having existed be-
forehand, touching simply on certain extraneous accidents. Such are
the differences apt to affect Dion during his life, leaving him nonethe-
less the same, for they in no way make another man. That he should
at first have worts on his face, for instance, and then later not have
any more, in no way makes him another man. They say that it is dif-
ferences of this sort that arise between the particular men of one world
and those of another."
There is no doubt that, as long as a difference subsists between
the numerical individual and his essential—although proper—55 de-
termination, it is impossible to assert the indiscernibility of their total
determinations. To have a wort or not, in any case, makes it possible to
assign two different moments of a same subject or two different palin-
geneses of the universe.56 Accidental and transitory differences enable
52
Simplicius, 1895, In Aristotelis Physic., p. 886; Arnim, S.V.F., II, n. 627, pp. 190-
191.
53
Alexander of Aphrodisias, In Aristotelis analyticorum priorum librum I com-
mentarium, 1883, p. 180; Arnim, S.V.F., n.624, vol. II, pp. 189-190.
54
Text attested to by the same Alexander, p. 181, 13.
55
On the theory of TO ifius TTOIOI/, which according to Chrysippus is what charac-
terizes each being in a permanent way, see: Brehier, 1910, pp. 111-112 and Gold-
schmidt, 1977, p. 17.
56
Arnim, S.V.F., II, n. 626, p. 190. Origen says that according to the Stoics "peri-
 THE THIRD CONJECTURE / 101

us to mark off the successive time of two states of a same individual

or of two universal palingeneses. Chrysippus could not, therefore, have
maintained their absolute numerical identity. Nor could he have re-
jected the linearity of time. Nor did he venture to reject the Master
Argument's first premise.
Extreme prudence is required then in interpreting the Stoic texts.
Where they claim the numerical identity of palingeneses, that iden-
tity may be imperfect. Tatian's testimony about Zeno does, of course,
seem to go further. "Zeno asserts57 that after the conflagration the
same men will be given to the same deeds, that Anytus and Meletus
will again be accusers, that Busiris will begin again killing his guests,
that Hercules will again accomplish athletic labors." Is there a per-
fect numerical identity here? We do know that Cleanthes remained
faithful to the Zenonian theses, whereas Chrysippus was an innovator,
not hesitating to oppose Cleanthes. It is difficult to grasp these points
of difference with exactitude. A text of Nemesius suggests however
that there is some connection between the insistence on astronomical
periodicity—itself astrologically interpreted and justifying divination—
and the interpretation of palingeneses in terms of absolute numerical
identity.58 And Cleanthes seems to have been especially interested in
astronomy. He constructed an original theory of the Sun which he
took to move in its sphere on a spiral path comprised between the two
tropics, with all the stars, both fixed and wandering, being transported
from East to West.59 While for Zeno and the other Stoics the ether
was supreme god, Cleanthes attributed that dignity to the Sun.60 It
was probably for that reason that he accused Aristarchus of impiety.
The nurturing warmth of the Sun sustains universal life.61 That fire is
the 'hegemon', self-moved and reasonable.62 The stars, beings of fire,
move by will and are divine.63 Probably an original development of
the theory of the fire artist as Zeno had taught it, Cleanthes' doctrine
was well suited to hellenizing the universal sympathy and the sun-cult
odically a universal conflagration takes place and after that conflagration is reborn
a disposition of the world exempt from any difference with respect to the dispo-
sition that had previously been realized. Many of them attenuated that teaching,
saying that the things of one period are subject to a slight difference, an extremely
weak difference in relation to the things of the preceding period." It is possible that
Chrysippus was the first to attenuate the ancient doctrine.
57
Tatianus, Adversus Graecos, cap. V; Arnim, S.V.F., I, n. 109.
58
Nemesius, De Nat. Horn., cap. 38, p. 277; Arnim, ibid., I, 109, p. 32.
"Tannery, 1912-1950, t. II, pp. 160-162.
60
Cicero, Academica, XLI, Brehier, p. 246; Cumont, 1909, p. 15 note 3.
61
Cicero, De Natura Deorum, IX, 24 sq., Brehier, p. 417.
62
Ibid., XI, 31, p. 419.
63
Ibid., XV, 40, p. 422.
102 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

that had come from the Chaldeans.64 Among all the Stoic doctrines it
is this one especially 'that represented the ethereal fire as the primor-
dial principle and held the stars to be the purest manifestation of its
power'.65
Cleanthes insisted on the astronomic basis of benevolence and fate.
His doctrine of the fire principle destined him to philosophically jus-
tify the sun-cult of the 'Chaldeans', as his Hymn to Zeus shows. It is
true that there is no extant text explicitly attributing a perfect numer-
ical interpretation of palingenesis to him, thereby setting him opposite
Chrysippus on this point. But all the remaining elements of his doc-
trine tend in this direction. It was thus within the logic of the system
to forgo the linearity of time and to draw from palingenesis what Ori-
gen considered to be the only consequence possible: perfect numerical
identity.66 But if time is symmetric, realizing the possible in the past,
and thereby inducing a reversal of time, is no longer absurd. What
scandalized Aristotle no longer presents a problem. Challenging the
Master Argument's first premise then-becomes legitimate.
In short, the Stoics extended to the universe as a whole the cyclical
return that had been restricted to the Heaven in Aristotle's system.
The ekpurosis realized a kind of synthesis of Heraclitus and Aristotle
and answered to Diodorus' negation of motion.
The fatalism of Cleanthes and Antipater 67 is no longer of a logical,
but of a physical, order. The total determination of the sublunar by
the supralunar world, in conformity with the astrological dogmas, and
the total stability of the system of astronomic equations, in conformity
with the dogma of eternal return, bring freedom down to the joyous
acceptance of destiny. It is this pantheism that supports the rejection
of Diodorus' first premise. As Stobaeus says, "As for Cleanthes, he
speaks as follows:... The tension that is in the substance of the universe
ceaselessly produces always the same revolution and same arrangement.
For just as all the parts of a same individual are born of seeds at the
proper times, so too do the parts of the universe, which include the
animals as well as the plants, come into being at the proper times.
And just as certain seminal reasons of the parts, condensing into a
seed, mix and separate again when the parts are born, so do all things
64
Cumont, 1929, pp. 160-162.
65
Cumont, 1912, p. 69; on Cleanthes' intelligible fire see Cumont, 1909, p. 15 note
2.
66
Origen, Contra Celsum, V, XX; Arnim, S.V.F., II, n.626, p. 190.
67
Who, in the Middle Stoa, maintained the doctrine of the conflagration.
 THE THIRD CONJECTURE / 103

come from one sole being and all condense into one, the revolution
having been accomplished, and this in conformity with the order."68

68
Stobaeus, Eel. I, 17, 3, p. 153, 7 W (Arii Did., fr.38 Diels), Arnim, S.V.F., I,
n.497, p. 111.
Freedom as an Element of Fate:
Chrysippus.

The uncertainties regarding Chrysippus' solution justify the establish-

ment of certain preliminary 'lemmas'.
Since it has been supposed that the Master Argument's second
premise is a thesis of pure modal logic, the discussions have developed
within this context. They gave sound results though their context itself
may and, according to the present interpretation, must be questioned.
This explains why Chrysippus' reform will be analysed as if it con-
cerned pure modal logic before it will be reinterpreted in the light of
temporal modal logic.
1. Basing ourselves on the Chrysippean theory of divination, it
will first be shown that, if Chrysippus does challenge a thesis of pure
modal logic, his doubts have to do with the literal expression of that
premise. What he contests is that from the possible, the impossible
does not logically follow. He admits, on the other hand, the validity of
the affirmative form one is tempted to give of that same premise: from
the necessary, the necessary logically follows.
2. The counter-example advanced to falsify the premise in its nega-
tive form in no way falsifies its positive form. The question it actually
raises is that of the validity of the interdefinability of the modalities.
3. The Stoic definitions of the modalities corroborate the view that
Chrysippus adopted a non-standard modal system.
4. This system is indeed related to Prior's 'system Q', which he
himself attributes to Chrysippus.

105
106 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

5.1 Were Chrysippus' doubts about the thesis of

pure modal logic according to which from the
possible the impossible does not follow, they
would be about its negative form, not about
its positive form.
At first glance, Cicero might be seen as asserting that Chrysippus chal-
lenges the Master Argument's second premise in its positive form, that
is to say, in that it posits that the necessary logically follows from the
necessary.
It is in the De Fato (VII, 14) that he analyzes the question of the
relation of divination to necessary futures.
"If the following is a true connection, 'If you were born at the rising
of the dogstar you will not die at sea' and if the antecedent in the con-
nection, 'you were born at the rising of the dogstar', is necessary (for
all true propositions about the past are necessary, as Chrysippus holds,
in disagreement with his master Cleanthes, because they are unchange-
able and cannot pass from true to false)- if, therefore, the antecedent
in the connection is necessary, the consequent becomes necessary too.
Although Chrysippus does not think that this principle holds in all
cases"1.
How is this passage to be interpreted?
The conditional is said to be genuine or sound, not simply because
it is true, but because it expresses a law of astrology: it therefore
expresses a necessity. The antecedent is itself necessary since true in
the past. Are we to suppose then that Chrysippus would refuse in such
a case to conclude that the consequent 'you will not die at sea' is itself
necessary?
If this were so, all science by inference would be impossible. But
Chrysippus, it must be remembered, is a dogmatist. Far from refusing
the preceding argumentation, he is so convinced of its formal legitimacy
and that if its premises were accepted divination would inevitably entail
necessity, that he recasts the very form of divination in order to avoid
that untoward conclusion. Divination will no longer be conditional,
but conjunctive.
"At this point Chrysippus gets nervous and hopes that the Chal-
deans and the other diviners will no longer employ expressions of the
form 'If anyone was born at the rising of the dogstar he will not die at
sea', but will say rather 'It is not the case both that someone was born
r, 1968, p. 478; Rackham, 1942, pp. 208-209.
 CHRYSIPPUS' DOUBTS / 107

at the rising of the dogstar and that he will die at sea' " (De Fato,
VIII, 15).
Just what does this rather odd-seeming transformation come down
to? If it is meant to reconcile fate and divination on the one hand
and to allow for non-necessary futures on the other, it must evidently
be that Chrysippus took this conjunctive form to be weaker than the
conditional one for which he had reproached the astrologers. As shown
by the passage quoted above (VII, 14), the astrological conditional ex-
presses a necessity: 'it is necessary that if anyone was born at the
rising of the dogstar, he will not die at sea'. As the antecedent is itself
necessary, since past, positing the necessity of the consequent, which
is future, becomes inevitable. It must be then that the change from
that conditional to the conjunctive form constitutes a weakening of as-
trological necessity. Since, for Chrysippus, the doubt cannot affect the
necessity attaching to the antecedent, it must be that the conjunctive
form weakens the very connection between antecedent and consequent.
What is the reason for that weakening in the force of the consequences?
Having distinguished 'sound' implication according to Philo and
according to Diodorus,2 Sextus goes on to examine a third theory
of the conditional, probably corresponding to the Chrysippean the-
sis. "Those who introduce the notion of connection (avvdpTrjaLt;) say
that the conditional is sound when the contradictory of its consequent
is incompatible with its antecedent. According to them, the condition-
als mentioned above to illustrate the doctrines of Philo and Diodorus
are unsound, but the following is true: 'If it is day, then it is day'."3
The conditional posited by Chrysippus is stronger then than that of
Diodorus and, a fortiori, than that of Philo. It is stronger than the
Philonian material implication and stronger than the Diodorean for-
mal implication, for Chrysippus adds a proper modal clause to the
conditional. The incompatibility between the negation of the conse-
quent and the truth of the antecedent transforms the conditional into
a sort of 'strict' implication,4 that is to say, into the statement of a
law. Since Chrysippus does not, however, spell out the nature of this
incompatibility—is it logical? physical?—it is not possible, without a
preliminary inquiry, to determine formally in just what the 'strictness'
of that implication consists. It must be added too that the example
given by Sextus could be misleading. When it is said that if it is day
2
See above, 3.3, p. 50.
3
Sextus Empiricus, PH II, 111-112.
4
A strict implication in Lewis' sense is one for which the conjunction of antecedent
and negation of the consequent is not possible. (Hughes and Cresswell, 1972,
p. 217).
108 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

then it is day, it seems that we could prefix the modal necessary to

the conditional, taking it to mean that it is necessary that if it is day
it should be day. But then it becomes difficult to make sense, on the
one hand, of the Ciceronian passage and, on the other, of the very
definition given by Sextus, enjoining us to understand the conditional
as the incompatibility of the contradictory of the consequent with the
antecedent in conjunctive form.
For Chrysippus, if the laws of divination were expressed in condi-
tional form in the astrologer's sense, the conditional would be necessary
and not simply true, as it would be if only the Philonian condition were
met, or even if one were to add the Diodorean condition, in so far in any
case as it is contested that simple permanence of the Philonian hypo-
thetical is enough to assure its necessity. Given a necessary antecedent,
the consequent would be necessary as well.5
Chrysippus' sound conditional must therefore be formally distin-
guished both from the conditionals of Philo and of Diodorus, since it is
stronger than they are, and from the astrologer's conditional criticized
by Chrysippus in the Ciceronian passage, since it must be weaker than
that one is. The astrological conditional signifies that it is necessary
that if an event p occurs the event q will occur; the Chrysippean 'con-
ditional' signifies only that it is impossible to verify at once both that
p occurs and that q does not.
But then how are we to explain that these predictions can lose their
necessity in taking on a conjunctive form?
One might first look for the reason for that weakening in the fact
that Chrysippus conjunction is a truth-function, whereas his condi-
tional and disjunction are not.6
But if it were due to the simple fact of taking on the conjunc-
tive form that the astrological argument lost its necessary force, it
would have to be because the (negated) conjunction is simply true.
5
As noted by M. Frede, 1974, p. 88.
6
M. Frede, 1974, p. 96. An analogous problem is met with in a discussion of
Ockham's reported by Prior, 1962, p. 242. Take the conditional 'If A is going to
occur, God knows that it is'. If the antecedent is false, the consequent is false and
the whole will be true. If the antecedent is true, the consequent is true. In either
case, the whole will be true. But if the antecedent is neuter, the consequent will be
false; and the whole will be neither true nor false. Taking the conjunctive form, by
contrast, 'A is going to occur' and 'God does not know that A is going to occur'
cannot both be true together. Therefore it will be true that it is not the case that
A is going to occur and God does not know it. In this case the conditional does
not follow from the corresponding conjunction.
That discussion nevertheless does not apply to our case where we are dealing
with modal and not simply assertoric propositions. But there is a notable analogy
in so far as the respective consequences are concerned.
 CHRYSIPPUS' DOUBTS / 109

For, in that case, a necessary consequence could not be drawn from

an assertoric proposition.7 But isn't it clear that that would be to
reduce the laws of astrology to simple repeated concomitances, and
to make Chrysippus the Hume of astrology? We must therefore de-
termine just what is the nature of the modality affecting conjunction.
Cicero too explicitly precluded this first solution in De Fato, VI, 12,
"... the propositions 'Fabius was born at the rising of the dogstar'
and 'Fabius will die at sea' are incompatible, and since it is posited
as certain about Fabius that he was born at the rising of the dogstar,
the propositions 'Fabius exists' and 'Fabius will die at sea' are also in-
compatible. Therefore the conjunctive proposition 'Both Fabius exists
and Fabius will die at sea' is composed of contradictory propositions,
and the event, according to the principle, cannot occur. Therefore
the event 'Fabius will die at sea' is of the kind of those which are
impossible. Therefore every false proposition about the future states
an impossible event".8 It is incontestable in this argumentation that
the conjunction, being contradictory, is impossible, for to contest that
would be to give up astrology and fate. But, in that conjunction, one
of the terms is true: namely, that Fabius exists. It must, therefore, be
the event designated by the other term that 'cannot occur'. Such is
Cicero's conclusion. Chrysippus, though, would have contested such a
conclusion. Indeed, from what Cicero expressly says in De Fato, VIII,
16, it can be seen that he assumes the right to go from the necessary
antecedent-consequent connection to the negation of the sound corre-
sponding conjunctive proposition, and naturally back again from such
a negation to such a connection.9 If the conjunctive form is meant to
save the possibles without destroying fate, it must be this last trans-
7
M. Frede, 1974, p. 88. Frede points out further that, given L ( p - q ) , the necessity
of p and of q would follow by the thesis: L(p • q) 3 (Lp • Lq). That thesis is in
fact demonstrable as a theorem of modal logic (Hughes and Cresswell, 1972, T. 3,
p. 34) without appeal to the Master Argument's second premise, which is required
only for the demonstration of its converse.
8
Brehier, 1962, p. 477; Rackham, 1942, pp. 206-207.
9
If 'Fabius was born at the rising of the dogstar' is designated by 'Pp' and 'Fabius
will die at sea' by 'Fq', Cicero's reasoning can be expressed as follows, where D
designates the sound Chrysippean conditional.
Pp D~ Fq =~ M(Pp- ~~ Fq)
=~ M ~ (~ PpV ~ Fq)
= ~ M ~ (Pp D~ Fq)
* = L(Pp D~ Fq)
3 LPp D L ~ Fq
But LPp. Therefore L ~ Fq.
The invalid argument step is marked by an asterisk (cf. note 11).
110 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT
formation that Chrysippus contests, and not the Master Argument's
second premise in general, for then no science would be possible.

5.2 Chrysippus' doubt about the interdefinability

of the modalities. From the non-possibility of
an event's occurrence it cannot be concluded
that its opposite is necessary.
This line of reasoning is supported by a seemingly peculiar argument
recorded by Alexander of Aphrodisias.10 Chrysippus constructs a
counter-example not, as is often said, to the second premise's positive
form, which maintains that the necessary follows from the necessary,
but to a negative form maintaining that the possible, and only the
possible, follows logically from the possible.11
10
In an. pr., 177, 25-33. "Chrysippus, in maintaining that there is nothing to
prevent an impossible's following from a possible, says nothing against the proof
given by Aristotle, but tries to show that it isn't so by way of examples which
are not correctly constructed. He says that in the conditional proposition 'if Dion
is dead, this man is dead', which is true when Dion is designated,the antecedent
'Dion is dead' is possible, for it can be true that Dion dies, while 'this man is dead'
is impossible, for after Dion's death the proposition 'this man is dead' destroys
itself, since the object designated no longer exists. For the designation concerns
something living and is attributed to a living being. Then when Dion is dead, if
'this man' is no longer possible and Dion no longer subsists, so that 'this man is
dead' cannot be said of him, then 'this man is dead' is impossible. That proposition
would not be impossible if, after Dion's death, it were still possible to attribute 'this
man' to precisely that to which was attributed 'this man is dead' in the conditional
proposition while Dion was alive. Since that cannot be done it follows that 'this
man is dead' is impossible". (Mignucci, 1978, p. 378).
11
If we abstract from the temporal indices, it is a question of the formula:
(B')* ~ M(p- ~ q) D (Mp D Mq)
That formula, rejected by Chrysippus, is equivalent to
* ~ M(p- ~ <j) 3> (~ Mq D~ Mp)
and
* ~ M ~ (p D q) D (Mp D Mq).
As for the formula:
(B")L(p 3q)D (Mp 3 Mq),
which, in spite of appearances, can be taken as another negative form of (Bj):
(Bi) L(p 3 q) D (Lp D Lq),
it can be derived from (Bi) in the following manner (Hughes and Cresswell, 1972,
T 8, p. 37):
1. h L(p D q) D (Lp D Lq) (axiom) (Bi)
2. h £(~ 9 3~ p) D (L ~ 9 3 Z, ~ p) (Sb[~ q/p,~ p/q])
3. I- L(p D q) D (~ L ~ p D~ L ~ q) (I— q D~ p = p D 9,1,2, Syll. M.P.,
and contraposition in second member)
4. I- L(p 3 q) D (Mp D Mq) (interdefinability of M and L).
 THE INTERDEFINABILITY OF THE MODALITIES / 111

The conditional given is 'if Dion is dead, this man is dead'. It is not
only true 12 but 'sound' as well, for there is incompatibility of negation
of consequent and affirmation of antecedent. This conditional that is
'sound' for Chrysippus is a fortiori so for Diodorus; it constitutes a
'strict' implication. Further, 'Dion is dead' is possible, for it can be
true one day that Dion should be dead. But 'this man is dead' is not
possible. Employment of 'this man' presupposes that one can show the
object in question, but precisely one can no longer show it once Dion
is dead.
We shall examine successively 1) just what Alexander's example
involves, 2) how it fits in with the question raised as regards the Master
Argument and 3) its connection with the conjunctive expression proper
to Chrysippean laws.
1) The argument could be taken as an ad hominem one.13 In that
case, it is the sentence 'This man is dead' that is impossible, because
one could never utter or formulate it.
But it is unnecessary to have recourse to the sentence and we can
suppose that the argument is formulated in terms of propositions and
is taken into account by Chrysippus himself. 'Dion is dead' or 'This
man is dead' will be taken then as a 'lecton', that is, as the class of
all sentences of the same form, completed by the specification of the
circumstances that determine their truth-values. Among the singular
propositions, two sub-classes may be distinguished. The simple mean
propositions are expressed by sentences of which the subject is a proper
The validity of the demonstration depends at step 4. on the validity of the impli-
cation
~ L ~ p D Mq,
that is to say, on the interdefinability of the modalities.
The sound Chrysippean implication is
p D q =~ M(p- ~ q),
which is equivalent in turn to ~ M ~ (p 3 q), but not to L(p D q).
Among the expressions used, it is surely (B') then that Chrysippus contests. The
care he takes in refusing to admit the 'conditional' form of the Chaldeans shows
that he accepts the validity of (Bi). Consequently, either 1) Chrysippus refuses
the implication ~ M ~D L in all cases and is driven, from contesting ( B 1 ) , to
contest 4 as well, or 2), which seems more probable, he limits the invalidation of
the implication to certain cases only, and postulates its validity in particular in
formulas beginning with L. He would then accept 4 in mathematical arguments,
for instance.
12
M. Frede, 1974, p. 116, says of this implicative statement that 4t is true, which
we must naturally understand in a Chrysippean sense. The statement:
(p D q) D (Mp D Mq)
can certainly be falsified.
13
Kneale and Kneale, 1962, p. 127.
112 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

name ('Dion is dead'). The simple definite propositions are expressed

by sentences in which the proper names have been replaced by egocen-
tric particulars or deictics ('This man is dead'). It is a logical law to
derive from a simple mean proposition the corresponding simple def-
inite proposition: 'If Dion is dead, then this man is dead', or 'If it is
night, then this day does not exist'. These conditionals must even be
sound in Chrysippus' sense, since there is incompatibility between the
antecedent and the negation of the consequent. For given that Dion is
dead, it cannot be, with 'this man' designating Dion, that 'this man
is not dead' should be true.14 The antecedent, on the other hand, is
possible. At the same time, the consequent is not possible. For, in
virtue of the Chrysippean definition,15 that is impossible which is such
that, if it is capable of being true, external circumstances prevent its
being true. And 'this man is dead' fits that definition exactly. Either,
the proposition is capable of being true, which is the case when Dion
is living and 'this man' does in fact designate him, but then external
circumstances prevent its being true. Or, with Dion dead, the same
proposition is no longer capable of being true, since one cannot indi-
cate Dion once he is dead. Once the designated subject has perished,
any proposition of deictic form (with 'this one' referring to the subject)
turns out to be automatically destroyed,16 since the indicator there can
no longer refer to the subject. That is not simply to say that we would
be unable to formulate or express such sentences: it is the propositions
themselves that do not exist,17 for want of corresponding facts.
2) Such an argument is marvelously suited as a response to the
Master Argument. Chrysippus, along with Diodorus, admits the valid-
ity of the following conditional, based on the first and third premises
together with the principle of conditional necessity: 'If there is a pos-
sible which is not and never will be realized, given that the past is
irrevocable and that a proposition is necessary while it is true, then a
contradictory consequence follows'. But Chrysippus, who diverges from
14
With p = 'Dion is dead' and q = 'this man is dead', the truth table would give
the conjunction (p- ~ q) as true only if p is true and q false, which is impossible.
15
See below, 5.3, p. 113 and 5.4.
17
Prior, 1967, p. 149. Prior is right in drawing a parallel here between Chrysippus
and Ryle. It is known that for Chrysippus the proper attitude to take when a sorites
was being propounded was to remain silent when the borderline cases arose (Cicero,
Ac. pr. II, XXIX, 93; Brehier, 1962, p. 229). The Sceptics interpreted that attitude
as one of suspended judgment, seeing in it the failure of dogmatism. But far from
being "intended less as a logical solution than as a procedural recommendation"
(Sedley, op.cit., p. 91), such a solution arose directly from the fact that in these
borderline cases the proposition is destroyed.
 THE INTERDEFINABILITY OF THE MODALITIES / 113

Cleanthes and agrees with Diodorus in accepting that conditional, di-

verges from both Cleanthes and Diodorus when they, remarking that a
contradictory consequence is impossible, add that the impossible does
not follow from the possible and declare the impossibility of the condi-
tional's antecedent, therefore, from the impossibility of its consequent.
For Chrysippus, one can, and must, admit that conditional. But it
can happen that, even when the consequent is recognized impossible,
the antecedent remains nevertheless possible. Indeed, the antecedent
has the form of an existential premise , 'there is a possible of such
and such a sort', comparable to a singular proposition of which the
subject is a proper name, 'Dion is dead'. Due to the realization of
the possible assumed as a means of proof in the Master Argument, the
consequent will have the form of a deictic proposition about a certain
possible, comparable to the deictic proposition 'this man is dead'. The
antecedent 'there is a possible of such and such a sort' or 'Dion is dead'
remain possible, even though their deictic ectheses, 'this possible is of
such and such a sort' or 'this man is dead', with the demonstrative
referring to nothing, be impossible.
There is this difference between the example advanced by Alexan-
der and the solution Chrysippus gives for the Master Argument. In the
first case it is a singular proposition having a proper name as subject,
in the second an existential proposition that is possible, while the cor-
responding deictic proposition is impossible. The deictic proposition
turns out to be 'destroyed' in the first case because its subject has
ceased existing, in the second because its subject does not, and never
will, exist. But the inexistence of the subject prohibits neither the for-
mulation nor the existence of a modal proposition. It does, however,
prohibit the formulation and existence of a deictic one.
What Chrysippus contests in the second premise, therefore, is the
principle according to which, when an antecedent with an inexistent
subject is the valid condition of a deictic consequent, the impossibility
proper to this second should entail the impossibility of the first. It is
not then the second premise's affirmative form that is at issue, but its
negative form and thereby the interdefinability of the modalities. For it
is the 'destruction' of the deictic propositions that invalidates that def-
inition. Thus for Chrysippus, those conditionals are valid of which the
antecedent is true (or even necessary, as in Alexander's example, since
it is a proposition true of the past) and the negation of the consequent
not true, so that there is a real contradiction between antecedent and
denial of consequent. Yet the non-truth of the consequent's negation
in all possible worlds is not enough to assure that consequent's truth
in any one of them at all, since that would require the formulability
114 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

or positive realization of the corresponding proposition in a possible

world, and this is precisely what is not given. As regards the conse-
quent, ('this man is dead' or 'this possible is not and never will be
realized'), either of two cases may arise. 1) The consequent is capa-
ble of being true in the worlds in which the antecedent is false (Dion
being living or given a possible of such and such a sort), but then it
is prevented from being true by the external circumstances expressed
precisely by the antecedent's negation. 2) It is not capable of being
true due to the fact that the proposition is destroyed in all the worlds
in which the antecedent is true. These two cases being exhaustive, the
consequent is impossible. Thus both in Alexander's example and in the
Master Argument (in the form 'If there is a possible that neither exists
nor ever will, then this possible neither exists nor ever will') there is a
sound conditional in which an impossible consequent follows logically
from a possible antecedent.
3) It still remains to determine just what the connection is between
Alexander's argument and Chrysippus' predilection for the conjunctive
form. Indeed, Alexander's example, just like Sextus', could be mislead-
ing. The conditional is sound, according to Chrysippus. because it is
impossible that Dion should be dead and this man not be dead. So
what is expressed by a sound conditional is the impossibility of the
conjunction of the putative antecedent with the contradictory of the
putative consequent. Yet that properly Chrysippean form can be given
an equivalent expression in terms of Philonian implication. The con-
ditional will be sound, from Chrysippus' point of view, when prefixed
by the modal operator 'it is impossible that not'. On the other hand,
prefixing the operator of positive necessity, which has a greater logical
force, is forbidden. Accordingly, Chrysippus' predilection for the con-
junctive form is due less to the form itself than to the negative and
weakened nature of the modal operator it supplants. The diviners are
guilty not of transforming the and into if, then but of going surrep-
titiously from the words 'it is impossible that not' to the words 'it is
necessary that' 18 in doing so.
18
Abstracting still from the temporal indices of the modalities, Chrysippus admits
I- L ~ p D~ Mp, h M ~ p 3~ Lp, I- Lp D~ M ~ p and I- Mp D~ L ~ p
but rejects their converses.
It is the rejection of
* ~ A/ ~ p 3 Lp
that invalidates ( B 1 ) , namely:
* ~ M ~ (p D q) 3 (~ Mq D~ Mp).
Were it legitimate, with the interdefinability of L and M, it would be possible to
obtain
 CHRYSIPPUS' NON-STANDARD MODAL SYSTEM / 115

It is easily seen how the form chosen by Chrysippus is imposed by

both Alexander's example and the Master Argument. In both cases
it is only the impossibility of having the conjunction of a proposition
p with the negation of a proposition q that can be shown. It cannot
be shown, on the other hand, that it is necessary that if p, then q.
For, on the supposition of the truth of p, it can be established that
the negation of q is not true; but it cannot be concluded, on the same
supposition, that q is true, since q is destroyed.

5.3 The non-standard modal system according to

Chrysippus.
The Stoic, and probably Chrysippean, definitions of the modalities
transmitted by Diogenes Laertius19 and Boethius20 are complex. The
commentators have all taken it for granted that interdefinability was
the rule, and their interpretations have accommodated that supposi-
tion. Independently of that supposition the texts are ambiguous.21
Yet no matter what interpretation one should settle on for the word
possible, interdefinability will be seen to be invalid. Consider the in-
terpretation the most favorable to its validity:22 'The possible is that
which can be true and which is not prevented by external circumstances
from being true'. It is immediately apparent that the necessary entails
the not-possibly-not, since the true entails the not-false, but that the
converse is not legitimate if the not-false fails to automatically entail
the true, as happens in the case where the proposition is 'destroyed'.23
Let us return now to the conjunctive form that Chrysippus would
give to the diviners' predictions instead of the usual conditional form.
Most often that conjunctive form consists in the impossibility of the
negation of a conjunction having to do with fate. That particular
expression evidently suggests a relationship between the rejection of
L(p D q), and then, by (B)",[L(p D q) D (~ Mq 3~ Mp)], the consequent
~ Mq D~ Mp.
It is remarkable that Chrysippus uses the sorites form in defense of Stoic theses. He
does not use them though in Carneades' hypothetical form, but rather as negated
conjunctions (Sedley, op.cit., p. 91, who rightly adds that "this formulation ... is
used to show that the relationship of the two propositions is something less than
one of strict entailment"; still, the non-necessary is not for all that identical with
the simply possible).
19
VII, p. 75.
20
In de int. II, p. 234.
21
Frede, 1974, pp. 107-114. On denial of the validity of the interdefinability of the
modalities see Vuillemin, 1983.
22
Frede's second interpretation, 1974, p. 108, p. 112.
23
Mignucci, 1978, pp. 327-330.
116 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

modal interdefinability, on the one hand, and the substitution of the

conjunctive for the conditional form of divination, on the other. What
is the nature of this relationship?
Cicero, setting the scene with Chrysippus at grips with the 'Idle
Argument', has him distinguish two sorts of assertion. "There are, in
reality, isolated assertions and assertions bound together. Here is an
isolated assertion: 'Socrates will die on a certain day'. Should he have
done some particular thing or have refrained from doing it, the day of
his death is determined. But if fate has it that Oedipus will be born
of Laius, one cannot say: 'should Laius have had intercourse with a
woman or should he not have', because the event is bound and 'confa-
tal', as he says, for fate has it both that Laius will have intercourse with
his wife and that he will beget Oedipus who will kill him" .24 Chaste,
Laius would not have been killed by his son; and that chastity was
dependent on his assent, that is to say, his nature. It is illusory to
think then that Laius is a plaything of external fates that dragged him
along in spite of himself, since the act from which his trouble stemmed
was his own. The Idle Argument posits right off that it is impossible
that Oedipus not be born of Laius and therefore not kill him, since the
impossibility is there whether Laius have intercourse with a woman or
not. All that we have the right to assert is that it is not possible that
Laius have intercourse with a specific woman and that Oedipus not
be born of him. For the edict of fate to be legitimate, that is, for the
conjunction of these two propositions to be impossible, it must be that,
if it is capable of being true, external circumstances prevent its being
true. To grant the prophecy its dramatic rather than its apotropaic
sense now, suppose that the first of the conjuncts is true (and even nec-
essary) and that Laius has had intercourse with a woman, and that the
second conjunct, namely that it will not be the case that both Oedipus
will be born of Laius and not kill him, is about the future. For the
conjunction to be simply impossible it has, if it is capable of being
true, to be prevented from being true by external circumstances. But
it is indeed capable of being true, since, Laius' intercourse with Jocaste
does not, if considered intrinsically or logically, contain in its analysis
the future occurrence of Oedipus' birth and his assassination of Laius.
However, there are external circumstances—foreign in themselves to
the essence of the past—precisely those precipitating the tragedy, that
prevent the conjunct from being true. What fate disallows is that, in
the fatal conjunction of two events, the anterior one should have oc-
curred without automatically being followed by the yet future one. But
24
Brehier, 1962, p. 484.
 CHRYSIPPUS' NON-STANDARD MODAL SYSTEM / 117

fate does not mean fatalism no more than physical connection means
moral constraint. The future event, that Oedipus will be born of Laius
and assassinate him, is not necessary. Its necessity would only be de-
rived from the necessity of the intercourse of Laius and Jocaste (which
is granted since it has already occurred) if it were positively necessary
that it be not the case that this intercourse has taken place and that
it will not happen that Oedipus be born of Laius and not assassinate
him. But this positive necessity requires not only that, were that con-
junction capable of being true, it would be prevented from being so by
external circumstances, but also that it be actually false. Then at least
one of the conjuncts must be false. It cannot be the past event of Laius'
intercourse with Jocaste, which is true and even necessary. Therefore
it is the future event of the non occurrence either of Oedipus' birth or
of his assassination of Laius. The negative impossibility differs from
the positive necessity and does not imply it, because the truth of the
future is not already given within the truth of the past. Formulated as
an astrological conditional, fate entails not only the impossibility that
Oedipus will not be born of Laius, but the necessity that he will be.
The Idle Argument becomes legitimate then, since the necessity of the
consequent can be detached.25
The prediction concerning Oedipus ought therefore not to have been
of the form 'It is necessary that Laius have intercourse with a woman
and beget Oedipus' but only 'it cannot possibly not happen both that
Laius have intercourse with a woman and that he not beget Oedipus'.
It is probable that if the Stoics insisted on the clauses of non-prevention
in their definitions of the modalities it is because they expressed the
laws of nature in the form: it is not possible not to have such and such
a conjunction.
Fate has to do with condestinates. The illusion of necessity is due to
the fact that one of the assertions—the birth of Oedipus—was isolated
and made to support the entire causal chain. This is how the cele-
brated example of the cylinder can be explained. "Just as in pushing
a cylinder one has given it a beginning of motion, but has not given
it the capacity to roll, so a sense-presentation will surely impress and
25
Taking Pp — 'Laius had intercourse with a woman'
and
Fq = 'Oedipus will be born of Laius',
the law of condestinates will be written: ~ M(Pp- ~ Fq). Given Pp, and thus LPp,
all that can be concluded without appeal to the interdefinability of the modalities
is:
~ M(Pp- ~ Fq)- ~ M ~ Pp. The conditional law L(Pp D Fq), on the other hand,
together with the given Pp, allows the detachment of LFq as the Idle Argument
would have it.
118 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

mark its form in the soul, but the act of assent will be in our power;
pushed from without, as we said of the cylinder, it will move by its
own force and nature. If something were produced without antecedent
cause, it would be false to say that everything happens by fate; but if
it is likely that everything that happens has an antecedent cause, what
reason could be adduced for not recognizing that everything happens
by fate provided that the distinction and difference between the causes
is well understood?" 26
It cannot be that a certain cylinder should not have a certain mo-
tion, if an external impulsion is impressed upon it. It is not for all
that that this motion bespeaks a brute necessity. For from the point of
view of providence it is simply a question of not possibly not bringing
about a certain conjunction, given the maximum perfection of the work
in conformity with the principle of the best. And from man's point of
view it is simply a question of not being able to avoid the consequences
bound to his acts, without its being for all that that the assent he gives
to moving mental representations is a brute and inevitable consequence
of the impression they make.
Leibniz made a famous commentary on this passage and its contin-
uation by AuliusGellius 27 . Chrysippus' cylinder is similar to Leibniz'
boat swept along at a greater or lesser speed in a river current. On the
one hand such a possible has its own nature and spontaneity which are
formal and not material, as is misleadingly suggested by the metaphor
of the cylinder and of the boat too; and they constitute the perfec-
tion of the individual in which is encompassed its assent28. And that
is what accounts for freedom. From the point of view of theodicy on
the other hand, if it is objected that the sheer fact that the cylinder
is by nature rough, the boat heavy or ungainly, are so many argu-
ments against providence, the reply must be that the partial evil is
in view of the general good, like the vulgar epigrams and inscriptions
in ancient comedy.29 Cicero's development of these two themes in the
De Natura Deorum is proof enough of the connection between Leibniz
and Chrysippus.30 Neither for the one nor the other do providential
26
De Fato XIX, 43; Brehier, 1962, pp. 489-490; assent is not assigned by the specific
chain of mental images (Cherniss, 1976, II, p. 591, note e).
27
Noct. Alt. L. VI, c. 2, quoted by Leibniz, Gerhardt, VI, p. 321; Jalabert, p. 326.
28
Leibniz, Gerhardt, VI, p. 314; Jalabert, p. 328.
29
Leibniz, Gerhardt, VI, p. 313; Jalabert, p. 327. Plutarch takes up the Stoic view
of comedy (Moralia, XIII, Part II, 1065; Cherniss, 1976, II, pp. 709-713) and turns
it against Chrysippus as impious. "But that refutation doesn't amount to much,"
says Leibniz (Gerhardt, VI, p. 313; Jalabert, p. 326).
30
Especially L. II, ch. XIV and XXXIII; Hamelin, 1978, pp. 94-95.
 CHRYSIPPUS' NON-STANDARD MODAL SYSTEM / 119

foreknowledge and preordination entail necessitarianism.31 Both ad-

mit that mental representations incline one but do not compel. To the
condestinates correspond the compossibles, and the origins of Leib-
nizian preformation are to be found in the order of natures according
to Zeno and Cleanthes followed by Chrysippus. It even happens often
enough that Leibniz expresses the conditional in negative form: "They
say that what is foreknown cannot fail to exist, and they are right;
but it doesn't follow that it is necessary".32 Leibniz, like Chrysippus,
explains evil by concomitance and repeats the adage of jurisprudence:
incivile est nisi tota lege inspecta judicare. There is a difference between
them however. Leibniz retains the interdefinability of the modalities
and escapes necessitarianism in distinguishing two kinds of necessity.
The first or metaphysical and brute necessity goes back to the princi-
ple of non contradiction, while the second which may, and in the final
analysis must, be moral is based on the principle of the best. Chrysip-
pus, on the other hand, does not appear to make that distinction. But
consider just what he does do. Suppose he distinguishes two sorts of
laws. The first, the mathematical ones, include the affirmative form of
the second premise, according to which the necessary follows logically
from the necessary, along with the interdefinability of the modalities.
Here the second premise is therefore valid in its negative form: the
impossible does not follow logically from the possible. For physical
reality, by contrast, other laws obtain. Here the second premise is still
valid in its affirmative form, but the interdefinability of the modali-
ties and therefore the negative form of the second premise no longer
hold. The exclusion of interdefinability then would serve to separate
the purely logical brand of incompatibility, which falls under it, from
physical incompatibility or causal law, which are not subject to it.
There is a passage in Diogenes Laertius (75) that bears this out.
"The necessary is a proposition which being true is not capable of being
false, or which is capable of being false but is such that it is prevented
by external circumstances from being false". Of this complex alter-
native characterization it is clearly the second term that concerns us
here. Just what is a proposition which is capable of being false but is
prevented by external circumstances from being false? It is a proposi-
tion such that it is simply impossible that it be not the case. Without
ever being able to be false it can happen that this last proposition is
not true, if it comes to be destroyed.
31
Leibniz, Gerhardt, VI, p. 330; Jalabert, p. 345, who quotes Cicero: "Sequitur
porro nihil deos ignorare, quod omnia ab Us sint constituta".
32
Leibniz, Gerhardt, VI, p. 123; Jalabert, p. 130.
120 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

What difference there is then between Chrysippus and Leibniz on

this point can only be attributed to the following principle (which is
nevertheless a fundamental one). True to the pantheism of the Stoa's
founders, Chrysippus cannot conceive of a distinction between divine
will and divine understanding. There is then but one necessity of fate.
To limit its harshness for freedom as well as for theodicy, it will be
necessary to refuse the right of passing from the impossibility of not
being to the necessity of being, from conjunctions to conditionals. By
contrast, the trans-worldly God of Leibniz allows of such a distinction.
To Chrysippus 'sound' conditionals correspond Leibniz' brute neces-
sities, to his simple 'astrological' conjunctions correspond conditional
necessities. But these last give rise to yet a further distinction which
is missing in Stoic philosophy. For with Leibniz there is a science of
simple intelligence having to do with the conditional necessity of corn-
possibles, and there is a science of vision having to do with the decree
of creation considered as being in view of the best.

5.4 A system related to Prior's system Q; the

double logical square of Chrysippean
modalities and the double temporal index in
the Master Argument's second premise.
It was Prior who constructed a system of modal or temporal logic,
called system Q, to account for situations in which, the propositions
not existing, it can be said neither that they are true nor that they are
false.33.
33
Prior, 1957, p. 49, shows how the system Q enables us to avoid the paradox: 'If
it is not possible that I don't exist (in the sense that it is not possible that there
are no facts about me), then it is necessary that I exist'. The paradox is a bit
forced though, for who would admit the antecedent? It is in 1967 that Prior makes
specific reference to Chrysippus and the argument about Dion (p. 152)
"...'If nothing exists, this man doesn't' is never false [~ M ~ (~ (3x)E\x D~ E\a)
where 'E!x' stands for 'x exists'], for it is true whenever there is such a proposition.
And it is possible that nothing should exist [M ~ (3x)E\x]. So here we have a
[~ M ~ (a D ())} and a Ma which are true, though the corresponding M/3, namely
[M ~ E\a] is false, i.e. [B'\ does not universally hold. And the example is almost
Chrysippus' own, except that I have replaced his 'Dion doesn't exist' by 'Nothing
exists', the entailment by which of 'This man doesn't' is perhaps clearer. It is
perhaps a little contentious to say that it could be that nothing exists, but if one
held that being of the basic sort one is, e.g. being a man, is 'essential' or 'necessary'
in anything that is of that sort, one could say that it could not be false that if no
•man exists, then this man doesn't, that it could be that no man exists, and that
it couldn't be (isn't the case in any possible state of affairs) that precisely this
man doesn't exist". (Prior, 1967 p. 152). One might think that the idea of the
destruction of propositions in the Chrysippean theory would allow us by the same
 A SYSTEM RELATED TO PRIOR'S SYSTEM Q / 121

System Q as conceived of by Prior, however, diverges from the

Chrysippean model in that 1) despite its temporal character, it sup-
poses that the second premise of the Master Argument is a thesis of
pure modal logic, 2) it implies the discontinuity of time, 3) it entails an
assimilation of Chrysippus and Diodorus on the fundamental question
of denning the possible as that which is sometimes realized, even if it
distinguishes them by means of the destroyed propositions. Alexander
has formally rejected the Diodorean interpretation of Stoicism, which
the addition of destroyed propositions would not essentially change. He
writes: "But, it will be said, by affirming that it is fate that produces
everything, neither the possible nor the eventual are denied. This in-
deed can be that nothing prevents from being though it not be".34
Cicero and Plutarch confirm Alexander on this point.35 System Q and
the system of Chrysippus can be said to be related, but not identical.
System Q has been axiomatized36 and studied from the point of
view of semantic models.37 If modally and temporally non-standard
in its refusal of interdefinability, it does conserve the logical laws both
of propositions and quantification. If the modalities are interpreted as
having a causal sense—and this is close to the Stoic intuitions about
fate—it is seen that the 'standard' logic of causality leads38 to I/ a
collapse of the modal distinctions if substitutivity of identity is main-
tained, as would be normal given a univocal sense of identity, and 2/
causal necessitarianism. To remedy the second fault the interdefinabil-
token to deny the validity of the retrogradation of futures. Indeed, the quantified
form of p D PFp will read (3x)<px D P(3x)Fipx.
The antecedent can be true and the consequent false if the consequent is 'de-
stroyed', that is to say, where for want of facts the proposition doesn't exist. But
that way out is blocked by divination: the ineluctability of the consequent will be ex-
pressed simply in negative (~ M ~ P(3x)F<px) and not affirmative (LP(3x)Fipx)
form. One does not have the interdefinability of G ('it will always be the case
that') and F nor of H ('it has always been the case that') and P and ~ P ~;
~ P ~ (p D <?) 3 (Pp D Pq) is not a theorem. (Prior, 1967, p. 156).
One might be tempted here to draw a parallel between Chrysippus and Aris-
totle, comparing the 'destroyed' propositions of the first with the truth-valueless
propositions of the second. But the parallel would be misleading.
I/ All Stoic propositions have a truth-value. A destroyed proposition that is, one
such that the conditions necessary for its use have disappeared (M. Frede, p. 49),
is altogether different from a proposition without truth-value.
2/ Destroyed propositions have no assigned relation to futures.
3/ The impossibility of a conjunction is in no way equivalent to a contingency;
although it cannot be assimilated to a positive necessity
34
Arnim, S.V.F., II, n° 959.
35
Hamelin, 1978, p. 87.
36
By Bull (Prior, 1967, p. 154; Hughes and Cresswell, 1972, p. 304).
37
F011esdal, 1965 and 1966.
38
Proposed by Burks; see F011esdal, 1966, p. 12.
122 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

ity of the causal modalities will be given up. To remedy the first, the
only singular terms admitted will be descriptions that keep the same
reference in all physically possible worlds. But this comes down to
saying that if it is causally necessary that a, and further that a = 6,
then it is causally necessary that b as well, where causal necessity has
nothing to do with our subjective manner of describing individuals, but
has to do rather with their proper essence.39 This 'causal essentialism'
is quite in keeping with the Chrysippean notion of fate.40
What is still missing in such systems is the explicit reference of
the Chrysippean modalities to a double temporal index and the corre-
sponding interpretation of the Master Argument's second premise.
The definitions, 41 or rather, elucidations (of the modal terms figur-
ing in the definiens) of Chrysippus are complex. Following M. Mignucci
we shall lay down the following conventions:
A means: 'p' is true
B means: 'p' can be false
C means: external circumstances prevent p's being false.
Replacing the word 'false' by 'true' and vice versa in these stipulations
we get:
A' means: 'p' is false
B' means: 'p' can be true
C' means: external circumstances prevent p's being true.
The definitions of necessity and of not-possibly-not (which will be dis-
tinguished from that of necessity), for example, will read as follows:
p is necessary (Lp) if and only if p is true and if p can be false it
is prevented by external circumstances from being so, that is
Lp = A.(B D C).
By contrast, the negative form 'it is not possible that not p' will be
written simply:
~ M ~ P = BDC,
where the assertion that A has disappeared. It is asserted only that if
'p' can be false, external circumstances prevent its being so. But there
39
F011esdal, 1965, pp. 272-273, who thus finds Prior's example again.
40
In such a system past and future are symmetric. But adjunction of the Master
Argument's first premise, admitted by Chrysippus, re-establishes the asymmetry of
time (that tempers the notions of absolute numerical identity in the eternal return
and thus opposes Chrysippus to Cleanthes). As for the system Q, it would seem to
have a good deal of affinity with the Aristotelian logic of possibles ad unum (the
virtues).
41
Mignucci, 1978; Vuillemin, 1983.
 A SYSTEM RELATED TO PRIOR'S SYSTEM Q / 123

is nothing to oblige supposing that the proposition p exists. It may

be 'destroyed' as happens for propositions containing a deictic, when
the object designated by the deictic no longer is. In the case of affir-
mative necessity Lp this circumstance is ruled out by the occurrence
of A. According to Diodorus the necessary was that which, being true,
never became false. Chrysippus modifies that definition in two ways.
He dissociates the positively necessary from that of which the nega-
tion is impossible, the second term being deprived of the truth-clause.
He introduces the notion of external circumstances as constituents of
compossibles and fate.
Because it rejects interdefmability the Chrysippean system will have
eight modalities and differentiated relations of implication and contra-
diction.

LpsA.(Bz)C) L-psA'.(B'z)C')
contraries

//
-subcontraries-
M-peB.-C
It is readily seen from this square of opposition why an impossible,
~ Mp, is compatible with the negation of a negative necessity, ~ L ~ p,
for there is opposition of contradiction between 'it is impossible that
124 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

p' and 'it is possible that p\ but not between 'it is impossible that p'
and 'it is not necessary that not p\
The question that arises at this point is that of the relation between
the double logical square of Chrysippean modalities and the double
temporal index figuring in the Master Argument's second premise. The
preceding arguments about the interdefinability of the modalities are
not sufficient to determine their meanings. But the occurrence of the
letters B and C in these definitions are enough to show that these
meanings do not belong to pure modal logic.
When Chrysippus says that from the strong Chrysippean antece-
dent
'If Dion is dead, then this man is dead'
the conditional consequent
'If it is possible that Dion is dead, then it is possible that this man
is dead',
logically follows, the strong antecedent would mean, according to the
definition of positive necessity, that 'If Dion is dead , then this man is
dead' is true and that if it is capable of being false external circum-
stances will prevent it from being false. But these strong antecedents
make the consequent false, in the case where 'Dion is dead' is true and
'this man is dead' is destroyed. We must therefore accept only the
weaker antecedent that it is not possible that Dion be dead and this
man not be dead, a conjunction which is indeed always true. Unfortu-
nately this conjunction does not authorize the consequent.
In the strong rejected antecedent as well as in the weak authorized
conjunction, we find two synchronic sentences, while the consequent is
a conditional whose antecedent states a diachronic modality and the
consequent a synchronic modality. This suggests that there must be
a connection between the Chrysippean non standard square and the
second premise.
Let us first examine what would mean for Chrysippus himself as
well as for the whole Stoic-Megaric school the second premise of the
Master Argument as put under its complete Aristotelian form, if no
distinction were made between weak and strong modalities:42
(B)cs(t)(MNPt D {(3*i)[MtlPtl • (« < ^ < Nv N < ^ < t)} •
(MtPt
42
The complete second premise says:
(t){MNPt D (Bti)Mtlptl • (t < ti < N V N < ti < t)} • (t)(Mtpt D pt)
This formula is equivalent to:
(t){(MNPt D (3ti)[MtlPtl -(t<t!<NvN<ti< t)]) • (Mtpt D pt)}
and entails the consequence by the law: [(P 3 Q) • R] D [P D (Q • R)}).
 A SYSTEM RELATED TO PRIOR'S SYSTEM Q / 125

In conformity with the Chrysippean square, the strong diachronic pos-

sible (MNpt) entails the following conjunction (1) that 'this man is
dead' can be true ( B ' ) , and (2) that external circumstances do not
prevent it from being true (~ C'). The first conjunct is true if and
only if there is an intermediary instant t\ such that Mtjpt,, since then
pt 'can be true'. The second conjunct is true if and only if no cir-
cumstance prevents ptl from being true, which would unmistakably be
the case would ~ ptl obtain. The second premise of the De Caelo
text had been analyzed into two parts: the proper second premise of
Epictetus—the synchronic contraction of the diachronic possible—and
the principle of conditional necessity. Both parts are again found in
the Chrysippean definition of the strong possible.
What singles out Chrysippus in the school is his distinction between
a strong possible (B1- ~ C") and a weak one [A1 D (B1- ~ C')}. A
possible simply contradicts its corresponding impossible in a standard
square of modalities. A weak possible still contradicts its corresponding
strong impossible in the Chrysippean square. But the conjunction of a
weak possible with the weak corresponding impossible does not express
here a contradiction. For example, the conjunction (~ L ~ p- ~ Mp),
or, expressed in the terms of the square:
[A1 D (£' • C1)} • (B' D C')
may be compatible. For it to be compatible, A' must be false and
the conditional (B1 D C') be true. And A' is false when it is not the
case that 'this man is dead' is false, a situation that obtains when the
proposition 'this man is dead' is destroyed.
Therefore, if a weak possible is substituted for the strong one in
the antecedent of the complete strong second premise of the Master
Argument:
(B)&(t)(~ LN ~ pt D {(3<i)[M tl p tl • (t < ti < TV V N < ti < t)}
•(Mtpt Dpt)}),
the conditional becomes false for Chrysippus, since the consequent may
be denied without contradiction.43
Though Chrysippus' model may seem artificial, it expresses an im-
43
For Chrysippus, the formula:
(3t){~ LN ~ p t - [ ( t i )(t < t i < NV N < ti < t) D ~ M t l p t l ] V ( M t p f ~ p t ) }
is compatible if one of the disjuncts or the consequent is true and if 'pt' is destroyed.
It is to be observed that for the first disjunct of the consequent being true ptl must
be destroyed at all ti past or future comprised in the intervals [tN] or [Nt], The
fitness of Chrysippus' example shows itself: Dion's death is either future and the
proposition "this man is dead" is "beforehand" destroyed (since it has never been
existing) or past and the same proposition becomes then "normally" destroyed.
126 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

portant representation of natural law as prediction. Every prediction

connects two different instants of time through a necessary relation.
This necessity, however, cannot be positive, since, in that case, it should
suppose the actual truth of the connection. It must therefore be neg-
ative, stating only the impossibility of the conjunction between the
capacity for truth of some proposition and its prevention from being
true by the whole of the confatalia. This weak necessity is apt for
connecting two terms to which their respective temporal indices assign
different roles. The first predictive term has, by definition, two dis-
tinct temporal indexes. It therefore belongs to the class of theoretical
terms, which no finite class of observational terms implies. The second
term consists in a connection between two simultaneous observations;
one observation is about the general state of affairs at that moment,
the other observation is about the actualization of the capacity at that
same moment for a singular deictic proposition to be true at that mo-
ment. Now it is the fate of these propositions of observation to be
destroyed. Chrysippus' counterexample was well chosen, because only
a natural law thus expressed may generate a theoretical or diachronic
prediction with destroyable synchronic verifiers.
Human perception, except when liberated by dream44 or divina-
tory 45 vision, is confined within the present.46 The chain of fate must
therefore contract time into the conjunction of the sound 'conditional',
as is shown by the theorem or argument of the sign. In the theorem:
"If somebody is wounded at heart, he will die", the wound at heart
is a sign not that he will have to die, but that he is having to die
"the present sign being the sign of a present thing" ,47 The theorem
delimits the share of fatalism, a fatalism which is broken as soon as the
temporal identity of a sign is destroyed.

5.5 A Philonian doubt about the second premise?

Two simple objections are made against the Master Argument's second
premise.
Suppose, first, that it is now possible for me to catch a train at the
station tomorrow at 8 a.m. and that going to the station needs half an
hour. If I neglect to leave my house tomorrow before 7.30 a.m., it will be
true that what had been possible until 7.30 has ceased to be so, though
no instant before 8 has offered the occasion of a synchronic possible.
The objection, however, forgets that the conditions to be filled for doing
44
Goldschmitt, 1977, p. 85.
45
Goldschmidt, 1977, p. 81.
46
Goldschmidt, 1977, p. 97.
47
Sextus, Adv. Math., VIII, 254-256; Goldschmidt, 1977, p. 44.
 A PHILONIAN DOUBT ABOUT THE SECOND PREMISE / 127

an action are explicitly mentioned, when the Ancients require that

external circumstances do not hinder the realization of the possible,
failing which the verbal modal judgment should not be given its genuine
sense. Aristotle agrees with this general understanding. Speaking of
the rational potencies, he specifies: "To add the qualification 'if nothing
external prevents it' is not further necessary; for it has a potency of
acting, and it is this not in all circumstances but on certain conditions,
among which will be the exclusion of external hindrances; for these are
barred by some of the qualifications of our definition".48
More pertinent reservations seem to arise from the reflection that
it is possible now that a certain bit of radium emit radiation at t,
though no emission occurs between now and t. But what we mean
then is that there is a very high probability P that, during the space
of time |TV — t\ < e, the given bit of radium will emit a radiation.
We imagine a very large number N of such identical states, i.e. of
similar bits of radium. The more the number N grows up, the more
next to P will be the relative frequency of emissions during the time
\N — t\ < e. At the same time, we shall verify that no emission has
been made by this particular bit, while the next one has emitted three
times. The probability is a property belonging to a class of individuals
and it must not be confused with individual potency. It is therefore not
contradictory to maintain that this bit of radium which has emitted
no radiation during the time |N — t\ < e had no potency of doing it, in
conformity with the second premise, although the probability P of its
emitting was estimated very high.
Could then the second premise still be doubted, while maintain-
ing the individual character of the possible? If such a hypothesis was
formulated it could only have been by a philosopher who would have
stripped away a maximum of their reality from the real possibilities
imbedded in time that the Master Argument is all about. In other
words, that philosopher would have had to abstract from the external
circumstances of the temporal realization of the modalities to concen-
trate solely on their logical aspect. He would not have challenged con-
ditional necessity, as Plato has done, throwing suspicion on the reality
of sensible existence. He would simply have purged the modal concepts
of their temporal adherences, separating them from what involves their
realization in time.

48
Metaphysique, & 5 1048° 16-21; see 2.4, p. 22; on this objection and the following
one see below, 10.1.
128 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Philo, the Megarian, was surely a philosopher of the like. It is only

a conjecture, but not a groundless one, to attribute doubts about the
second aspect of the second premise to him.
Philonian or material implication reduces natural laws to formal
implications and therefore to simple universals lacking in any modal
surplus. So it diverges from the conception Diodorus as well as Chrysip-
pus and Aristotle had of laws. One can presume that after having dis-
sociated assertoric logic from all modal engagement Philo should have
continued his effort in trying to dissociate the temporal modalities as
much as possible from their temporal engagement. He says that "the
possible (which is moreover none other than the non-necessary) is that
which is susceptible of being by virtue of its own essence, so that a
thing that isn't and will not be, but is prevented from being by exter-
nal circumstances only, is none the less possible than that which is or
will be. Straw and a shell at the bottom of a well have the capacity,
the one of being burned, the other of being seen, wood in the middle
of the ocean remains combustible. In a word, possibility consists in
mere aptitude or virtuality". 49 The clause identifying possible with
non-necessary blocks Chrysippus' escape route. There is nothing to in-
dicate, on the other hand, that Philo contested either the first premise
or conditional necessity. This last principle has to do only with the
relation of present existence to present necessary existence, and it is
doubtful that a Megarian would abandon their connection.
We must therefore examine the second aspect of the second premise.
The unique index of the synchronically possible takes its place between
the two indices of the diachronically possible. As truistic as it seems,
that operation supposes that the second diachronic index is not shifted
to infinity. Removing that supposition, the Diodorean definition 'the
possible is that which is or will be true' gives way to the Philonian 'the
possible is that which is or will be true or of which the truth will be
deferred to infinity by external circumstances'. Never assignable, the
ever prevented possible of the third premise would not give rise to a
synchronic possible that would contradict the simultaneous conditional
necessity given rise to by the hypothesis that the possible will not
be realized. This shift to infinity will simply push to the extreme
Diodorus' solution, of which it can be considered the limiting case. It
would resolve the aporia economically without for all that giving up
conditional necessity which seems indispensable once one dogmatically
identifies being and event. It would still fit in too with the general
49
Hamelin, 1978, p. 84; Doring, 1972, fr. 135-137, pp. 41-42.
 A PHILONIAN DOUBT ABOUT THE SECOND PREMISE / 129

necessitarianism to which Philo, like his Megarian masters, seems to

have remained faithful. 50
That elegant and economic solution from the logical point of view
has the drawback of obliterating at the limit the distinction between
real and logical possibility just as Philo's extensional definition of sound
conditionals had the particularity of eliminating any consideration of
necessity from the concept of law. It is understandable that the Stoics,
attentive as they were to the relation of the modalities to physical
reality, should have paid little attention to a solution that suppressed
rather than resolved the question raised.51

50
Hamelin, 1978, p. 84.
51
"Although that definition [the one Philo gives of the possible] does retain some-
thing Megarian, not only ... in that it leaves no room for genuine contingency,
but also in that it makes no appeal to external circumstances for weighing their
possibility, it is seen nevertheless to invoke these circumstances for determining the
real. This is meeting Stoicism midway" (Hamelin, 1978, p. 84). But Philo mentions
external circumstances only to eliminate them in his definition of the possible, if]
iirirriSeioTrjTi HOVR ('by virtuality alone'), so goes the Philonian definition accord-
ing to Simplicius (In cat., 195, 31-196, 24 Kalbfl.; Doring, 1972, p. 41, fr. 137).
There we have the point of view of a logician concerned with separating his science
from physics.
Part III

Systems of Contingency: The

Lyceum, The Garden, The
Academy
6

Towards Rehabilitating Opinion as

Probable Knowledge of Contingent
Things. Aristotle.

The Epictetus passage reporting the Master Argument makes mention

only of the Megarian School and the Porch.
Yet the other schools of philosophy met Diodorus' challenge as well.
Cicero's De Fato attests to this as far as the Academy and the Epi-
cureans are concerned, and in all likelihood Chapter IX of the De In-
terpretatione contains Aristotle's own response.
Cleanthes and Chrysippus, in contesting the argument's first and
second premises respectively, were not acting solely as logicians. Both
the thesis of numerically identical eternal recurrence and that of con-
fatalia presuppose a physical image of the world and consequently a
certain representation of causality. In the one case this was the sym-
pathy of connection where there was room left for spontaneity, in the
other it was the constantly renewed order of palingenesis.
The solutions proposed by the other schools have that same char-
acteristic. When an Aristotle, an Epicurus, a Carneades or a Plato
challenge one or another of the Master Argument's implicit logical
premises, they do so in the name of a certain conception of physics and
a determinate principle of causality. Chance and fortune, the clina-
men, the specificity of deliberate cause and the specificity of spiritual
movement are all brought in for freedom's sake.

6.1 De Interpretations, Chapter IX.

Chapter IX of the De Interpretatione is one of Aristotle's most difficult
and most disputed texts. A translation will be given and the intro-
duction analyzed to specifically determine the problem raised: that of

133
134 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

future contingents. According to Aristotle, the solution must conform

to the two logical principles of non-contradiction and the excluded mid-
dle. On the other hand, it requires abandoning the Megarian theory
and in particular the Diodorean definition of the possible. Aristo-
tle's explicit solution is grounded in two principles: a distinction must
be made between absolute and conditional necessity and the validity
of the principle of bivalence is subject to limitation. The Stagirite's
overall conception confirms the lesson of De Interpretation, Chapter
IX. He did not attempt to reform logic, whose principles he had just
established, but to rehabilitate the notion of opinion as constituting
knowledge of the contingent. Such knowledge would have a probabilis-
tic value.
Translation of the text
18a28. Applying to what is or has been, it is necessary for the affirma-
tion or the negation to be true or false. Applying to universal things
taken universally it is always so that one is true, the other false, and
applying to singular things it is the same, as we have said. But apply-
ing to universal things not taken universally it is not necessary; this
we have also spoken of. Applying to future singular things, however,
it is not the same.
18a34. For if every affirmation or negation is either true or false it is
necessary too for everything to be the case or not be the case. For if
one person says that something will be while another says that that
same thing will not be, it is clearly necessary that only one of them
is saying what is true, since every affirmation is either true or false.
For applying to this sort of thing it will not be the case that both are
simultaneously saying what is true.
18a39. For if it is true to say that something is white or not white,
then it is necessary that it be white or not white (186); and if it is
white or is not white, it was true to affirm it or to deny it. And if it
is not the case it is false, and if it is false it is not the case. It follows
that it is necessary that either the affirmation or the negation be true.
1865. Nothing then either is or happens either by chance or inde-
terminately, nothing that will or will not be, but everything happens
necessarily and without any indetermination. For either he who affirms
says what is true, or he who denies. Otherwise it is indifferently that an
event would occur or would not occur. For that which is indeterminate
does not or will not happen rather in one way than another.
18fe9. Again, if something is white now it was true before to say that
it will be white, so that it was always true to say of anything that has
happened that it will occur. But if it was always true to say that it
is or that it will be, it is not possible that it not be or that it will not
 DE INTERPRETATIONS, CHAPTER IX / 135

be. But it is impossible that that which cannot not occur should not
occur. And that of which it is impossible that it not happen happens
necessarily. Therefore all futures come about necessarily.
18615. Consequently nothing will be the case indeterminately or as a
result of chance, for that which depends on chance is not of necessity.
18617. Nor is it possible to say that neither the affirmation nor the
negation is true: that an event for instance neither will take place nor
will not take place. Firstly, if the affirmation is false the negation is
not true and if the negation is false it happens that the affirmation is
not true. Secondly, if it is true to say that a thing is white and black,
both qualities must belong to it. Should they belong to it until tomor-
row, they will belong to it until tomorrow. Suppose, by contrast, that
tomorrow the event neither will nor will not occur: nothing indetermi-
nate, such as a naval-battle for instance, would then take place. For
it would be necessary that the naval-battle neither take place nor not
take place.
18626. Such, among others, are the absurdities that arise if for every
affirmation and negation, either applying to universals as universals or
applying to singular things, it is necessary that one of the opposites be
true, the other false, and if there is nothing indeterminate in events but
that everything is and happens as a result of necessity. Consequently
there would no longer be any point in deliberation nor in taking any
pains, with the idea that if we accomplish a certain action a particular
result will follow, whereas if we do not accomplish it the result will not
follow.
18634. For there is nothing to prevent one man's saying ten thousand
years beforehand that something will be the case, another that it will
not, so that it necessarily will be that of one of the two cases it was true
to predict it then. Nor does it really matter whether people did make
the affirmation or not. For it is clear that reality is what it is even if
there was not this one to formulate the affirmation, this other one the
negation. For it is not because of the affirming or the denying that the
event will or will not occur, even (19a) had the announcement been ten
thousand years beforehand rather than at any other moment. Hence, if
it has been from all time that one of the contradictory propositions said
the truth, it was necessary that that should happen and every event
has always come about then in such a way as to happen necessarily.
For it is not possible that what anyone has truly said will be the case
should not happen; and as to what has happened, it was always true
to say that it will be.
19a6. But what if these consequences are impossible? For we see that
there is also an origin of what will be both in deliberation and in action
136 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

and, that the potentiality of being and of not being resides entirely in
things that do not always exist in act, things that, since they may be
or not be, the one as well as the other, may therefore also come to be
and not come to be. There are many manifest cases of this kind. This
garment, for example, may be cut in two and yet will not be cut in
two, but will wear out first. In the same way, it may not be cut, for
it could not wear out first were it not possible for it not to be cut in
two. This holds for all other events as well which are mentioned as
having the same kind of potentiality. Clearly then not everything is
or happens of necessity but some things come about indeterminately,
and for them neither the affirmation nor the negation is more true, the
one than the other, whereas for some other things one of the two is
true more frequently, though it does come about that the other should
happen and not it.
19"23. It is necessary that what is be, when it is, and that what is
not not be, when it is not. Yet it is not of necessity that everything
that is is or that everything that is not is not. For it is one matter for
everything that is to be necessarily, when it is, and another for it to be
necessarily simply. The same holds for everything that is not.
19"27. The same argument applies to contradiction as well. Everything
necessarily is or is not, will or will not be, without saying for all that,
if we divide, that one of the two is necessary.
19a30. Let me take an example. It is of necessity that tomorrow there
will or will not be a naval-battle. Yet it is neither that the naval-
battle takes place necessarily tomorrow nor that it does not. What is
necessary is for either to take place or not take place.
19a32. Consequently, since propositions are true in so far as they
conform to the things themselves, it is clear that whenever the things
behave indeterminately, and have a potentiality for contraries, the same
necessarily goes for the contradiction as well. This is the case for things
that are not always existent or that are not always non-existent. For
it is necessary then that one of the two contradictory propositions be
true or false, but it is not this one or that one but either, and when one
is more true than the other it is nonetheless not already true or false.
It is clear then that it is not necessary for every affirmation or negation
taken from among opposite propositions that the one be true, the other
false. For what is non-existent but has the potentiality of being or not
being does not behave after the fashion of what is existent, but in the
manner just explained.
 OUTLINE OF THE PASSAGE: INTRODUCTION (18°28-34) / 137

6.2 Outline of the passage: Introduction

(18a28-34): The problem raised.
The passage quite naturally falls into two parts: the first (18a34-19a22)
in which the Megarian theory is set forth and refuted, together with
an interlude on the validity of the excluded middle (18i)17-25), and the
second in which Aristotle sets forth his own solution (19a23-1964). A
remark of Pacius1 who classifies all the possible theories concerning
contradictory pairs of singular statements about the future into four
groups, suggests a more finely tuned division of the text and of its
logical articulation. (The different possibilities according to Pacius are
that both are true, that both are false, that one is true and the other
false now, and finally, that one is true and the other false, but only
potentially). The passage is then reduced to a chain of implications. If
one admits the principle of non-contradiction—the two statements are
not both true—then, if one admits the law of the excluded middle—
the two statements are not both false—then if one further admits the
universal validity of the principle of bivalence—one of the statements is
true, the other false now—one cannot avoid holding that all statements
about the future are necessary. By contraposition, the existence of
future contingents will require that if non-contradiction and excluded
middle are to be maintained the principle of bivalence will be called
into question and, consequently, one of the statements will be true and
the other false, but only potentially.
At first Aristotle singles out three sorts of contradictory statements
that are not to figure in the subject of his inquiry. The first two
sorts, singular statements about the present or the past and universal
statements taken universally, are such that the one is presently true, the
other presently false. The third class consists of universal statements
not taken universally. These are indeterminate or indefinite opposites
(man is white/man is not white). When treated as two particular
opposites they are subcontraries and can both be true, but not both
false, at the same time. No difficulty arises with respect to them for
the one is not really the negation of the other.2
While universal statements, taken universally or not, follow the
same laws whether they are about the future or not, singular statements
are different with respect to the future. The quantitative distinction is
a necessary criterion for determining the fourth class of statements. To
obtain a sufficient criterion the matter involved must also be taken into
Quoted by Edghill in The Works of Aristotle, 1928, p. 18°, footnote 5.
2
Aristotle, De Interpretation, VII, 17*28-33.
138 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

account, as St. Thomas points out.3 The word designating the future
here is usually opposed to the future participial of the verb to be like
what is in contingent matter is opposed to that which is in necessary
or impossible matter.4 Assertions resulting from an essential predi-
cation, such as 'Socrates will be a man' or 'Socrates will not be an ass',
affirm or deny a property of a subject not in so far as it is a particu-
lar singular subject but after the manner of universals,5 as is required
by science. The opposition between these singular future assertions is
therefore treated as an opposition between a universal and its contra-
dictory, where the opposite truth-values obtain presently. The inquiry
will be limited then to deciding whether it is necessary in the case
of singular statements about the future in contingent matter that one
of the opposites be true and the other false, presently.6 These state-
ments may be of two different grammatical forms. There are singular
predicative statements such as 'this will be white', and pseudo-dated
existential statements, e.g. 'there will be a naval-battle tomorrow'. Not
every future singular predication or future singular existence statement
is accidental. Socrates will be essentially reasonable and the Heavens
will necessarily have a given motion. But a particular man will be
standing or sitting by accident, and he will exist by accident. It is to
this twofold domain that the inquiry is limited
3
St. Thomas, De Interpretation, Liber Primus, Lectio XIII, 1-3 (Oesterle, De
Interpretatione, 1962, 6, pp. 102-103).
4
The text says " im. 84 toiv xat? ' ixaercaxal jiEXXovctov" in singularibus et futuris.
Ammonius (quoted from William of Moerbecke's Latin translation, In De Int., 138-
139, Corpus Latinum Commentariorum in Aristotelem Graecum, t. I, pp. 264-265)
specifies that what is to be understood by 'future' here is what is thought of as of
contingent matter. He reminds us that in the De Generatione et Corruptione (II,
11, 337fc3) Aristotle opposes 'mellon' (or 'future') and 'esomenon' (future participle
of the verb ei^t). The latter signifies what will happen in any case, as when we
say 'winter or summer will come' or 'an eclipse will take place', whereas the former
involves a future that can happen or not, as 'I'll go for a walk' or 'I'll sail'.
5
St. Thomas, Peri Hermeneias. p. 43: "secundum universalium rationes".
6
As St. Thomas points out: "Aristotle has not mentioned contingent matter
until now because those things that take place contingently pertain exclusively to
singulars, whereas those that per se are inherent or are repugnant are attributed to
singulars according to the notions of their universals". "Aristotle is therefore wholly
concerned here with this question: whether in singular enunciations about the
future in contingent matter it is necessary that one of the opposites be determinately
true and the other determinately false" (Oesterle, op.cit., p. 104).
PRINCIPLE OF NON-CONTRADICTION AND LAW OP EXCLUDED MIDDLE / 139

6.3 Validity of the principle of non-contradiction

and the law of excluded middle (18a38 and
18617-25).
A simple remark suffices to exclude the possibility of making an excep-
tion to the principle of non-contradiction for statements in contingent
matter about the future. Its application to the future adds no further
plausibility to the negation of this principle—a negation some attribute
to Heraclitus and that corresponds to the thesis of universal motion in
physics.7
On the other hand, Aristotle spells out his reasons for maintaining
the validity of the excluded middle for statements about the future. He
does this after having set forth the Megarian theory, where he argues
that not only would rejection of the law fail to enable us to escape the
difficulty proper to that theory as regards the modalities, but would
further open the door to the specific aporias regarding truth, analyzed
in the Metaphysics in connection with the doctrine of Anaxagoras.8
As to truth, since, if the true and the false do not form a disjunction,
once the futures will have to come to be so as to verify one of the two
opposite statements, the other will still not be falsified.9
As to modality, suppose in conformity with the doctrine of uni-
versal mixture, which is the physical counterpart of the negation of
the excluded middle, that two present contraries are true; it is true
that something is simultaneously black and white10 (principle of ho-
moeomeries). The conjunction of these two states is thus necessary
(Set). But with nothing to distinguish the logical behavior of present
from future singular statements, it must be concluded from the fact
that it is true that a certain thing will be white tomorrow and that
that same thing will be black that necessarily that thing will be simul-
taneously black and white tomorrow. The only thing specific to the
doctrine of homoeomeries is that it expresses the necessitarianism of
7
Metaphysics, F, 3, 1005624; 7, 1012a25.
8
Metaphysics, T, 7, 1012°26.
9
This is a particular case of the general refutation given at Metaphysics, F, 7,
1011625-29.
10
The principal manuscripts read /j.e~/ai; (large) instead of ne\av (black); and
Ross, following Boethius and Moerbeke, adopts this reading which renders the text
unintelligible. The correction of pi-iav adopted by Minio Paluello, 1949, is neces-
sary. What is more, it is in agreement with the Aristotelian passages treating of
Anaxagoras' theory of homoeomeries. According to him all was mixed in the begin-
ning (Metaphysics, A, 8, 989a30-989620). As an example of a real intermediate—
therefore essential when treating, as is the case here, of the logical conditions of the
thought of change—Aristotle specifically give grey between black and white (T, 7,
and 36).
140 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

the future in terms of conjunction instead of expressing it in terms of

disjunction.
Let us substitute 'non-white' for 'black' in the above example in
order to make it deal with just a single predicate. It follows that if
it is true that a certain thing will be white tomorrow and that that
same thing will be non-white tomorrow then it is necessary that it will
be white and non-white tomorrow. Doesn't such an expression come
down to denying the principle of non-contradiction? As shown at the
end of Book F of the Metaphysics, the specific absurdity produced by
the negation of the excluded middle becomes apparent only when one
goes from the affirmation that all is true to the affirmation that all is
false. The example above then, with negations substituted for affir-
mations, will read 'If it is true that a certain thing will not be white
tomorrow and that that same thing will not be non-white tomorrow,
then it is necessary that it will be neither white nor non-white tomor-
row'. Aristotle then introduces the example of the naval-battle. If the
excluded middle is wanting it will have to be that the battle neither
will take place nor will not take place tomorrow. Intuitionism,11 in
abandoning one of the fundamental principles of logic, hoped to make
room for indeterminacy. All it issues in is a conjunction of impossibles.
With indeterminacy giving way to necessity, by way of impossibility,
the difficulties of necessitarianism are added to those of intuitionism.
It is not known whether there were philosophers in Aristotle's day
who rejected the excluded middle to escape necessitarianism, as Epi-
curus was later to do; in any case the Stagirite rejects such a solution
as illusory.
n
c( 8e jjuit' eotai ^irJTC HTJ ?aiai aupiov oux av elr) to orcoicp' etux£ v (18fc22-23).
Those Aristotle has in mind here are the partisans of the intuitionistic logic peculiar
to Anaxagoras, where one hasn't the right to substitute an affirmation (white) for
a double negation (not-not white). The conjunction of a simply negated statement
and the same statement doubly negated
(~P- ~~p)
does not violate the principle of non-contradiction. For that to happen it would
have to be possible to substitute 'p' for '~~ p', which is precisely what intuitionistic
logic disallows. In Metaphysics, T, 8, Aristotle remarks "Again, there are obviously
contradictories which cannot be at the same time true. Nor on the other hand can
all statements be false, yet this would seem more possible in view of what has been
said". (101262-4) (Ross translation). On Anaxagoras' intuitionism and theory of
continuum, see H. Weyl, Principles of mathematics and natural science, New York,
Atheneum, 1963, p. 41.
 CRITICAL EXAMINATION OF THE MEGARIAN THEORY / 141

6.4 Critical examination of the Megarian theory

(18a34-18617 and 18625-19a22).
Aristotle records two demonstrations of the Megarian thesis. The first
is of a rather logical nature, the second refers more to reality.12
The first demonstration (18a34-1869) consists of two stages. The
first of these is brief and general (18a34-39). It expresses the principle
of correspondence, foundation of Aristotle's semantics, though retain-
ing only the passage from truth to being. But it gives this passage an
apodictic force and a modal aspect heralding the principle of condi-
tional necessity.
(C) If every affirmation or negation is true or false, it is necessary or
it is impossible for the corresponding state of affairs to exist.
The overall sense here requires us to take necessity in the divided
sense.13 Since the thesis is universal we are free to apply it to an
affirmation or a negation about the future, as long as its truth-value is
present:
(Cf) If every affirmation or negation about a future is true or false it is
necessary or it will be impossible for the corresponding state of affairs
to have to exist.14

12
Ammonius, op.cit., 1961, pp. 251-252.
13
ei yap aXr)i?E<; etnav on Xeuxov ?) o-u ou Xeuxov eativ, dvdyxT) rival XEUXOV TJ ou
XEUXOV (IS^g-lS6!).
From the grammatical point of view there is no decisive argument for choosing
between the division (Lp V L ~ p) and the composition L(pV ~ p) of necessity. (On
this point see Mrs. Frede, 1968, pp. 16-17). One might then reject our decision as
unfounded and follow Edghill (op. cit., 19a30-32) in taking the consequent of (C)
in the composite sense. The formula (F)
*L(p\J ~ p) D (Lp V L ~ p)
is not a thesis of modal logic.
(Notice Hughes and Cresswell, 1972, p. 38, who are mistaken in saying that the
antecedent is false and the consequent true). On the composed-sense interpretation,
the denial of (F) would suffice for avoiding Megarian necessitarianism. But that
would mean supposing that Aristotle had attributed to the Megarians the thesis
(F) whose modal invalidity is patent.
14
Ammonius presents this passage as follows. "Suppose two individuals feign di-
vining with respect to a particular event, attempting to predict of an invalid for
example, one that he will recover, the other that he won't. Clearly it is necessary to
hold one of the two cases true, the other false. If the one predicting recovery there-
fore says what is true, it is necessary that the invalid recover (for it has already been
postulated that the truth of discourse is followed by the occurrence of the thing in
any case); but if the one denying it says what is true, clearly it is impossible that
the invalid recover. That is why either it is necessary that the thing happen or it is
impossible that the event take place. Therefore contingency is excluded..." (1961,
pp. 267-268).
142 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

The second stage of the first demonstration (18a39-1869) has to do

with an example in contingent matter, the white. One would have ex-
pected that, from the explicit position of bivalence (curiously deduced
at line 1864 as necessary) and by detachment, the conclusion of Megar-
ian necessity would have been drawn, as the rest of the text which
describes necessitarianism seems to demand. But these lines, perhaps
corrupted, simply apply (C) to the example, rounding off the analysis
of the true by that of the false. In addition, they evoke the passage
from being to truth, and more precisely from the present being of a
state of affairs to its past truth. 15 This is probably an encroachment
on the second demonstration.
The second demonstration, from what Ammonius says, proceeds
from "more evident and more generally accepted theses".16 In stating
the principle of the retrogradation of truth, it shows why the Megarians
accord themselves the right to affirm the antecedent of (C) and of
(CF), i-e. the principle of bivalence. Aristotle gives successively three
different forms to the statement of the principle of retrogradation: the
first indefinite (R/, 1.10), the second quantified (Kg, 1.11), the third
Diodorean (R D , 1.12).
(R/): If something is the case now it was true at any moment of
the past whatsoever that it will be the case.
(RQ): If something is the case now then it always was true that
that thing will be the case.
(R/j): If something is the case now then it always was true that
that thing is or will be the case.
Given that Diodorus defines the possible as that which is or will be
the case, (Ro) says that the Diodorean possible is necessary.17 Con-
sequently, if it is granted that that which was true and foreseeable
from all eternity cannot not happen, then everything that is in act
is necessary. But just as ( C F ) could be concluded from (C) in the
first demonstration, if it is legitimate to posit (R/j), retrograding from
15
The transfer of time is expressed by the imperfect 9)v. If there is white or if there
is not white it was true that there will be white now or it was true that there will
not be white now. Ammonius insists on the imperfect (1961, p. 269): "it is not
only according to the time in which the things are and subsist that it is true to say
they are as they are, but the prediction with respect to them is true also before the
event".
16
1961, pp. 274-275.
17
D. Sedley, 1977, p. 80, gives a span of twelve years (334-322) for the possible
influence of Diodorus on Aristotle. This would mean assigning a rather late date
to Chapter IX of the De Interpretations. The question of relative chronology here
remains at least an open one. (cf. Celluprica, 1977, p. 16).
 CRITICAL EXAMINATION OF THE MEGARIAN THEORY / 143

present to past, it is also legitimate to posit (Rf), retrograding from

future to past.
(Rf): Supposing that something will be the case, it always was
true that that thing is or will be the case.
For, since (Ro) is supposed valid for anything whatsoever, regardless
of the modality of its matter, a refusal to apply retrogradation to the
future would amount to the postulation that only futures can be con-
tingent. The future world would then be subject to different laws from
the past world. This supposition is all the more ridiculous in that, with
the scission between past and future in constant displacement, one and
the same event which would not always have been true in the past as
long as still future would suddenly come to have always been true in
the past at the moment of its actualization. If there can be contingency
in the future it must have been that there could be contingency in the
past.
The retrogradation thesis, which according to Aristotle is character-
istic of Megarian philosophy, implies and expresses the second aspect
of the correspondence principle, though without respecting its limits.
From the existence of a state of affairs we have the right to conclude
the truth of the statement of that existence. But the retrogradation
principle extends that truth to time in its entirety, consequently entail-
ing necessitarianism on the sole supposition that what was always true
is necessary. It allows positing the principle of bivalence on the basis
of the law of excluded middle. In virtue of this latter, it can be said of
a future event that it will be the case or it won't. In virtue of retrogra-
dation, it was therefore always true or it was always false that it is or
will be the case: hence bivalence. It is this detachment of bivalence,
once given retrogradation, that justifies the second demonstration. It
is this detachment that allowed the steps from (Cp) to the Megarian
conclusion:
(M) All futures are necessary or impossible.
Necessity is communicated then to all becoming (18625-19Q6). It
is apparent for events foretold by divination. It remains valid, all
prophecy aside, for it is truth that depends on existence, not the other
way around. The necessitation of being by truth is therefore indepen-
dent of divination.18 These consequences are in contradiction with the
18
It is the truth of the thing—in the future as in other times—that allows the
statement of the thing, not the other way around. (Metaphysics, I'l, 1053a33). St.
Thomas, who accepts the principle for human science, rejects it for Divine science
(Summa Theologiae, I q!4, a 8: "scientia Dei est causa rerum"). This position
is understandable in terms of creation. St. Thomas avoids necessitarianism in
qualifying the science of God.
144 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

experience of becoming, with contingency and with the capacity for

contraries both as regards human volition and as regards events due
to chance or to non-necessitating causes. They lead to inaction and
to the fatalistic acceptation of what happens (Idle Argument). With-
out the contingency of things not always existing in act (19a6-22) the
sublunary world would not be what it is.

6.5 Aristotle's solution (19a22-1964); conditional

necessity and exceptions to the principle of
bivalence.
Before leaving the Megarians Aristotle concluded that since the conse-
quent of (Cp) is inadmissible the antecedent will have to be rejected.
From the denial of necessitarianism the invalidity of bivalence will be
concluded (19a16-22). It is thus natural for Aristotle to set forth his
own solution in the form of a contraposition. Since not all futures are
necessary, there will not be an already given truth-value corresponding
to every proposition about the future. Corresponding to the antecedent
of this contrapositive is the distinction between two sorts of necessity
(19°23-27), corresponding to its consequent are the exceptions to the
principle of bivalence (19a27-1964).
In distinguishing two sorts of necessity Aristotle abstracts from the
difference between past and future. 19 He in no way abstracts from the
temporal condition in general.20
"The science of God is the cause of things insofar as the things are in the science.
But it was not in the science of God that the things would be from all eternity.
Therefore, although the science of God is eternal, it nevertheless does not follow
that the creatures are from all eternity".
Divination was not simply a matter of vulgar superstition in antiquity. It was
also an axiom of Stoicism, as has been seen. Some Peripatetics like Dicaearchus
and Cratippus admitted it (Pauly, 1842, Bd II, p. 1118). Aristotle himself counts
melancholy as one of the predispositions for divination by extasy (Probl. XXX,
1 and 14, and ibid., p. 1121 and 1123 on dreams). But the Peripatetics rejoin
the Cynics and Epicureans in opposing the oracles. Overall, the scaling back of
the importance of divination in the De Interpretatione fits in with the general
rationalistic tendency of Aristotelianism.
19
As Ammonius remarks, 1961, p. 289.
20
William of Moerbecke translates Ammonius commenting on Aristotle (p. 290)
otav ?j by quamdiu existerit. The text has TO n^v ouv clvou TO ov OTIXV ?) xal
TO nf) ov jiT) etvai oTav HTJ fj ocvayxr). For Aristotle, OTCXV fj signifies either an
iteration involving indeterminate futures or a general cause. Boethius translates
by quando est (quando non est), Moerbeke by cum fuerit and St. Thomas by dum
est. (According to Hintikka this phrase would mean that the statements are dated,
1964, pp. 472-473, reprinted as Chapter VIII of Time and Necessity, 1973; cf. also
Boudot, 1973, 4, p. 467). Looking at Metaphysics, -0 3, 1046629, which treats of
the reality of possibilities and the refutation of the Megarians through the analysis
 ARISTOTLE'S SOLUTION / 145

Three interpretations of conditional necessity are possible. The first

two express respectively the necessity of the consequent:
If p takes place then it is necessary that p take place,
and the necessity of the consequence:
It is necessary that if p take place, then p take place.
The second interpretation is a tautology.21 Necessitarianism cannot
be derived from it, but it fails to translate Aristotle's text. The first, in
spite of its hypothetical form, asserts the simple or absolute necessity of
the event. According to Aristotle it corresponds to the Megarian theses.
The necessity has to do in the consequent with an already temporally
determinate proposition. It is therefore de dicto and corresponds to
the proposition's being always true.
We must therefore come to a third interpretation, positing a con-
ditional connection between an event and its necessity, as in the first
case, but subordinating the necessity to the duration of the act of the
event. Necessity, having itself become the object of a temporal con-
dition, governs an open statement containing a temporal variable and
is itself subject to that same variable: it is de re. The corresponding
principle can be expressed in the following quantified form:
For any time, t, if p takes place during time t, it is necessary during
time t that p take place during time t.
The quantification can be over the future as well as the present or past,
but the time of the necessity cannot be different from that of the event
described by the statement.
Ammonius' examples22 of simple necessity, 'the sum of the angles
of a triangle is equal to two right angles', 'the Heavens are always in
motion', 'fire is hot', 'Socrates is mortal', express cases of essential
predication in which the predicate's necessarily belonging to the sub-
ject is either eternal or is co-extensive with the subject's duration. The
statements 'the sun is eclipsed by the moon' and 'Socrates is walking'
afford examples of conditional necessity, in which the predicate's be-
of motion, we find the expression oiav cvEpyj} tiovov 8uva(Ti9ai. ("there is potency
only whenever there is act"), where the repetitive applies to any indeterminate
time whatsoever. This is the sense I shall retain, thus identifying oiav f) and otctv
£vepYT)- The term 'being simpliciter' has to do with what is in general, therefore in
potency or in act (which is eminently the case for future contingents), whereas the
phrase 'whenever it is' indicates that the subject has passed to act.
21
According to Lukasiewicz, 1957, p. 131, it is in this way that Alexander interprets
the De Interpretations and many a Scholastic is in agreement with him (Prior, 1962,
p. 211).
22
153(289).
146 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

longing to the subject is necessary solely during the actuality of the

event: the eclipse or the walk.
As the eclipse example shows (the eclipse was the model of foresee-
able and necessary events), conditional necessity alone is no guarantee
of contingency. It is a necessary but not a sufficient condition of it. In
the world of Diodorus the modalities depend on the logical status of
the propositions, that is to say, on the behavior of their truth-values
over the course of time. It is not the nature of reality and the temporal
relation of predicate to subject then that make for the necessity of a
state of affairs. It is the stability of the truth-values of a dictum. The
retrogradation of truth on the one hand, the eternity and simplicity
of necessity on the other, go together. By contrast, the Aristotelian
modalities depend on the kinds of predication and on the real relation
of predicate to subject. When this relation is accidental it imposes on
necessity, conceived of de re, a merely temporal validity. Taken by it-
self, such a conditional necessity, inseparable as it is from its temporal
context, would not be able to retrograde. There are however extrinsic
considerations having to do with causality that may, as in the case of
the eclipse, provide a ground for retrogradation.
We must now turn to the consequence conditional necessity has on
the contrapositive of (Cp): abandonment of the principle of bivalence
(19a27-32). For as far as future accidents are concerned, all that can be
said about the corresponding state of affairs is that it will be necessary
for it to be while it will be or that it will be impossible for it not to
be while it will not be. On the other hand—except where an extrinsic
retrogradation is invoked—one cannot say of the corresponding state of
affairs that it is necessary simpliciter or that it is impossible simpliciter
that it must be. Consequently, it will no longer be possible to maintain
either that every statement about the future is true or false.
The conditional necessity of the eclipse degenerates, for the laws of
astronomy guarantee the retrogradation of truth, since the respective
positions and motions of the sun and moon are given. It is therefore
already true that there will be an eclipse at a certain moment. Simple
necessity applies to this kind of accident. But if there is nothing actu-
ally in the present causes to render the naval-battle ineluctable, that
battle will be necessary only if it takes place and while it will be taking
place. It is therefore not already true that it will take place. It can be
said in the composed sense that it is true that tomorrow a battle will
take place or not, for time occurs only vacuously in the disjunction.
On the other hand, there is no more ground for distributing truth,
than there was for distributing necessity, in saying that it is already
 ARISTOTLE'S SOLUTION / 147

true that there will be a naval-battle tomorrow or that it is already

false.23
23
The example of the naval-battle is taken up in three different propositions. In the
first the necessity is composed, in the second distributed and in the third composed
again. Aristotle rejects the second while accepting the two others.
Bekker's text at 19" 30 contains the word zaecrdotL that Minio Paluello corrects to

The first choice was that of Boethius who translated both terms by juturum esse.
But then the third proposition (line 32) would be but a repetition of the first (line
30). If we want to avoid such a shocking redundancy within just a two-line interval
we must take Minio Paluello's lesson to heart, recalling the passage at IS'lS. It is
thus that William of Moerbecke, in his Latin translation, opposed fore and fieri.
In Metaphysics, K, 8, 1065° 14-17, Aristotle discusses the thesis of universal de-
terminism based on the arguments of physics borrowed from the notion of causal
regression. All chance, all contingency will be impossible if one admits in the case
of an accidental being the regression of causes all the way back to a cause which
is itself necessary. But there are two ways of conceiving of that regression, accord-
ingly as it is applied to what is or to what is becoming. "Even if the cause were
supposed no longer what is (ov) but what is becoming (fiyvotJ.ei'oi'), the conse-
quences would be the same: everything would happen necessarily; for the eclipse
will take place tomorrow if such-and-such happened, and such-and-such happens if
some other thing happens in turn, and that other thing if a third thing happens".
We could in no way agree with Bonitz, vol. II, 1849, p. 464, and Tricot, II, p.
611, in taking Aristotle in the Metaphysics passage to be opposing present or past
being to future being. That was St. Thomas' interpretation who commented (p.
540, number 2282): "Someone might object to the argument saying that the cause
of future contingents is not already posited as the present and the past but is up
to now contingent like the future. For the result is that then everything happens
necessarily as was the case before [where we reasoned about what is, not about what
is becoming]. For if a certain cause is future it must be future in some determinate
time and determinately distinct from the actual present. Suppose it to be tomorrow.
If therefore the eclipse, which is itself the cause of certain accidental futures, is a
future that will take place tomorrow, and if all that happens happens through some
cause, it must be that that eclipse itself, as a future that will take place tomorrow,
happen if this will have happened, that is to say, as a result of something pre-
existing it..." St. Thomas thinks then that after having reasoned about present
or past causes Aristotle goes on to reason about causes to come, thus extending
to the future a causality that that extension does not preserve from determinism.
But it is clear that Aristotle considers the eclipse phenomenon to be determinate.
He is not concerned with the effects that will result from this phenomenon posited
as future. He is inquiring whether for this phenomenon, itself posited as a future
effect, there exists a causal chain, an already given regression, which is precisely
the case. Moreover, the word -yi-yvoiiEvov refers to the present just as the word 6v
does. The opposition of the two cases distinguished by Aristotle has nothing to do
with the temporal index of causality, but with the question of whether causality
applies to beings or to events (as Ross well notes in his Commentary, t. II, p. 322).
And if one wonders why Aristotle introduces that opposition when he defends the
possibility of contingency both in the context of causality and in that of logic, the
example of the deterministic eclipse will have to be taken into account for finding
the answer. He was clearly aware that any solution of the difficulties that would
limit itself to reserving contingency to events, even future ones, abandoning beings
148 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

When a property belongs essentially to a subject it belongs to it

of simple necessity. The statement expressing that belonging is nec-
essarily true, its contradictory impossible. Division of the modality is
therefore legitimate. When a property belongs to a subject acciden-
tally it belongs to it according to conditional necessity. Division of
the modality in the disjunction of contradictories is then illegitimate,
since the attribution of a truth-value to a determinate one of the parts
cannot be made in the absence of a condition that, by the very nature
of things, can be specified only at the moment of the event. This is
precisely what contingency in the universe means.
There is a correspondence between the truth of statements and the
reality of states of affairs (19a32-19(>4). Contingency lies in what is
not always existent but not always non-existent either. As for the
corresponding contradictory statements, although each of them is nec-
essarily true or false, it must be said that it is not determinately either,
whether it be that the chances are equal for both events or again that
one of them has a greater probability of occurring than the other. Even
in this last case one could not already (T]§r))24 attribute a definite truth-
value to the stronger statement. Ontologically speaking, contingency
is proper to beings in potency (5uv<rrc5v), susceptible of only a condi-
tional necessity, whereas simple necessity characterizes beings properly
speaking (OVTCOV) which are always in act. Logically speaking, given
that all disjunctions of contradictories are true, it is only those disjunc-
tions dealing with essence, that is to say, with beings, that allow the
distribution of the true and the false that can then retrograde. Disjunc-
tions dealing with beings in potency admit neither of retrogradation
nor of the division of yet indeterminate truth-values. 25
to determinism, would fly in the face of astronomical observations. In Metaphysics,
K, he shows that the regression of causes enchains both beings and events in a
same determinism. In the De Interpretatione he shows that what allows for a
future event's being contingent, such as the naval-battle tomorrow, is the non-
distributivity of the modal operator. And it is this same non-distributivity that
allows for events in general that happen presently to happen contingently. If one
of the factual proofs of contingency is voluntary deliberation, since we deliberate
in the present, contingency must not be reserved to futures but it must issue from
the very present taken as potency of the future, that is, becoming. Here we come
again to the lesson of line 18: "it is evident that not everything either is or happens
as a result of necessity".
24
G. H. von Wright, 1979 6 , pp. 237-250. The principle contested by Aristotle is
therefore that of the retrogradation of truth under the form:
(t')t' <tD(pt =Tt,pt).
For this principle applied to that of the excluded middle would legitimate the
universality of the principle of bivalence.
25
By his substantialist conception of duration, Bergson is in opposition to all Greek
 THE DIFFERENCE BETWEEN ARISTOTLE AND DlODORUS / 149

6.6 Aristotle's general conception confirms the De

Interpretations solution; the difference between
Aristotle and Diodorus.
In De Interpretatione, Chapter IX, Aristotle distinguishes two sorts of
logical necessity. By the same token, he must distinguish two corre-
sponding sorts of causality from the ontological point of view. It is
Metaphysics, K, Chapter 8, that furnishes this complement of proof
and shows, in the limitation of the principle of causality, the reason for
rejecting the universality of retrogradation and bivalence.
"Evidently there are no causes and principles of the accidental, of
the same kind as there are of the essential; for if there were, everything
would be of necessity. If A is when B is, and B is when C is, and if C
exists not by chance but of necessity, that also of which C was cause
will exist of necessity, down to the last causatum as it is called (but
this was supposed to be accidental). Therefore all things will be of
necessity, and chance and the possibility of a thing's occurring or not
occurring are removed entirely from the range of events. And if the
cause be supposed not to exist but to be coming to be, the same results
will follow , everything will occur of necessity. For tomorrow's eclipse
will occur if A occurs, and A if B occurs, and B if C occurs; and in
this way if we subtract time from the limited time between now and
tomorrow we shall come sometime to the already existing condition.
Therefore since this exists, everything after this will occur of necessity,
so that all things occur of necessity".26
thinkers and in particular to Aristotle, who identified freedom with choice among
many possibles. Nevertheless, Bergson bases his conception on the rejection of
the principle of the retrogradation of truth, which for Aristotle is none other than
the refusal to attribute a determinate truth-value to future contingents. "Yet the
conviction always persists that, even if [a given state of the universe] was not
conceived of before coming to be it could have been, and that in this sense it figures
as a possible from all eternity in some real or virtual intelligence. Looking deeper
into this illusion, it will be seen to depend on the very essence of our understanding.
Things and events come about at determinate moments: our judgement recognizing
the apparition of the thing or of the event can only come after it; it has therefore
a date. But that date is immediately effaced in virtue of that principle, anchored
in our intelligence, that all truth is eternal. If the judgement is true at present it
must, so it seems to us, always have been true. Little does it matter that it should
not yet have been formulated: it posited itself by right before being posited in fact.
In this way we attribute a retroactive effect to every true affirmation; or rather, we
impress upon it a retrograde movement."(1934. p. 21).
26
1065"6-21 (Ross, Works of Aristotle, vol. VIII). St. Thomas, quoted by Tricot
(II, p. 611), winds up his commentary on the eclipse example as follows. "If the
eclipse must occur tomorrow, and if everything that occurs results from some cause,
it must be that the eclipse itself occurs because of something that pre-exists it,
and that thing in turn because of another; and in this way, always because of the
150 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Aristotle responds to this argument by making a distinction. Tak-

ing over for the eternal beings the Platonic expression of being per
se, Aristotle declares them to be determinate and necessary, identify-
ing them with being qua true (we; dArjtfec; 6v). Accidental beings are
not necessary but indeterminate (dopioiov), and their causes are un-
ordered and infinite. 27 But the principle of correspondence transmits
the properties of the things and their causes to the statements about
them. It is because of the causal indetermination of accidents that it is
not already true, from all eternity, to say they are so; and since once p
is true it is always true, a statement about an accident cannot be true
before it happens.

anticipation or removal of causes, we shall take something away from the time there
is between the present moment and the future eclipse. Therefore, since this is a
finite time, and since everything finite is exhausted when something is taken away
from it, everything that will occur will be due to a cause existing now". This is the
reasoning used by Lukasiewicz (op.cit., 1967, p. 29). Its principle is the following:
"...every fact G occurring at instant t has its cause in some fact F occurring at
instant s earlier than t, and at every instant later than s and earlier than t there
occur facts which are both effects of fact F and causes of fact G" (ibid, p.28).
27
1065° 21-26. To the argument of necessity Aristotle replies that for contingent
beings there is a limit to the search for causes (Ross, 1924, vol. I, 363). This
is precisely the point of view defended by Lukasiewicz, which he does moreover
by recourse to the analysis of infinity and continuity. Causal chains have inferior
limits beyond which it is pure imagination to want to go back (op.cit., pp. 30-33).
For Aristotle the limit to the regression of causes is due to the fact that "some
condition will arise, if it does arise, not by a process but instantaneously" (Ross,
ibid.). The Stoics in particular blamed Aristotle for thus having introduced an
a.i/aiTio<; Kii/r)<ru; that would liken him to Epicurus. Mrs. Frede (pp. 115-117)
attributes this objection to a misunderstanding which was quite natural given the
obscurity of the terms Aristotle uses. The only difference between 'essential' and
'accidental' causality is that the former involves a unique and immanent teleology
whereas the latter involves several distinct teleologies. The accident is therefore
necessary once reinserted in its proper teleology (someone hid a treasure in the
field) and is properly accidental only when referred to an extraneous teleology (a
man plowing the field finds the treasure). Aristotle's doctrine is therefore not
unlike that of Cournot when he defines chance as the meeting of two independent
causal series. That is a classic and plausible interpretation. The Stoics in fact will
insist on the universal connection of the causal chain, the fatum. In this sense, the
limit in the regression of causes as conceived by Lukasiewicz would be Epicurean,
not Aristotelian, in spite of the mention made of instantaneous and ungenerated
conditions. The instantaneous character of the accidental event would be due to the
instantaneity of the meeting of the two independent series, each of which by itself
would be indefinitely generated. There is still, besides, a fundamental difference
between Aristotle and Cournot. For Cournot chance results from the meeting of
two series determined only by efficient causality, whereas for Aristotle it results
from the meeting of two teleological series.
 THE DIFFERENCE BETWEEN ARISTOTLE AND DlODORUS / 151

In declaring in De Interpretatione, IX that the truth-value of propo-

sitions about future contingents is not already assigned,28 Aristotle is
not then introducing a thesis that is to be without echo in the rest of
his work. It remains to be shown that, as a consequence of his adop-
tion of conditional necessity, this thesis is indispensable if we hope to
preserve what is specifically Aristotelian as over and against Diodorean
necessitarianism.
Considered by itself, the principle of conditional necessity brings
the two doctrines dangerously close to one another. Although Aris-
totle explicitly mentions possibles that will not be realized,29 the
principle of conditional necessity would seem to exclude them a priori.
Since it is supposed that the act contrary to a given potentiality, the
inexistence of something supposed contingent for example, lasts indefi-
28
Mrs. Kneale, 1962, p. 51, thinks that it is the right we assume to speak of a dated
truth that vitiates the whole of De Interpretatione, IX. She says that Aristotle's
definition of truth "gives the most important fact about the predicate 'true', namely
that, if 'P' is any propositional sign, the proposition that — P and the proposition
that it is true that —P mutually entail one another. This holds also when the
propositional sign is a sentence in the future tense. For it is true that there will be
a naval battle tomorrow if, and only if, there will be a naval battle tomorrow. By
introducing the phrase 'it is true that' we make no assumption about determinism
which is not made by use of the simple sentence in the future tense. We mislead
ourselves, however, when we speak, as Aristotle does, of its being true now that
there will be a naval battle tomorrow, for we thereby induce ourselves to suppose
that this will not be true tomorrow evening, when the battle is over, but something
else will, i.e. 'There has been a naval battle today'. Two different sentences are
plainly involved here, but they both express the same proposition in the sense
that to convict any person who uttered either of error would also be to convict of
error any person who uttered the other at the appropriate time." Aristotle's error
here would come down, therefore, to attributing the predicates 'true' and 'false'
to sentences, whose tenses vary, instead of reserving them for propositions, whose
tense does not vary since the propositional content bears no relation to the moment
of utterance. But the question remains whether a proposition about the future has
a truth-value before the event, even when one has decided with Mrs Kneale to
eliminate sentences in favor of propositions.
29
As pointed out by Chevalier, 1915, p. 273, the third proposition of the Master
Argument, SuvaTov elvai o OUT' lativ cxXT)$£c; OUT' laTai, takes over the letter of
Metaphysics, 1047*8-9, ouSev xoXuEi Suvorcov TI ov etvai ?) yevEenJai nr) elvai nr]8'
siaE^iJai. (Cf. Hamelin, 1978, p. 60). St. Thomas comments on the De Caelo in the
following way. "It is therefore clear that everything destructible will be destroyed
one day. And the same holds if something is generable in its nature, it is necessary
that it should have been generated. This is not to be understood, however, as if
everything generable was generated one day. For many things can come about which
never do. But this cannot be the case: namely, that something already existing in
its nature (jam in sua natura existens) be generable, and nevertheless should not
have been generated but should have pre-existed eternally". (St. Thomas, 1952, ad
283i>8, p. 138).
152 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

nitely, throughout all this time it is necessary for that act to be,30 and
consequently the potentiality in question is reduced to an impossibil-
ity. Mustn't we conclude then with Diodorus that the possible is what
is or will be?31 The repeated oppositions concerning modality, acci-
dent, non-being, motion and even language,32 between the ontology
30
De Caelo, I, 28161, 281t15-25, 282 a 22-25, 283=24-29. See Cherniss, 1962, p. 416,
on these passages.
31
At De Caelo, I, 283°27, Aristotle writes that "the destructible is at some time
destroyed".
J. Hintikka, 1973, pp. 93-113, systematically assimilates the Aristotelian and the
Diodorean conceptions of modality. The Metaphysics passage (0, 4, 1047*3-14)
that he alleges in favor of his thesis enjoins us only not to consider as semantically
possible that which is in principle prevented from passing to actuality, whether by
an already given material or efficient cause or by a logical principle. It in no way
says that a possible will have to be, but simply that a possible that certainly will
not be is not an authentic possible.
32
According to Aristotle "potency and actuality extend beyond the cases that in-
volve a reference only to motion" (Metaphysics, Q, 1, 1046"l-2), and they therefore
go beyond the range of the corresponding notions according to Diodorus. Time can
fall within the scope of an Aristotelian modality. Moreover, infinity, the void, and
matter are eternally in potency and never pass to actuality. When Diodorus criti-
cizes the possibility of motion regarded as a passage (Fr. 128 in Doring, 1972, p. 37:
"...so something living dies neither in the time in which it lives nor in the time in
which it does not live, therefore it never dies") he is in fact denying generation.
He is to be counted therefore among those who, through awkwardness, imitate
the Ancients and according to whom "no being is generated or destroyed, because
whatever is generated must necessarily be so either from being or from non-being,
two equally impossible solutions; indeed being cannot be generated for it existed
already, and nothing can be generated from non-being for there must be something
to underlie it" (Physics, I, 8, 191Q27-32). But how does Aristotle reply to those
who simply deny motion, like the ancient Megarians, or to those like Diodorus who
reduce it to a cinematographic succession of discontinuous states? He invokes the
distinction between essence and accident (Physics, I, 8, 191612-17; Metaphysics, A,
10691'14-34). Accidental non-being, that is to say, privation, gives rise to genera-
tion. But all material beings, as such, suffer privation in virtue of their contrariety,
and the missing contrary has a sort of ghostly existence that Aristotle calls poten-
tiality and that, because of its incompleteness, produces an uneasiness calling for
change. As regards quality for example, potentiality is the ghostly presence of the
contrary (white) in something having a given quality (black). Of course when the
potentiality passes to actuality, if ever it does, it will be a development in present or
future time just as much for Aristotle as for Diodorus. But for Aristotle it is there
at the very core of the thing before becoming actuality. In the same way, motion
for Aristotle is the fulfillment of what is in potentiality in so far as it is in poten-
tiality. The act of motion, which would be expressed grammatically by the present
progressive tense (this is moving), could in no way therefore be confused with a
succession of immobilities, that is to say, with the positions occupied successively
by the mobile in the course of discontinuous time. The 'now', says Aristotle, is a
limit, not a part of time (Physics, IV, 10, 218a6) and motion cannot be reduced
to a correlation between the points occupied by the mobile and the 'nows' dividing
time.
 THE DIFFERENCE BETWEEN ARISTOTLE AND DlODORUS / 153

of Aristotle and that of Diodorus, would be reduced then to so many

delusions, completely in Diodorus' favor.
It is precisely this closeness due to the principle of conditional ne-
cessity, that suggested the Master Argument to Diodorus on the basis
of a passage of the De Caelo. To extricate Aristotle from such a com-
promising companionship it will suffice to show that the exceptions to
the principle of bivalence allow him to answer the Master Argument
without invalidating the De Caelo passage for all that.
The De Caelo demonstrates that nothing created, that is to say
nothing having the possibility of being destroyed, can avoid destruc-
tion. For suppose a world of which it would be true that it would
forever be preserved. By the principle of conditional necessity it would
be necessary that it should be preserved forever. It would therefore for-
ever be impossible that it should be destroyed, contrary to the conse-
quence of the premise by which it was posited that it had been created.
Why the Master Argument, which is nevertheless modeled on the De
Caelo, is not valid according to Aristotle can readily be seen through
a comparison of the premises. The De Caelo reasons about the eternal
or destructible existence of the world and therefore about an essential
property that it is either true or false to attribute to the subject. The
Master Argument's third premise, on the other hand, says that what
For the Megarians, Diodorus included, there is no distinction to be made between
non-being as such and accidental non-being. There is consequently no distinction
that can mark off an essential disposition predicate that, once supposed the thing
exists, will inevitably pass to actuality within a given limit of time, from an ac-
cidental disposition predicate that could, once the thing is supposed to exist, not
pass to actuality. The only difference between essence and accident or between
possibles ad unum and possibles ad utraque that Diodorus' philosophy allows us to
retain is that regarding the irrevocability of a unique event that takes place in time
(growing old, dying) and the succession of contrary states (to be sitting, standing),
but which also inevitably pass to actuality in time.
Then again, either motion as succession is an illusion or in any case, understood as
the fulfillment of what is in potentiality in so far as it is in potentiality, understood
that is as the passage of becoming, it melts away into chimera: it is only to the
trace that it leaves in space and time that any reality can be accorded.
This opposition between Aristotle and Diodorus is borne out further by their
theories of language. Euclid, followed by Diodorus on this point (Fr. 30 in Doring,
p. 10) rejected all reasoning by analogy, and Diodorus himself admitted of no ambi-
guity as regards words (Fr. 112 in Doring, p. 31). For Aristotle, by contrast, it is one
of the essential mistakes of the Eleatics, so like the Megarians, to take 'being' in an
absolute sense when its uses are numerous (Physics, I, 3, 186a24-25) and it is only
by analogy that matter can be known, since it is potentiality that cannot become
actualized (Physics, I, 7, 191a7, just as for Plato it was knowable only by a bastard
reasoning). Thus Diodorus maintains the synonymity of'being' and 'potentiality',
Aristotle their ambiguity. Diodorus reduces potentiality to occurrence in present
or future time, Aristotle sees it as the simultaneous coexistence of contraries.
154 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

neither is true now nor will be so is possible. But such a potentiality,

though contradictory for Diodorus, would not be so for Aristotle on the
supposition (the only interesting one) that we are reasoning about an
accident, unless there were already existent the causal chain allowing
us to say that the statement about the future is already false now and
will remain so. But it is precisely this present assignation of truth-
value that is missing when the futures are in contingent matter, and it
is this absence that invalidates the Master Argument.

6.7 First interpretative hypothesis: More than

two truth-values.
It is therefore in limiting the validity of the principle of bivalence that
Aristotle resolves the aporia of the Master Argument. We must try now
to determine whether his solution is logically satisfying and whether
the intuition it is based on can be legitimately systematized. With this
inquiry we leave the realm of history and fact for that of interpretation
and hypothesis.
There are three hypotheses that come rather naturally to mind.
The first is that Aristotle's system demands the construction of a logic
of more than two truth-values. The second requires modifying the
standard definition of truth. The third suggests the introduction of
probabilities.
Leibniz, in spite of his overall sympathy for Aristotle and his con-
ciliatory approach in interpreting the texts, often criticized calling the
principle of bivalence into question. Such a challenge fails to recognize
"the very nature of truth which is determinate in enunciations we can
form about future events as it is in all other enunciations, since the
enunciation must always be true or false in itself, although we don't
always know which it is".33 Taking a tip from this final concessive
clause, we can specify the nature of the confusion that would have
crept into Aristotle's conception: he would have taken the uncertain
for the indeterminate 34 and fallaciously derived an objective property
33
Leibniz, Gerhardt, 1875-1890, VI, p. 30; Jalabert, 1962, p. 30.
34
"Philosophers agree today that the truth of future contingents is determinate,
that is to say that future contingents are future, or that they will be, they will
come to pass, for it is just as sure that the future will be as it is sure that the
past has been. It was already true a hundred years ago that I would write today,
as it will be true after a hundred years that I have written. So the contingent is
no less contingent for being future; and the determination, which would be called
certainty if it were known, is not incompatible with contingency. The certain and
the determinate are often taken for one and the same thing since a determinate
truth is one that is susceptible of being known, so that determination might be said
to be an objective certainty" (pp. 129-130). Like the Megarians, Leibniz accepts
 FIRST INTERPRETATIVE HYPOTHESIS / 155

regarding the order of things from a subjective property regarding our

knowledge. Cavalier in questions of freedom, Leibniz takes aim here at
Epicurus just as much as at Aristotle.35 His explanation applies better
in fact to the former, who contested the excluded middle, than to the
latter, who simply rejected retrogradation.
What about the logical question? Wouldn't it be only out of ex-
treme desperation that one would come to imagine that the truth of
'p or q* is not a sufficient condition for the truth of 'p' or the truth
of V ?36 But the difficulty is perhaps that the reconciliation of truth
and freedom requires more than two truth-values.37 It is this consid-
eration that prompted Lukasiewicz to propose first a system of three
truth-values, then another of four. Neither of them however can be re-
garded as Aristotelian. The first makes the possibility of p equivalent
to the assertion that if not-p then p. This corresponds to none of the
senses of the word 'possible' according to the Stagirite and, above all,
it invalidates the principle of the excluded middle which is contrary to
the letter of both the De Interpretatione and the Metaphysics.38 The
the retrogradation of the truth of future statements, which is nothing other than
the determination of their truth.
35
Leibniz, Gerhardt, VI, p. 211; Jalabert, p. 222; the text is quoted pp. 167-168.
36
Quine, 1966, p. 21; Hamelin, 1978, p. 35.
37
Von Fritz, 1972, I, pp. 241-250.
"Quine's criticism seems right—and very strongly so at first—only because the
decisive factor, namely the relation of the statements to the states of affairs stated,
is left out of the picture. The proposition 'it is true that p or q' holds without
limitation for all events, past and future, when 'p' and 'q' stand for the contradictory
opposition of 'be' and 'not-be'. At each instant a determinate event can but occur or
not occur, or a (transitory) state can but obtain or not obtain: Tertium non datur.
The principle of the excluded middle (afi'u>/ia ai/Ti<pa(?£u<;) has in that measure
an absolute validity, even though the corresponding state of affairs is realized only
when the event in question or the temporary state takes place or does not take
place. The proposition 'it is true that p or it is true that q' follows from the
proposition 'it is true that p or </' in the following sense: with the occurrence or
the non-occurrence of the event in question, necessarily either the proposition 'it
is true that p' or the proposition 'it is true that q' is verified (wahr wird). In this
sense the first proposition really is, as Quine supposes, the 'sufficient condition' for
the second, but not in the sense in which the second must be valid already before
the realization of the state of affairs corresponding to p or to q, since it is just this
that determines the appropriateness of p or q."
Though justified in showing the possible inadequacy of one or another given
logical formalization, this argument is not justified in abolishing all symbolism, as
the author himself points out in adopting a logic of more than two truth-values.
38
Lukasiewicz, 1967, pp. 40-65.
Three axioms are formulated:
1. ~ Mp D~ p (axiom of necessity)
2. ~ p D~ Mp (axiom of conditional necessity)
3. (3p)(Mp • M ~ p) (axiom of contingency)
156 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

second system preserves the excluded middle, but there is no echo in

Aristotle's work39 of the twin definitions of the possible in which it
Lukasiewicz gave an altogether different interpretation of the axiom of conditional
necessity in 1950 and 1955 (Aristotle's Syllogistic, Oxford, Clarendon Press) as
signifying either L(p 3 p), in which case no difficulty is raised, or the rule (x)a D
La, which, without further restriction, leads to paradoxes (1955, pp. 153-154).
If it is posited then that 1) the modal operators are truth-functors, 2) there are
two and only two truth-values, the true and the false, it can be shown that a)
the axiom of conditional necessity leads to a collapse of the modalities, b) the
axiom of contingency leads to the admission that everything is possible, c) there
is incompatibility between the last two axioms whose simultaneous assertion is
inconsistent. Still taking the modal operators to be truth-functors, if a third truth-
value is admitted so that a proposition can be neither true nor false, it can be shown
with a fairly intuitive assignment of matrices for the propositional connectives that
a) the axiom of conditional necessity no longer leads to a collapse of the modalities,
b) the axiom of contingency no longer leads to the admission that everything is
possible, c) the three 'Aristotelian' axioms then form a coherent system. The only
truth-functional definition of the possible that can be given in such a system is one
leading to the equivalence of the possibility of p to the assertion that if not-p, then
p (Tarski, in Lukasiewicz, 1967, p. 55: Mp = (~ p 3 p); Prior, 1962, pp. 246-247).
In the system in question (p. 59) disjunction is defined in terms of implication:
p V q ~Df (p 3 <?) 3 9- For the truth-value 1/2 (neither true nor false) the value of
the principle of excluded middle differs from true (i.e. 1).
(pv~p) = ( p D ~ p ) o ~ p = (^ D i ) 3 i = i 3 l = i.
As Prior remarks (1962, p. 244), "At this point there is therefore a divergence
between Lukasiewicz's system and that suggested in the De Interpretations; for
Aristotle held that even when neither 'There will be a sea-battle tomorrow' nor its
negation is definitely true, the alternation 'Either there will be a sea-battle or there
will not' is definitely true. ... It is not because its components are both neuter, but
because they are contradictories, that Aristotle affirms this disjunction".
(On a similar difficulty, but regarding conjunction rather than disjunction, cf. Prior,
op.cit., 1962, pp. 246-247).
39
In his Aristotle's Syllogistic (op.cit., 1951 (1957)) Lukasiewicz surmises that Aris-
totle rejected the axiom of conditional necessity (1957, p. 136) on the interpretation
already proposed but in fact gave it another form through his acceptance of true
individual apodictic propositions (pp. 153-154), whereas acceptance of the axiom
of contingency, i.e., of true contingent propositions, always leads to the possibility
or any proposition whatsoever (pp. 154-157).
Lukasiewicz then constructs a four-valued system in which there are no true apod-
ictic propositions. As a result, the distinction between truths of reason and truths
of fact simply vanishes. But within this system it is also possible to formulate
twin definitions of possibility. Using the two definitions simultaneously makes it
possible to establish, without paradox now, the existence of contingent proposi-
tions (pp. 158-180, 205-208). Lukasiewicz likens the two possibilities 'Mp' and
'Wp' (p. 173) to a pair of identical twins that cannot be distinguished if met with
separately but can be distinguished when seen together. In other words, there are
sequences such as
MWp and W'Mp
 SECOND INTERPRETATIVE HYPOTHESIS / 157

results. A more general objection still is that Aristotle is nowhere seen

to evoke more than two truth-values.

6.8 Second interpretative hypothesis: propositions

without a determinate truth-value.
We must then, without leaving the realm of two-valued logic, construct
a system that validates conditional necessity and excluded middle while
not validating the principle of bivalence.
This is an impossible task so long as one admits an intuitive crite-
rion of truth maintaining the equivalence of a statement's assertion and
the assertion of the statement's truth (the Tarski criterion). 40 But in
the analysis he gives of The Concept of Truth in Formalized Languages,
Tarski cites the example of Aristotle as one of those to have given an
intuitively acceptable semantic definition of truth for the statements
of everyday language. It will be seen that this allegation is not well-
founded. 41
According to Aristotle, "to say of what is that it is not, or of what is
not that it is, is false, while to say of what is that it is, or of what is not
that it is not, is true; so that he who says of anything that it is, or that
it is not, will say either what is true or what is false, whereas [the par-
tisan of an intermediate between contradictories] says neither of what
is or of what is not either that it is not or that it is" , 42 So goes the first
which are valid without our being able to substitute the twins, for the formulas
MMp and WWp
are invalid. Such a system has its proper paradoxes though (Prior, 1957, pp. 127-
130; Hughes and Cresswell, op.cit., 1972, pp. 307-310).
If we keep the definition of disjunction: p V < 7 = ( p 3 g ) 3 q and we choose the
value p = 2, 3, in table M9 (Lukasiewicz, 1957, p. 168), the following facts will be
noticed:
pV ~ p = (2 3 3) 33 = 3 3 3 = 1
p V ~ p = ( 3 3 2 ) 3 2 = 232 = l.
The principle of the excluded middle therefore, verified for the values p = 0,1, is
verified for the non-classical values as well.
40
"... assume that 'A or not A' is logically true, but that 'A' is truth-valueless. If
'A or not A' is logically true, it is true. By the Tarski truth-criterion, therefore, we
have A or not A. Further, the Tarski truth-criterion, and the convention that "A is
false" means "(not -A) is true" entails the conditional: if A or not-A, then 'A' is true
or 'A' is false. But we already have A or not-A. Hence, the consequence that 'A'
is true or 'A' is false follows immediately by Modus Ponens, a consequence which
contradicts the assumption that 'A' is truth-valueless". (Lambert, 1969, p. 96; also
Haack, 1974, p. 68).
41
Tarski, A., 1956, p. 155.
"Metaphysics, T, 7, 1011626-29; I follow Tricot, especially for the bracketed addi-
tion, I, p. 235.
158 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

of the arguments Aristotle brings to bear against those who challenge

the universal validity of the excluded middle and postulate an interme-
diate between contradictory statements. This argument is "evident for
anyone who defines what the true and the false are" (1011(>25). Since
this same appeal to evidence is also made in Aristotle's third argument
(1012a2) we shall join it to the first: "Again, discursive thought affirms
or denies every object of discursive or intuitive thought; this much is
evident, by definition, whenever it says what is true or is false. When
it connects in a certain way, either through affirmation or negation, it
says what is true; and when it connects in some other way, what it says
is false".43
The entire set of these arguments aims at reducing the adversaries
of the principle of excluded middle to the absurd.44 It proceeds as
follows:
(1) Every proposition is the connection or union of a subject (S) and
a predicate (P) (De Interpretatione, 5, 17"12).
(2) Such a connection, therefore, can have any of four possible forms:
S is P, non-5 is P, S is non-P, non-5 is non-P.
(3) A proposition is true if it connects 5 and P as they are connected
in reality; it is false if it connects them otherwise (Metaphysics,
0, 10, 1051625).
(4) The definition of true and false is most easily applied when sub-
ject and predicate designate the same thing. The four forms
enumerated in (2) then become: that which is is, that which is
43
Metaphysics, T, 7, 1012"2-5.
44
That this is a question-begging reduction can be seen by referring to the truth-
value matrix for negation by Lukasiewicz in 1930 and to the three-valued logic of
Bochvar, in which the excluded middle is invalidated for the internal connectives
(~ i — i\ iV ~ i = iVi — i) and the assertion of the principle of the excluded middle
(a(i) = /) is false. (S. Haack, 1974, p. 170).
The intuitionist says that "TO fj,£Ta£v di/Ti<£>d<7£u;c" is. Hence he says neither
of that which is nor of that which is not that it is or that it is not, which is the
condition for saving truly or falsely. His statement is therefore neither true nor false,
which is absurd. Ross (1924, I, pp. 284-285) justly remarks "(1) that Aristotle does
not assume merely that to say of what is that it is not or of what is not that it is
is false, and that to say of what is that it is or of what is not that it is not is true,
but these are the definitions of falsity and truth, i.e. are convertible propositions.
It is only on this assumption that it follows that the opponent, who maintains the
existence of what neither is nor is not, is saying what is neither true nor false.
(2) That the opponent is assumed to admit (a) the correctness of the definition of
truth and falsity, and (b) that every judgement must be either true or false. Thus
Aristotle is inferring the metaphysical form of the law of excluded middle—that
there is no objective intermediate between contradictories—from the logical form.
The argument thus has value only ad hominem. But of this Aristotle is well aware;
he knows that first principles cannot be demonstrated".
 SECOND INTERPRETATIVE HYPOTHESIS / 159

not is, that which is is not, that which is not is not. Applying
(3) it is evident that the two extreme forms are true, whereas the
two middle ones are false (1011625-27).
(5) Therefore, to say that what is is, or that what is is not, is to say
what is true or what is false and there is no third (1011627-28).
To say that S is P is true if S is P and false if non-5 is P, and
to say that non-51 is P is true if non-5 is P and false if 5 is P,
and there is no third possibility.
(6) On the other hand, to say that the intermediate between 5 and
non-5 either is or is not P, is not to say what is true or what is
false (1011628-29).
(7) Therefore, the negation of the excluded middle, taken as disjunc-
tion of the subject (intermediate between what is and what is
not), is neither true nor false.
(8) To say that 5 is P is true if 5 is P, and false if 5 is not-P; and
to say that 5 is not-P is true if 5 is not-P, and false if 5 is P, in
virtue of (3)(1012a4-5).
(9) Therefore, to say that 5 is P or that 5 is not-P is to say what is
true or what is false, and there is no third possibility (1012a2-4).
(10) On the other hand, to say that 5 is an intermediate between P
and not-P is not to say what is true or what is false.
(11) Therefore, the negation of the excluded middle taken as disjunc-
tion of the predicate (to be intermediate between what is and
what is not) is neither true nor false.
(12) The negation of the excluded middle, therefore, whether taken
as disjunction of the subject (first argument) or as disjunction of
the predicate (third argument), is neither true nor false.
This argument uses the definition of truth through its most evident
instance (4) or through its consequences (5) and (8). It does not con-
tain it explicitly and we had to import it from another passage of the
Metaphysics (3). The text where Tarski thinks he sees the Aristotelian
definition of truth is in fact not one. Just what relation then does
the Aristotelian definition have to the Tarski criterion? According to
Aristotle, "truth and falsity, insofar as things are concerned, depend
on their combination or separation, so that he who thinks that what
is separated in fact is separated, or that what is combined in fact is
combined, thinks truly, while he who thinks what is-contrary to the
nature of the things is in error."45 But that definition differs from
Tarski's criterion, not only by the subject to which true and false are
45
Metaphysics, 0, 10, 1051t2-5.
160 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

attributed—the statement for Tarski, thought and its affections for

Aristotle46—but above all, by the range of that attribution and also
by the relation posited between the thing and that to which true and
false are attributed.
Aristotle attributes truth and falsity to composites, either contin-
gent (the wood is white) or necessary (the diagonal is incommensu-
rable), but also to simple natures as well. Whereas the composites are
objects of discursive knowledge, the simples are known through intu-
ition. Aristotelian correspondence, which is between two structures for
composites, reduces to the simple presence of the thing for truth, to its
absence for falsity, in the case of simple natures. In the last case error
is impossible: here you either see or you don't see.47
Tarski's criterion leaves no room for this difference which Aristotle
holds to be fundamental once the truth of logical principles is at issue.
But the shortcoming of Tarski's criterion can already be shown by
insisting on the priority of the thing with respect to the very existence
of an intuitive thought. Where there is a question of simple natures,
Aristotle specifies, "if the object is existent it exists in a particular way,
and if it does not exist in this way it does not exist at all; and truth
means thinking these objects, and falsity does not exist, nor error, but
only ignorance".48 Where the object is inexistent, there is no thought,
therefore no falsity. So, contrary to Tarski's criterion, there is a truth
for simple natures that no falsity can come to deny.
In De Interpretatione, Chapter nine, the same priority of the ob-
ject with respect to thought and truth has the consequence that, as
future contingents are still indeterminate as to the object, it is in no-
wise plausible that the corresponding affirmations and negations have
a determinate truth-value before the occurrence of the event.
The formal expression of such a conception requires a system that
1) preserves all the logical tautologies, 2) does not, however, make an
already-determinate truth assignment to every statement, 3) does not
introduce a third truth-value. Van Fraassen's languages of superval-
uations meet these three conditions. To a molecular proposition of
which some elements lack a determinate truth-value, a supervaluation
will assign what would be the ordinary valuation in cases where that
determination is unique (as with tautologies and contradictions) and
no determinate truth-value where that is not the case. For example,
a molecular proposition 'p V q\ which is sometimes true and some-
46
Metaphysics, E, 4, 1027625-1028a2.
^Metaphysics, Q, 10, 1051fc30-1052all.
48
Metaphysics, 0, 10, 1051i>35-1052°2.
 THIRD INTERPRETATIVE HYPOTHESIS / 161

times false on a classical evaluation, will not receive a determinate

truth-value on a supervaluation. On the other hand, the proposition
'pV ~ p: and 'p- ~ p1 will receive respectively the supervaluations
'true' and 'false', which they are assigned universally on the ordinary
evaluations.49 Some inferences considered valid in classical logic cease
to be so in this system. Such is the case for the 'disjunctive dilemma'
(if A \- C and B h C1, then A VB h C). It is right to contest such infer-
ences when they involve formulas lacking determinate truth-values, but
once that indetermination is removed the classical rules are restored.50
Van Fraassen's system fits into the program of free logics, i.e. log-
ics without existential presuppositions, as formulated by K. Lambert.
Adopting it to Aristotle would mean effectively fixing the status of fic-
tive objects, so important in the poetics and the rhetoric, and treating
singular statements with non-referential subjects as truth-valueless.51
In that respect, they would be assimilated to future contingents. On
the other hand, in freeing the whole of logic from existential presuppo-
sitions we would surely come into conflict with the assumption of the
non-empty universe of discourse which is at the very base of the logic
of terms and therefore of the entire Aristotelian syllogistic.

6.9 Third interpretative hypothesis: probability.

Chapter IX of the De Interpretatione and Chapters 4 and 6 of book
II of the Physics enable us to separate the sets of contingent events—
those that are neither necessary nor impossible—into three classes: I/
those with a low frequency (ut in paucioribus) and involving chance or
fortune, chance and fortune resulting from the meeting of two indepen-
dent final causal series, the first not accompanied by the representation
of the end, the second accompanied by it; 2/ those that have as much
chance of coming about as not (ad utrumlibet) and which depend on
deliberate choice; 3/ those with a high frequency (ut in pluribus) which
represent the unimpeded effect of natural causality, though neverthe-
less in the sublunary world it doesn't always follow its course.52
49
Bas C. van Fraassen, in K. Lambert, 1969, pp. 67-91. Mr. Heinzmann points
out that it would be possible to imagine another model that would preserve the
syntactic validity of the excluded middle, but only on a non-standard interpretation
identifying truth with demonstrability. While such a model would satisfy Aristotle's
formal requisites it would be in flagrant violation of the objective conception of truth
that he proposed.
50
Haack, op.cit., 1968, p. 216.
51
Lambert, op.cit., 1969
52
"We say that everything either is always and of necessity, or is what happens for
the most part, or is neither what happens for the most part nor what is always and
of necessity, but what happens merely as it chances. That it might be cold in the
162 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

A representation such as the above is completely at odds with the

idea of probability, since it prevents bringing together under a single
heading the set of events whose foreseeability is neither null nor ab-
solute. It holds every law to be necessary, since deploying the finality
of nature. It is by principle then that the exception escapes the law.
Failing in finality, it could in no way be counted among the cases falling
under a law.53 That, according to Aristotle, is the reason why natural
selection can in no way simulate finality.54
In spite of all this, there is one circumstance that should have, and
in fact did, provoke an inquiry on the part of Aristotle and his school
into just what there is in common between chance, fortune, freedom,
and the approximative laws of the sublunary world. The privative
day-days, for example, is something that happens neither always and of necessity,
nor for the most part; it can happen only sometimes. The accidental, then, is what
occurs, but not always nor of necessity, nor for the most part" (1064i>32-1065a3).
Chance and fortune are defined in the Physics, II, 6, 197h18-21. "In the realm of
things that take place in the absolute sense for the sake of some end, when they
come about without having in view the result and having their final cause outside
of it, then we speak of the effects of chance; and for the effects of chance of the
kind susceptible of being chosen and touching beings capable of choice, we speak of
the effects of fortune". The examples given are that of a horse coming to a certain
place and finding safety there (197i>14-15), and the unexpected recovery of a debt
through meeting the debtor (196fc34-197"4).
53
Science cannot then content itself with assigning the probability of an event in a
series: it must explain its occurrence in a determinate way. "For all science deals
either with what is always or with what is for the most part. How else is one to
learn or to teach another? The thing must be determined as happening always or
for the most part. It can thus be said that honey-water is good for one having a
fever, for the most part; but one cannot give an account of exceptional cases, say
at what moment honey-water will not produce that effect, at the new moon for
example: for even what happens at the new moon happens either always or for the
most part, while the accidental is outside the always and the most part" (Meta-
physics, E, 2, 1027a20-26; A, 30, 1025a14-30). It can be seen here why Cournot
writes: "It follows from all this that Aristotle has an inkling, though in a most
confused way, of the applications of the doctrine of chances and of probabilities,
and of the future science of statistics, though not knowing whether to place it in
'science' or 'opinion' " (1975, p. 450). It is evident then that neither Aristotle
nor Aristotelianism developed a clear and distinct concept of probability and that
they weren't even aware of the questions that arise concerning its measure. It is
nevertheless so, as Cournot says, that the attention they paid to the imperfect de-
terminations of the sensible world and to accidents must—as in Chapter IX of the
De Interpretatione—have given rise to themes that enter into conflict with logic if
there is no room made in the system for a primitive theory of probabilities.
54
"And of course it is the beings in which everything was produced as if it had
been for a purpose that have been preserved, being organized in a fitting way; the
others, like the man-faced of oxen in the theory of Empedocles, have perished and
continue to perish. This, among other ways, is how those who raise this difficulty
argue, but it is impossible that this should be the case" (Physics, II, 8, 198i>29-32).
 THIRD INTERPRETATIVE HYPOTHESIS / 163

concepts that the determinist doctrines of Diodorus and the Stoics

denied or assimilated to fate55 took on, at the hands of the adversaries
and then consequently of the defenders of Aristotelianism, a unity that
was at first awkward but quickly assumed some consistency. Polemical
pressure made Aristotle and his disciples more and more attentive to
the individual, to opinion, to contingency: in a word, to probability.
It is not an unfounded fiction then to suppose a disciple of the
Lyceum to be wondering about the probable in itself. With a view to
harmonizing his principles, it will not have taken him long to posit the
probable in nature, and not in our imperfect knowledge.56 His attitude
can be found in taking the opposite view of the Stoics. In speaking of
them, Alexander says "whatever happens, and no matter how diversi-
fied the causes, our adversaries still think it equally true of all that it
is impossible that under the same circumstances, either with respect
to the cause or with respect to what follows from it, that things should
sometimes not happen in a certain way and sometimes happen in that
same way".57 Our philosopher will be committed therefore to recog-
nizing that, under the same circumstances with respect to the causes,
what follows from the causes will happen at one time in one way, at
another in another, where the set of these 'ways' or eventualities con-
stitutes an exhaustive set of contraries and with the probable resulting
from the contrariety among these eventualities.58 There will be 'iden-
tical circumstances' when the total combination of eventualities is the
same. Without denying the univocal and entirely determinate action
of a 'causal disposition' he attributes to essence, the Peripatetic will
also admit that the probable, an accidental cause as regards facts stem-
ming from an irrational causality, an essential cause as regards those
stemming from free choice, ends the supposedly infinite59 causal chain
in the fabric of a supposedly universal determinism. In this latter case,
that of free causality, he will bring to bear possible causes that will
not all become real, as the effect is reduced to the choice (even only
aleatory at that) of a single case.
The Ancients do not seem to have distinguished the concept of
event from that of eventuality, an event being defined as a determinate
subset of the set of eventualities. Having defined a set of eventualities
X, they reason about only a part of that set, a part corresponding to
55
Alexander of Aphrodisias, in Nourrisson, 1870, X, p. 211.
56
Alexander of Aphrodisias, op.cit., X, p. 211.
"Ibid., XXII, pp. 262-263.
5s
lbid., IX, pp. 208-209; XII, p. 223.
™Ibid., XXII, pp. 260-261.
164 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

a natural event, and the complement of that part.60 It would suffice

to generalize this process to meet the first requisite of the calculus of
probabilities. Instead of a natural part of X, extracted from the set of
its subsets, let us consider the set of subsets of X in its entirety. Let
us construct on the set X of eventualities what is called a field of sets
F: i.e. a non empty family of sets (in this case events) closed under
complementarity and logical sum.
A field of sets F has the following two properties, a/ Since when it
contains event A it also contains its complement non-A and their logical
sum, it contains the entire set X. By definition this latter set contained
all that will occur no matter what—it corresponds, so to speak, to the
total cause, b/ Since F contains the sum of A and non-.4 it contains
their logical product as well: i.e. the complement of that sum.61 This
product is the empty set and corresponds, so to speak, to the absolute
absence of cause. In other words, if the contingent proposition 'The two
processions will meet' corresponds to event A, while its contradictory
'The two processions will not meet' corresponds to event non-A, it will
be possible to form both the disjunction and the conjunction of these
two contradictory propositions, namely 'The two processions will meet
or they will not meet' and 'The two processions will meet and they will
not meet'.62
The Ancients' concept of the probable was a qualitative one. But
as is witnessed by the opposition of the words "rare, equal, frequent",
they conceived of an order of greater and lesser probabilities; and in the
case of two events having an equal probability, as well as in the frequent
situation of drawings, they conceived of a numerical assignment of
equal probabilities. Imagine our Peripatetic then thinking to himself
60
Consider the following elementary eventualities: there will be a procession a at
time t in one of the two places x or y and there will be a procession b at the same
time t in one of the same two places x or y. The set X of eventualities is formed by
combining these elementary eventualities. It will contain four elements, namely: I/
a will take place in x and b will take place in x, 2/ a will take place in x and b will
take place in y, 3/ a will take place in y and b will take place in x, 4/ a will take
place in y and b will take place in y. The set of subsets of X, say Y, has 2 4 = 16
elements, among which are to be counted the empty set and X itself. One might
for example conceive of the event {a takes place at x and b takes place at x or a
takes place at x and b takes place at y} which would be the set {1,2}, that is to
say, the event in which a always takes place at x. The Ancients would tend to raise
the question as to whether the two processions would meet or not. They would
focus then on the event A = {1,4}, the processions meet, and the complementary
event ~ A = {2, 3}, they do not meet.
61
~ (ALI ~ A) = An ~ A = 0.
62
7r(,4n ~ A) = TT(~ X). But ir(X) - ir(AU ~ A) = 1;
if(X) + TT(~ X) = 1. Therefore TT(~ X) = 1 - 1 = 0.
 THIRD INTERPRETATIVE HYPOTHESIS / 165

of a field of sets constructed on the set of eventualities X. If he defined

in this field an event A comprised of half the eventualities of X, he will
spontaneously assign the probability 1/2 to A and to non-A, since the
eventualities of X are symmetrically disposed, and the probability 1
to their sum and the probability 0 to their product. If he extends his
considerations to general cases he will quite naturally define an additive
finite probability space with every probability being a number between
0 and 1 and such that every event has a probability equal to or greater
than 0, the probability of the set X of all eventualities is equal to 1
and if two events are without a common part, as is always the case in
the examples of the Ancients, the probability of their sum is equal to
the sum of their probabilities. And this is the second requisite of the
calculus of probabilities.
Let us come back now to event A. To say with Alexander that from
a same set of causes contrary future contingents may occur is to say
that on the same set of eventualities X one can assign a probability
to A, 0 < ir(A) < 1, such that the probability of the complementary
event ~ A will be 1 - n(A) and therefore itself > 0. Thus A and
non-A can both occur, contrary to what the Stoics held, and one can
say that it is possible that A at the same time that it is possible that
non-A. A proposition such as this affirming the contingency of A will
be analyzed simply in terms of the conjunction of two propositions
assigning complementary probabilities to these two events. But the
sum of these probabilities, equal to the probability of their sum, is
none other than 1. As Aristotle said, it is not determinately certain
either that A will occur or that non-A will occur, if it was supposed that
neither Tr(A) nor ?r(~ ^4) was equal to 1 or to 0. On the other hand,
it is determinately certain that n(A\J ~ A) = 1, so that the excluded
middle is always verified. Nor is the principle of non-contradiction ever
violated since ir(Ar\ ~ A) = O.63
It is therefore possible to make sense of De Interpretatione, IX and
to interpret it speculatively without historical claim by translating it
into the language of probabilities with A being the event of which p
asserts the existence.
63
If we were to interpret conditional necessity as Leibniz does and as Lukasiewicz
tended to do in 1955 (viz. as meaning, it is necessary that if p, then p) its translation
in terms of probability would be simply ir(A(J ~ A) — 1, since 'if p, then p', is true
if and only if 'not-p or p' is true. Conditional necessity would be_indistinguishable
from excluded middle, another form of the principle of propositional identity.
166 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT
1. p is possible n(A) > 0
p is necessary if(A) — 1
p is impossible n(A) — 0

2. It is necessary that p or not-p Tr(A\J ~ A) = 1

It is impossible that p and not-p 7r(Ar\ ~ A) = 0
Even if it is necessary that p or not-p Tr(AU ~ A) = 1
it is not for all that that it is does not entail
necessary that p or that it is necessary either ir(A) = I
that not-p or TT(~ A) = 1

3. p is absolutely necessary ""(^4) = 1

p is conditionally necessary XB(A) = 1
(an incomplete expression which means
that p is necessary on condition that q,
B being the event attested by q)
This table must not be read as a translation since the probability
sentences do not preserve the truth-value of the corresponding modal
sentences. However modal sentences entail a fortiori the correspond-
ing probability sentences; while the probability sentences do not fix
sufficient conditions for their modal counterparts.
To go further it would be necessary to recast the modal expressions
with two temporal indices in terms of probabilities. If it does not seem
advisable to affect the probabilities themselves with temporal indices,
there is nevertheless an interpretation of probability theory, advanced
by certain physicists, suitable for producing the desired result. These
physicists define probability as a prediction or mathematical expecta-
tion of relative frequency.64 Such a definition, which for that matter
seems to be required by a reflection on the modalities of double tempo-
ral index65 and, more especially, on Aristotle's66 future statements in
contingent matter, makes it immediately possible to translate the Mas-
ter Argument's premise (A) and the principle of conditional necessity.67
The second premise (B) would say that if it is possible now that p at t
we predict for a large number of systems similar to p at ti(N <ti<t)
a relative frequency > 0. To say, finally, that there are possibles that
will not be realized is to assert that, in such a set of systems, there are
64
C.F. von Weizsacker, Aufbau der Physik, Carl Hanver, Mvinchen-Wien, 1985,
p. 110; M. Drieschner, Voraussage-Wahrscheinlichkeit-Objekt, Lecture Notes in
Physics 99, Springer, Berlin, 1979, p. 73.
65
von Weizsacker, 1985, p. 81.
66
/6td., pp. 52-53.
67
Ibid., p. 93. If probability is predictive (A) and (NH) are verified.
 THIRD INTERPRETATIVE HYPOTHESIS / 167

some that will not verify p at ti, although the relative frequency of p
at t\ is different from 0.
Looked at in this way, the intuition developed in the De Inter-
pretatione is not logically contradictory. It does nevertheless require,
that, within the domain of the possible as regards reality and which
is therefore already different from the pure logically or mathematically
possible, a distinction be made between essence and accident. Only
the accident is subject to probability. To be already true or false, for
an accident, is to have already fallen within the domain of reality, that
is to say, to have a probability equal to 1 or to 0. Not to have a deter-
minate truth-value is to have a probability falling between those two
extremes. Then all the modal expressions used in Chapter IX trans-
late into probabilist expressions and the paradoxes pointed out by the
logicians vanish.
A theoretical difficulty, however, dismisses this interpretation, at
least if we use it in its direct and literal sense. When Aristotle doubts
that bivalence is universally valid, his doubt is cast on individual con-
tingent factors. The premises (A), (B) and (NH) of the Master ar-
gument are considered as true. What is questionable in (C) is not
the notion of a possible deprived of realization but the assertion of
its actual truth before the non-realization has occured. On the other
hand, the probabilistic interpretation requires that probabilities be in-
terpreted as frequencies, since any hint of subjective probabilities is
barred in Aristotle's dogmatic system. Then there are not individual
events, but frequencies that are able to verify or to falsify probabilistic
sentences.
However, introducing frequencies into the Master Argument's in-
terpretation seems to recommend Diodorus' rather than Aristotle's
solution.68 Moreover, when physicists today make reference to Aris-
totelian potentialities, they indeed allude to second order probabilities
depending on probability amplitudes, all concepts that lead us far away
from the letter of the Master Argument. 69

68
See later, p. 249.
69
See later, 10.5, pp. 259-263.
Epicurus and Intuitionism.

Epicurus escaped the Master Argument by rejecting Dialectic.1 He in-

sured freedom, for the atoms and the soul, in admitting that the atoms
move2 with an uncaused swerve.3 Antiquity mocked him for these de-
cisions, nor were Bayle or Leibniz more favorable to him on this count.
The latter writes: "It seems that in order to preserve freedom and
avoid an absolute necessity, Epicurus, after Aristotle, held that future
contingents were in no way capable of a determinate truth. For if it
was true yesterday that I would write today, it could not therefore fail
to happen: it was already necessary; and, by the same reason, it was
so from all eternity. Thus everything that happens is necessary and it
is impossible that it should happen otherwise. But since that is by no
means the case, it would follow, according to him, that future contin-
gents have no determinate truth. To uphold this sentiment Epicurus
went so far as to deny the first and greatest principle of the truths of
reason; he denied that every statement was either true or false. For
here is how he was driven to that extremity: You deny that it was true
yesterday that I would write today, therefore it was false. The poor
'Cicero, Academica, II.XXX.97; Brehier, 1962, pp. 231-232. "For if they don't get
Epicurus, who scorned and ridiculed dialectic, to admit the truth of the following
statement to them: 'Either Hermarchus will be alive tomorrow or he will not', while
the dialecticians establish that every disjunctive statement such as 'yes or no' is not
only true but even necessary (see now how clever he is, that Epicurus you take to
be slow-minded: for they say that if he admits that the one or the other is necessary
it will be necessary that, tomorrow, Hermarchus be alive or that he not be alive;
but there is no such necessity as that in nature), then while the dialecticians, that
is to say, Antiochus and the Stoics combat Epicurus, Epicurus turns the whole of
dialectic upside down. For if a disjunctive proposition made up of contraries (I call
contraries statements of which the one affirms what the other denies), if then such
a disjunction can be false, then no other is true."
2
Cicero, De Fato, X(23); Brehier, 1969, p. 482.
3
Cicero, De Fato, IX(19), X(22); Brehier, pp. 480-481; De Fato, XX(46-48),
Brehier, p. 491.

169
170 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

man, unable to admit that conclusion, was forced to say that it was
neither true nor false. After that, there is no need to refute him". 4
Leibniz' judgment leaves the difference between Aristotle and Epi-
curus in the dark. Aristotle, along with the dialecticians, explicitly
admits the truth of "either Hermarchus will be alive tomorrow or he
will not"; he even establishes its necessity. What he denies is the
distribution of truth—and therefore of necessity too—over each of the
disjunction's components. Epicurus rejects the truth, and therefore the
necessity, of the disjunction itself.5 What Epicurus denies is therefore
not the determinate character of the truth-values of the disjunction's
components, but the truth of that very disjunction. His criticism is
directed not to the principle of bivalence but to the principle of the
excluded middle. Naturally, a challenge to this latter is a challenge to
the former, for then the excluded middle is neither true nor false.6
Historians of philosophy have paid little attention to Epicurus' de-
nial of the principle of the excluded middle, either for lack of interest
in logical questions or for the suspicion they bore towards Ciceronian
testimony. Logicians have given two different interpretations of that
denial. The only one that can be retained since it is compatible with
the texts will lead to considering Epicurus' doctrine as a form of in-
tuitionism. We shall have therefore to examine the nature of the Epi-
curean criteria. From this examination we shall draw consequences
regarding certain hypotheses as well as the principle of the excluded
middle, which will permit us to assign the probable attitude of Epi-
cureanism towards the Master Argument. Thereafter we shall look at
other intuitionist representations of modality in Descartes and Kant.
4
Leibniz, Gerhardt, VI, p. 211; Jalabert, p. 222. Leibniz based his remark on Ci-
cero's Academica reporting something Carneades said, perhaps ironically, in praise
of Epicurus (Bk. 2, XXX(97); Brehier, 1969, pp. 231-232).
5
See note 1.
6
On the contrary, denying the validity of the bivalence does not entail denying the
validity of the principle of the excluded middle, since the first denial only bears on
future contingents and does not question logical truths. Cicero makes mention of
Epicureans who, ashamed to have to declare that there are propositions that are
neither true nor false, "declare more impudently still that alternations of contra-
dictories are true, but that neither the one nor the other of their terms is true"
(De Fato, XVI. 37; Brehier, 1962, p. 487). This Ciceronian passage evokes that
of Quine on the fantastic character of Aristotle's position. The difficulty there is
in distinguishing logically between the principle of bivalence ('p' V '~ p') and the
principle of the excluded middle ('pV ~ p') explains how the Epicureans, natively
contemptuous of logic, could have taken over the Aristotelian position which per-
haps recommended itself to their eyes by virtue of the Stagirite's reputation as a
logician.
 FIRST LOGICAL INTERPRETATION / 171

7.1 First logical interpretation of the Epicurean

denial of the excluded middle: the
three-valued logic of Lukasiewicz; reasons for
rejecting this solution.
At the same time as he proposed a three-valued logic to explain De In-
terpretatione IX,7 Lukasiewicz emphasized Aristotle's reticencies and
obscurity. He added that in according a truth-value, albeit indefi-
nite, to propositions about future contingents, the Aristotelians had
retreated in face of the consequence of the principle gingerly advanced
by their master. To them he contrasted the Epicureans who had es-
poused Aristotle's indeterministic intuition.8
Let us look then at this three-valued system where characteristic
matrices have been defined for the prepositional connectives taking as
arguments these three truth-values and having as values these same
truth-values, whereas the characteristic matrices of the modal opera-
tors take the three truth-values as arguments but have as values only
the true or the false. On this first interpretation, 1) the principle of
the excluded middle is invalid,9 2) the interdefinability of the connec-
tives 'and', 'or' (laws of De Morgan) is respected, 3) the disjunction
(p V q) is no longer definable as (~ p D q), 4) the modal operators are
truth-functional 10 and 5) the possible is defined as that which is not
false.11

7
See above, 6.7, pp. 153-155.
8
Appendix to "Many-valued Systems of Propositional Logic" in Polish Logic,
edited by Storrs McCall, 1967, p. 64.
9
See above, p. 154, Note.
10
Lukasiewicz in Storrs McCall, 1967, pp. 40-66. With the three-values chosen as
1, j , 0 , the following must be verified: 1) ~ p = 1 — p, 2 ) p D < 7 = l i f p < g and
p D < 7 = l— p + 17 if p > g; see also Prior, 1962, pp. 240-250. Disjunction is defined:
p V q = (p D q) D q.
^Lukasiewicz, op.cit., p. 55. (Mp = £ > ^ ~ p 3 q).
The matrix is:
p ~p ~p D 9
1 0 1
1 1
2 2
0 1 0
whence MO = 0, M j = 1 and Ml = 1.
whence
LO S~ M ~ 0 =~ Ml =~ 1 = 0,
L\ = 0, LI = 1; M ~ 0 - 1, M ~ | = 1, M ~ 1 = 0;
~ MO = 1, ~ M| = 0, ~ Ml = 0.
The theorem of two-valued prepositional calculus (~ p D p) D p is falsified here for
172 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

But does this three-valued system of Lukasiewicz enable Epicurus

to refute the Master Argument? The refutation can evidently bear
only on the conclusion since the excluded middle figures in the rest of
the argument only to say that every moment of time belongs either
to the past or to the union of present and future, and that is hardly
contestable. Epicurus would have to have held then that the Master
Argument proved the incompatibility of its three premises but that,
even admitting this incompatibility, one was not forced for all that to
relinquish one or another of them, following the consequence drawn
by Diodorus, Cleanthes and Chrysippus. Formally, that comes down
to saying that the negation of the conjunction of several propositions
does not entail the disjunction of their negations. To escape Diodorus,
Epicurus had then to abandon the law of De Morgan.12 But that law
is specifically verified by the matrices of the first calculus Lukasiewicz
proposed.13 And this is so precisely because those matrices still verified
the interdefinability of 'and' and 'or'. If Epicurus' parry was to suffice
in saving him from necessitarianism, it had to be that in denying the
excluded middle he was assured of denying De Morgan's law as well,
and that he had thus already denied the interdefinability of 'and' and
'or' in adopting a new definition of negation. The choice Epicurus was
faced with was comparable to that of Chrysippus; and both of them
fell under the same reprobation on the part of Cicero and Bayle.14
Chrysippus gave up the interdefinability of the modalities.15 More
radical yet, it was at the level of the prepositional calculus itself that
Epicurus asserted the independence of the connectives. For that, it
It will be noted that the same goes too, for the same value, when a stronger
negation-matrix, ~ ^ = 0, is adopted.
There
ineic (~-~)-)~)-
^ 2 ^ 2 ' - > 2=
~ (0~)-)~)i
~ ^ - ^ 2 ' ^ 2=~l~)-
- J 2 ~=2-'
12
Here, for example, with p, q and r designating the three Diodorean premises:
~ (P ' 1' r) ^ ~ pV ~ gV ~ r.
13
Let us show this for the case of two variables:
P 9 p-q ~ (p • q) ~ pV ~ q
1 1 1 0 0
1 i 1 1 i
2 2 2 2
1 0 0 1 1
1 1 1 1 i
i 1 1 ! !
i
2
2
0
2
0
2
1
2
1
0 1 0 1 1
0 1 0 1 1
2
0 0 0 1 1

It is seen that there is reciprocal implication between the last two expressions.
14
Leibniz, Gerhardt, VI, pp. 211-212; Jalabert, pp. 222-223.
15
See above, 5.3, p. 113.
 SECOND LOGICAL INTERPRETATION / 173

suffices to change the matrix for negation. It will take its arguments
from the three truth-values but, as is the case with the modal operators,
it will have as values only the true and the false: namely, the true for the
false and the false for the other two argument-values.16. Negation then
has the same matrix as did the impossible in the previous system.17

7.2 Second logical interpretation of the Epicurean

negation of the excluded middle: The
Intuitionist System.
With this truly 'modal' definition of negation and the old characteristic
matrices of the propositional connectives, some of De Morgan's laws18
are falsified but not the one that interests us. It can be shown that
the system of axioms verifying these new matrices, on the condition
that one of the axioms is omitted, corresponds to Heyting's intuition-
ist propositional calculus.19 Should we try to construct a model of
this last mentioned calculus, no set of matrices with a finite number
of elements renders it valid; but validation may be obtained from an
infinite number of elements20 and therefore with an infinite number of
16
Prior, 1962, pp. 250-259.
P ~P
1 0
I 0
0 1

17
See note 11.
18
P '1 ~ (P • g) ~ PV ~ q ~ (~ pV ~ q)
1 1 1 0 0 1
1 1 1 0 0 1
2 2
1 0 0 1 1 0
i 1 1
rf
1
2
I
2
0
0
0
0
1
1
2
0 0 1 1 0
0 1 0 1 1 0
0
1 0 1 1 0
2
0 0 0 1 1 0
Consequently:
1— (p • q) =~ pV ~ ,

but not
* ~ (~pV~g) D (p-g).

19
The axiom in question is h (~ p 3 q) 3 {[(q D p) D q] 3 q} which is verified by
the three-truth-valued matrices.
20
Prior, 1969, p. 253.
174 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

irreducible degrees of possibility.21 But then it becomes impossible to

interdefine any of the prepositional connectives no, and, or, if then. In
particular, the conditional of which the antecedent is the negation of a
conjunction of propositions and the consequent the disjunction of the
negations of those propositions is invalid.22 Then, and only then, can
Epicurus escape the Master Argument's conclusions.

7.3 Are the Epicurean 'criteria' compatible with

intuitionism?
But is it reasonable to foist off such a conjecture on Epicurus, who
held dialectic in contempt? Anaxagoras is the only Greek philosopher
mentioned by Aristotle as having challenged the excluded middle. But
then his theory of homoiomereia and of infinite divisibility gives ad-
ditional good reason for considering him an ancestor of intuitionism.
Epicurus, by contrast, adheres to the finite divisibility of matter and
the existence of atoms. Both these dogmatic themes inherited from
Democritus are contrary to the spirit of intuitionism, at least if we
give the word 'intuitionism' its full sense: a constructive methodol-
ogy applied to the mathematical continuum and to the whole building
of our theoretical and practical ideas. As we do not know what was
Epicurus' conception of mathematics, it will then be safer to regard a
priori as doubtful the logical interpretation that has been advanced to
explain the Epicurean denial of the excluded middle.
Let us consider however just what it was that distinguished Epi-
curus from Democritus. It is what made him formerly the object of
universal derision: the swerve in physics that prompted Cicero23 to ask
whether the atoms "draw straws for which one will deviate and which
one will not deviate", and his forsaking of the excluded middle in logic.
It is easy to imagine the unexpected new vigor modern atomism has
21
Lukasiewicz in Storrs McCall, p. 61. That infinity of possibles will be defined by
the unique formula:
Mp=~p D p
in giving p first two, then three, then four, etc. distinct truth-values, thus with the
matrices
Mi(0) = 0, Mi(|) = 1, Mi(l) = 1
M 2 (0) = 0, M 2 (l/3) = 1, M 2 (2/3) = 1, M 2 (l) = 1, etc.
22
This is thesis 55b in Kleene, 1971, p. 24 which is not valid in intuitionist logic:
(~
23
(P ' 1} D (~ PV ~ ?)).
Cicero, De Fato, XX(46), Brehier, p. 491. Sedley (op.cit., note 137, pp. 115-116)
thinks that in his Letter to Herodotus (307/6) Epicurus "still seems quite unaware
of the determinist threat". His clinamen theory and his wariness of necessity are
explicit in the Letter to Menoeceus and the fragments of his book on free will written
after 296/5.
 ARE THE EPICUREAN 'CRITERIA' COMPATIBLE WITH INTUITIONISM? / 175

given to Epicurean physics and to its discontinuity theses.24 Then

again, his Canonic that accords a fundamental importance, though in
a quite different sense from Anaxagoras, to the activity of the subject
in the formation of the criterion of truth, becomes clearer in the light
of intuitionism.
Intuitionism has often been considered to admit a definition of truth
reduced to verification or, in mathematics, to construction. From the
logical point of view that would mean, consequently, that in any for-
mal system expressive enough, since not all that is true in the classical
sense is demonstrable, this truth that goes beyond demonstration is
only chimerical and it would be wise not to linger on it. More gener-
ally, it would emphasize this theme of reasonableness: given that not
all that is true in the classical sense is verifiable, it is important to
restrict oneself to the verifiable only, at least to the extent that only
it produces direct or indirect consequences in our present experience.
But what is it that the Letter to Herodotus (51) says, for example, when
it comes to distinguishing between illusion and error? There is illusion
when our disposition receives mental images (phantasms) resembling
external objects from sensible material. But we can do nothing about
an illusion; it arises therefore in connection with an 'immobile' appre-
hension. Error, on the contrary, requires a sort of action on our part, a
'mobile' apprehension. Epicurus says, "Error would not exist if we did
not receive also in ourselves another associated but different motion.
It is from this motion associated with the imaginative apprehension,
but different from it, that, if there is no confirmatory evidence or there
is counterevidence, error arises, and truth if there is confirmatory evi-
dence or no counterevidence".
Even though the commentators on Epicurus seem to be at one, for
once, on the criteria of error (lack of confirmatory evidence or coun-
terevidence) and truth (confirmatory evidence or no counterevidence),
this part of the doctrine is not without problem. It is repeated in other
texts25 that throw no light upon it. What they do seem to agree on is
that "the control of opinion by experience, positive or negative, splits in
two here, leaving open a margin of uncertainty: the non-confirmation
may be only temporary, and the absence of falsification is not equiva-
24
Mugler, 1953; Vlastos, 1965.
25
Cf. the translation of the Letter to Herodotus (50-51) in Bailey, p. 29 and in
Furley, p. 207. There are other similar texts on the criteria in Epicurus: Vita
Epicuri (34). "Opinion they also call supposition (6x6Xr)<|Hv), and say that it may
be true or false: if it is confirmed or not contradicted, it is true; if it is not confirmed
or is contradicted, it is false". (Bailey, p. 163).
176 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

lent to a direct confirmation" ,26 But that margin of uncertainty may,

on the one hand, lead to a resuscitation of fears and myths, 27 and on
the other hand, it squares poorly with Epicurean empiricism in gen-
eral as well as with other explicit texts. For Epicurus denounces as
the principal source of error the temptation to put on an equal footing
that which awaits confirmation (i.e. opinion about sensation) and that
which does not await confirmation (i.e. sensation itself).28 The Life
of Epicurus passage setting forth the definitions of true and false that
seem "to leave open a margin of uncertainty" is immediately followed
by this empiricist declaration: "For this reason was introduced the
notion of the problem awaiting confirmation: for example, waiting to
come near the tower and seeing how it looks to the near view" ,29 Such
'awaiting' would be senseless if every non-falsified opinion were true.
If the empiricist principle is fundamental, if "all investigations must
be controlled by sensations" ,30 we must first inquire just to what the
given definitions of truth and falsity apply. They cannot apply to
opinions about sensible things which are liable to confirmation or falsi-
fication, since such opinions are only problems;31 'clear vision' confirms
or falsifies: it leaves no room for a margin of uncertainty. If this is so,
we must look for another level of knowledge upon which to apply the
definitions. There is a passage in Sextus Empiricus that gives us a
lead in explaining what is to be understood by non-falsification. "Non-
falsification is conformity of the invisible thing, posited by hypothesis
and conjecture, to the phenomenon. For example, Epicurus asserts the
existence of the void, that which specifically is invisible (adelon). But
26
Rodis-Lewis, 1975, p. 111.
27
Principal Doctrines, XII, in Bailey, p. 97.
28
Principal Doctrines, XXIV, in Bailey, p. 101 and commentary, p. 363. "XXIV.
If you reject any single sensation and fail to distinguish between the conclusion of
opinion as to the appearance awaiting confirmation and that which is actually given
by the sensation or feeling, or each intuitive apprehension of the mind, you will con-
found all other sensations as well with the same groundless opinion, so that you
will reject every standard of judgement. And if among the mental images created
by your opinion you affirm both that which awaits confirmation and that which
does not, you will not escape error, since you will have preserved the whole cause
of doubt in every judgement between what is right and what is wrong." (Bailey
Translation).
Cf. also Goldschmidt, 1977, pp. 270 and 272 as well as his commentary on confir-
mation, falsification and prenotion.
29
Life of Epicurus (34) in Bailey, p. 163.
30
Letter to Herodotus (50). Commentary in Bailey, p. 195.
31
Bailey, p. 197, writes "The vision of an object at a distance should always be
regarded as a problem awaiting the confirmation of the nearer view, by which, if
it is either confirmed or not contradicted, it is true." But if the nearer vision is
possible there is necessarily confirmation or falsification!
 ARE THE EPICUREAN 'CRITERIA' COMPATIBLE WITH INTUITIONISM? / 177

one believes in it because of the evident fact of motion; for if there is no

void, motion should not have existed either, if there is no place towards
which a body in motion can direct itself, since the whole would be full
and compact. So the (evident) phenomenon of motion does not falsify
the invisible thing which is conjectured".32 Non-falsification is there-
fore a relation between an hypothesis, unverifiable in itself because not
amenable to clear vision, and the clear vision of a phenomenon. Such
a relation can only be of a logical order, namely, one of implication.
Mrs. Rodis-Lewis goes on with the following commentary: "On the
contrary, the existence of motion falsifies the Stoic denial of the void. It
is in this way that use of reasoning by the absurd, which has been noted
to be at the base of atomism, is justified. Without the void, there is
no motion; without the principle 'nothing is generated from nothing',
we fall into the absurdity that everything comes from anything at all.
If the principle denied has but a single contradictory, the negation of
the negation amounts to confirmation".33 Even if there is a complete
parallelism in the definitions given of the true and the false, are we able
to say for all that that the existence of motion at once falsifies the Stoic
negation of the void while confirming the Epicurean affirmation of it? It
doesn't seem so. For if it were admitted that a double negation amounts
to an affirmation, the atomist thesis would be susceptible not only of
non-falsification in Sextus' sense, but of confirmation. Distinction must
be made then, in the definitions given of the true and the false, between
what applies to opinions expressed in observation statements and what
applies only to theoretical principles and therefore to hypotheses either
about invisible things or about things such as heavenly bodies which,
though visible, are beyond the reach of a 'clear view'.
The interpreters of Epicurus34 have discussed the obscure notion of
empoXr) ifj<; Siavoiat;, projection of the mind. It seems to correspond to
the notion of scientific hypothesis. But of that there is no direct confir-
mation if it is true, but only non-falsification through something sen-
sibly evident resulting from it by clearly understood principles. Take,
for example, the question of the speed of the atoms and the eidola in
the void. Epicurus says, "Next, no observation falsifies the theory that
the eidola (idols) are of maximum fineness: hence they have maximum
speed, since they all have a path big enough to ensure that nothing
32
<7. Dogm., 1, 213, quoted by Rodis-Lewis, 1975, p. 111.
33
Op.cit., p. 112. Likewise, Furley (quoted by Goldschmidt) and Goldschmidt
speaking of the inverse supposition (P, for if ~ P, then Q which is observed to be
false; therefore ~~ P, that is to say, P), 1977, p. 219, note 4.
34
Bailey, pp. 259-274.
178 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

collides, or few things collide...".35 Furley rightly draws a connection

between this passage and the Aristotelian doctrine. "Aristotle held
that the speed with which bodies move through a medium varies with
the weight of the bodies and the density of the medium. Void has no
density at all, so that the speed of motion through a void can have no
ratio to other speeds at all. Since this is impossible, Aristotle argues
that there is no void. Epicurus turns this argument upside-down...:
there is a void, so atoms move through it at 'inconceivable' speed.
They do not, of course, move at infinite speed; but their speed is such
that it cannot be related at all to the speed of observable motion".36
This is a remarkable argument. Let p be 'there is a void' and q be
'the speed of the motion of atoms through a void has no ratio to the
speed of observable motion'. Epicurus and Aristotle both admit the
thesis:
If p, then q.
Aristotle then gives a reductio argument:
If 'if p, then <?' is a thesis and it is not possible that q,
then it is not possible that p,
and from the impossibility of a speed having no ratio to observable
speeds, he deduces the impossibility of the void. This is precisely an
application of the Master Argument's second premise interpreted as
a thesis of pure modal logic. But both the premises, p and q, the
condition and its consequence, belong to the realm of invisible things.
For Epicurus then, both of them must be 'inferred' from phenomena
without its being possible for a reduction to the absurd to eliminate the
condition on the ground of some supposed internal impossibility that
would characterize the consequent. The consequent cannot be declared
impossible, nor possible either, simply on an examination of its nature.
For Epicurus there is something dialectical about Aristotle's rea-
soning, for the internal possibility of a concept such as that of incon-
ceivable speed is a dubious one. We must therefore come back to the
relation of the invisible to phenomena. Just what is it, on Aristotle's
account, that is given in sensation? Motion is. It is therefore on the
relation of the void to motion, and only on it, that it will be possible
to ground a subsequent inference as to the legitimacy or illegitimacy of
the concept of inconceivable speed, itself a consequence of the concept
of the void.

35
Letter to Herodotus (47), Furley Translation, p. 127.
36
Furley. p. 127.
 ARE THE EPICUREAN 'CRITERIA' COMPATIBLE WITH INTUITIONISM? / 179

Let there be the following principle then: no void (not-p), no motion

(not-r)—change of place supposing the existence of place and hence of
the void.37
If 'if not-p, then not-r' is a thesis and if r, then it is not the case
that not-p.
There is motion, therefore by modus tollens it is false that there is
no void. There is refutation or falsification of the denial of the void
by the phenomenon of motion. Therefore the Aristotelian hypothesis
of that denial is false. One might think that from there one would be
justified, according to Epicurus, to conclude that there is void. But
it would be a futile conclusion for, being based only on a refutation
by the absurd of a dialectical thesis, it would be subject to the same
uncertainty as that thesis is. To give grounding to the existence of the
void there must therefore be a positive argument of such a nature that
a sensation or a phenomenon could give support to the hypothesis.
Epicurus contents himself with saying that the hypothesis of the void
is not falsified by the phenomena. In other terms,
1) the modus ponens if p, then r and p therefore r is not employed
because, while the 'if not-p, then not-r' is a clear principle for the
£7i(.poAr| Tf)<; Siavotag, it is not at all the same for the affirmative condi-
tional 'if p, then r'. For it is not sufficient that there be a void for there
to be motion. We don't know exactly in what sense Epicurus inter-
preted his conditionals, but in the weakest sense of material implication
p might be true and r false (in a possible, contrary to factual, immo-
bile world). All one can say is that the existence of the void renders
motion possible. From the existence of motion as given, the necessity
of the existence of the void cannot be concluded; but the hypothesis
of the void is not falsified by the fact of motion. On the contrary,
once the void is supposed, the incommensurable speed of the atoms
follows (p D q), or rather, that speed is the necessary consequence of
the conjoint supposition of the void and of the motion of atoms.
2) It is not justified to pass from a double negation to an affirmation,
or, what comes down to the same thing, the excluded middle cannot
be applied to principles having to do with invisibles. The most that
can be attained is a relation of non-falsification. Posit the void, and
the phenomena do not falsify it, since motion, an evident phenomenon,
would not exist without it. And since the void entails inconceivable
37
Letter to Herodotus (40). "If there were not what we call void, space and in-
tangible nature, bodies would have nowhere to be nor anywhere to move, as they
do indeed appear to move". This is of course not a reduction to the absurd of
Aristotle, since he explicitly rejected this implication.
180 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

speed, this last hypothesis, just as its condition, should be regarded as

not falsified.
So in the definition of true and false two aspects, or even two inde-
pendent elements will be distinguished according to whether we have
to do with opinions consisting in anticipations of perception or, on the
contrary, in theoretical statements. But this distinction requires at
the same time a clarification of the relations obtaining among the four
terms: confirmation, non-confirmation, falsification, non-falsification,
and between the terms of this set and the two terms: true and false.
An observation statement is true if it is confirmed. For it to be false
it suffices that it be non-confirmed. When I remark that the tower is
not square, the opinion that I had had, 'the tower is square', is seen
to be false. Here, non-confirmation amounts to falsification. There is
no room for a margin of uncertainty. By contrast, when the question
is one of a mental projection about sensible but inaccessible things,
or invisible things, there will be neither confirmation nor direct con-
firmability. The true will be reduced to non-falsification. Falsification
of the hypothesis remains possible, however, through the phenomenal
consequences that can be attached to it. Theoretical non-falsification
must be regarded as a kind of verification, as the passage quoted from
Sextus proves it. As the concepts and principles they deal with are
not amenable to direct comparison with experience, not being prop-
erly empirical concepts, the decision about them is made in connection
with the entire set of empirical consequences associated with them by
virtue of 'projections of thought'.
Consequently, whereas truth and falsity are subject to a homoge-
neous and symmetric criterion in the case of opinion about the 'near'
sensible, and there is conformity with the canons of empiricism with
true meaning 'verifiable' and false 'verifiable-that-not', that is, con-
firmed and falsified, it is a different story for theoretical hypotheses.
These may be falsified although they can never be confirmed. Epi-
curus then had an inkling of the asymmetry between falsification and
confirmation in this domain. His Canonic announces Popper. If it
doesn't reach exactly the same conclusions as Popper, it is that for
Popper non-falsification does not amount to truth, and further, every
protocol observation statement is subject to being ulteriorly corrected.
This last aspect is understandable in the light of Popper's polemics
against psychologism and the conventionalist interpretation of 'basic
statements' that is rightly and resolutely rejected by the empiricist
canon.38. As for the first, Epicurean intuitionism tends to efface the
38
Popper, 1973. (There is a note, p. 103(1), to this often obscure text that
 ARE THE EPICUREAN 'CRITERIA' COMPATIBLE WITH INTUITIONISM? / 181

difference between the truth it attributes its hypotheses and the ab-
sence of positive truth proper to hypotheses according to Popper. On
the one hand, any perceptive anticipation is in itself simply a problem
awaiting confirmation or falsification by direct and 'near' experience.
As such it is therefore without truth-value, since it is only experience
that will confer that upon it. An hypothesis, on the other hand, is ver-
ified or falsified by the phenomena following from it. There is no truth
or falsity apart from these procedures of confirmation. Moreover, the
verification of an hypothesis being limited to non-falsification is asym-
metric to falsification, and strictly speaking, given the invalidity of the
excluded middle, it is not absolutely positive nor does it in any way
authorize the conclusion to truth in the dogmatic sense of adequation
to the object.
The Epicurean theory of law gives a remarkable illustration of the
precepts of the Canonic. For in the procedures of verification and fal-
sification of laws, more complex than those operative on the level of
sensible opinion since they can deal only with the far-off consequences
of the laws, the method of anticipation has to do no longer with simu-
lacra passively received but with effects "whose cause we ourselves have
posited" ,39 There will be two distinct cases. Either confirmation will
be inherited from the past as with a body of laws left by the ancestors.
In this case the law brings its confirmation (hence, non-falsification)
with it and can fulfill its essential function of establishing security, just
as confirmation and non-falsification establish security in the order of
knowledge.40 Or again, there may be a proposed law of which the
confirmation or non-confirmation is 'in suspense'. Goldschmidt has
likened this to the Attic procedure of 'revision of laws' and the action
taken against the author of a detrimental law.41 In this last case the
law is like a hypothesis about the future. If there is often confirmation
and hence justification of a rule by its present conformity to the com-
mon interest, or non-confirmation and rejection of its legality because
its consequences do not conform to the interest of the community, the
immediate confirmation assimilating the juridical criterion to the cri-
terion of opinion about 'near' sensible things, it also happens that law
changes as a result of constitutional upheavals. What was just ceases
seems to lessen this first difference.) In the other direction, as Goldschmidt aptly
points out (1977, p. 217), Epicurean sensation differs from Humean impression
by the "fundamentum in re accorded it in Epicurean realism by the theory of
simulacra".
39
Goldschmidt, 1977, p. 217, pp. 221-222.
40
Id., ibid., pp. 218-219.
41
Id., ibid., pp. 196-197, 222-223 ; Goldschmidt, 1981, p. 87.
182 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

to be useful. In this case it will have to be admitted that the conse-

quences of a law that was accepted and obeyed for some length of time
did not really confirm that law, but simply did not falsify it temporar-
ily. When untoward consequences arise and are more and more clearly
recognized as such, they falsify the law and oblige its revision. In other
words, from the non-falsification that justified the law up to that time,
we go to the falsification which takes that justification away.42
A proposed law conceived of as a juridical hypothesis is different
from a scientific hypothesis about invisible or out-of-reach things, be-
cause the future will assimilate it one day to an opinion about prox-
imate sensible things. That is why the criterion of what is just will
change, as required by positive law. Let us note one last analogy be-
tween legal theory and the Canonic. Suspension of the positive legal
order or anomia inevitably produces a state of insecurity where ev-
eryone harms everyone else. Security, which consists simply in not
mutually harming one another, therefore supposes a state of law. The
non-falsification of an hypothesis would produce its truth not in the
strong, dogmatic sense of correspondence, but in the weak sense of a
substitute for verification for that which cannot be verified. Likewise,
the non-falsification of a rule produces its justification not in the sense
of an ideal justice, but in the weak sense of what plays that role in
positive law: namely, peace in the city.

7.4 Consequences of the Epicurean criteria:

Plurality of hypotheses and rejection of the
excluded middle.
Two consequences follow from this canon: the indeterminacy of certain
hypotheses and the rejection of the excluded middle.
In the first place, to be sure, Epicurus admits that both for search-
ing out the hidden causes of phenomena, the invisible (aSr)Xov), includ-
ing both the atoms and the things of the heavens and meteors, and to
reach the future (upoajasvov), opinion must inevitably go beyond the
bounds of perception. But the truth-criterion of these mechanisms
42
Epicurus, Principal Doctrines, XXXVII and XXXVIII, Goldschmidt, 1977,
pp. 280-282. Epicurean legal theory could be compared to that of today's neo-
positivists. If the annulability of a legal norm is not generally the same as its being
void ab initio, and if specific legal bodies are required to decide of it, it is that a
norm may be subject to revision only in the name of another norm (Kelsen, 1945,
XI He). But in order that an explicit contradiction does not upset the edifice of law,
the older norm must be regarded precisely as a non-invalidated hypothesis that was
active as long as it was accepted and whose annulation coincides with the formal
act of invalidation.
 CONSEQUENCES OF THE EPICUREAN CRITERIA / 183

of projection is still perception itself. Hypotheses, whether about the

imperceptible reasons for things or about inaccessible or future things,
have their criteria in perception alone in so far as it confirms or at least
does not falsify them. In the case of explanatory hypotheses, it can
but not falsify them, in the sense we have given to this term. But then
we are in no way obliged to choose between incompatible hypotheses
meant to explain one and the same natural phenomenon.43 Epicurus
condemns those "who for not having clung to the method of the pos-
sible have fallen into that (impious) vanity, because they think that
the phenomena come about only in one way and reject all the other
ways in harmony with the possible, tending towards the unintelligible
and incapable of bringing together in one theory the phenomena that
must be accepted as signs".44 Is that reason enough however to make
a skeptic out of Epicurus and to say that "the Epicureans shouldn't
have regarded their own atomist metaphysics as any more than an
hypothesis, not refuted but not confirmed either by the facts—an hy-
pothesis which only served them moreover to repel other hypotheses
that seemed morally dubious to them"? 45
Are we to completely subordinate physics to ethics and judge that
"Epicurus recognizes that an indefinite number of different explana-
tions could be as true as materialism", the 'truth' of this last residing
in its pragmatic consequences, that is to say in its capacity to "more
completely liberate man from his fears"?46 That would be a com-
plete misinterpretation. First of all, it is within the unique, categorical
framework of atomism that these multiple explanations arise.47 Then
too, precisely in the case of celestial phenomena, Epicurus was seen
to require unification in a single theory of the diverse explanations,
that is, to absolutely reject the dualism some of his adversaries held
there to be between supra- and sublunary worlds,48 and the astrolog-
ical consequences they derived from this representation.49 One could
in no way then invoke the plurality of certain hypotheses to infer that
Epicurus had a skeptical attitude. The eminently peremptory thesis
of the non-divinity of celestial things is fundamental in his system.
Skepticism attracted some of Democritus' disciples, like Nausiphanes,
43
Letter to Herodotus (79-80) in Bailey, p. 51; Sextus Empiricus, VII, 211.
44
Letter to Pythodes (97), translated by Rodis-Lewis (pp. 59-60); Lucretius, De
Rerum Natura, V, 526-533.
45
Windelband-Heimsoeth, 1935, p. 171.
46
Sartre, 1949, p. 191.
47
Boyance, 1953, pp. 428-429.
48
Letter to Herodotus (77), in Bailey, p. 49.
^Principal Doctrines (XI), in Bailey, p. 97.
184 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

because Democritus depreciated sense knowledge. Epicurus, by con-

trast, reestablishes the credentials of that knowledge and, through it,
of phenomena,50 the only things proper to providing a criterion. What
is more, when Epicurus speaks of the plurality of hypotheses, he often
seems to be defending simply the plurality of causes. In particular,
that is the case in the Letter to Pythocles (98), when he looks for the
explanation of the inequality of days and nights.51 The only conclusion
that can be drawn then from the plurality of certain hypotheses has
to do with logic. It is experience, never the logical organization of the
premises, that lets us decide between hypotheses. Logical reasoning is
empty when it assigns things essences such that immateriality, eternity
or necessity should derive logically from them.
This negative conclusion entails in the second place, another con-
cerning the very principles of logic. It is its non-falsification, that is
to say, the confirmation of its experimental consequences that assures
us of the truth of atomism. It is not simply the fact that the contrary
hypothesis is false. That is why the excluded middle and the law of
double negation are themselves powerless to determine the truth of an
hypothesis, whatever its nature. But now take the case of the future.
How could it be said that a prophecy about it is true or false, if divina-
tion is ridiculous, as Democritus had already shown? Since there is no
confirmation or falsification to be had before the awaited event itself,
the excluded middle could have no application to the future.
It has been said that Democritus' system suffered in having been
transmitted through the system of Epicurus which subordinated theory
to practice and introduced the metaphysical concept of freedom into
philosophy.52 And indeed this concept of freedom of indifference or
balance of will, that was the admiration of a Marcus Aurelius,53 is
the corner-stone of the philosophy of Epicurus. But that freedom is
in the first place the freedom to refuse the solicitations of opinion,
the representation of future evils for example, in order to accept only
the present, i.e. sensation, cut off from the active movement of error.
50
Rodis-Lewis, 1975, p. 65; Bailey, p. 404.
51
Bailey, p. 67, p. 277 and p. 251.
52
Windelband-Heimsoeth, ibid., p. 162.
53
Thoughts, Bk. IX(41); Brehier, 1962, p. 1220. "Whoever uses the Democritean
argumentation, saying that there is no free motion of the atoms, because of their
respective impacts, and that it seems from that that everything moves according
to necessity, we shall say to him: Don't you know, whoever you are (Diogenes
addressing the passers-by of Oenoanda), that the atoms too have a free motion,
not discovered by Democritus but brought to light by Epicurus; it is the existence
of a swerve, as he shows from phenomena." (fr. 32, in Rodis-Lewis, p. 79); Lucretius,
De Rerum Natura, II, 252-293.
 EPICUREANISM AND THE MASTER ARGUMENT / 185

The connection with the denial of the principle of the excluded middle
becomes apparent once we consider the weight of representations or
motives in voluntary decision making. Both adversaries and partisans
of the freedom of indifference agree in denying that a motive can be
at once sufficient and insufficient for entailing decision. But it is not
for all that that the partisans of will admit that a motive is either
sufficient or insufficient for entailing decision. For it is the indifference
of freedom that bestows upon the motive the power of determination in
which the representation in itself, no matter what its weight, is lacking;
and that is why it isn't true that such a motive in itself either is or is
not sufficient for entailing decision.54

7.5 Epicureanism and the Master Argument.

What then is Epicurus' possible relation to the Master Argument?
And why, as Cicero has it, did he invoke the invalidity of the excluded
middle in order to avoid the necessity he thought resulted from that
argument? Diodorus' influence on Epicurus has been played down as
far as the theory of elementary motion is concerned.55 The influence
of the Master Argument on the other hand is great: the Ciceronian
testimony is unimpeachable. We know too that Epicurus wrote a trea-
tise Against the Megarians and another On Fate.56 In his Letter to
Menoeceus57 (133-134) he protests against the fate of the philosophers,
judging it to be worse than what the common people believe in. Then
too, he explicitly admits Diodorus' first premise.58 In so far as the
third premise is concerned, it has been noted in connection with Aris-
totle's potential infinity that Epicurus "would be impatient with the
idea of a potentiality which never can be actualized" .59
At the same time, we should expect the same impatience with a
Diodorean potentiality that is actualized at a time as far removed in
54
It is this indifference of equilibrium that Leibniz refuses above all in Epicurus
(Gerhardt, IV, p. 297; Jalabert, pp. 310-311). In short, t/there is an uncaused phys-
ical or psychological movement, there are exceptions to the principle of the excluded
middle. According to Carneades, both Epicurus and his adversary Chrysippus agree
in accepting that proposition (Cicero, De Fato, X (20-21); see below, 8.1, p. 206).
Let A be the proposition 'it is possible that the motive p should not be accompanied
by the decision q' (case of insufficient motive) and B the proposition 'it is impossible
that the motive p should not be accompanied by the decision q'. What the parti-
sans of free will refuse is the validity of the conditional ' ~ (A • B) D (~ AV ~ B)1.
(See above, note 22).
55
Furley, p. 134.
56
Vita Epicuri (27-28) in Bailey, p. 159.
"Bailey, p. 91.
58
Fragments (LV) in Bailey, p. 115.
59
Furley, p. 155.
186 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

the future as one might wish. If we take into consideration the infinity
of time, we will not be able then, according to the Canonic, either to
affirm or deny the third premise. Moreover, the overall accommodation
made in Epicurean thought for the free play of what is the aleatory
would hardly allow discarding the third premise, whose truth is an
experiential given for free will. As for the second premise taken as a
thesis of pure modal logic, it is employed by Epicurus in the Letter to
Herodotus (57), for example. He starts from the thesis (5fjXov) "if there
is an infinity of parts, however small they may be, having whatever size
they do, the body composed of them will itself be infinite in size".60
The consequence is impossible: "how then could a size like that be
limited?" Therefore the antecedent is impossible as well.61 We have
reason to presume then that Epicurus escapes the Master Argument's
necessitarian consequence by indirectly challenging the dogmatic scope
of the third premise, and that to this end, he denied the validity of the
excluded middle, given the overall intuitionistic style and the exigencies
of his canonic.
When serious attention is payed to the fundamental innovations
Epicurus made in Democritean atomism, it is difficult to take Epicure-
anism as a kind of model of dogmatic empiricism as Kant does in the
Antithetic of the Critique of Pure Reason. The characteristics of the
Kantian antitheses: infinity of the world in space and time, inexistence
of a simple substance, denial of freedom and total subordination of the
world to the laws of nature, denial of the existence of an absolutely
necessary being are proper to Democritus, though it is difficult to rec-
oncile atomism with the second.62 Epicurus tempers all but the last
of these theses. To the eternity of the atoms he adds the theme of the
plurality of worlds that come to be and pass away and have therefore
beginning and end.63 If the soul is a perishable composite, it has, for a
certain time at least, a sort of formal permanence that fits in well with
its independence, parallel to that of the atoms. Finally, the swerve and
freedom break the chain of natural causes. In short, with the excep-
60 "|?or it is clear that the unlimited number of particles have a determinate size
and, whatever the size of these components, the magnitude [of the body] would be
unlimited as well." (Bollack, Bollack, Wisemann, 1971, p. 113).
61
There is no contradiction between the existence of an infinity of atoms and the
fact that an infinite number of atoms cannot enter into the makeup of a limited
body. For as the Letter to Herodotus (40) says, "the whole is unlimited in extension
and in the number of bodies"; and it goes on to add (45b), "the number of worlds,
some like and some unlike ours, is equally infinite".
62
See the difference Kant makes between monadism and atomism, Critique of Pure
Reason (B478/A460).
63
Lucretius, De Rerum Natura, II, 1047-1089.
 INTUITIONIST CONCEPTIONS OF REALITY: DESCARTES AND KANT / 187

tion of the existence of the necessary being,64 one would put Epicurus
rather on the side of the Kantian theses than the antitheses, that is to
say, precisely on the side of practical interest as over and against the
speculative interest of reason.65
Nevertheless, while placing Platonism (the thesis) decidedly among
the dogmatic systems, Kant hesitates about Epicureanism. In an im-
portant note he says: "There is still doubt, however, whether Epicurus
ever propounded these principles as objective affirmations. If they were
perhaps nothing more than maxims for the speculative exercise of rea-
son, he would show a more genuinely philosophical spirit in that than
any other of the sages of antiquity. That, in the explanation of phe-
nomena we must proceed as if the field of inquiry was not bounded
by limits nor a beginning of the world; that the material of the world
must be accepted by experience, if we are to hope to learn from it; that
we must determine no production of events, if not by the immutable
laws of nature; and finally, that we should have no recourse to a cause
distinct from the world—those are even today very good, though little
followed, principles for extending speculative philosophy and for dis-
covering the principles of morals, without an appeal to outside help.
Nor for that reason can anyone desirous of ignoring these principles,
as long as mere speculation is involved, be accused of wanting to deny
them".66 Had Kant taken into account the Epicurean innovations in
the domain of physics with the swerve as well as in that of the canonic
with his rejection of the excluded middle and of morals with the as-
sertion of free will, he would no doubt have given more weight to this
remark and would not have hesitated to see in Epicureanism a sort of
precritical philosophy.

7.6 Other intuitionist conceptions of reality:

Descartes and Kant.
If Kant is right in noticing an affinity between Epicureanism and criti-
cal philosophy, it is still possible to generalize his remark. For since this
affinity results from their common intuitionist tendencies, one might
well wonder what philosophical traits in general are delineated by such
shared tendencies. To find an answer it would be best to compare sev-
eral philosophical systems differing both in their particular choices and
architectonics, but all complying with these same intuitionist tenden-
64
Which the Epicureans were the only philosophers of antiquity to deny as the
prime mover (B478/A450).
65
B494/A466.
66
B499/A471.
188 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

cies. We shall thus compare Epicurus, Descartes and Kant, examining

their systems successively from the point of view of method, of anthro-
pology and of theodicy.
The dogmatic systems accept the Aristotelian adequation of the
affirmation to the thing as their criterion of truth. But this adequation
is extraneous to the ways by which it is established. In particular, an
entity's existence is independent of the mode of its demonstration. So
the irrationality of TT could be asserted by apagogical demonstrations
reducing the contrary supposition to the absurd, long before Hermite
was able to give a constructive demonstration of it. Dogmatism is
insensitive to this distinction: TT was all the same irrational in itself
before any demonstration, direct or indirect. The difference disappears
because the logical means employed are posited as being genuinely
equivalent, since this equivalence is taken either to be given a priori,
to be founded on the pragmatic success that collective experience has
assured equally to these means, or to be as it were programmed in
the laws of the language we all use for communicating our experience.
It is postulated, therefore, that the laws of ordinary logic are valid
for all possible worlds and that the choice of one or another mode of
inference from among them is simply a question of will. The principle of
the excluded middle and that of the equivalence of double negation to
affirmation are among those laws that the order of things, the success of
human action, or linguistic habits have selected as being prerequisites
of any possible knowledge.
This is the postulate intuitionism challenges in its refusal to ex-
empt the laws of logic from the inquiry it insists on into the crite-
rion of truth. The adequation it requires, therefore, is no longer that
between the thing and its true representation, but that between the
representation and the canon proper to guaranteeing its truth. For
Epicurus this canon is confirmation through perception of all that goes
beyond perception in opinion, and consequently, rejection of all that is
not confirmed or at least of all that is not 'non-falsified' in the sense
Sextus gives to this term. For Descartes an idea is true if it is clear
and distinct, or short of that, as happens in the case of the experience
of compound substance, if it does not overstep what can be necessarily
concluded from truths demonstrated on the basis of clear and distinct
ideas. For Kant a concept has content if it is constructible in sensible
intuition, and is without objective validity, if not without regulative
utility, once it oversteps the bounds of possible experience. In all these
cases then, opposed to the dogmatic notion of adequation is a nar-
rower concept of truth based on a specific test. Granted, the test is
very differently conceived of by each of the three; but for each of them
 INTUITIONIST CONCEPTIONS OF REALITY: DESCARTES AND KANT / 189

though, it is independent of logic and even a precondition of it. This

is why Epicurus confines what must be kept of dialectic to within the
narrow limits of the canonic and challenges the excluded middle and
the necessity of choosing between exclusive hypotheses forming a com-
plete disjunction. Descartes scorns the logic of the Ancients. True, he
replaces it by a geometric algebra, but the theory of proportions gov-
erning that algebra excludes expressions requiring an infinite number
of algebraic operations. Trigonometry and calculus are thus banished
from the method; 67 the continuum remains beyond the reach of rea-
son, the object of simple mechanical processes, for the only clear and
distinct idea we have of the infinite is the intellectual one of order, not
the imaginative one of a collection. Kant expressly limits the extension
of concepts to finite sets. If he does admit the validity of the excluded
middle and of apagogical reasoning in mathematics, it is that he thinks
that the empirical character of the intuition in which such reasoning
necessarily finds its application guarantees a priori the possibility of a
constructive demonstration that will sooner or later produce the evi-
dence that was lacking and replace the reduction to the absurd. When
dealing, on the contrary, with a pure concept of the understanding
to which there corresponds no empirical intuition, the concept of the
world for example, apagogical reasoning that would assign or refuse
to assign it an object is inadmissible, and simply goes to show that
some of reason's pretentions are illegitimate in themselves, even if the
concerns behind them are not without ground. Consequently, from the
fact that the assumption that the world is infinite or that it is finite
leads to a contradiction, we have no right to conclude to the necessity
of choosing. This would only arise in the case of an analytic opposition
where one and the same existent subject, or one constructible in intu-
ition, is assigned two contradictory properties.68 In the cosmological
antithetic on the other hand, reason is enveloped in a sophism stem-
ming from the fact that in the major premise (when the conditioned
is given the entire series of all its conditions is also given) the word
condition is taken in another sense than in the minor premise (where
the objects of sense are given as conditioned). In the minor premise
the word is given an empirical sense, in the major a transcendental
one. But the empirical sense requires construction of the synthesis
of the conditions regressively in time, so to suppose this achieved is
67
Vuillemin, 1960, pp. 9-73.
68
Such is the case for the opposition: IT is a rational number, TT is not a rational
number, before Lambert proved the truth of the second proposition (by the absurd).
No one would have concluded that they both were false for want of the existence
of the subject IT.
190 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

contradictory. The transcendental sense, by contrast, requires posit-

ing the totality of the conditions as realized and independent of any
succession. The confusion results from hypostatization of the subject:
the world conceived of as the totality of the conditions of phenomena.
There is dialectical opposition precisely when contradictory properties,
such as finite and infinite, are attributed to a subject without reality:
the world in itself. Kant does not explicitly challenge either the ex-
cluded middle or the equivalence of double negation and affirmation.
But why does the given totality—in itself—of the conditions of a phe-
nomenon fail to define a possible subject of affirmation, when it denies
the conditions of the infinite regressive synthesis it implies, and ille-
gitimately frees itself from the bondage of intuition? Why does Kant
confine not only the continuous but even the infinite to the province
of sensible intuition, which can and must be given (gegeben) but could
never be delimited (abgegrenzt)! It is that the conceptual synthesis of
the 'I think', far from being creative, presupposes a manifold extrane-
ous to the understanding and proper to the sensibility, even when that
sensibility is taken as pure. A set theory that would countenance a
given infinite number, and o fortiori cardinal numbers of higher infin-
ity, would run afoul of the same criticism as that made of the world,
since it supposes actualized the totality of conditions given only by the
progression of numerical synthesis. But there is more. Not only does
the entirely negative Kantian dialectic manifest an affinity with the
results of the intuitionist criticism of classical mathematics, but it is
seen to be indistinguishable from them, once we examine the nature
of the transcendental sense given to the idea of the totality of con-
ditions. For how is it known that this sense is empty and provides
only an imaginary object? The answer given is that it is because it
corresponds to no intuitive given. But how do we make sure of this?
Consider two entire series of all the conditions, one of phenomenon B,
today's sunrise, the other of phenomenon A included in the first series,
yesterday's sunrise. The second series is a proper part of the first and
it can be shown that there is a one-to-one correspondence between the
elements of the two series, which shows that the part is as 'great' as the
whole. In short, when concepts of proven utility within the province of
sensible intuition are applied beyond the limits of that province, con-
cepts such as part, whole, equivalence lose their primitive meaning. It
is this experiment, already mentioned by Galileo, that makes us leery
of the objective validity of the concept of an entire series of conditions.
There is however something ambiguous about the experiment. It may
indicate that there is no possible object corresponding to the entire
series of conditions. But then again, it may suggest that certain prop-
 INTUITIONIST CONCEPTIONS OF REALITY: DESCARTES AND KANT / 191

erties of concepts must be modified when the field of their application is

switched from conditions taken in the series to the series in its entirety.
How is one to decide between these two interpretations? Kant
clearly opts for the first. No positive property can be assigned to
the entire series of conditions. The first critical answer to the ques-
tion of the world's magnitude is thus negative: "the world has no first
beginning in time and no extreme limit in space".69 The dogmatic
antithesis concluded that the world is infinite as regards both time and
space. This affirmative answer, which was dialectical, is now to be
replaced by an affirmation that is acceptable because it has to do with
the constitution of the subjective synthesis of the world and not with
the world as a thing in itself. "The regression in the series of worldly
phenomena as determining the magnitude of the world continues in
indefinitum" .70 What distinguishes the world in itself from the empir-
ical world then, is that the former admits of the illusory attribution of
absolute magnitude whereas the latter does not. But how do we get
from the empirical world to the world in itself? The empirical world
is encompassed within the process of the subjective synthesis whose
incompleteness is one of the givens of experience. The thing is itself is
posited as the total series of the synthesis. But the only difference that
constitutes the act of dogmatic projection lies in the application of the
excluded middle. It is recognized that the empirical world does not
have the property P (that of having limits). Nevertheless one does not
assume the right to conclude that the world has the property non-P,
which, in attributing infinity to it, would amount to taking it dog-
matically as a thing in itself. The property non-P holds only for the
subjective series and is expressed in the experience we have of being
able to continue the series indefinitely. The subjective series falls un-
der the excluded middle; it is finite or infinite, and we conclude by the
absurd that is is infinite. 71 Application of the excluded middle outside
the subjective synthesis results in hypostatizing the thing in itself. To
maintain that the empirical world is not a thing in itself is in fact to
reject the universal validity of the disjunction of finite and infinite, i.e.
its validity independent of the conditions of intuition and construction.
But this just is the intuitionist principle.72
69
B548/A520. This answer differs from the first proposition of the antithesis only
by the adjectives erster Anfang and aiisserste Grenze, marking that the dogmatic
confusion is avoided and that the world is taken, both in the major and minor
premise, as a thing in itself.
70
The same remark would apply in the case of the second antinomy, on the one
condition of substituting the words in infinitum for in indefinitum (B551/A523).
71
Indefinite for the first cosmological question, infinite for the second.
72
In the letter to Chanut of June 6, 1647, Descartes adopts a similar position.
192 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

From its required subordination of the logical principles to the di-

rect and positive test of fact results the inevitable introduction of some
skeptical traits into the intuitionist method. Epicurean freedom often
differs little from suspension of judgment. 73 This is the case when-
ever opinion goes beyond the ken of sense. Cartesian doubt becomes
metaphysical and forges fictions going beyond the natural errors and
illusions, in order to arrive at an unshakable certainty. Kant praises
the skeptical method,74 which is in fact one with transcendental ideal-
ism. All of these skeptical traits in effect assign limits to the domain
of necessity. Epicurus admits that what is true is necessary, but there
are only truths of fact; whatever does not fall under sensation imposes
no necessity either. Descartes distinguishes the uncreated first truths,
all having to do with the divine attributes and the consequences fol-
lowing from the idea of the infinite being, from eternal truths which,
though they force their necessity upon us, were nevertheless instituted
by divine free will and therefore have a necessity which is not primi-
tive or intrinsic but derives from the constancy of the divine decrees.75
What is necessary for our understanding is in no way necessary for
God himself.76 It is remarkable that Descartes considers as one of
the absolute impossibilities that God should be able to make what is
or what has been not be.77 The Master Argument's second premise
taken as a thesis of pure modal logic seems to fall in with the cre-
Only God is infinite, he says; the world is but indefinite. "In which there's a very
remarkable difference: for to say that something is infinite one must have some
reason that makes it known as such, which one can only have in the case of God;
but to say that a thing is indefinite, it suffices to have no reason by which it could be
proved that it has limits." The remainder of Descartes' reasoning there is altogether
comparable to Kant's (B548/A520).
73
See Gassendi, Disquisitio Metaphysica, p. 68.
74
B535/A507.
75
Gueroult, 1953, II, p. 30.
76
"As for the difficulty of conceiving how it was free and indifferent for God to
make it not have been true that the three angles of triangle should be equal to
two right angles or, generally, that contradictories could not be together, it can
be easily removed by considering that God's power can have no limits, and then
further, by considering that our mind is finite and created in such a way as to be
able to conceive possible those things that God willed to be truly possible, but not
in such a way as to be able to so conceive what God could have made possible
yet nevertheless willed to make impossible. For the first consideration lets us know
that God can not have been constrained to make it true that contradictories cannot
be together, and that He could therefore have done the contrary, while the other
assures us that, even though this be true, we should in no way try to understand
it, as our nature is simply incapable of that". (Letter to Mesland, May 1644, IV
p. 118).
"Letter to Morus, Feb. 5, 1649, V, p. 273: reference to St. Anselm's Proslogion
(VII) given by Gueroult, 1953, II, p. 29.
 INTUITIONIST CONCEPTIONS OF REALITY: DESCARTES AND KANT / 193

ated eternal truths. As for the third, it seems to result directly from
the existence of our freedom. Furthermore, these truths for our un-
derstanding are not unfounded once that, given the demonstration of
God's existence, His veracity is engaged in the clear and distinct idea
we have of necessity.78 Two sorts of possibility79 and necessity are
thus distinguished. Assuredly the distinction is reminiscent of Aristo-
tle and foreshadows Leibniz.80 But Leibniz found the creation of good
and evil,81 and of the eternal truths generally, so shocking that he sus-
pected there was some 'philosophical trick'82 behind the idea: what
was traditionally attributed to the divine understanding was made out
to be an object of the divine will, endowed with freedom. The two sorts
of modalities distinguished by Descartes are altogether different from
the dogmatic opposition between the order of essence and the order of
existence or between the principle of contradiction and the principle
of the best. Descartes writes to Mersenne83 that since some men "un-
derstand the truths of mathematics but not that of the existence of
God, it is not surprising if they fail to believe that the former depend
on the latter. On the other hand, they should judge that, since God
is a cause whose power surpasses the limits of human understanding,
and since the necessity of these truths does not outstrip our knowl-
edge, these truths must then be something lesser and subject to that
incomprehensible power".
So Descartes' absolute or uncreated necessities are not those of
essences, objects of the understanding, precisely because these essences
are created. They have to do only with God's truth, which lies beyond
our understanding because of His immensity that we cannot encom-
pass. As for created necessities, though on the same scale as our un-
derstanding, there is nonetheless something about them that shields
78
Gueroult, 1953, II, p. 34.
79
Gueroult, 1953, II, p. 39. "God in His omnipotence can do everything in
principle—even what we judge to be positively impossible—as long as it is not
repugnant to His very omnipotence, that is to say, as long as it is not a question
of absolute impossibility. But having instituted as eternal truth whatever our un-
derstanding perceives of as positive impossibilities, being immutable and truthful,
God will do nothing in the universe that we judge excluded by them. If He should
happen to do so however, He will call it to our attention in an indubitable way.
Moreover, He can always do what we fail to understand the possibility of, in the
case where we haven't a sufficient reason to judge it necessarily as positively impos-
sible; there is nothing, on the other hand, that could make Him unable to realize
whatever our understanding conceives of as possible".
80
Gueroult, 1953, II, p. 39.
81
Leibniz, Gerhardt, VI, p. 219; Jalabert, p. 230.
82
Leibniz, Gerhardt, VI, p. 227; Jalabert, p. 239.
83
Letter of May 6; Correspondence, A.T., I, p. 150.
194 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

their modality from the penetration of reason. And that is precisely

the arbitrary and dependent state that is theirs as being created. For
to maintain the gulf separating the infinite from us it must be that
what is de jure for us is in some inscrutable way simply a matter of
fact for God. These necessities therefore remain founded on the expe-
rience of a given that imposes itself upon us; and the claims of several
such givens may tangle themselves in such a knot that we are unable
either to unravel or cut it. This is precisely what happens in the case
of freedom and providence,84 a difficulty that seemed strange to Leib-
niz for whom the modalities, whichever they may be, are dogmatically
bound up with being and are fathomable by reason. "Could he have
been unaware," he asked, "that it is not possible for there to be an
invincible objection to the truth? since such an objection could only
be a necessary logical sequence of other truths whose result would be
contrary to the truth one holds, and consequently there would be a
contradiction among truths, which is of the utmost absurdity".85
Kant explicitly reduces the modalities to the connections of the
cognitive faculties with the formal conditions of experience, depriving
them of any synthetic force as regards the object.86 There is no differ-
ence of content between a hundred real thalers and a hundred possible
ones. This conception, that rules out the ontological proof, does not
fall under the classical disjunction of words and things. Kant takes the
modalities neither in a de dicto nor in a de rei sense, but de cogita-
tione, so to speak. Such is the logical consequence of intuitionism. If
the object of knowledge 'revolves' about the knowing subject, modality,
divorced from the object, does no more than describe the connection
of object to subject. Necessity therefore is not distinct from what con-
stitutes the complete system of the possibility of experience: that is
to say, the phenomenon. It makes no sense when supposedly applied
to things in themselves. There is no longer the opposition of absolute
and created necessity. Only the latter subsists in the form of tran-
scendental idealism. It is one with the act of cognition, tied as that is
to phenomena, and gives rise only to illusion when supposed to have
bearing on things.
84
"However the power and knowledge of God should not keep us from believing
that we have a free will, for it would be wrong for us to doubt of what we perceive
interiorly and know through experience to be in us because we do not understand
something else which we know to be incomprehensible by its nature." (Principes,
A.T., I, vol. 9, art. 41). In opposition to the absolute necessity of divine preordina-
tion is the created necessity, irresistible given the interior feeling, and guaranteed
by divine veracity, of my freedom.
85
Leibniz, Gerhardt, VI, p. 89; Jalabert, p. 94.
86
B296/A233.
 INTUITIONIST CONCEPTIONS OF REALITY: DESCARTES AND KANT / 195

The inseparability of the modalities from the experiential givens

in general, which is characteristic of the intuitionistic method, makes
the importance of psychology or at least of anthropology in such a
framework immediately clear. This is manifest in the account of error
and in the experience of freedom.
Since the criterion of truth is not adequation, but confirmation as
witnessed to by the well-ordered relation among the faculties, error
is to be accounted for as an intrinsic derangement of this relation.
Excessive expectation above and beyond the sense image for Epicurus,
prejudice or precipitation of the free will making judgments that go
beyond the passively received ideas of the understanding for Descartes,
illusion of reason overstepping the limits imposed by sensibility on the
understanding as it constructs its objects for Kant, the mechanism
of error is different in each of the three cases. It depends in each of
them however on the excessiveness of some faculty's emancipation and
disregard for the servitudes imposed upon it by the human condition:
heed sensation, admit only of the clear and distinct as true, admit of
objects only within the limits of possible experience.
But that faculty's overstepping its bounds is nothing other than an
expression—albeit a vicious one—of our freedom. Epicurus is said to
have introduced the transcendental concept of freedom into philosophy.
The intuitionist philosophers are the only ones to have retained it.
Leibniz made the remark: "It is funny that a man like Epicurus,
who had set aside the gods and all incorporeal things, should suppose
that the will, which he himself makes out to be composed of atoms,
could have empire over the atoms and deflect them from their path
without it's being possible to say how".87 But he adds that this man-
ifest absurdity is preferable to the subtlety of Carneades that muddles
things up in situating the freedom of indifference in the soul88 where
it seems to be more at home. According to Descartes it is true that
the freedom of indifference is but the lowest degree of freedom, still de-
prived of clarity, yet essential to the possibility of merit and demerit,
and human action is "entirely free and indeterminate" ,89 Leibniz said
that indetermination spoiled freedom;90 but Descartes founded free-
87
Leibniz, Gerhardt, VI, pp. 307-308; Jalabert, p. 321. This ascendancy of the soul
can be compared with the power of the horseman to alter the direction of move-
ment. Leibniz will object that the direction too is preserved (Leibniz, Gerhardt,
VI, pp. 135-136; Jalabert, p. 143).
ss
lbid., see also Gassendi, p. 452, on this point.
89
Principes, Part I, art. 41.
90
Leibniz, Gerhardt, VI, p. 331; Jalabert, p. 346.
196 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

dom on an acute internal feeling (equally contested by Leibniz)91 that

was one with the faculty of doubt, the very principle of his method.92
Such a conception of freedom, whose certainty would be guaranteed
by an interior sentiment, leaves much to be desired; and Kant sided
with Leibniz' criticisms on this point. "But there are many who be-
lieve they can explain this freedom with empirical principles, just as
they can explain other natural abilities. They regard it as a psycho-
logical property, the explanation of which turns solely upon a more
exact investigation of the nature of the soul and of the incentives of
the will and not as the transcendental predicate of the causality of a
being which belongs to the world of sense; but it is this latter which
is what really counts. Thus they deprive us of the great revelation
which we experience through pure practical reason by means of the
moral law—the revelation of an intelligible world through realization
of the otherwise transcendent concept of freedom; they deprive us of
the moral law itself, which assumes absolutely no empirical ground of
determination. Therefore it will be necessary to add something here
as a protection against this delusion and to expose empiricism in its
naked superficiality" ,93
This objection intuitionism makes to some of its expressions is
worth attending to. The proof required for the ground of truth cannot
consist in a brute fact, since the presence of the fact could always be
contested; it consists in the methodical production of that fact. Nor
can that methodical production be reduced in turn, without sophistry,
to a fact. The activity of control required by intuitionism is enough to
distinguish it from pure empiricism. But then what exactly is the con-
nection of that activity to fact? Kant established that what is given as
fact on the moral plane is the consciousness of duty. If that conscious-
ness makes us posit freedom it is that there would be a contradiction
in our being obliged were we not free. Freedom is the ratio essendi
of the consciousness of duty, which is itself the ratio cognoscendi of
freedom. We are therefore led to postulate our freedom without over
being able to grasp it directly in an act of consciousness, because we
are faced with the specific fact of duty which refers us beyond the realm
of experience. Freedom has the status of an a priori synthesis made
necessary by the confrontation with a sui generis experience requiring
the relation of the phenomenon to the thing in itself.

91
Leibniz, Gerhardt, VI, p. 130; Jalabert, p. 137.
92
Descartes, Principes, Part I, art. 39.
93
Kant, Critique of Practical Reason, L.W. Beck, 1976, p. 200.
 INTUITIONIST CONCEPTIONS OF REALITY: DESCARTES AND KANT / 197

This is a characteristic trait of Kantianism and of Kantian moral

philosophy. In examining Kant's own treatment of the question of dog-
matic illusion however, we can see its general bearing for intuitionism.
The Critique of Pure Reason teaches how transcendental idealism al-
lows us to free ourselves from the dogmatic illusion. It does not instruct
us why we ought to do so, given that that illusion is natural for our
reason. It says simply that the fact of this illusion, which would oth-
erwise be without any way out, is an indirect proof of transcendental
idealism.94 But supposing we were to remain bound to that illusion,
what would the consequences be? What would result would be the
most absolute necessitarianism. Kant agrees with Diodorus and ad-
mits the consequence he draws from the Master Argument. As long as
we are considering a thing under the condition of time and are dogmat-
ically confusing this condition with a property of things in themselves,
natural necessity radically excludes any freedom and therefore any pos-
sibility that would not be realized in time. For necessity "implies that
every event, and consequently every action, which occurs at a certain
point of time, is necessary under the condition of what preceded it.
Since the past is no longer in my power, every action which I per-
form is necessary because of determining grounds which are not in my
power. That means that at the time I act I am never free".95 The
Master Argument's first premise, together with the principle of causal-
ity and the confusion of the phenomenon with the thing in itself, leads
inevitably to necessitarianism. As long as the dogmatic confusion is
accepted it is impossible to escape the consequences of necessity. In
particular, recourse to a comparative96 notion of freedom, noting with
the Stoics and Leibniz that the determining grounds of our actions
are interior ones, resulting from our spontaneity, in no way exempts
us from the domination of necessity. This is but useless subterfuge,
for "here reference is made only to the necessity of the connection of
events in a temporal series as they develop according to natural law,
whether the subject in which this evolution occurs be called automaton
materiale when the machinery is impelled by matter, or, with Leibniz,
automaton spirituale97 when it is impelled by ideas. And if the free-
dom of our will were nothing else than the latter, i.e., psychological
and comparative and not at the same time transcendental or absolute,
it would in essence be no better than the freedom of a turnspit, which
94
Ibid., P. 203.
95
Ibid., p. 200.
9&
Ibid., p. 201.
9r
Leibniz, Gerhardt, VI, p. 131; Jalabert, p. 137.
 198 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

when once wound up also carries out its motions of itself .98 Hence
the dogmatic illusion results in necessitarianism. But necessitarianism
runs against our moral conscience. "A man may dissemble as much as
he will in order to paint his unlawful behavior, which he remembers, as
an unintentional error, as mere oversight, which can never be entirely
avoided, and consequently as something to which he was carried along
by the stream of natural necessity, and in this way to make himself
out as innocent. But he finds that the advocate who speaks in his be-
half cannot silence the accuser in him when he is conscious that at the
time when he committed the wrong he was in his senses, i.e., he was in
possession of his freedom".99 What is revealed through the paradoxi-
cal sentiment of repentance is not the actual psychological presence of
freedom but its obligatory postulated presence at the base of our past
responsibility taken as the ratio essendi of the imputation of our acts
to a moral agent.
The dogmatic illusion then is really a moral fault, or, at best, an in-
tellectual habit got for the purpose of casting the veil of necessity over
our moral faults. To deliver us from this illusion, it is not necessary
for intuitionism to appeal directly to the experience of freedom, but it
must simply indicate the phenomenon of that freedom, self-blame,100
which alone is the object of experience in the matter. But Kant's tran-
scendental theory of freedom has a universal bearing for intuitionism
precisely because, according to it, all our intellectual errors are faults
in that reflection leads us to feel them as such. For Epicurus' impa-
tience of opinion and Descartes' prejudice and precipitation of judg-
ment confront us with facts that accuse us. Even if there is nothing
positive about error, formally it still bears witness of our imperfection
and poses a problem for us that postulates our freedom by implication.
Thus the firm resolve to take only experience into account appeals to a
faculty going beyond that experience, and it has been rightly said that
it was Epicurus who introduced the transcendental concept of freedom
into philosophy. And this very concept is the keystone of intuitionism.
Psychology and anthropology give rise to it, though it does not fall
within their jurisdiction. It is characteristic of that paradoxical exi-
gency imposed by our finite reason. It is Fichte who best described it
in saying: "To be free is nothing, to become free, everything".101
This anthropological exigency may be seen as intuitionism's ulti-
mate requisite. That is how Epicurus took it in placing the gods in
98
Critique of Practical Reason, op.cit., p. 203.
"Ibid., p. 204.
loo
lbid., p .204.
101
Quoted by Gueroult, 1930, I, p. 269.
 INTUITIONIST CONCEPTIONS OF REALITY: DESCARTES AND KANT / 199

the inter-worlds and in depriving them of all care and knowledge as

regards our actions. He thus insured himself against the difficulties of
any possible theodicy by extirpating them at their root. In limiting
necessity and modality to the experiences that make us sense them,
intuitionism in general tends to reject questions of theodicy because
the necessity linking essences as well as existents has, in itself, only an
artificial ground. In other words, as long as intuitionism limits itself to
a description of knowledge, it systematically rejects the idea of final-
ity that would give necessity a basis in divine right or noumenal law.
Descartes' physics is therefore altogether mechanistic.102 It is true that
Kant's physics, as Newtonian, does admit forces; but these forces have
lost the metaphysical and finalistic character given them by Leibniz:103
they become the organs of universal mechanism.104
Yet far from being incompatible with a reflection on the finality
of error and evil, intuitionism even invites such a reflection. For it
explains error as a conflict of faculties: opinion's out-distancing sensa-
tion, the reciprocal encroachments of the understanding and the faculty
of judgment, the confusion of pure reason and the understanding ap-
plied to empirical intuition. But even if we admit with Epicurus that
everything there is results from the interplay of the atoms, once that
interplay has issued in the formation of the soul, the soul must surely
owe the conservation of its arrangement for a certain time to its internal
viability, that is, to the absence of incompatibility that that conserva-
tion reveals between the relative stability of that arrangement and the
global situation of the other atoms and bodies. It has been pointed out
that Lucretius borrowed the notion of natural selection from Empedo-
cles to explain the persistence of the living.105 But then what natural
selection accounts for the conflict between opinion and sensation? For
it can be presumed that if the conflict were without survival value na-
ture would not have maintained it. And here, through the fact of the
existence of error, the problem of theodicy arises again.
But this problem can only be solved, if at all, by a comparison of
the drawbacks of error, formidable ones since error is an obstacle to
wisdom and happiness, with the advantages procured as a general rule
102
"Preoccupied with founding the autonomy of a purely mathematical physics,
[Descartes] puts the accent on what God's incomprehensibility comports in the
guise of absolute freedom taken as a power or decision subject to no rule of any
sort, a sovereign power of yes and of no, of doing and not doing, of doing thus or
otherwise. He thus radically separates science and theology, the material world and
teleology." (Gueroult, 1953, II, p. 215).
103
Leibniz, Gerhardt, VI, p. 321; Jalabert, pp. 335-6
104
Vuillemin, 1955, passim.
105
Furley, 1970.
 200 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

by the cause of error, i.e., the excessive broadness of opinion over sen-
sation. It will be said for example that that excess allows the living
to foresee dangers and therefore aids in conserving life. But it is seen
that theodicy engages in speculation going beyond sensations. It is
this then that represents the extreme limit of intuitionism, going fur-
ther even than the adventure through which credence was granted to
atomism.
Descartes reaches the same limit in connection with his own system
when he tries to clear God of the responsibility for error. "Preoccupied
with integrating the human experience (formal error, feeling) into the
total system of philosophy, and affective psychology and ethics into the
system of the sciences, he reintroduced into the created world, through
the principle of the best, finality, theodicy, and consequently theology,
although it was natural theology... Divine freedom's infinitude can no
longer be conceived of then as the absoluteness of a decisive power, of a
power of arbitrary pronouncement of yes and no, for God can not sub-
scribe to nothingness: He cannot deny being. There is an ascendancy
therefore, in the species, of the idea of the perfect or of the infinite
essence over pure freedom. Whether this primacy comes from the di-
vine will itself, arising by virtue of its nature as unlimited omnipotence,
or is imposed from without under the constraint of a dominant under-
standing, as for Malebranche or Leibniz, makes certainly more than
just a nuance of difference. But on either hypothesis the fact of that
primacy remains indubitable".106
That balance of freedom and perfection in God dominates the subtle
and complicated interplay of psychology and metaphysics throughout
the Cartesian theodicy.107 On the one hand, the incomprehensibility of
the divine decrees would have as limit the imperfection of the work and
therefore the formal reality of error, which psychology requires. On the
other hand, Divine Perfection demands that that error be an absence
of being and that the formal imperfection it introduces in nature cancel
out as an ingredient of a perfect whole subject to the principle of the
best.108 But if in the final analysis divine freedom submits, so to
speak, to its positive perfection like the incomprehensibility of creation
submits to the assent we owe to the visible finality of that creation in
virtue of the divine veracity, this double subordination is rather induced
from signs than produced by the pseudo-necessity of the principle of
the best. It is an object of rational faith rather than one of reason.
106
Gueroult, 1953, II, p. 215-6.
107
Gueroult, 1953, I, p. 300 sq.
los
lbid.,pp. 306-18.
 INTUITIONIST CONCEPTIONS OF REALITY: DESCARTES AND KANT / 201

In this respect it is reminiscent of Epicurus' position on finality.

Epicurus' greatest disciple, Darwin, also subordinates the mechanism
(still called chance and having to do with hereditary mutations) to
finality (also called necessity and having to do with the selection or
conservation of mutant strains). Without nature's sanctions in conse-
crating these changes as biologically adapted and useful, they would
disappear without a trace and there would be no evolution. But how
do we ascertain the utility and finality of natural selection? There are
two criteria given: fecundity and ecological extension. But these crite-
ria are the external signs by which we recognize what life has judged
useful. It is from them we induce that a mutation will have been use-
ful. On the other hand, we understand strictly nothing of the nature
of that finality. We know only the mechanisms. But from the pres-
ence of signs, proliferation or disappearance of a species, we conclude
that there must be some finality. The degree of an individual organ-
ism's complexity, often taken as a measure of evolution and as capable
of furnishing an internal criterion of finality—as Spencer mistakenly
believed—plays no direct role in Darwin's appraisal of the situation. It
is possible that very complex and efficient organisms should have been
incapable of forming fecund and durable species: selection will have
eliminated them. Thus, in spite of the ontological subordination of
mechanism to finality, Darwinism maintains the epistemological sub-
ordination of finality to mechanism. All that we know of the former
is taught us by signs taken from the latter. We are not privy to the
council of nature.
The same thing goes for Descartes' system as well. Divine freedom
is ontologically subordinate to God's perfection, but epistemologically
there remains that unfathomableness that results in my having to be-
lieve that what is obeys a purpose because it is God's creation, not for
some intrinsic perfection that I could find in it. That is why error—
a point that scandalized Gassendi109—may be said to have a sort of
perfection, because it belongs to this created world. Once it has been
demonstrated that God exists and is the Creator it follows that the
creation must be a perfect work. But no direct examination of that
work would lead us, as through a Thomistic analysis of effects, to posit
its finality and therefore its dependence on divine will. Recognition of
finality in no way detracts from its incomprehensibility, for they are
one and the same. And here we are poles apart from Leibnizian dog-
matism. A propos of this incomprehensibility Leibniz jibes at Nicole
that if faith and reason are both gifts of God, then the combat of faith
9
Gassendi, pp. 418-20.
 202 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

against reason is a combat of God against God.110 For Descartes there

is a faith within reason, and that is what posits the principle of the
best that we cannot penetrate, whereas according to Leibniz it would
have to be said that it is this principle, which is the highest reason,
that enables us to ground faith on reason.
This difference shows up clearly in connection with the example of
the duel that Descartes gave Princess Elisabeth to explain how God's
omnipotence and human freedom are reconcilable.111 "If a king who
has outlawed dueling and who quite well knows that two gentlemen of
his realm who live in different cities are having a dispute and are so
excited against one another that nothing could stop them from fighting
if they were to meet; if, I say, this king sends one of them on business
on a certain day towards the town where the other is, and he also
sends the other on business the same day towards the place where the
first is, he very well knows that they can't help meeting, and fighting,
and thereby breaking his law, but he doesn't thus force them to do so,
and his knowledge and even the willfulness he had to determine them
in that manner do not prevent its being as voluntarily and as freely
that they fight as they would have done on meeting had he known
nothing of their difference and had it been on some other occasion
that they crossed paths. And they can be just as justly punished
because they have transgressed his prohibition. But what a king can
do in the matter regarding some free actions of his subjects, God, who
has infinite foreknowledge and power, does infallibly regarding all the
actions of man. Before sending us into the world he knew exactly
what all the inclinations of our will would be; it is he himself who
put them in us; he is also the one who disposed all the other things
outside us so that certain objects would be present to us at certain
times, knowing that on those occasions our free will would determine
us to some one thing or another; and he willed it so, but he did not
will to compel it for all that. And as we can distinguish two different
degrees of will in the king, one by which he willed that these gentlemen
fight, since he brought it about that they meet, and the other by which
he did not will it, since he prohibited dueling; so do the theologians
distinguish in God an absolute and independent will by which he wills
that all things should happen as they do, and another which is relative
and has to do with the merit or demerite of men, by which he wills
that his laws should be obeyed." Some time earlier Descartes had
110
Leibniz, Gerhardt, VI, p. 73; Jalabert, p. 77 and applied to Descartes, Leibniz,
Gerhardt, VI, p. 89; Jalabert, p. 94.
111
Letter of January, 1646, Correspondence, A.T., IV, pp. 353-4; see also Principes,
part I (41).
 INTUITIONIST CONCEPTIONS OF REALITY: DESCARTES AND KANT / 203

written to the same princess that "philosophy alone suffices for knowing
that the least thought could not enter the mind of a man, that God
should not will and should not have willed there from all eternity".112
Leibniz113 thought that even Calvin had said nothing harder than this
last assertion of God's "total" causality over our least thoughts and he
excuses Descartes only if divine will is taken in a "permissive" and not
in an absolute sense, as Descartes however explicitly maintains.
The conflict comes to a head when Leibniz corrects Descartes' ex-
ample in order to make it acceptable. He says that it would be nec-
essary "to invent a reason that obliged the prince to arrange or allow
that the two enemies should meet".114 That reason is the principle of
the best. On the contrary, for Descartes, from the fact that God did
something it follows that it was the best. In another letter to Elisabeth
he advances considerations on our freedom and on divine providence115
that are contrary by our lights but both equally well founded, and goes
on to conclude that that freedom "is not incompatible with a depen-
dency of another nature, according to which all things are subject to
God". That other nature is a truth of simple consequence following
upon the proof of the existence of God, hence altogether indirect and
already resemblant of a rational postulate, of an object of faith.
Just as the usefulness of survival value of a mutation has only an ex
post facto justification, the finality and insertion of human freedom in
the plans of providence has but an indirect and, as it were, retrospective
rationale. As Descartes admits, this is because the truths proportionate
to our cognitive faculty in so far as it moves spontaneously and of itself
in the bosom of truth are of a different nature from those met with
by that same faculty when it has to account for error and for the
remedies error calls for. Just what is this difference of nature? It is
to Kant's glory to have discerned it and made it the touchstone of his
philosophical system.
Taking up a remark of Leibniz' on the difficulties added to the ques-
tion of freedom by the physical concurrence of God and creatures with
the will,116 Kant posits the ideality of time as a necessary condition
for the solution of the problem. For if divine action were accomplished
under the condition of time, necessity would be inescapable. "There-
fore, if the ideality of time and space is not assumed, only Spinozism
112
Letter of Oct. 6, 1645, Correspondence, A.T., IV, p. 314.
113
Leibniz, Gerhardt, VI, p. 207; Jalabert, p. 218.
115
Letter of Nov. 3, 1645, Correspondence, A.T., IV. p. 333.
116
Leibniz, Gerhardt, VI, p. 122; Jalabert, p. 128; Kant, Critique of Practical
Reason, p. 206.
 204 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

remains, which holds space and time to be essential attributes of the

First Being itself and the things dependent upon it (ourselves included)
not to be substances but merely accidents inhering in it. For if these
things exist only as its effects in time, which would then be the condi-
tion of their existence itself, the actions of these beings would have to
be merely its actions, which it performs anywhere and at any time".117
The ideality of time therefore leaves the way open to a solution of the
theodicy. It is the noumena, not the phenomena, that God creates. He
cannot therefore be the cause of the actions of beings in so far as they
are conditioned by time. Creation then doesn't really add a new diffi-
culty to the problem of creatures' freedom, "because creation concerns
their intelligible but not their sensible existence, and therefore creation
cannot be regarded as the determining ground of appearance".118
The Critique of Pure Reason had restricted the usage of the modal-
ities to the capacity of objects being given in an intuition. But it is
possible to go further, as in the Critique of Judgment,119 and to inquire
into the transcendental origin of the modalities, in tracing their relation
to our faculty of knowledge. Leibniz distinguished between essences,
residing in the divine understanding and representing in themselves
the different possibilities along with their measure of degree of perfec-
tion, and existents, objects of divine decree and therefore subject to
the principle of the best. Kant reduces this ontological opposition to a
transcendental one. The understanding's limitation to the conditions
of intuition is the condition for our knowledge's being objective. "If
our understanding were intuitive it would have no objects but such
as are actual. Conceptions, which are merely directed to the possibil-
ity of an object, and sensible intuitions, which give us something and
yet do not thereby let us cognize it as an object, would both cease
to exist".120 Distinction of the modalities is therefore founded on the
opposition between our two sources of knowledge. If there is a modal
logic, if the modalities do not collapse into one, it is not because there
is anything objective in the Leibnizian hierarchy of possible worlds, but
simply because our faculty of knowing is irremediably discursive and
can reach through its synthesis only phenomena given in the ideality
of time. "To say, therefore, that things may be possible without being
actual, that from mere possibility, therefore, no conclusion whatever as
to actuality can be drawn, is to state propositions that hold true for
117
Kant, Ibid., p. 207.
U8
Ibid., p. 208.
119
§76.
120
Ibid., p. 570.
 INTUITIONIST CONCEPTIONS OF REALITY: DESCARTES AND KANT / 205

human reason, without such validity proving that this distinction lies
in the things themselves".121
In taking the fact of error to be the result of divergence among the
faculties, intuitionism finds the remedy in critically assigning these fac-
ulties a domain of validity. But what error shows is that the possibility
of transgressing the bounds set is inevitably tied to our human nature.
Opinion outstrips sensation, understanding and judgment encroach on
one another's domain, reason inexorably disengages itself from the lim-
its of intuition. But the fact is that these proscribed excesses are not of
uniquely negative bearing. A classically demonstrated theorem poses
a problem for mathematicians just as selective value and finality do
for the intuitionist philosopher. Subordinations arise that, seen from
the outside, seem to run counter to the method. Finality, for example,
imposes itself as a principle of natural science.122 But insurmount-
able difficulties would arise if it were of the same order as mechanical
causality whose sway extends over the phenomena. For teleology is
inexplicable. Hence, in so far as applied to objects, it does not af-
ford a principle of truly determinant judgment, but only one of merely
reflective judgment, even though this last be universal.123
Kant tells us that the same can be said of the connection between
freedom, presenting us with the law as an imperative, and actual ac-
tion. If our will were holy there would be no such distinction. Thus
the requisite subordination of mechanism to finality and of particular
to universal,124 does not express a dogmatic subordination in the ob-
jects, of which we could form a determinant conception. It expresses
only a regulative principle, necessarily valid for our human faculty of
judgment. Far from governing creation like the Leibnizian principle of
the best, it is a subjective principle depending on the contingent nature
of our faculty of knowledge. The illusion common to Diodorus and his
dogmatic adversaries, and probably to all of modern modal logic, lies in
the confusion that attributes a constitutive use to principles that have
no meaning once divorced from cognizing activity and whose legitimate
and simply regulative use is to police that cognizance from the inside
without presuming to legislate with respect to things in themselves.

121
Ibid., pp. 570-1.
122
Ibid., §68.
123
Ibid., §74.
124
Ibid., §76.
8

Carneades and the Skeptical

Nominalism of the Modalities,

In the De Fato Cicero sets a scene with Carneades at grips with

Chrysippus and Epicurus, dissociating an implication that his two
adversaries both admitted, even though they did draw contrary con-
clusions from it. First the debate itself will be analysed. Then it
will be shown that the dissociation Carneades proposes comes down
to challenging the fundamental postulate of dogmatism, namely, the
Aristotelian definition of truth. He is consequently able to retain all
of the Master Argument's premises without burdening himself with
Diodorean necessitarianism for all that. Such a conception leads to
a non-existential interpretation of the quantifiers as in the theory of
ampliation as it is found in Buridan.

8.1 What is the relation between the principle of

the excluded middle and the principle of
causality (De Fato, X-XII)?
It has been said that the discussion of Carneades and Chrysippus "was
over two propositions that formed the basis of the latter's conception:
(1) there is no motion without cause; but, he said, an act that would
be free and contingent without at the same time being determined
would be a movement without cause; (2) of two contradictory propo-
sitions about the future, one must well be true and the other false, for
the rule of contradictories holds in a parallel way of past and present;
but a free act supposes that the true can become false, and inversely.
Therefore, in one way as in the other, freedom and determination must
be reconciled and, since there is an uninterrupted causal chain, the de-
termination must be an eternal pre-determination, involving a plurality
of diverse causal series however so that the necessity of the whole leaves

207
208 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

room for individual spontaneity in each of these series and fate does
not wrong freedom".1 Epicurus denied both propositions asserted by
Chrysippus. Carneades took a more balanced stance, typical of his phi-
losophy in general, accepting the first of these propositions but limiting
his acceptance of the second.
Such an interpretation fails to draw out all that is in the texts. The
dispute between Chrysippus and Epicurus becomes a dispute over the
materiality of the two principles of causality and the excluded middle,2
as if these two principles were admitted as independent. But that is
not the case. For here is Chrysippus' reasoning: "If uncaused motion
exists, it will not be the case that every proposition (termed by the
logicians an axioma) is either true or false, for a thing not possessing
efficient causes will be neither true nor false; but every proposition is
either true or false; therefore uncaused motion does not exist. If this
is so, all things that take place take place by precedent causes; if this
is so, all take place by fate; it therefore follows that all things that
take place take place by fate" .3 The polysyllogism with contraposition
takes as its starting point the following Epicurean principle:
If there is motion without cause (Pi), there are exceptions to the
principle of the excluded middle (P?),
and immediately furnishes an application of the principle: the proposi-
tion saying that 'there will be a certain swerve' is neither true nor false.
The principle is true and the consequent P2 false. In order to criticize
his two adversaries then, Carneades begins by specifying the hypothet-
ical principle they both take to be true. If there is opposition between
Chrysippus and Epicurus, it is because the latter, accepting the an-
tecedent as true (the existence of motion without cause), is obliged
to admit the consequent (invalidity of the excluded middle), whereas
Chrysippus, admitting the falsity of the consequent (not-P 2 ), the ex-
cluded middle being always true, is obliged to reject the antecedent,
the invalidity of the principle of causality, (not-PI). As the univer-
sal principle of causality (not-PI) entails the universal precedence of
causes (Ps), and that precedence in turn entails fate (Pi), the truth of
1
Robin. 1944, p.127.
2
In spite of the fact that the principle is here expressed in the meta-language, it
will be spoken of throughout the paragraph as the principle of the excluded middle
and not as that of bivalence. For in the debate before us none of the speakers
establishes a difference between those two principles and the discussion at no point
touches on Aristotle's solution.
3
Cicero, De Fato, X.20-21, trans, by H. Rackham in the Loeb Classical Library,
1942, pp. 216-217.
PRINCIPLE OF THE EXCLUDED MIDDLE AND PRINCIPLE OF CAUSALITY / 209

the Epicurean principle together with the universality of the excluded

middle entails fate.4
Carneades contests neither the Epicurean principle nor the falsity
of its antecedent. If he refuses to accept fate it is that he rejects one of
the particular premises of Chrysippus,5 according to which universal
causality entails universal precedence of causes. For neither voluntary
motion nor even the inert motion of an atom can be properly said to
be without cause: it is simply that their cause is not antecedent or
exterior.6 Chrysippus replies to Carneades' objection that in denying
the universality of antecedent causes one is inevitably open to denying
the excluded middle, since "future events that have not got causes
to produce them cannot be the objects of the assertions" .7 But the
validity of the excluded middle is admitted.
This last contestation shows that Chrysippus takes the contrapos-
itive of the Epicurean principle in a very particular sense: from the
truth of a proposition that he rightly holds to be eternal and that he
accordingly retrogrades he concludes the eternity of the event's causes.
Carneades therefore proposes another interpretation of the contrapos-
itive, according to which the eternity of the truth in no way entails the
eternity of the cause. Thus, with Chrysippus and against Epicurus,
he admits the universality of the excluded middle and the principle of
retrogradation. 8 With Chrysippus again, he admits the universality of
causality. But against him, he rejects the universality of antecedence:
"Yet it does not immediately follow from the fact that every statement
is either true or false that there are immutable causes, eternally ex-
isting, that forbid anything to fall out otherwise than it will fall out.
The causes which bring it about that statements of the form 'Cato will
come into the Senate' are true statements, are fortuitous, they are not
inherent in the nature of things and the order of the universe; and nev-
ertheless 'he will come', when true, is as immutable as 'he has come'
(though we need not on that account be haunted by fear of fate or
necessity)".9
4
Chrysippus' reasoning may be schematized as follows:
PI 3 ?2, ~ P2, ~ Pi 3 Pa, f*3 3
~ Pi D A

5
~ Pi D P3.
6
De Fato, XI. 24 "...voluntary motion possesses the intrinsic property of being in
our power and of obeying us, and its obedience is not uncaused, for its nature is
itself the cause of this" (XI.25, ibid., pp. 220-221).
7
XI.26, ibid., pp. 220-222. Translation altered.
8
XII.27.
9
XII.28, ibid., pp. 222-223.
210 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

8.2 Aristotle's dogmatic definition of truth called

into question (De Fato, XIV).
Carneades' parry, which did not touch Chrysippus, did affect Aristotle.
For according to Aristotle, it is certain that the truth of a future entails
the actual existence of its cause.
The discord between Carneades and Aristotle is all the more re-
markable in that their doctrines of cause seem to be very close.
When Carneades asserts that "there is a difference between causes
accidentally precedent and causes intrinsically containing a natural
efficiency",10 or again "the causes which bring it about that statements
of the form 'Cato will come into the Senate' are true statements, are
fortuitous, they are not inherent in the nature of things and the or-
der of the universe",11 it is the very language of Aristotle he employs.
When he mocks the Stoics for going back and back and back in infinite
regress instead of assigning causes,12 of being ignorant of proximate
causes and of confusing the effective cause bringing something about
with that without which it could not have been, it is again an Aris-
totelian argument and distinction he makes his own.
But if in agreement with the Stagirite on the theory of causes, of
freedom and of chance, Carneades differs with him on the theory of
truth. For he considers the truth of a future event to be conceivable
without the immutable existence of its cause. For to maintain with
Chrysippus, but also with Aristotle, that the eternal truth of a propo-
sition about the future entails the necessity of fate, is to say nothing.
"For it makes a great deal of difference whether a natural cause, ex-
isting from all eternity, renders future things true, or things that are
going to be in the future can be understood to be true even without
any natural eternity. Accordingly, Carneades used to say that not even
Apollo could tell any future events except those whose causes were so
held together by nature that they must necessarily happen. For what
consideration could lead the god himself to say that the Marcellus
who was three times consul was going to die at sea? This had indeed
been true from all eternity, but it had no efficient causes. Therefore
Carneades held the view that Apollo had no knowledge even of these
past events which had left behind them no trace of their passage—how
much less had he knowledge of future events, for only by knowing the
efficient causes of all things was it possible to know the future; there-
fore it was impossible for Apollo to foretell the fate of Oedipus when
10
IX.19, ibid., pp. 214-215.
U
XII.28, ibid., pp. 222-223.
12
XV. ibid., pp. 228-233.
 ARISTOTLE'S DEFINITION OF TRUTH CALLED INTO QUESTION / 211

there were no causes fore-ordained in the nature of things making it

necessary for him to murder his father, nor could he fortell anything of
the sort".13 The truth of a future- or past-tense proposition, dissoci-
ated then from the existence of its cause or of its trace, allows neither
prediction nor retrodiction. There is a complete separation between
the determination of events and the truth of propositions.14
Truth is therefore only a formal property of propositions, and not
a real one.15 But if one refuses, as against Chrysippus and Aristotle,
to make the formal and the real coincide, one no longer need fear
necessity. "Moreover, those who say that things that are going to
be are immutable and that a true future event cannot be changed
into a false one are not asserting the necessity of fate but explaining
the meaning of terms".16 Between Carneades and Aristotle there is
conflict over the relation of truth and actuality, and the first conflict
leads to another over the possibility of knowledge. For Aristotle the
truth of a statement results from its agreement with the thing in act
and therefore with the actuality of its cause. If a statement about the
future is true, it is that the determination of that future is actual. For
Carneades the truth of a statement, being formal, in no way entails
the existence in act of the thing or of its cause. A statement about
the future can therefore be true without that future's determination
being actual. Both philosophers of course consider that knowledge of
the truth implies the pre-existence of causes: Aristotle, because there
is no truth without act, Carneades, because the knowledge of truth is
impossible in the absence of a pre-existent cause. But because truth
entails act for Aristotle, he is able to save the contingency of the future
only by depriving propositions about future contingents of the actual
possession of a truth value, whereas Carneades says simply that, such
propositions being true or false in themselves, that truth and that
falsity are inaccessible to our knowledge which is limited to probability.
Imagine a wise man embarking for Puteoli which is four miles away,
with a good helmsman and a tranquil sea, and wondering whether he
will get there safe and sound. The proposition 'the wise man will get
there safely' is true or false from all eternity, just by virtue of the
very definition of the notion of proposition. But is the wise man able
to know that truth-value? "... surely he has not got the knowledge
already grasped in his mind and perceived that he will make the voy-
13
XIV.32-33, ibid., pp. 228-229; Robin, 1944, p. 125.
"Robin, ibid., p. 127.
15
Robin, ibid., p. 128.
l6
De Fato, IX.20, Rackham, 1942, pp. 214-215.
212 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

age as he intends? How can he have it"?17 All that his knowledge
can attain is the probability of a safe passage given what he believes
he presently perceives, and that probability is enough for determin-
ing action; but there is a gulf between it and truth. The outcome of
Carneades' skeptical pragmatism is the probable, as it would have been
for Aristotelian dogmatism had it classed all contingent events in a sin-
gle category. But it is because there are propositions about the future
that are neither true not false that the probable exists for Aristotle,
while for Carneades the probable results from the gulf there is between
the self-containment of truth and the act of our apprehension.
If a theory of degrees of belief were to be developed for Carneades
as one of degrees of determination has been for Aristotle, it might be
said that the Carneadean probable opens the way to 'subjective' prob-
ability while Aristotelian contingency suggests 'objective' probability.
We shall refrain from speculating about these developments however,
especially since for Carneades, for whom a statement's truth implies
no actual determination of the thing, the accent put on the subject in
no way reduces the probable to ignorance of what is determinate.

8.3 Carneades and the Master Argument (De

Fato, IX).
De Fato, IX contains the following argument. 1) Diodorus' necessi-
tarianism renders the difference between past and future as regards
necessity purely apparent (IX, 17). 2) Nor is invoking the notion of
degrees in the precision of propositions about the future of any avail
for giving that appearance some reality: the more general propositions
having to do with essence are no more necessary than the more partic-
ular ones having to do with accident (IX, 18). 3) If there is no more
necessity in past than in future or in essence than in accident, that goes
to show that the question is one of grammar, not of ontology. Rather
than having to do with the nature of events it has to do with the nature
of statements and with the fact that once true, they always have been
or will be so. There is nothing for Epicurus to fear therefore from a
grammatical necessity of this sort and it is not necessary to invalidate
the excluded middle to establish freedom (IX, 19).
Those today who accuse Cicero of having misunderstood Diodorus
in failing to recognize that the necessary does not merge with the
Diodorean possible, and who seek a grammatical, that is to say formal,
way out of the debate, have simply not read this chapter. For in the
17
Cicero, Academica, II.XXXI, trans, by H. Rackham in the Loeb Classical Library,
1933, pp. 594-595.
 CARNEADES AND THE MASTER ARGUMENT / 213

first two parts of the argument, the Orator shows that the Diodorean
distinctions, from Diodorus' point of view, not only fail to avoid but
even establish necessitarianism. The last part of the argument shows
that he knew of one philosophy that maintained that that Diodorean
necessity was only apparent or formal and that was the philosophy of
Carneades. The mainstay of the argument's first part is the consid-
eration of the change of truth-value. "... it is no more possible for
things that will be to alter than it is for things that have happened;
but...whereas in the things that have happened this immutability is
manifest, in some things that are going to happen, because there im-
mutability is not manifest, it does not appear to be there at all, and
consequently, while 'This man will die of this disease' is true in the
case of a man suffering from a deadly disease, if this same is said truly
in the case of a man in whom so violent an attack of the disease is not
manifest, none the less it will happen. It follows that no change from
true to false can occur even in the case of the future. For 'Scipio will
die' has such validity that although it is said of the future it cannot
be converted into a falsehood, for it is said about a human being, who
must inevitably die".18
It is clear that Cicero is here speaking of propositions, that is to
say of the equivalence classes of all possible statements about a certain
presently future event. Diodorus, on the contrary, limited himself to
statements or to propositions as partial equivalence classes modulo a
given lapse of time, formulated or not, according to the state of our
knowledge. Cicero maintains that in going from the latter to the former
necessitarianism in forma cannot be avoided.
But let us leave grammar and come back to reality. In other words,
let us eliminate these purely exterior variations of truth-value, in taking
the equivalence class of statements modulo a determinate date. Accord-
ing to Diodorus a proposition is possible if it is true or will be true. In
this case its truth retrogrades; it is necessary as soon as one takes for
granted, with Diodorus, the Aristotelian principle of correspondence.
For Cicero then, Diodorus does not escape the necessitarianism
for which the Megarians were reproached. He would only avoid it in
sticking to a consideration of grammatical forms alone and in refusing
to consider the corresponding propositions.
Some have compared Diodorus to Quine. Just as Quine is skep-
tical of the contemporary modal logics, so was Diodorus skeptical of
the Aristotelian modal theory "but offered nevertheless some 'harm-
l8
De Fato, DC.17; Rackham, 1942, pp. 212-213. [I have slightly revised the Rack-
ham translation of this passage. Translator's note.]
214 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

less' senses that might be attached to modal words".19 It is known

that Quine does not admit the 'propositions' that Cicero brandishes
against Diodorus. Yet he proposes converting variable statements into
eternal ones by incorporating into the statement the date upon which
it is made.20 But in the system of Diodorus this substitution will
have the same effect as would the introduction of Ciceronian proposi-
tions. Eternal 'future' statements will retrograde since they are deter-
minate, unlike the corresponding statements and propositions of Aris-
totle. Therefore they are necessary.
Now for the second part of the argument. Cicero has already shown
that the Diodorean future has the same modal status as the past. Will
there be a way out then in distinguishing, among futures, those having
to do with essence (Scipio will die because all men are mortal) from
those having to do with accident (Scipio will die a violent death)?21
Such a way out is precluded in the Diodorean perspective in which there
is no distinction between essence and accident. Hence the accident is
every bit as necessary as the essence.
Carneades' conclusion follows from this universal modal collapse.
If we are no more able to separate future from past than accident from
essence, if every proposition is subject to Diodorean necessity, it must
be because that necessity is a fact of language and not a trait of nature,
and that the immutability of the truth-values of propositions has no
ontological bearing.
It is only in appearance that Diodorean necessitarianism follows
from dialectic, i.e. logic, which is a neutral art. The illusion is due
to a dogmatic interpretation of logic mixing the formal with the real.
Interpreted without dogmatism the Master Argument has no force.
Carneades though, in the Ciceronian passage, fails to specify in
just which sense the Master Argument is without force. Does he mean
that, since Diodorus' necessitarianism has only grammatical bearing,
we should accept his conception of the possible as what is or what
will be? Or are we on the contrary to keep the third premise without
fearing contradiction with the other two? There is one sentence of
Cicero's that suggests by implication that Carneades favored the second
19
Prior, 1967, p. 16.
20
Quine, 1960, §40, pp. 191-195.
21
It seems impossible to determine precisely whom the second part of Carneades'
argument was aimed at. It will be remembered however that for Aristotle there is
no possibility of a science of the accidental. The accidental is that which happens
sometimes, as its being cold in the dog-days; and it is precisely its specificity,
with respect to essence, that preserves against necessitarianism (Metaphysics, K,
8, 1064fc30-1065°14).
 THE LOGICS OF "FICTIVE" NAMES: BURIDAN'S AMPLIATION / 215

position. "Hence if, while it is consistent for the Stoics, who say that
all things happen by fate, to accept oracles of this sort and all other
things connected with divination, yet the same position cannot be held
by those who say that the things which are going to happen in the
future have been true from all eternity, observe that their case is not
the same as that of the Stoics".22 The partisans of Carneades admit
the eternal truth of propositions about future contingents. It would
seem then—and this impression is reinforced by the entire context of
the discussion between Carneades and the Stoics—that these partisans
go along with the Stoics in admitting the eternal truth of a possible that
will not be realized, but without running the risk of necessitarianism for
all that. What is dangerous about the third premise on the dogmatic
interpretation is that the truth of the negation entails the act of the
contrary event. Once that act is suspended, necessity vanishes.
We must now turn to the logical consequence of Carneades' skepti-
cism. The late scholastic philosopher, Buridan, will provide the occa-
sion for doing so.

8.4 From Carneades to the logics of "fictive"

names: Buridan's ampliation.
There is a passage in Buridan that specifies what could have been
Carneades' position regarding the Master Argument's third premise:
"There is a third distinction, namely, that something is said to be
corruptible, either because it is already existent and can not exist, or,
in another sense, because it can exist and then afterwards no longer
exist. But, in this second sense, there are infinitely many corruptible
things that never will be corrupted; for there are infinitely many things
that can be engendered but will never be engendered, and even though
all things that can be engendered are, by the same token, corruptible,
they will never have to be corrupted if they have not been previously
engendered. And that is why the only question that arises here arises
with respect to corruptible things that already exist" ,23
The words 'there are' in this passage have no existential import. A
particular proposition can be true without there existing any subject
that the asserted property is attributed to. If, in fact, a proposition
must be able to be true without for all that there existing or having
existed or having to exist in act the state of things that verifies the
proposition, then it must be that actual existence is- completely dis-
sociated from a property's being predicated of a subject and that, as
22
De Fato, XV, 33; Rackham, 1942, pp. 228-231.
23
Buridan, 1942, p. 119.
216 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Carneades would have had it against Chrysippus and Aristotle, truth

entails no ontological commitment.
In his theory of ampliation, Buridan systematically developed this
conception of a logic thus freed of existential involvement or rather
of one rendered explicitly sensitive to the conditions of existence and
inexistence.
The theory of ampliation (ampliatio) seems to meet the logical need
of specifying the ontological status of subjects in two sorts of proposi-
tions that can be true when their subject has no actual supposition at
the time when the proposition is asserted. These two sorts of proposi-
tion are I/ those about the future and the past as well as those about
the possible, and 2/ those depending on a verb expressing a "preposi-
tional attitude" (believe, imagine).
Buridan remarks, in his Consequentiae: "Some maintain that every
true proposition is true because, whatever its manner of meaning, the
thing or things are as is signified. As for me, I believe that the following
proposition is not true by virtue of what is said, namely, that if Colin's
horse that walked well is dead this proposition is true 'Colin's horse
walked well', and it is not in the thing as the proposition signifies, since
the thing is corrupted... but rather that if that proposition is true it
is because it was in the thing as the proposition signifies that it was...
Likewise this is true 'Something that never will be can be', not because
it is as the proposition signifies but because it can be as the proposition
signifies it can be; and so it is evident that according to the diverse
sorts of proposition it is appropriate to assign the causes of their truth
in different ways".24
Ampliation consists formally in extending the supposition of the
subject term to things not existing at the time the proposition is as-
serted, the extension being made by way of an inclusive disjunction
with the present time. 'Some man will die', for example, is to be trans-
lated as 'For some x, x is a man or x will be a man, and x will die'.25
But how are the quantifiers to be interpreted? Some have thought
that ampliation came down to hypostatizing all subjects that have ex-
isted, will exist or can exist by putting them into a pool from which
the speaker or the Creator pulls them out as need be.26 Interpreted
in this way, the theory of ampliation would inevitably lead to positing
the existence of inexistent beings. Indeed, the second category of verbs
involving ampliation includes verbs in the present tense which have
24
Id., Consequentiae, I, 1; quoted by Moody, 1953, pp. 53-54.
25
Moody, 1953, p. 54 and p. 56 where Albert of Saxony, Logic, II, ch. 10, is quoted.
26
Prior, 1967, p. 143.
 THE LOGICS OF "FICTIVE" NAMES: BURIDAN'S AMPLIATION / 217

the power of being transitive with respect to past, future or possible

things as well as with respect to present things;27 such are the verbs 'to
understand', 'to be acquainted with', 'to think'. The reason for this,
as Albert of Saxony explains, is that when a thing is understood, "the
act of understanding terminates in that thing just as well, when the
thing is something which has existed, or will exist, or can exist, as it
does when the thing exists at the same time as the act of understand-
ing it".28 The second use of ampliation suggests that its mechanism
would lead to taking terms such as 'chimera' and 'golden mountain' as
designating objects having a being distinct from existence. It would
thus evoke, in the Middle Ages, the "extravagances" Russell imputed
to Meinong and attribute to his own early writings.29
Without a doubt, such an interpretation must often have tempted
philosophers. Not all of them, however, succumbed. In contrasting the
legitimate consequence, "If I am now eating bread, bread now exists",
to the illegitimate one, "If I am thinking now of a rose, a rose now
exists", Albert of Saxony clearly shows his disregard for the notion of
a pool of beings which would have included the imaginary rose.30 And
Buridan explicitly rejects such an interpretation. When the supposition
of the subject does not presently exist, we haven't the right to go from
truth to existence, though we always have the right to go from present
existence to truth. We must therefore reject conditionals of the sort
'If it is true that Colin's horse has run, Colin's horse exists', and, 'if it
is true that a possible will never be realized, a possible exists that will
never be realized'.
Truth is not a ground of existential import. But truth can com-
bine with what today's ordinary logic calls "existential quantification".
Buridan himself declares that "there are an infinite number of corrupt-
ible things that will never be corrupted". It's the "there are", i.e., the
existential quantification, that is the issue here and gives rise to the
difficulties mentioned. The point then is to determine just what sense
to give to this notion of logical existence which is altogether other than
actual existence.
It obviously does not have its standard or objective sense,31 but a
purely nominal one. If it is the ways of signifying that dominate ampli-
27
Moody, 1953, p. 56.
28
Moody, 1953, p. 57.
29
Russell, 1956, p. 45; Kneale and Kneale, 1962, p. 262 as regards William of
Shyreswood.
30
Moody, 1953, p. 57.
31
Moody is right in rendering the existential quantifier by "for some x" rather than
by "there is an x".
218 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

ation it is obviously impossible to give it an objective and homogeneous

scope, which seems rather reserved to the way of signifying of the past.
In Diodorus' third premise then, the existence of a possible that will
never be realized and that must be assured to ground the assertion that
it is true that a possible will never be realized is an existence merely in
the sense in which a mathematician would say that something exists if
it is non-contradictory. To say here that there is a possible is simply
to say that it is possible.32
For Aristotle, the truth of a temporal proposition entails the assign-
able character of its act at the appropriate time. Such a consequence
is homogeneous. It applies to all times indifferently. That is why
Aristotle is obliged, in order to preserve the asymmetry of time, to
consider the truth-value of propositions about the future in contingent
matter to be indeterminate. By contrast, ampliation guarantees the
asymmetry from the outset in utilizing a quantification whose actual
scope must be reexamined for each mode of signification. Truth here
becomes inoffensive in the sense that no conclusions can be drawn from
it mechanically. On the contrary, certain "received" laws of modal and
temporal logic must be called into question.33 One law of modal logic,
for example, is the following:34
If it is possible for all x that x have a certain property, then every
x possibly has that property.
Buridan objects that it could be that everything was God, in the case
in which He had created nothing, without it's being the case that every-
thing could be God. In like manner, he objects to the law of temporal
logic:
If it has been the case that everything had a certain property, then
everything has had that property,
where the antecedent 'It was true (before the creation) that everything
was God' can be true, and the consequent 'it is true that everything
has been God' false.35
Theories of ampliation therefore deviate from standard quantifi-
cational logics. The modern logic that seems the most suited to ex-
32
Moody, 1953, pp. 57-58; thus for Albert of Saxony, the ampliation required for
the necessity of scientific propositions indicates that the quantification is over what
exists and what can be.
33
Prior, 1967, p. 138.
34
Hughes and Cresswell, 1972, p. 144. The demonstration of the law in question
(t- M(x)<fx D (x)M(fx) does not involve the Barcan formula.
3b
Sophismata, ch. 4, sophism 13; quoted by Prior, 1967, p. 138: I- P(x)ipx D
(x)Ptpx.
 CARNEADES AND THE PRINCIPLE OF CONDITIONAL NECESSITY / 219

pressing their requisites is Lesniewski's "ontology" or logic of names.

Besides singular names naming extant individuals and common nouns,
it admits fictive names. It does not maintain the standard interpreta-
tion of the quantifiers, but distinguishes between 'There is an x such
that jx" and Tor some x, fx\ As a consequence, it is obliged to intro-
duce new logical functors and arrives at theorems such as 'For some x,
x does not exist' and at laws of identity which are very deviant with
respect to those of standard logic.36 It is also possible, in the so-called
"free" logics,37 to maintain the usual interpretation of the quantifiers
and identity, in modifying the laws of quantification in such a way as
to purge logic of all existential presupposition attaching to the use of
names. The complications accepted in diverging from the standard
notion of quantification, or in maintaining it while rejecting the onto-
logical interpretation of proper names and of the word 'all', have the
advantage of purifying the logical instrument of all compromise with
existence and of clearly expressing the separation of logic and ontology.
But consider the meaning of the logical quantifier 'there is' devoid
of any ontological involvement. Saying that there is a possible that will
not be realized must be taken as being about a possible that assuredly
cannot be realized in the real world, though it can be "realized" in a
possible world which is an ideal alternative to this real world. But what
can a possible realization as opposed to a real realization amount to if
not to that phantom of existence, to that logical 'there is' extraneous to
ontology and capable of coexisting with the actual existence of the op-
posite reality? Does the non-dogmatic interpretation of truth, in other
words, automatically constitute a challenge to the principle of condi-
tional necessity? To clarify the positions of Buridan and Carneades,
we must take up the question of this challenge again.

8.5 Carneades does not abandon the principle of

conditional necessity; he simply deprives it of
the ontological involvement conferred upon it
by the dogmatic interpretation of truth.
The apparent ambiguity of the skeptical position regarding conditional
necessity arises from the fact that, if truth and reality in themselves
are inaccessible, it seems difficult to maintain that reality entails neces-
sity while it obtains, though it can be maintained, on the other hand,
36
Lejewski, 1957-1958, passim.
37
In the sense developed by Lambert, for instance in "On Logic and Existence",
1965, pp. 135-141. On the relation between "free logics" and Lesniewski's ontology
see Lambert and Scharle, 1967.
220 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

that phenomenal or apparent reality, which alone is accessible for the

skeptic, does entail an apparent or phenomenal necessity in its turn,
so that the principle of conditional necessity is maintained without its
entailing dogmatism. Briefly, there are two possible interpretations of
skepticism: the first, as a rupture between reality or truth and neces-
sity; the second, as a rupture between reality, truth or necessity for us
and reality, truth or necessity in themselves. The first case leads to
a rejection of the principle, whereas the second can accommodate it.
We will show that it is the second interpretation that must be chosen.
In depriving the existential quantifier of its objective signification, the
skeptic is not telling us that being is not necessary but that being, and
the supposed necessity it entails, is inaccessible.
In denying the existence of cataleptic presentations, what Arcesilas,
whom we know to have been influenced by Diodorus,38 is contesting
is that sense presentations should be able to necessarily command our
assent. On the one hand, it is impossible to distinguish absolutely
between the true and the false; on the other, our assent remains always
free. When Carneades, starting from this critique, develops his theory
of degrees of probability, he too is combating the idea of an absolute
necessity revealing itself to us in the true.
This allows us to make better sense of the paradoxical transition
from Platonism to skepticism in the New Academy. For Plato himself
had also contested the view that presentation could attain to certainty.
Of the sensible we have only opinion, and opinion can never become
science.39 It would seem that the New Academy, more and more at-
tentive to sensation and less and less preoccupied with the ideas, was
able to find inspiration in a mutilated but still authentic Plato.
The apparent ambiguity of skepticism results from the following
dilemma. On the one hand, Carneades says that the truth of a propo-
sition about the future does not entail its necessity. This is one of the
lessons of the Ciceronian de Fato. To be consistent, as he does claim to
be in his adherence to the principles of dialectics, he must then reject
the principle of conditional or hypothetical necessity, for if it is already
true at t that a certain thing will happen at t'(t' > t), it will be true
at t' that that thing will happen at t'. But suppose the validity of the
38
Sedley, op. cit., pp. 82-83.
39
Thus according to Cicero (Academica, II, XXIII (74)), Socrates and Plato are
the ancestors of skepticism.
The same Platonic theme is found in Aristotle too: of the sensible there is no science,
but only opinion. But as a consequence of the immanence of Aristotelian form and
of his theory of first substances, excepting supra-lunar beings, the significance of
this Platonic theme is limited (Cherniss, 1962, p. 239, pp. 340-343 sq.).
 CARNEADES AND THE PRINCIPLE OF CONDITIONAL NECESSITY / 221

principle of conditional or hypothetical necessity: if it is already true

at t that a certain thing will happen at t', it will be necessary at t'
that, while t', that thing will happen at t'. But from the truth at t
of the occurrence of that thing at t', we should not have been able to
conclude to its necessity, even if only conditionally.
On the other hand, Carneades holds that we cannot distinguish the
true from the false. But this thesis, which is deliberately skeptical,
enables us to avoid necessitarianism while at the same time retaining
the principle of conditional necessity. For what difference would this
principle in fact make, since, not being able to attain truth, we are not
able to attain the necessity supposed to follow from it either? Here we
can conclude from its being "true" at t that a certain thing will happen
at t'(t' > t) that it will be necessary at t', while t', that that thing
happen. But since that truth is only "for us" and we cannot conclude
from it anything about the nature of things in themselves, the necessity
here has only to do with our representation of the occurrence of the
thing and not with that occurrence itself.
It is possible to decide between these two interpretations and we
have every right to acquit skepticism of the accusation of ambiguity
leveled against it.
In the first place there is no passage in the skeptics attesting to
abandoning the principle. If it was abandoned it only would be so by
implication. But Carneades never ceases reminding us ironically that,
as for its contents, the world of the skeptic is not different from that
of the Stoic. "For this wise man of whom I am speaking", says Cicero,
"will behold the sky and earth and sea with the same eyes as the wise
man of your school, and will perceive with the same senses the rest of
the objects that fall under each of them. Yonder sea that now with
the west wind rising looks purple, will look the same to our wise man,
though at the same time he will not 'assent' to the sensation, because
even to ourselves it looked blue just now and tomorrow it will look
grey.-".40
In fact, abandonment of the principle is not required by skepticism.
If in saying that propositions about the future have an immutable
truth-value we are not insisting on necessity but simply defining the
sense of the terms, will it not be the same in the case of saying of what
is that it cannot not be, while it is? The conditional necessity of what
is is a trait of language, not affecting reality itself.

40
Academica, II, XXXIII (105); Brehier, 1962, pp. 235-236.
222 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

There is a passage in the Academica41 that can help us show how

Carneades was able to formally accommodate conditional necessity
without being forced into dogmatism.
As the passage on the sea just quoted shows, all our sense know-
ledge-and it can only be this that is involved where the principle of
conditional necessity is at stake— is given to us by way of appearances
or sense-presentations. A sense-presentation is a witness, veridical or
not, 42 that is to say, naturally and immediately associated with a mark
that makes it the presentation of an object represented. If the object
represented is effectively associated with the presentation, it will be
said, as with the Stoics, that that presentation can be perceived; Zeno
defined it as "a presentation which, coming from a real object, bears
the mark, the impression, the image of the object".43 If the real object
is missing it will be said that the presentation cannot be perceived.
The argument can now be reconstructed as follows.
(1) There is such a thing as a false presentation. Since it is a matter of
sensation here, Epicurus contests this assertion, though it is admitted
by Stoics and Academics alike. Epicurus maintained that if a single
presentation were false the whole of science would collapse, and con-
cluded from this that all presentations are true. The Academics, who
were in agreement with Epicurus on his conditional and with the Stoics
on the existence of the illusions of sense, will prove the impossibility of
science.
(2) A false presentation cannot be perceived. This second principle is
admitted by Epicureans. Stoics and Academics
Let us now reason on the hypothesis:
H. There is a true presentation, A.
Since it is true it can be perceived and a real object corresponds to
it.
It is here that the principle of conditional or hypothetical necessity
comes into play.
N. H. While true presentation A persists, it is necessary that there
exists the real object corresponding to it.
(3) There is no true presentation coming from the senses to which there
can not be compared a presentation differing in no way from it but
which cannot be perceived.44 The Stoics denied this third proposition.
d., XXVI (83); Brehier, 1962, p. 225.
™Ibid., XXV (81); Brehier, 1962, p. 224.
43
/MU, XXIV (77); Brehier, 1962, p. 222.
44
For the sake of the argument I reverse the order here of these last two propositions
reported by Cicero.
 CARNEADES AND THE PRINCIPLE OF CONDITIONAL NECESSITY / 223

To illustrate its meaning we can consider the perception of identical

twins. "If therefore a person looking at Publius Servilius Geminus used
to think it was Quintus, he was encountering a presentation of a sort
that could not be perceived, because there was no mark to distinguish a
true presentation from a false one" .45 If A is a supposed presentation of
a, there is a presentation B as close as one would to A and, ultimately,
indiscernible from A, such that B represents a b different from a. But B
presents itself as a mark or an impression or an image of o and therefore
cannot be perceived. The dispute between Stoic and Academic centers
on this point. As Cicero said "I imagine that Zeno was sharp enough
to see that if a presentation proceeding from a real thing was of such
a nature that one proceeding from a non-existent thing could be of
the same form, there was no presentation that could be perceived".46
"Proceeding from a non-existent thing" here is obviously to be taken
in the sense of proceeding from another thing than from that which is
represented as real. The skeptical position gets its force from the fact
that the Stoics admitted the existence of false presentations.47
In virtue of H and NH, it is possible to compare to A which is
necessary while it is true, a presentation B, indiscernible from A, and
which cannot be perceived.
(4) Among those presentations which do not differ from one another,
it is impossible that some should be perceived and others not. This is
nothing but an application of the principle of indiscernibles to presen-
tations. If A is indiscernible from B, B should have all the properties
A has. Stoics, Epicureans and Academics agree on this.
Therefore, on the supposition of (1), of H and NH, since B cannot
be perceived, A cannot either. If there exists, therefore, a true pre-
sentation and consequently a conditionally necessary one, it cannot be
perceived. What will be possible then is to make conjectures about the
probable, not to give one's assent to the true. There is therefore no
presentation such that a perception of the real follows from it48 and
there is therefore no presentation either such that a perception of what
is conditionally necessary follows from it. What threatened freedom
was dogmatism, not conditional necessity. The latter may be retained
once the former is renounced.

45
Academica, II, XXVI (84); Brehier, 1962, p. 225.
46
Ibid., XXIV (77); Brehier, 1962, p. 222.
"Ibid., XXV (79); Brehier, 1962, p. 223.
48
/&«*., XXXI (99); Brehier, 1962, p. 232. As Hamelin remarks (1978, p. 30),
in agreement with Aristotle on the definition of truth, "Carneades denatures the
notions of truth and falsity".
9

Platonism and Conditional

Necessity.

If it is Aristotle's refutation of the Timaeus that occasioned Diodorus'

formulation of the Master Argument, it is probable that Plato should
have denied, by implication at least, one of the principles advanced by
Aristotle and tacitly used by Diodorus.
A candidate fitting this bill is the principle of conditional necessity,
whose justification is to be found in the analysis of local motion. It
is in reflecting on the soul and its pure movements that Plato and the
Platonists were led to challenge conditional necessity. In another con-
text, but for similar reasons, Duns Scotus systematized these doubts.
We shall also analyze his conception.

9.1 Platonism and the principle of conditional

necessity.
In the Timaeus Plato himself denied the principle of conditional neces-
sity by implication. The world, being produced and composed by the
demiurge out of the primitive disorder, must be susceptible as is every-
thing composite, to being destroyed, even if it is indissoluble except by
the power that united it.1 It is the demiurge's will, constantly oppos-
ing this destructive force in virtue of his love of order, that insures the
perenniality of the world. The demiurge addresses the created gods as
follows: "Gods, children of gods, who are my works, and of whom I am
the artificer and father, my creations are indissoluble, if so I will. All
that is bound may be undone, but only an evil being would wish to
undo that which is harmonious and happy. Wherefore, since ye are but
creatures, ye are not altogether immortal and indissoluble, but ye shall
certainly not be dissolved, nor be liable to the fate of death, having
1
Timaeus, 32C.

225
226 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

in my will a greater and mightier bond than those with which ye were
bound at the time of your birth" .2
Plato, as will Aristotle, posits a sort of permanence of the modal
status of beings. A composite being is by nature perishable and will
therefore always remain so. From the fact that it was engendered it
always will be able to perish. Yet in the case at hand that force of
destruction will never be realized because the will of the demiurge will
constantly keep it in abeyance. So far from excluding the capacity of
not existing at the same moment, as conformity to the principle of
conditional necessity would require, existence is compatible with that
capacity, and in virtue of divine will it will remain so for the rest of
time. The demiurge's address to the gods implies therefore, on the part
of Plato, a revocation of the principle of conditional necessity3 that was
nearly unanimously espoused by the Ancients. This stems, for Plato,
from the gap there is between the ideas and the sensible world. In so
far as it is sensible, necessity characterizes the strictures of mechanism
and materiality and therefore expresses only the nature of the image.4
It would be foolish to conclude from the strictures on the image the
existence of corresponding strictures regarding the ideas.
The Platonic refusal of conditional necessity is therefore indepen-
dent of the particular way in which the creation of the world is inter-
2
Timaeus, 40E-41C; Jowett trans, in Hamilton, 1963, p. 1170.
3
In the Platonic questions, VIII, 1007 (Plutarch's Moralia, XIII, Part I, Cherniss,
1976, I, p. 89), Plutarch writes in the same vein "... as they came into being together
[the universe and time], together they will also be dissolved again if any dissolution
overtake them, for what is subject to generation cannot (be) apart from time just
as what is intelligible cannot apart from eternity either if the latter is always to
remain fixed and the former never to be dissolved in its process of becoming". Since
they were born together the universe and time will perish together. That is to say,
they conserve the capacity of decaying together, even if (by divine will) they are
maintained perpetually in being. The reasoning here differs from the demiurge's
address above only in that here is posited the compatibility of a constant association
of being with a no less constant association of the corresponding contrary capacity.
As Cherniss has pointed out to me, the principle of the conservation of modal status
is implied in a passage of the Laws (818 A7-E2; and Cherniss, 1962, pp. 608-9)
where Plato distinguishes two sorts of necessity: a divine necessity to which the
gods are subject and which has to do with the objects of arithmetic, geometry
and astronomy, and a human necessity having to do with what exists sensibly. A
consequence of this distinction is the invalidation of conditional necessity, for the
demiurge is not subject to "human" necessity.
4
"The nature of this world is blended of intelligence and necessity. What is good
in it comes from god, what is evil from the primordial nature as Plato says for
referring to matter as a simple substance as yet unadorned by god." (Plotinus,
First Ennead, VIII, 7, 4-7). Cf. also Plato, Timaeus, 48A.
 PLATONISM AND THE PRINCIPLE OF CONDITIONAL NECESSITY / 227

preted, which from what Atticus5 says was already a matter of dispute
among the ancients. If, with Atticus himself, the literal interpreta-
tion is adopted,6 the productive power of the demiurge will be placed
above the obstacles materiality set up against it. If, with the Platonic
Statesman, creation is taken in a symbolic sense with the accent put
on an indefinite succession of such "creations",7 the deity will be seen
as indefinitely thwarting the capacity of disorder and destruction that
haunts the images by their very nature. Such a capacity, as will be
pointed out by Aristotle in the De Caelo (281615-25), will have to be
posited at the same time as the contrary actuality, for the ascendancy
of form applies to an infinite time.
That such really was the doctrine of the Platonists finds verification
in a passage of the City of God in which St. Augustine, criticizing
them, shows that starting from their principles one cannot refuse the
5
A second-century A.D. Platonist engaged in the general polemic of the day with
the Aristotelians. Cf. Taylor, 1969, p. 443.
6
On the opposition between Plato and Aristotle, fr. 4 (1977, pp. 50-4 (801a-
804b)).
7
"Then, when a long time had passed, it began to rest from this tumult and
confusion; convulsion gave place to calm, and the world went on its way, settling
down into its normal course with superintendence and sovereignty over itself and all
it contained, mindful to the best of its power of the lessons of its maker and father.
Now it discharged this task at first with much exactness, but more carelessly as time
went on, the reason of this being the corporeal strain inwrought in the primal fabric
of its structure, which, before it was brought to the order we see today, was steeped
in all manner of lawlessness. From its fashioner, indeed, the world has received only
good; all the violence and wrong it has within it or breeds in its creatures, it has as
a relic of that its ancient state. So while the captain was still there to help in rearing
the creatures within it, it brought forth little that was amiss and much that was
good; when parted from him, in the years just after the severance, it still orders all
things excellently well, but in process of time, as forgetfulness comes over it, the old
discord prevails ever more and more till, in the fullness of the days, the world runs
wild, producing little that is good with a great admixture of the contrary, and so
comes in peril of perishing with all its contents. So there comes the moment when
God, who first made an ordered world of it, beholds it in these straits; concerned
that it shall not be broken by the buffetings of the storms of disorder, and so founder
in the boundless ocean of unlikeness, he takes his place again at the tiller, turns
back all that had run to disease and dissolution in the age just past while the world
was left to its own devices, restores its order once more, and makes it safe against
death or senescence." (The Statesman, 273 A-C, trans. Taylor). Mugler, 1960,
p. 170 sq., draws a radical line of distinction between the demiurge's redressment
of the world in The Statesman symbolizing only "the structural information of the
world and the regulative power it has over the diffuse forces of necessity" and the
diachronic intervention of the World-soul of the Timaeus and panpsychism of the
Laws. But then he is obliged to construe the expressions in The Statesman making
mention of the world's memory of the demiurge's instructions, for instance, as a
simple anthropomorphism (p. 192), whereas it may have been for Plato the sign of
the presence of a soul.
228 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

resurrection and eternity of the beatified body. "Will it be said that

this dissolution does not take place [for the created gods] because God,
whose will surmounts all obstacles as Plato says, does not will it so?
Who then is to prevent God from not willing it for earthly bodies too,
since He can ensure that what has a beginning should exist without end,
that what is made up of parts should remain indissoluble, that what
is derived from the elements should not return to them? Why would
He not ensure that earthly bodies be imperishable"?8 As Macrobius
will say, we must distinguish between two different immortalities: one
proper to the soul and belonging to it by nature, the other proper
to the world and thus to bodies which are capable of death but are
defended from it by the intervention of another.9
The Platonists thus challenged the principle of conditional neces-
sity universally admitted by their adversaries. Are we to conclude from
that that their doctrine of creation, taken figuratively or not, opened
the way for the Christian doctrine, and that they alone among the
Ancients established a clear-cut distinction between "real" and "logi-
cal" possibility that would be systematized by Duns Scotus and would
make modal logic as we know it today possible?10 That would mean
ignoring that "real possibility" expresses a plausible view of things. It
is not formal logic but philosophy that explains the different sense that
it took on.
The thesis underlying the principle of conditional necessity for the
adversaries of creation is that of the substantiality of the sensible. How
would a spontaneous and sempiternal movement of heavenly bodies
and atoms be conceivable if the "real possibility" of their rest was sup-
posed? When Aristotle postulates the eternity of the world—identical
to the universe for him—and when Epicurus postulates the eternity of
the universe—compatible with the vicissitudes of the different worlds
according to him—, such an eternity exclusive of all real possibility
of corruption, is in no way distinguishable from perenniality in time,
which is itself indefinite, being nothing more than the measure or ac-
cident of an indefinitely continued movement.11 Suppose then that we
take away every sort of real necessity from corruptible things and from
the particular and partial movement underlying it. In this case, the
s
De Civitate Dei, XIII, 17, St. Augustine, VII, 1695, pp. 336-7.
s
Aut enim idea est immortale quid, quia per se non est capax mortis, aut quia
procurations alterius a morte defenditur (Macrobius, 1952, II. 13).
10
Such is the upshot of Faust's otherwise erudite and penetrating monograph
(1931-1932), but which leaves out Plato and reduces the Platonists to obligatory in-
termediaries between the Greek confusion and the Christian clarity and distinction.
u
Plotinus, Third Ennead, VII, 7-10.
 PLATONISM AND THE PRINCIPLE OF CONDITIONAL NECESSITY / 229

necessity attributed to indefinitely lasting beings (as a result of the

substantiality of the sensible) would have to be completely and uncon-
ditionally withdrawn from generable and corruptible beings. But then
the causality of the heavenly bodies and the atoms would lose its effi-
cacy and would no longer save the phenomena in view of which their
sempiternality had been posited. Although the Stoics admit of no per-
manence with the exception of the divine logos, they nevertheless reach
the same result. For this logos lives indestructibly in a time, conceived
of as an interval of motion. To eliminate conditional necessity, that is
to say, the necessary existence of any finite phase in the history of the
world, would be to destroy divine continuity and to usher the void into
this world.
It is in virtue of the connections he establishes between eternity and
time that the Platonist will refuse these consequences. Since there is
no such thing as sensible substance, we cannot assimilate eternity to
perenniality in time, 12 and the atemporality of the intelligible bears but
a distant analogy to time which, even when universally quantified, 13
affords only a mobile image of itself. Everything in space and time is
generable and corruptible,-and what it is could in no way be founded
on an immaterial sensible, on invisible atoms or even on a divine intra-
mundane breath. Since what is sensible is the image of an extramun-
dane reality extraneous to time, it must be possible for it, so as not to
transgress its status as image, to be other than it is while it is.
Let us begin with the substantial and eternal intelligible. The world
Soul must then produce time and the universe together.14 Hence the
Soul will bear to time the same relation which the intelligible Being
bears to eternity.15 As Plotinus says, "the universal Soul, in producing
the sensible world, moved, not by the intelligible movement, but by one
that is only an image of it, and in striving to render this movement
similar to the first, first rendered itself temporal in engendering time
instead of eternity, and then submitted its work to time in embracing
12
"When we say that Being always is, that there is not a time in which it is and
another in which it is not, it is only for the sake of expressing ourselves more clearly.
In saying always we don't take the word in its absolute sense; but if we use it to
show that Being is incorruptible it may mislead the mind, taking it out of unity to
make it scan the diverse." (Ibid. VII, 6).
13
"But Being, whose nature it is to have no need of the future nor to be related
to another time, whether measurable or indefinite and existing indefinitely, Being,
which is already self-sufficient, is that very Being our intelligence seeks and not a
simple equivocation . Its existence does not come from some quantity or other, but
it exists prior to all quantity and admits of no sort of quantity in itself." (Ibid. VII,
6).
14
Tmiaeus, 38.
15
Plotinus, Third Ennead, VII, 11.
230 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

the whole of existence and all the revolutions of the world within it" ,16
Suppose then that the Soul ceases to act and to exercise its power and
withdraws within itself and in eternity. Succession, time, all images
would disappear at once.17 Such is the secondary being of images, of
time, and therefore of sensible movement and of all that partakes of
quantity. There is a possible world in which the Soul would return
into its life of pure contemplation and where that secondary being
would disappear. It must therefore not be necessary. Even the sensible
things that will last forever, dependent as they are on the activity of the
Soul, are without necessity in spite of their perenniality.18 There is no
absolute sensible necessity. A fortiori there is no conditional sensible
necessity.

9.2 Consequences of the connection between

conditional necessity and the substantiality of
the sensible for modality, causality and
freedom.
To admit conditional necessity is to suppose the substantiality of the
sensible. The only motion one is then obliged to admit is sensible
motion. Aristotle says: "there is no such thing as motion over and
above the things". 19 Both Stoics and Epicureans followed him on this
point. The analysis in this paragraph will be limited to Aristotle's
own doctrine and to that of the Peripatetics in general, since they are
the ones who had to systematically articulate the consequences of a
decidedly anti-Platonic principle and the consequences to be expected
from a return to Platonism as well.
If all motion is sensible and subject to the principle of conditional
necessity then there is a real distinction, i.e. an incompatibility, be-
tween real potentiality and achieved reality. For in virtue of condi-
tional necessity, a substance at time i has the capacity of being at a
later time t' in a different situation or state from the one it is presently
in at t. This capacity of being belongs to substance therefore only in
the form of privation. "For each thing of this kind is capable of being
at one time actual, at another not. Take for instance the buildable as
buildable. The actuality of the buildable as buildable is the process of
building. For the actuality of the buildable must be either this or the
I6
lbid., VII, 11.
17
Ibid., VII, 12.
18
Ibid., VII, 12. "The universal sphere would not exist either as it does not exist
prior to time, because it is in time that it exists and moves."
19
Physics, III, 1, 200632.
CONDITIONAL NECESSITY AND THE SUBSTANTIALITY OF THE SENSIBLE / 231

house. But when there is a house, the buildable is no longer buildable.

On the other hand, it is the buildable which is being built. The process
then of being built must be the kind of actuality required. But build-
ing is a kind of motion...".20 Since the formulation expressing the real
potential contains both the present privation and the present capacity
of future realization, it would be contradictory to posit that realization
presently.
This disposition of the modalities making motion a real 'imper-
fect'21 actuality, subjects the universe to an extrinsic causality. Take
the point of view of the movable. It is presently deprived of a property
or a situation that it is capable of having later. Since what it is in
actuality now is only a privation with respect to what it is in potency,
this cannot be the cause that will actualize what it potentially is, and
the onset that will be the source of that actualization has its origin in
something else. The moved which is deprived of what it is in potency
and the movent that actually possesses what that potentiality is a
potency for 22 are not identical then, unless it be accidentally. In other
words, the movable, in so far as it is moved, is not to be confused with
what moves it, otherwise there would be complete actuality and not
motion. But if that is the case, then everything that is moved is moved
20
Ibid., Ill, 1, 201*7-13.
21
Ibid., Ill, 2, 201!>29. Motion is an actuality, though an imperfect one. Saint
Thomas comments (Marietti, 1965, 2305, p. 546): "The reason is that it is the
actuality of something incomplete, that is to say, of what is possible or what is in
potency. For if it were a complete actuality it would eliminate any potentiality,
which is the capacity in matter for a given determination. That is why complete
actuality is the actuality of what is in act and not of what is potentially. But motion
is the actuality of something potential since it does not suppress the potentiality.
In so far as there is motion, there is a potentiality in what is moving for that to
which it tends by way of the motion. But it is only the potentiality that was the
potentiality of being moved that motion suppresses. And even this is not totally
suppressed since the moved is still capable of being moved, given that all that is
moved will be moved in virtue of the division of continuous motion".
22
Physics, VIII, 5, 25769-10. "The movent on the other hand is already in activity:
e.g. it is that which is hot that produces heat: in fact, that which produces the
form is always something that possesses it."
232 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

by something else. There is no self-mover.23 The known desirable is

an unmoved mover; desire is a moved mover.24
The nature attributed to the soul derives from these principles. The
soul is not a self-mover. Nor can it be moved except accidentally, for
otherwise, since all motion is divisible, its simplicity would disappear
along with its essential immobility.25 For the human soul, sensation
and knowledge are to be accounted for by its being moved to what it is
only potentially by an agent that is actually the object of that poten-
tiality. For sensation, which is potentially all the sensible forms, this
actualizing agent is provided by the external sensible objects. In the
case of thought, it is necessary to posit an internal efficient cause to
actualize all the intelligible forms of which the agent is capable. Such
is the agent intellect—whatever difficulties for Aristotelianism may be
entailed by the necessary actuality of the forms through which that in-
tellect produces knowledge.26 The freedom of such a soul, if admitted,
23
Physics, VIII, 5, 257a31-!>13.
"Aristotle's reason for refusing to be content, as Plato was, with the notion of a
self-mover, is that in so far as it moves, it must already have a certain character,
while in so far as it is moved, it must have that character only potentially, and
actually not have it. E.g. that which warms itself must be warm in order to impart
warmth, and cold in order to receive it. The law of contradiction, therefore, forces
us to analyze the self-warming into a part which is warm and a part which is cold,
i.e. self-imposed change turns out to be change imposed by one thing on another."
(Ross, Aristotle's Metaphysics, I, CXXXIII, n.5).
The formula for the moved at t is:

At the same moment t the movent has the property expressed by 'pt'- What leads
to the contradiction pointed out by Ross is conditional necessity, since if the formula
for the moved at t could be
Mtpt • ~ pt ,
there would be no need to appeal to a movent other than itself for it to produce
the action. Aristotle's argument in Physics VIII, 5 is a dilemma (Cherniss, 1962,
pp. 390-1, n.210). The continuity and therefore the divisibility of motion renders
a self-mover impossible. Either the self-mover moves itself as a whole, but then
it would both undergo and cause the same motion at the same time, and motion
would no longer be an incomplete actuality. Or self-motion would result from a
partial motion which would be reciprocal (with A moving B and B moving A, in
effect reproducing the previous confusion of action and passion) or ordered. If self-
motion is partial and ordered, a part A will move a part B without being moved
by it. But if A moves itself in moving B the same difficulty will arise again: A
would undergo the motion it causes. Therefore A must move B while remaining
immobile itself. The mainstay of the argument is the essential separation of movent
and moved, in other words, conditional necessity. For the Thomistic doctrine on
this point see Effler, 1965, pp. 180-91.
24
Saint Thomas, Summa Theologiae, Part I, quest. 80, art. 2; Gilson, 1952, p. 583.
25
De Anima, 408a34-(>30; Cherniss, 1962, p. 402 sq.
26
I'm following Cherniss quite literally here, ibid., p. 169.
 THE CONSEQUENCES OF ABANDONING THE PRINCIPLES / 233

cannot be conceived of as a capacity for non-motivated decision. Free

action then will consist simply in the play of intellectual representa-
tions informing the will, the explicative motive of our actions being the
representation of the good.

9.3 The consequences of abandoning the principle

of conditional necessity and the substantiality
of the sensible world for the Platonic and
Platonistic theories of modality, causality and
freedom. The same abandonment entails
similar consequences for Duns Scotus.
For Plato, sensible motion, which is attributed to the unorganized ag-
itation of the receptacle, far from representing the actuality, even if
only imperfect, of substances, expresses the simple deficiency that pre-
vents the sensible image from reflecting the idea exactly. It lacks the
minimum consistency required of an image for reflecting an idea. By
contrast, there are other motions of a spiritual nature—thoughts, de-
sires and feelings—which are characteristic of the human or universal
soul and prior with respect to physical motions.27 Soul is at once con-
templation of the ideas and principle of motion. Its motive force, which
is immediately manifest in the case of spiritual motions, mediately so
in that of sensible ones, depends for its regeneration on the energy got
from contemplating the ideas. It can also happen that the soul, which
normally animates a body, spends at least some time in the purely in-
telligible world. At such a time it exercises its contemplative function
without exercising its motive function; and this seems to establish a hi-
erarchy between the two functions. Even within the act of intellection
a distinction must be made between thought in motion and thought
at rest once contact with the intelligible is attained. It is this spiritual
motion and rest, recognized as positive phenomena, that must reflect
the two ideas of motion and rest.28
The Platonists will modify this doctrine. Plotinus, for example,
insisting on the superiority of the contemplative life of the soul over its
motive activity, will put the accent on the insubstantial character of
27
For instance Laws, 898 C; Cornford, 1960 (1935), p. 246; Cherniss, 1962, p. 402.
28
The Sophist, 254 D-F; Parmenides, 129 D-E; Cherniss, 1962, p. 439.
234 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

the sensible.29 The invalidity of conditional necessity here will be all

the more evident.30
29
The organizational activity of the soul presupposes an immutable contemplation
of order (Brehier, 1961, p. 34). "The part of the soul which is the first is on high;
always near the summit in an eternal plenitude and illumination, it remains there
and participates, the first, in the intelligible; the second part of the soul, which
participates in the first, proceeds eternally, a second life issuing from the first,
activity projecting in all directions and nowhere absent. In proceeding, the soul
leaves its superior part in the realm of the intelligible, that the inferior part leaves;
for if its course caused it to abandon that superior part it would no longer be
everywhere but only in the place its course reached." (Third Ennead, VIII, 5). In
comparing the life of the One to a cosmic tree, Plotinus distinguishes the principle of
life which remains in the roots, immobile, from the life which circulates throughout
the whole tree (Ibid., 10). The same metaphor can apply to the two "parts" of the
soul. It is because it lives in intelligible eternity, out of all time and space, that the
soul can act ubiquitously and animate an extended body. It is only the successive
manifestations of this body in sensible motion that are subject to the principle of
conditional necessity.
This is because, following the Platonic teaching, motion, which is a supreme
intelligible genus along with rest (Sixth Ennead, II, 7), exists qua intelligible prior
to time and to space (Ibid., II, 16). For an idea could not admit of the conditions of
possibility and a fortiori of definition that characterize only sensible being. "The
Motion which acts upon Sensible objects enters from without, and so shakes, drives,
rouses and thrusts its participants that they may neither rest nor preserve their
identity—and all to the end that they may be caught into that restlessness, that
flustering excitability which is but an image of Life.
"We must avoid identifying Motion with the objects moved: by walking we do
not mean the feet but the activity springing from a potentiality in the feet. Since
the potentiality is invisible, we see of necessity only the active feet—that is to say,
not feet simply, as would be the case if they were at rest, but something besides
feet, something invisible but indirectly seen as an accompaniment by the fact that
we observe the fact to be in ever-changing positions and no longer at rest." (Sixth
Ennead, III, 23, trans. Stephen MacKenna).
The relation of the sensible to the intelligible is therefore comparable to the
relation between the two functions of the soul. The soul moves but even in moving
remains in contemplation of the idea. The self-mover, which is none other than the
soul then, 'comes' to the sensible but does not 'remain' in it. Motion goes toward
the mobile but does not inhere in it as to be severed from its mover, the invisible
soul, that diffuses it like a breath upon the agent it animates.
30
The categories of the sensible world are not to be conceived of as being on the
same simple pattern as those of the intelligible world. In the intelligible world
motion and rest are two distinct and even contrary supreme genera; soul and idea,
whose meeting is defined by contemplation, both belong to the realm of being.
But is there any ground for positing the same opposition or even distinction in the
sensible world? Are sensible immobility (i^pE(jua) and motion opposed in the same
way that intelligible repose (cruaaK;) and motion are? Not at all.
"Stability in that [the intelligible] realm does not arise from the fact that what
is of a nature to move is not moving; what is There is stable because Stability
has taken hold of it; in so far as it has Motion, it will never cease to move: thus,
it is stationary under the influence of Stability, and moves under the influence of
Motion. In the lower realm, too, a thing moves in virtue of Motion, but its Rest
 THE CONSEQUENCES OF ABANDONING THE PRINCIPLES / 235

Duns Scotus' system lies outside the orthodox Platonic tradition,

and is even radically hostile to it. He never relents criticizing Avicenna,
the neo-Platonist for whom salvation is to be expected from philosophy
as a matter of course. It is not that Avicenna is wrong in appealing to
the dignity of human knowledge, to our soul's affinity with the intelli-
gible, in short, to its need for God. The error lies in attributing to the
possible intellect a natural light which it does not have in act pro statu
isto but for which it is owing to Christian revelation and theology.31
Aristotle described man as he is, Avicenna as he ought to be. The
first was unaware of grace; the second confused it with nature. Both
were unaware of freedom, the one confining it to the miscarriages of
nature, the other to the supposed necessity of creation. But in dissi-
pating the Platonic illusion that attributes a supernature to nature,
Scotus denounces Aristotle's mistake of reducing nature to its present
state and exalting that state to the level of essence. For the Platonists,
the invalidity of conditional necessity required the compatibility of the
actual with the simultaneous contrary potential which transcended na-
ture. With Scotus this requirement finds a ground which resides in the
excess of theology over metaphysics.
Saint Augustine extended to the union of human body and soul
that perpetuity that the Platonic demiurge had established for the
union of the world soul and the world. What is dissoluble by nature
was therefore to coexist with an elective indissolubility. Scotus takes
over the same thesis in other terms. To deny the validity of conditional
necessity is to assert that a state of fact brought about by God can
forever defy the natural possibility of the contrary state. But if we are
bound by theology to believe in the reality of that state of fact, we
is caused by a deficiency; it has been deprived of its due motion." (Plotinus, Sixth
Ennead, III, 27, trans. S. MacKenna, pp. 517-8).
Sensible immobility is therefore a limit of sensible motion: it is not its contrary.
But in so far as the motion is an ordered and regular one, in so far as it is of the sort
that only a soul can impart to a body, we find Plotinus in the same situation as that
described by Plato in the Timaeus for the motion of the Heavens. This motion can
come to an end in virtue of the possible dissolution of the world, but the Demiurge
does not allow that dissolution. For Plotinus, the case is the same for the organized
movements of a living being: they can cease, but in so far as animated by the soul,
the soul will not allow them to cease. The inert world of Platonic necessity has been
compared to entropy (Mugler, op. cit.): the demiurge must constantly reintroduce
information to maintain the world in dynamic equilibrium. The Plotinian soul
exercises the same office. The universe is constantly prey to a possible dissolution
or immobilization; but the energy supplement afforded it by the soul, nurtured on
ideas, prevents that's happening. There is a possible then that is never realized,
because of the will of the demiurge and the soul; and as it never can be realized it
must coexist in the instant with the contrary actuality.
31
Gilson, 1952, pp. 17-24, pp. 38-43.
236 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

are bound by logic to postulate that such a state is possible by na-

ture. There is therefore in the human composite, just as in the cosmic
one, a natural passive potential for incorruptibility. No natural ac-
tive potentiality however comes to activate that passive one. From the
philosophical point of view then that passive potentiality will remain a
dead letter. The theologian, whose views are not limited simply to na-
ture, postulates then that to every passive potentiality—understanding
by that what a thing is capable of at best, given its essence as union
with the soul for the body—there corresponds a natural active or free
potentiality.32 In short, what thwarts conditional necessity are perfec-
tions not inscribed in the nature of beings; and the passive possibilities
instituted by these perfections will only become real through the actu-
ality of a free cause.
Instead of invoking spiritual motions as does Plato in saying what it
is that retains Socrates in his prison, the point of departure for Scotus
will be the passive potentialities revealed in contingent unions. The su-
perior part of something contingent will include the passive potentiality
of not being subject to the constraints of the union. Thus, even if the
human soul was created in union with the body, it must be conceived
of as capable of being created for itself in a first instant of nature,
with its role as part of the union relegated to a second instant.33 As
for the inferior part, the fact of the union assures it of its capacity of
being united to the superior part. Since a free active potency is re-
quired for actualizing these passive capacities, the immortality of the
soul as well as the resurrection of the body will depend on such. The
theologian makes the metaphysician stipulate the passive capacities of
beings. Matter itself as receptive of form is not reducible to a simple
indeterminate capacity. Its natural inclination to another form confers
positive being on it.34 From this disposition results the multiplication
of formal distinctions. The body, for instance, must be capable of being
32
"According to the Theologians, to say that there is an active natural potentiality
corresponding to every passive natural one is to say something false, for the nature
of the higher beings is capable of more perfection than that within the power of
active natural potentiality. Nor does this passive potentiality exist in vain, for it
can be actualized by a free agent as well as by a natural one; and if it is a matter of
acting outside of itself the free agent is even more efficacious and has more power
than the natural one, since the free agent is infinite, which the natural agent is
not. That proposition must therefore be taken as follows: to every natural passive
potentiality there corresponds either a natural or a free potentiality that reduces it
to act; and to this I agree." (Rep. Par., L IV, d.43, q.3, n.18; Gilson, 1952, p. 645;
Vives, XXIV, p. 519).
33
Gilson, 1952, pp. 480-1; see 4.2, p. 80 sq.
34
Op. Ox., L II, d.12, q.l, n.ll, Vives XII, p. 558; Rep. Par., I II, d.12, q.l, n.10,
Vives, XXIII, p. 6.
 THE CONSEQUENCES OF ABANDONING THE PRINCIPLES / 237

united to the intellective soul which is its 'essential form'. But there
is such a gulf between them that an 'organic bodily form' seems nec-
essary as the ultimate disposition of matter to receive the intellect.35
The multiplication of forms will result from the requisite specification
of every potentiality become positive. But the hierarchy ordering that
matter and those forms increases, rather than diminishes the distance
between them. As strong as the bond between them is, there is nothing
that could prevent God from creating matter without form. For Plato
the bonds established by the demiurge's creation took precedence over
the constraints of the sensible world. Divine power, for Scotus, takes
precedence over the bonds formed by creation. How could conditional
necessity subsist in the face of that power?
In connecting sensible motion with a mobile, Aristotle is drawing
the consequence of his modal definition of motion as imperfect act,
identical with the process of realizing form in matter. But let us, with
Plato, consider the motion of things as secondary. A thing can be
moved by other things that move, but the ultimate cause of its motion
cannot exist in any other thing.36 We must posit its model in the idea
of motion and its realization in the life of the soul. But if motion cannot
be reduced to any other given than motion itself, there is no longer a
reason for its modal analysis in terms of imperfect actuality. On the one
hand, the idea of motion is pure actuality in virtue of its immateriality.
On the other hand, the soul, which takes its model from that idea
and produces the spiritual motion and secondarily the physical one, is
free from all potentiality for it is not a subject susceptible of contrary
determinations: it is the self-mover in act. Only physical motion then
falls under the Aristotelian analysis. It alone requires the distinction
between the mover and the moved.37
Scotus has similar thoughts on the matter. He says that it seems
impossible for an angel to move if the motor and mobile are always dif-
ferent and if what is in act cannot be in potency. But the impossibility
is only an apparent one stemming from the application to spiritual
things of what is only valid for material things in general.38 We must
Par., L IV, d.ll, q.3, n.22, Vives, XXIV, pp. 125-6.
36
Cherniss, 1962, p. 453.
37
Ibid., p. 441, p. 453.
38
Op. Ox., L II, d.2, q.10, n.l, Vives, XI, p. 523; ibid., d.25, q.l, n.12, Vives, XIII,
pp. 207-8. "When it is said that mover and moved must necessarily be distinct by
their subject, what is said is true only of corporeal things: I think that even here
moreover it is not necessarily true. But I say that it is simply false for spiritual
things. Otherwise God could not create a single Angel, bare in its nature, and
which, left to its nature, could comprehend its essence, thereby being a same mover
and moved indistinctly as to the subject."
238 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

distinguish between the modal incompatibility and the substantial in-

compatibility of potency and act in the instant.39 As a mode, potency
has to do with a being that doesn't actually exist, but that can exist. In
this case there is incompatibility between being potentially and being
actually at the same instant.40 By contrast, as substance or principle,
the potency is either a material or receptive cause or an efficient cause.
But once the causal point of view is taken the incompatibility of po-
tency and act no longer holds in general. Potentiality and actuality can
coincide in the same instant precisely because they are not superposed
exactly in all respects. It is only in univocal agents that there is an
incompatibility between potency and act. In equivocal agents on the
other hand this does not obtain.41 For as regards these, the subject in
39
Effler, 1965, p. 182.
40
Op. Ox., L. II, d.25, q.l, n.12, Vives XIII, p. 208. "When you say that this is the
agent in act, that the patient in potency, etc., I say that if act and potency are taken
as two first differences of being, they divide also everything that is. Consequently,
they divide any given numerically identical being. Thus it is a contradiction for one
and the same thing to be simultaneously in act and in potency relatively to a same
thing, because the potency, in so far as it is a distinct difference of being from act,
is necessarily inclusive of the opposition or negation of that with respect to which
it is posited a potency, as the white in potency is not the white in act, as long as
it is in potency. And so it is impossible that something thus existing in potency
should actualize itself."
41
Ibid., n.13, p. 208. "But act and potency may be taken in an entirely different
way, namely as dividing the potency or the active principle into the univocal and
the equivocal. For as regards univocal agents it is true that one is the patient
in potency, another the agent in act, and it is therefore impossible that a same
thing should be both formally such-and-so in act and simultaneously such-and-so
in potency. As regards equivocal action however, this need never be the case, for
here the agent must be more noble and virtually so."
Consider the example of 'intensive quality', i.e., of the intension and remission
of forms. Albert the Great and St. Thomas had held that in the case of intension
or remission the original form or degree was destroyed and a new form, more or
less intense, was created in its place. Their position was based specifically on
the principle of conditional necessity. The moments of motion are incompatible:
"each ubi is destroyed when another ubi is occupied" (Maier, 1951, pp. 60-1). It
is this analogy with local motion that Duns Scotus combats when he asks whether
the pre-existing charity perishes in its entirety when a different degree of charity
is attained, so that there is no numerically identical reality that remains in the
different degrees of charity (op. cit.). The subject receiving a form in different
degrees is equivocally one: it is more or less perfect. What it has then to one
degree it can, at the same moment, have to a higher degree. Consequently, "... if
you said that nothing can be identically and at the same time potentially other and
have in actuality that other thing either eminently or virtually, I say that that is
false, for that comes down to saying that such a thing is not capable of its perfection.
For we see that nature has given to that to which it gave the potentiality of being
augmented the active potency of effecting the augmentation and of sustaining it
that it might be conserved" (Scotus, ibid., p. 207). The intellective soul, therefore,
has the potentiality of different perfections which are not distinct as to the subject.
 THE CONSEQUENCES OF ABANDONING THE PRINCIPLES / 239

potency and subject in act are only equivocally the same, as is required
by the imputability of action.
The objection against the existence of a self-mover is therefore
removed.42 "Soul, however, as self-motion in which there is no dis-
tinction of substrate and activity can have no potency of motion in
this Aristotelian sense which involves the possibility of contrary deter-
minations but as a positive process must have a pattern in reality of
which it is the manifestation" ,43 The life of the soul is not the attribute
of a substance. We must not say that soul moves, but rather that it is
self-mover or self-motion.
Once the principle of omne agens est praesens passo is given up,
the way is clear for the freedom of the human will.44 The proposition
assigning an external cause to every volition had been condemned45
in 1277. The faith posits that "nothing other than the will is total
cause of the will in the will" ,46 The will continues therefore to exist
"Therefore it is absurd that a very noble form, such as the intellective soul, should
not have the active as well as the receptive potentialities of its accidental perfection.
And because in such forms there cannot be given an active and passive potency
distinct from the subject, as they are not organic potencies, they are therefore not
distinct by the subject and will therefore be united here without distinction as to
the subject, without however being deprived of formal distinction." (Ibid., n.13,
Vives XIII, p. 208).
42
Duns Scotus, ibid., n.14, pp. 208-9 "Likewise, it has been shown in book four
that a separated soul has the power of moving itself to another place, for to move
thus is suited to a very imperfect being, and what are distinct perfections in an
inferior nature must be unified in a more perfect nature to which the inferior one
is ordered."
43
Cherniss, 1962, pp. 441-442.
44
Op. Ox., L II, d.25, q.l, n.2, Vives XIII, pp. 197-198.
45
Gilson, 1952, pp. 377-379.
46
Op. Ox., L.II, d.25, q.l, n.22, Vives XIII, p. 221. The condemnation was of
"the proposition that the soul wills nothing unless it is moved by something else.
Whence it follows that it is false that the soul moves itself. This is wrong if we
understand moved by something else, namely the desirable or the object, to mean
that the desirable or the object is the entire reason for the movement of the will."
That the soul is necessarily moved by something else was the position of Scotus'
adversary, Godfrey of Fontaines. From this argument he distinguishes that of "an-
other Doctor" who is St. Thomas: "As concerns the proof that it is the same thing
to say that mover and moved are indistinct as to subject as to say that there is
a self-mover, it seems that another Doctor would concede that the will is moved
by an object in so far as apprehended by the intellect, and yet that the object as
apprehended by the intellect or the intellect showing said object are not distinct
from the will as to subject; and yet he would not concede that mover and moved
here are one and the same thing." Scotus refutes both opinions. "As for me, I say
not only that they are indistinct as to subject, but that one and the same thing
simply can be mover and moved." (Op. Ox., L.II, d.25, q.l, n.12, Vives XIII, p. 208;
see above Chapter 4.2, note 30, p. 81).
240 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

as a power at the same moment in which it produces its acts. For it

is the equivocal, not the univocal, cause of the acts of volition, being
more noble and eminent than they. Abandonment of the principle of
conditional necessity and of its consequence, the principle of external
causality, revolutionizes the theory of freedom in substantializing the
will as the equivocal and autonomous principle of its acts. Conditional
necessity, even there where it does apply, i. e. in the world of phenome-
nal effects, does not then have an absolute validity, but only a validity
secundum quid.47 It is a sort of necessity taken on contingently by
what is contingent by nature.
The Master Argument is ineffectual here. Duns Scotus and Plato
do give different explanations of the phenomenal world, and no matter
how the descent of the ideas into things is represented it is not com-
parable to the contingent creation of the Christian God. But the Idea
does have this much in common with the Creator: what necessity there
is, whether logical or real, remains within it; and whatever is produced
outside of it by any mode whatsoever, i.e. be it necessarily or con-
tingently, is itself devoid of necessity. Conditional necessity belongs to
the laws of nature; but the laws of nature are decidedly contingent.
Aristotle's hylemorphism supposed the validity of the principle of
conditional necessity: (1) the ideas are never separated from the sen-
sible, (2) every motion is a sensible motion, (3) the mover is never
identical with the moved, (4) self-motion is a contradictory concept,
(5) the hierarchy of the causes within motion points at a prime, im-
mobile mover, (6) finally every motion is the act of a potency as such
(potential teleology).
Abandoning the principle of conditional necessity (1) makes sense
of a world of separate ideas or of separate or separable spiritual sub-
47
Gilson, 1952, p. 586. The consequences of that assertion have been examined.
To admit extrinsic causality only is to leave the way open to infinite regression. "In
accepting that the first action in the will should be caused by an object, whatever
be the object that must be posited according to you, that action is purely natural.
Therefore it is not within the power of the will. And indeed as Augustine says, it
is not in our power not to be touched by what we see. If therefore after that act
I can move my intellect to consider this or that, I ask by what act. Not by that
which I have just spoken of, since it is not within the power of the will. Therefore
it must be by another. I ask where that other is from, if it is through the will itself
or through an object or phantasm? If it is through the will, then I have what I
advanced since that will is within its power and effectively is through it, and for the
same reason the same is true of the first volition. Or again it is through something
else, namely an object or a phantasm. And if this is so it will then be a natural act
and consequently, it will be no more within the power of the will to thus command
the intellect to consider this or that than it was to produce the first act" (Op. Ox.,
L. II, d. 25, q. 1, n. 7, Vives XIII, p. 201).
 THE CONSEQUENCES OF ABANDONING THE PRINCIPLES / 241

stances,(2) pure spiritual motions then become the paradigms, for

which (3) the mover can be identical with the moved, (4) self-motion
defines freedom, (5) the hierarchy of the ideas or of the spiritual sub-
stances supposes a system of reflection or emanation or creation from
the One or the Perfect Being to the least images or creatures, (6) finally
every motion is the act of such a power (actual teleology).
10

Epilogue

To write the history of a problem is to reconstruct, insofar as possible,

the arguments whose conjunction it is that raises the difficulty and to
examine the solutions that have been proposed in challenging either
one of these arguments or an implicit and generally accepted premise.
The historian, as such, does not judge history. But to refuse giving
an epilogue would be to do violence to thought. Was the problem raised
a genuine one and were the solutions proposed in fact acceptable ones?
An epilogue is not a conclusion. It goes beyond the limits of what
is strictly historical. Once this step has been taken we should not be
in the least surprised to find ourselves at grips with the philosophers'
disputes.
The historical analysis has allowed us to conclude that the Master
Argument rests on the incompatibility of three premises (four if we
distinguish conditional necessity from the contraction of the diachron-
ically possible) which all have to do more with physics than with logic,
since they are all analyses of movement in the most general sense, that
is to say, of change. The clear sign of this is the double temporal index
affecting them all.
This being the case, one might expect on the part of the philosopher
a properly physical inquiry upon which to base his judgment. It is in
this way that we shall proceed.
There are two things however that complicate matters here. The
first is in connection with the very nature of the Master Argument.
For the movements and capacities that the argument addresses itself
to in the first place are "rational" movements and capacities that bring
deliberation into play. The study of such objects as these belongs
to ethics before belonging to physics. Let us note nevertheless that
human decisions inevitably fit themselves into the course of nature,
secondly that nothing prevents nature, insofar as it is recognized to be

243
244 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

contingent, from posing the same problem for analysis that freedom
does, and finally, that the naturally contingent, without pretending to
impose its own particular features on freedom, may suggest analogical
ways of rendering this latter possible.
The second source of complication, extraneous to the Master Ar-
gument, is one of circumstance. The majority of philosophers today,
either because of their lack of familiarity with the sciences or because
they deny that there is any separation between scientific and ordi-
nary languages, postulate that it is sufficient to correctly analyze the
common language with which they express themselves and which has
incorporated wisdom and truth in order to dissipate the confusions, the
paradoxes, the pseudo problems and imagined solutions transmitted by
the history of philosophy.
Aristotle, already, led into the real examination of a question by
a preliminary linguistic or dialectical examination. Let us follow his
lead.

10.1 The impasse of natural language.

Faced with the three premises of the Master Argument, common sense,
along with the philosophers who follow it, rejects the solution of Clean-
thes and that of Diodorus out of hand, since natural language seems to
block any application of the possible to the past and to require in the
domain of what is naturally contingent, just as well as in that of human
decision, that there be some possibles left unrealized. Little inclined
to enter into the logical quibbles of Chrysippus or to flee towards the
Platonic ideas, not to mention the other solutions of the paradox, given
that all these solutions do violence to common sense and to language,
they are invariably brought back to the second premise. Brief exam-
ination makes them reject it. Don't we ordinarily say, after all, that
what is possible today for us to do the day after tomorrow will turn
out tomorrow to be no longer possible?1
On the other hand, the axiom of conditional necessity must be re-
jected because it precludes consideration of the only thing upon which
we act, of the somewhat extended or specious present, in favor of ab-
stractions such as those of spatio-temporal points of situation, those of
the instantaneous velocities and impulsions of kinematics or dynamics,
all illusions, that obstruct the experience of freedom.
The first objection seems to be so strong that one wonders why
Chrysippus, who goes to such lengths to avoid the premise of contrac-
tion did not think of it. But then just what was the sense of the word
above, 5.5, pp. 124-125.
 THE IMPASSE OF NATURAL LANGUAGE / 245

possible in the context of physics about which the Ancients reason? The
analysis of the Chrysippean modalities has shown what it was. The ab-
sence of internal contradiction is not generally sufficient for recognizing
a possible; there must be conditions of reality as well, conditions about
which the philosophers are not in agreement but which require, in any
case, that the realization of the possible should not be prevented by
obvious external hindrances. As Aristotle puts it, "a thing is capable
of doing something if there will be nothing impossible in its having the
actuality of that of which it is said to have the capacity".2 As the
condition has to do with a being that, not walking, has the capacity
of walking, or that, walking, has the capacity of not walking, the most
general impossibility we can imagine is that of a possible that would
come to be contracted and realized at the very instant at which the
contradictory state of affairs would be the case.
Let us then follow the suggestions of ordinary language. Let us
forget the requirements of reality that the Ancients attached to the
notion of the possible. Let us consider now a possible at t, but such
that from an instant t' < t on it will be clear that it will not be reduced
to a contracted possible: we will not have Mtipt>. Since the contraction
of this possible between now and t' has not even been envisaged, this
amounts to a denial of premise (B). We then posit, subject to the
validity of (A):
(A) and (~B) (3t) ~LN~pt- (t^N < ti < t D~ MtlPtl).
What is the meaning of this assertion that formally contradicts Aris-
totle? It will be seen in examining the two limiting cases that arise
when N and t are identified and when t\ = t is precisely fixed. In the
first case we reach a contradiction: pN is impossible and its negation is
not necessary. In the second case we assert, for example, that it is not
impossible, if it is 6 o'clock, that I should not take a certain train at
8:02, although at no intermediate instant am I faced with the concrete
possibility of taking it. In short, the event considered is a verbal pos-
sible, a word that makes no commitment. For Aristotle, for Diodorus,
for Chrysippus, even though there is no logical contradiction in the
concept this word expresses, even if we consider the concept possible,
it "is" not a possible as long as the impossibility of the contraction at
the stated instant is not excluded and as long as the diverse hindrances
that are liable to intervene have not been removed.
This is what we sometimes express in explicitly adding the clause
of non-hindrance to the statement of diachronic possibility. Aristotle,
2
Metaphysics, Q, 3, 1047°24-25.
246 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

who knows what saying something means, has been seen to mention
this given.3
A being has a potency insofar as this is a potency to act, not an
absolute power, but subject to certain conditions, among which will
be included the absence of external obstacles. For a possible to be
authenticated such, the hindrances to its realization must first be ruled
out. It is because it receives them instead of ruling them out that the
first objection invalidates (B) so readily.
Henri Bergson and William James, in the name of metaphysics
and psychology, and often in thinking that they were criticizing logic
and everyday language, defended an idea similar to that expressed by
the objection made against conditional necessity. It may be doubted
whether the testimony of language can be cited in an unequivocal fash-
ion on the matter. Aristotle, in any case, to whom we owe the dis-
tinction between the limits and the parts of time and who on other
occasions rested arguments on the distinction, did not look to it for a
solution to the aporia on freedom but reserved the use of it to extri-
cating us from the labyrinth of the continuum. The fact is that the
axiom of conditional necessity does not seem on its own to commit
us to an interpretation in terms of instants. All that is required by
contraction is the identity of the two indices. While and for as long as
the possible lasts, the actuality lasts. The validity of the axiom in no
way depends on the particular conception one adopts as regards the
relation between the continuum and its elements.

10.2 A probabilistic reconstruction of the Master

Argument: Diodorus' solution.
Let us leave grammar behind and go on to physics.
The aporia of Diodorus imposes no condition relative to the distinc-
tion between natural events and human decisions. When determinism
and universal causality are assumed, as they are in classical physics,
the most natural and simplest solution seems to be that proposed by
Diodorus. From this point of view, a possible destined not to be real-
ized is not a possible, but only a word.
But since the aporia imposes no conditions either as regards de-
terminism or causality, it must be applicable to all events, including
those we take to be contingent and to which we are able at most to
attribute a probability. It can be seen from statistical mechanics, and
from the reflections of a Laplace before that, that classical physics
does not reject such a notion. We have shown that it was called for
3
Metaphysics, 0, 5, 1048°16-21, quoted on page 125.
 A PROBABILISTIC RECONSTRUCTION OF THE MASTER ARGUMENT / 247

in ancient physics by how dangerously close Aristotle's position came

to Diodorus'.4 That is why we had to try to interpret the Aristotelian
notion of contingency in terms of probability.
One of the reasons that kept Aristotle from reaching the notion
of probability lies in the special teratological status he accords to the
notion of "rare" event.5 To profitably investigate the relations between
the possible and the probable, while at the same time giving all their
weight to the objections raised against the Master Argument's second
premise, it will be of interest to reflect on the Poisson probabilities
and the Poisson law, which have to do specifically with such events as
the emission of a radioactive particle, the arrival of a customer at a
ticket window, the occurrence of a typing error on a printed page. The
Poisson model applies when the three following conditions are satisfied:
I/ The numbers of events considered in a given region are indepen-
dent (The region in the example to be taken up here is a time interval
with the year as unit).
2/ The probability of an event's occurring in a region of size h
is approximately proportional to h for small h, independently of the
location of the region.
3/ The probability of more than one event's occurring in a region
of size h is negligible in comparison with the probability of one event's
occurring, for small h.6
Taking r as the average value of time in which a system remains in
its state or its life span (r being the inverse of the expected number
of events per unit time), it is demonstrated that then the number of
events in the interval t follows the Poisson distribution of parameter:
t_
r
TT (n events in the interval t) = e~^ ~^~
n = 0, 1, 2...
It will be observed that once n becomes notably greater than £,
ir(n) tends rapidly to 0. We have:
n(n + 1) = 7r(n) •

4
See above, 6.6, p. 149 sq.
5
See above, 6.9, pp. 159-160.
6
P.W. Lindgren, G.W. McElrath, Introduction to Probability and Statistics3,
MacMillan, London, 1969, p. 68, where the authors give mathematical precision
to the sense of the terms "approximately proportional" and "negligible". E. Borel,
Elements de la Theorie des Probabilites, Paris, Albin-Michel, 1950, p. 186-187.
248 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

When 7 = 1, and in forcing the last decimal place:7

TT(O) = l/e = 0.368
TT(l) = 7T(0) = 0.368
*(2) = 1/27T(1) = 0.184
7T(3) = 1/3 7T(2) = 0.061
7T(4) = 1/4 7T(3) = 0.015
*(5) = 1/5 7T(4) = 0.003
7T(6) = 1/6 7T(5) = 0.0005
7T(7) = 1/7 7T(6) = 0.00007
*(8) = 1/8 7T(7) = 0.000008
Total: 0.999578
Proponents of the Master Argument's third premise, those who
dissociate the possible from its realization, will think they find in these
figures an argument in their favor. They will say that whatever the
number n of occurrences of the event in the unit of time considered, the
other numbers of occurrences k ^ n will be excluded. The probability
of k occurrences however is possible in the unit of time. But there being
k occurrences is excluded, which falsifies premise (B) in its conjunction
with (A) and (NH).
It will suffice, however, to consult the list of purely theoretical num-
bers corresponding to the probabilities of the occurrences of the event
in question to rule out this sophism and reestablish the conditions
called for when we want to translate a statement about the possible
with double temporal indexation into terms of probabilities. There
are indeed two differences that are obvious. I/ We are not at all sur-
prised at the occurrence first of the event which was far from having
the greatest probability. 2/ In repeating trials, we would be surprised
to find that the statistics exactly verified the theoretical probabilities:
the existence of deviation, and increasing deviation in absolute value
as the number of trials increases, belongs analytically to the concept
of probability.
The thing is, the possible is ordinarily an individual event, the naval
battle tomorrow, without the set of naval battles entering explicitly into
its concept. By contrast, the probable has to to with repeatable trials
and its concept is inseparable from that of the set of like events. It loses
all its significance once divorced from what constitutes its specifically
collective relation to possible experience.
What is more, the theory fixes this relation ideally to an infinite
number of trials. Such is the role of the law of large numbers that
7
Borel, op.cit., p. 189.
 A PROBABILISTIC RECONSTRUCTION OF THE MASTER ARGUMENT / 249

determines the relation between probability and observed frequency.

In its "weak" form, this law states, with / standing for the observed
frequency of the expected event in jV trials, that the probability that
the relative frequency j j differs from the probability of the event by
more than a small fixed number e can be made as small as one likes by
increasing the number of trials:

lim p(\4r - n\ > e) = 0 for all e > 0.

Af-»oo N
Keeping in mind the collective feature of probability and its essential
relation of corroboration to a frequency, it is to the law of large numbers
that we must compare Diodorus' law of the possible. That law, which
is equivalent to the conjunction of (A, B and NH), says that if it is
possible at TV that p at t, then there is a t\ such that N < t\ < t
and p at ti. But just what are we asserting when we use the Poisson
distribution to attribute the probability 7r(2) = 0.184 to the occurrence
of two events in the unit of time? In the list of probabilities given
above the total did not reach 1. To obtain a total of 1 as the theory
requires, we shall follow the Table of values of the Poisson probabilities
in forcing the last decimal place of 7r(7), the final probability of the
list. (We shall return later to the error thus introduced, which does
not affect the theoretical argument). We obtain then the following list:
7r(0) = 0.3679
7r(l) = 0.3679
7r(2) = 0.1839
?r(3) = 0.0613
7r(4) = 0.0153
7r(5) = 0.0031
7r(6) = 0.0005
n(7) = 0.0001
Total : 1.0000
With the total equal to one we can now imagine an urn in which
there have been mixed 10.000 tickets, proportionately distributed over
the eight designated classes and appropriately marked. If we make
10.000 drawings without replacement, necessarily 1.839 tickets of the
third class will be drawn. If, on the other hand, the urn contains a mil-
lion tickets marked according to the required proportion and suitably
mixed, 10.000 drawings will give a frequency that differs from 1.839
and a relative frequency slightly different from 1.839/10.000. To assign
that probability to the event of class 3, even in leaving room for the
absolute deviation that can grow as N grows, is to assure that if the
relative frequency deviates from the probability by a number e > 0,
250 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

increasing J\f will suffice for reducing that deviation. The annual occur-
rence of two events has the probability 0.1839 if, and only if, in a long
series of years and for like populations, it is probable that the limit of
the difference between relative frequency and probability tends toward
0. There is, therefore, no probability without a determination of the
conditions of realization of the probable event; and these conditions of
non-hindrance affect the probabilities of rare events as much as those
of frequent ones.
What then, in probabilistic terms, would correspond to the state-
ment (C) which asserts the existence of a never realized possible? It
would be a statement asserting that, no matter how long the sequence
of trials chosen relatively to a positive probability nor how great the
number of trials become, the probability that the relative frequency of
the expected event should differ from the probability of the event by
more than a small number, e, cannot be made as small as one likes.
And this statement contradicts the theorem of Jacques Bernoulli.
Thus when proper care is taken to translate the Master Argument's
premises into probabilistic terms, the law of large numbers leads an-
alytically to an acceptance of Diodorus' solution and incrimination of
premise (C). To apply probabilities to the natural sciences is to bind
these latter to frequencies. The sense in which the words "real" and
"virtual" are to be taken here still remains to be specified. For in
the probabilistic reconstruction of the Master Argument, the two con-
tradictory statements expressing it are both asymptotic. Their truth
depends on a passage to the limit for an infinite number of trials. A
probability is not defined by a frequency after the manner in which
Diodorus defined a possible by one or more realizations; and the pas-
sage of the relative frequency to the probability is itself only probable.
The realizability to which the possible is reduced is to be understood
as a virtual sequence of virtual events on a plane of abstract repre-
sentation. As happens every time a scientific concept is defined, we
have changed the sense of the term from that which it had in ordi-
nary language. The individual possible takes its place in a "reality"
going beyond experience, since it is made up of a virtual sequence of
virtual events. But the infinite future postulated by Diodorus also has
a transcendence comparable to that of this sequence.
 THE SPECIAL STATUS OF PREMISE (C) / 251

10.3 The special status of premise (C):

Chrysippus' solution and the 'Unique Law of
Chance'.
Is the question settled then?
It would be if we confined ourselves to the theoretical domain of
the probability calculus, translating the Master Argument's premises
into propositions of that calculus, as we have done. It would be pos-
sible, however, while still maintaining that the law of large numbers
translates the conjunction of premises (A), (B) and (NH), to refuse
the translation proposed for premise (C). It might be argued that that
premise is not on the same level as the others. It bears not on a the-
oretical relation of probability to experience, but on a practical one.
What is true in theory need not to be true in practice, as otherwise, our
experience being finite by definition, we could not apply the probability
calculus.
Aristotelians could be tempted by the following argument. The law
of large numbers reduces probability to frequency only if there exists
an irreducible theoretical probability. Because of its irreducibility the
relevant sentence is comparable to the premise (C) and is not "already"
provided with a truth value.
The temptation must be resisted. First the probable validity of the
law of large numbers means that no experience suffices for its legitima-
tion while the realization of the possible falsifies (C). Secondly, while
Aristotle denies that (C) as well as (~C) have actual truth value, the
law of large numbers—even if it is a theoretical law—applies to expe-
rience only by choosing (C) against (~C). In themselves probabilities
justify rather Diodorus' than Aristotle's contingency.
Let us therefore turn to statistics. Consider those given for births
of quadruplets in France8 from 1945 to 1962.
8
Quoted by J.L. Boursin-P. Caussat, Autopsie du hasard, Bordas, Paris, 1979.
The mean of the distribution, equal here to the variance (square of the standard
deviation), is equal to j.
252 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Year Number of Year Number of

quadruplet births quadruplet births
1946 4 1955 1
1947 1 1956 0
1948 1 1957 0
1949 2 1958 0
1950 1 1959 0
1951 2 1960 2
1952 0 1961 2
1953 0 1962 0
1954 1
The Poisson Law solves the problem of repeated trials when, with
a large number of trials, the probability is quite weak for each of them.
We shall naturally adopt the hypothesis then that the statistic at hand
obeys the Poisson Law. Moreover, there are 17 births of quadruplets in
17 years. Since the average is equal to 1, we shall specify the hypothesis
in setting t — T in the Poisson distribution. What remains to be done
now is to compare the frequencies to the theoretical probabilities just
presented.
Let us now consider one of the statistical tests devised for answering
the question as to whether or not to accept a Poisson distribution of
parameter 1 in the case of our statistic.9 For each class of events
the frequencies fi(i = 1,...,17) are noted. These are compared in
turn to the corresponding mathematical expectations (products of the
corresponding probabilities pot furnished by the null hypothesis HQ
—that is to say, the hypothesis of the Poisson distribution—by the
number of years, 17). From this the difference between the observed
and expected frequencies is deduced:
\fi ~
9
Lindgren, McElrath, op. cit., p. 240.
 THE SPECIAL STATUS OF PREMISE (C) / 253

Number of
quadruplet births 0 1 2 3 4 5 6 7
in 17 years

Observed
frequencies 7 5 4 0 1 0 0 0

Expected
values 6.25 6.25 3.13 1.04 0.26 0.07 0 0

Difference 0.75 1.25 0.87 1.04 0.74 0.07 0 0

It will be noted that except for the numbers 6 and 7 the expected values
all differ from the observed frequencies with the deviation in excess of
1, the value of the standard deviation, for the numbers 1 and 3. The
test will be based on the value of a mean square deviation:
i=7 / ,.

which gives:
(0.75)2 (1.25)2 (0.87)2 (1.04)2 (0.74)2 (0.07)2 =
6.25 6.25 3.13 1.04 0.26 0.07
Under asymptotic conditions for the number of trials,10 the statistic
would obey a distribution \ 2 with 6 degrees of freedom. If ^ 2 is too
large there will be reason to cast doubt on the hypothesis H0. The
question that arises is for which set of values x2 > M (M being con-
stant) the hypothesis HO is to be rejected, with the constant M devised
to assign the probability a of rejecting H0 when H0 is true. 11 In other
words, to fix M is to fix the probability a that one will accept of being
mistaken in rejecting a true hypothesis:
Q = 7r(reject H0\H0)

10
"The exact distribution of this statistic

=E :
,,2 _ V^ (/« --A/po;) 2

is not simple and it depends on the p0's in the null distribution (Ho)', but it was
found that this dependence disappears as the sample size becomes infinite, and that
the distribution for n = oo can be computed. It is a 'chi-square distribution with
n — 1 degrees of freedom' ". (Lindgren, McElrath, op. cit., p. 240).
n
Lindgren, McElrath, op. cit., p. 240.
254 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Let us decide, for example, to run the risk of being mistaken with a
probability of a = 0.10. In the table for \2 with 7 - 1 = 6 degrees of
freedom, we shall then choose M such that FxiM = 0.90, which gives
M = 10.6. Since x2 = 3.92 < 10.6, the test calls for acceptance of the
hypothesis of a Poisson distribution.
Even apart from asymptotic considerations, the decision to accept
or reject a statistical hypothesis is never exempt from risk. All that
can be demanded of it is that it be informed by the measure of the
risk involved. Then too, this measure should be exact. But what is it
in fact that we see? First we calculated the Poisson Law of parame-
ter 1, in forcing the last decimal place, for 9 degrees of freedom: the
approximative total of the probabilities fell short of 1. Next we con-
sulted a table in order to specify the sense of the word probability in
such a distribution. That table was limited to 7 degrees of freedom.
Since the values not listed by the table were smaller than 1/2(0.001),
it "rounded off' the value of 7r(7) in conformity with the table's degree
of precision.12 Whatever the degree of precision required of a table,
all values going beyond that degree of precision will be disregarded.
Without such disregard, and thus without there being imposed some
limit to the degree of precision, the very formulation of a statistical
test would become impossible and the probability calculus would have
no application to experience.
This necessary approximation has attracted little attention from
the men of science, with the exception of Emile Borel. It is, so to
speak, the counterpart of the law of large numbers. As an idealiza-
tion of a theoretical order, the law of large numbers embeds a finite
sequence of trials in an infinite one. As a practical idealization, the
"law" of approximation enjoins us to disregard a part of probability
or a sufficiently small probability. The law of large numbers translates
Diodorus' definition of the possible into probabilistic terms. It will
be seen that the law bearing on approximation gives consistency to
Chrysippus' definition of the probable.
Borel writes: "The practical applications of the results [bearing on
repeated trials] are dominated by an empirical law that ... must be
regarded as absolutely certain and that might be called the unique law
of chance. That law, which is altogether simple, can be expressed as
follows: Events with an extremely weak probability are never realized,
but this statement should still be made explicit by indicating just what
12
Boursin, Caussat, op. cit., p. 324. Instead of ir(7) = 0.00007 we take 7r(7) =
0.0001, in introducing an error of 0.00003 < 1/2(0.001).
 THE SPECIAL STATUS OF PREMISE (C) / 255

is meant by extremely weak probability"?13 The degree of precision

involved is naturally relative to a given scale. What is negligible on
a human scale is one thing, what is negligible on a terrestrial scale
another, and what is negligible on a cosmic or supercosmic scale is yet
other. The first list of probabilities given stopped at IT(8) = 0.000008.
The tables stop at 7r(7). They disregard ?r(8) and introduce an error
of the order of 10~6. This is the order of precision that Borel says is
negligible on the human scale.
We shall not go into the justification of these scales. In doing so
we would find ourselves confronted with the Megarian paradox of the
heap and with undecidable concepts.14 We shall limit ourselves to
pointing out that the theoretically true statement, affirming that it
suffices to reach a sufficiently large number n of trials for the weakest
probability to be realized,15 has every chance of being devoid of any
practical signification. For "if we turn our attention to the simplest
and most elementary experiments, which are by far the most frequent
(impacts of infinitesimal particles, the ultimate elements of matter and
energy) and to time intervals far surpassing the life of our solar system,
the number of times that an experiment can be repeated is no more
than a power of 10 with an exponent of less than a thousand".16 The
realization of an event corresponding to a very small probability would
require, in order for the number of repeatable trials to be sufficient,
a cosmic homogeneity rendering possible the formation of the same
combinations; and this is nothing but a puerile and anthropocentric
representation of an infinite universe.17
In short, an event with a small enough probability will never occur
because the physical conditions rendering possible the repeatability
of the trial will have disappeared. Such destruction of the physical
conditions of the probability falsifies the conjunction of the axioms (A),
(B) and (NH); it reinstates premise (C). It is in conformity with the
doctrine of Chrysippus and with the destruction of deictic statements
that doctrine invoked18 and for which the unique law of chance provides
an unexpected warranty.
13
Borel, op. cit., p. 100; Boursin, Caussat, op. cit., p. 288.
14
See their formulation by Diodorus, a probable critic ot the Stoics, above, 3.5,
pp. 64-66; on the unique law of chance and the paradox of the heap, see Boursin,
Caussat, op cit., pp. 290-291.
15
This is the "mathematician's argument" referred to by Borel, op. cit., p. 106.
The mathematician in question seems to be Poincare; see, for example, La Valeur
de la Science, Paris, Flammarion, 1970, pp. 173-174.
16
Borel, op. cit., p. 107.
17
Borel, op. cit., p. 107.
18
See above, section 5.2, p. 107 sq.
256 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Chrysippus' recourse to a non-standard modal system is at once

explained and justified. According to the law of large numbers, the re-
alization of the probable, as is required by the conjunction of premises
(A), (B) and (NH), is validated, but weakly so. This much is required
by the theory of probability: it is impossible for an event with a posi-
tive probability, no matter how slight, not to occur some day at least
once; but this does not imply that it is necessary for it to occur some
day at least once, if the circumstances rendering possible the repeated
trials from which the event can result no longer obtain. On the other
hand, according to the unique law of chance, one can, and even must,
disregard the realization of an event whose probability is sufficiently
small. Thus premise (C) is validated, but weakly so too. And this is a
requisite for the application of probability theory to experience. It is
not necessary for an event with a sufficiently small, although positive,
probability to occur some day at least once, which does not imply that
it is positively possible for it not to occur any day at least once, for
this latter would usher in a conflict between the unique law of chance
and the law of large numbers. 19
Whatever their respective merits, the two solutions of Diodorus
and Chrysippus are not of equal value. They haven't the same value
in that the second can be reduced to the first. It will be possible to
see this in taking a closer look at the law of chance. The elimination
of 7r(8) from the Poisson table for parameter 1 means in effect that the
table admits a precision of less than 10~6. We have a right then to re-
place the approximation statement '?r(8) = 8.10~6 may be disregarded'
by the complete and theoretically exact statement 'the mathematical
expectation that there will be at least one occurrence of 8 favorable
events in the course of 106 units if time is equal to 8'.20 The effect
of this latter statement, if brought to bear on pure probability the-
ory, would be to eliminate from Diodorus' urn containing all possibles
those for which external circumstances of hindrance would postpone
the realization for too long a time. Chrysippus, who has recourse to
incomplete statements and who distinguishes the impossibility that not
from necessity, calls the events thus excluded possibles that will never
be realized. Once the statements are completed the distinction is dis-
solved and the excluded events are called impossibles of a scale. The
argument was thought to give a practical validation of contingency. In
fact it has only succeeded in narrowing its domain.
19
See above, p. 121. The conjunction (Lp • M ~ p) is contradictory because M ~
p D~ Lp. By contrast, the conjunction (~ M ~ p- ~ Lp) is consistent because we
do not have the right to conclude Lp from ~ M ~ p nor M ~ p from ~ Lp.
20
1 x 8.10-6 x 10-6 = 8.
 CONTINGENCY AND IGNORANCE / 257

10.4 Contingency and ignorance: The statistical

mix.
The fact that it is possible to reconstruct an equivalent of the Master
Argument in probabilistic terms goes to show that there is some notion
of contingency that is not incompatible with the Diodorean premises
and that leaves the principles of logic intact. It is along these lines,
moreover, that Diodorus resolved the aporia. He denied axiom (C)
and gave a sui generis21 definition of contingency. What remains to be
done within that conception transposed into terms of probability is to
determine the part due to nature and that due to human ignorance.
Let us replace the image of two mutually exclusive possibles, one of
which is but a shadow of reality, by a dichotomous contrivance that will
force nature to recognize a degree of possibility for them both. To make
the image concrete think of a stream of projectiles shot from a canon
too inaccurate to control their angular dispersion and subjected to
passing through a screen pierced by two holes in the manner of Young's
experiment and then reaching a detector that absorbs and counts them.
But any other apparatus would do just as well on condition that there
be passage from an initial state i to a final state / by means of a well
defined alternative: either passage by way of A or passage by way of
B, with distinct and exclusive A and B.
When it is waves—not particles—that are submitted to the test
of Young's apparatus, the experiment gives rise to the remarkable and
paradoxical phenomenon of interference: crests and troughs cancel each
other out in superposition, as Posidonius had understood. But wave
propagation doesn't shock our physical intuitions. It is continuous and
non local. The wave passes through both holes and that is why there
is interference. There is no analogue here of a choice between two
mutually exclusive possibles. The resultant intensity IAB measured
at the detector is not reducible to the sum of the intensities I A and
IB that would have been obtained respectively in obturating first hole
B, then hole A. It contains an interference term that depends on the
difference of phase between the two waves diffracted at A and B.
Let us return now to particles.
The detector registers the arrival of a particle at point x on the
detection screen. This event can be described by the following diagram
(read from right to left):
(MO-

There is a transition from the initial state (emission of the projec-

21
See above, p. 42. The infinite time there assigned for the realization of the
possible evokes the abstract and virtual character of statistical series.
258 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

tile by the canon) to the final state (arrival on the detector at point
x). Furthermore, this transition can be analyzed. It is the sum of
two possible transitions: one in which the particle passes through the
intermediary state \ua > (passage through hole A), the other in which
it passes through the intermediary state \Ub > (passage through hole
B):
(/x|i> = (fx\Ua)(Ua\i) + (fx\Ub)(ub\i).
Attribution of corresponding probabilities involves, in all cases, the
elaboration of a density function making is possible to speak of the
probability of the detection of a particle at point x and of conventions
that allow of assigning numbers to the elements of the diagram of
transitions that will be apt to suitably define the probabilities sought.
Of these conventions we shall retain only what is of importance for the
discussion of the Master Argument. The expression for the transition
(fx\i) is generally a complex number, and its probability is equal to
the square of that expression, |(/i|«)| 2 , which is always a real number.
Experiment shows that every time we observe, or can observe, by
which branch of the alternative the particle has passed, the total prob-
ability is the sum of the probabilities of the two branches:
(1) *«/*!»)) - *((/x|tt«X««l')) + T((/»l«6>(« 6 |t».
22
This equality corresponds to the classical hypothesis according to
which every transition is observable and corresponds to a well defined
path. We shall see that, with the validity of the Master Argument,
it entails Diodorus' solution. Quantum particles, on the other hand,
obey a different law. The total probability is the square of the sum
of the transitions — we are speaking now of probability amplitudes — so
that here there appears a term of interference that did not figure in
expression (1):
(2) Jr((/*|i» = *({/,K){tt«|») + {/*K>(«fc|i».
This equality23 disallows observation and the observability of the in-
termediary state. It is impossible to determine by which branch of the
alternative the transition is effected. Consequently, there is not a well
defined path corresponding to a transition. We shall see that in this
22
That is to say:

23
That is to say:
K/. Wl 2 = (!(/« + «a}|2 • IK + i>| 2 ) + (K/x + «fc)| 2 • ItoWl 2 ) +
2(/x|u a )(u 0 |t){/ x K){u f c |i).

It is this last term, called the interference term, that makes the difference.
 CONTINGENCY AND IGNORANCE / 259

case the Master Argument loses its cogency and physics can do without
Diodorus' solution.
Let us first consider the case of classical particles. The frequency fx
(the number of particles reaching the detector at x] is measured, and in
shuttering hole B the frequency JXA (the number of particles reaching
x after passing through hole A) is measured, as is the frequency fxB
with hole A shuttered. If M is the number of particles having passed
through one of the two holes, then fx\J\f, fxA\N and fxB\-N" measure
the corresponding relative frequencies that will have to be compared to
the respective theoretical probabilities KX,KXA,'!J'XB- There are three
weak laws of large numbers that follow from this, the first resulting
from the sum of the other two. When the particles are projectiles in
a Young apparatus, it is a simple matter to reconstruct the Master
Argument. The projectiles collected on the detector screen at x fall
into two disjoint classes, though without our being able to distinguish
those having passed through A from those having passed through B.
We are nevertheless certain, now that the event has taken place, that
each individual projectile considered has passed either through A or
through B. In the first case the probability that it passed through B is
zero, and vice versa in the second case. Thus axiom (A) is verified. We
are justified then, as regards the consequent of axiom (B), in retaining
only that disjunct that has to do with the future. But at the moment
of firing, the probability TTX attributed to a certain particle now col-
lected at x may be broken down into the two non null components TTXA
and TTXB . The weak law of large numbers applied to TTXA and -KXB
guarantees that the probabilistic version of the conjunction of axioms
(A.B.NH) holds for both of these probabilities, notwithstanding the
fact that they are contraries: each of the two "degrees" of the possible
contracts and is realized with the frequency sought as the number of
trials is increased at will.24

24
Let us shutter B and count /XA at x. The law says that
lim P(|^--7r^|> £ ) = 0.
JV—>oo J\

The contrary event (passage through B) is excluded due to the shuttering of B.

This is what verifies the second term of the conjunction in axiom (C). But TTXB is
then equal to 0 and the first term of the conjunction ((3t)(~ L^Pt • N < t)) is
therefore falsified. It could likewise be shown, in shuttering A, that the law of large
numbers for KXA verifies the axioms (A.B.NH) while falsifying axiom (C).
260 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

One might think that with the two holes A and B left open and
the two probabilities TTXA and TTXB each greater than zero the Master
Argument would cease to apply. It would be argued that since the law
of large numbers verifies the conjunction of axioms (A.B.NH) for TTXA
in probabilistic terms, the contrary event is excluded though without
its probability being null, and this is a state of things that pleads in
favor of axiom (C). But this would be to confuse two experiments that
are in fact independent. For just what does it mean to say of a particle
collected at x that at the moment of its emission it had probability
KXA of passing through A and probability TTXB of passing through B, A
and B both being open? Take one such particle. At the very latest at
the moment of passage through the holes, the possible will have had to
synchronically contract, in conformity with axiom (B), and will have
had simultaneously to be realized, in conformity with axiom (NH).
The possibility of the contrary choice is nothing but a word once the
contraction is made. Are we obliged then to recognize that before the
contraction, in the time interval separating the emission of the particle
from its passage through the hole, there was a contrary possibility and
that since that contrary possibility was destined never to be realized,
the probabilistic model is seen to verify axiom (C)? The symmetry of
the Young apparatus requires that TTXA = TTXB for x equidistant from A
and B. Isn't this an argument of weight for attributing equi-possibility
to both the excluded and the realized events?
But what is the meaning of the concession made? Have we really
the right to put off the choice—that is to say, the synchronic contrac-
tion and realization—until passage through the holes? This would be
to misunderstand the meaning of equality (1). For indeed, with all due
precaution taken as regards times of emission and the counting of par-
ticles, we can postulate that collecting the particles simultaneously at
x that have come through one or the other hole with both holes open
comes down to the same thing as collecting successively at x those
particles having come through A with B shuttered, then those having
come through B with A shuttered. As this second experimental case
decides for Diodorus' solution, the first will then do so as well. In other
words we can imagine that the particles are already partitioned at the
source into two exclusive classes much as would be the balls of two dif-
ferent colors in an urn. The emission of a particle is like a drawing in a
statistical mix. If M is the total number of particles, N.TTXA of them are
marked "destination x via A" and A(TT X B of them "destination x via
B". Reduced to its simplest expression the Master Argument, together
with Diodorus' solution, asserts only that it is impossible for one ball
 CONTINGENCY AND NATURE / 261

in the urn to be two different colors or for one particle to be marked

for two destinations and two distinct paths.
Classical probability theory, expressed by (1), assigns each parti-
cle to one of the two exclusive classes fxA, /XB at the outset. This
assignation is another way of incriminating axiom (C). The Diodorean
solution it implies is characteristic of classical physics. I/ It is exten-
sional and thereby compatible with the language of physics. 2/ It is
simple, as can be seen by comparing it with the other solutions ex-
amined. 3/ It excludes no elementary laws, though without requiring
a knowledge or application of them. It is thus compatible with the
principle of determination and with the principle of causality which
were defended by all ancient philosophers with the exception of Epi-
curus. 4/ It leaves room in science for ignorance. Laplace, one of the
founders of the probability calculus, put it in charge of that ignorance.
A superior intelligence, aware of initial conditions and of the limiting
conditions of a phenomenon would be able to foresee its future exactly.
Probabilities realize "the most felicitous supplement to the ignorance
or weakness of the human mind". 25

10.5 Contingency and nature: The state of

superposition.
The fact that probability in the classical sense, that is to say relative
to our ignorance only, does not challenge the validity of the Master
Argument but even calls for the solution of Diodorus is something that
becomes clear as soon as we try to imagine a remedy for that ignorance.
If we observe the passages through hole A or if we go back to follow
the paths taken at the moment of their emission, assignment of the
particles either to class fXA or to class JXB for any x is immediately
given: probabilities become useless.
If it were possible then to go back to the state of each particle at the
moment of its emission we could in principle (or a superior intelligence
in the place our own could) answer the two questions we ask: 'where
is the particle going?' 'through which hole does it pass?'. These two
questions may be represented in the form of two alternatives with the
first admitting an infinite and even continuous number of choices, the
second being dichotonous. When we imagine a superior observer ca-
pable of observing all the conditions of a phenomenon, the hypothesis
that all physical magnitudes are in principle simultaneously observable
25
Laplace, Essai philosophique sur les probabilites. "Were it not for our igno-
rance", said, Poincare, "there would be no probability, there would only be room
for certainty". (La science et I'hypothese, Paris, Flammarion, 1968, p. 196).
262 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

comes down to an up-front distribution of the different particles into

their class fxA or /XB (with x indexed over a continuous set of values).
Since the hypothesis remains impractical for a finite observer such as
us though, incapable as we are of answering the two questions asked
exactly and of assigning each particle its point of destination and place
of passage, we are reduced to assigning probabilities to these two mag-
nitudes. Instead of answering 'this particle has destination x by way
of hole A' we must content ourselves with answering 'this particle has
the probability TT^'. The difference between the two answers reflects
no difference with respect to the state of nature. It simply reflects
a difference relative to our knowledge of that state. And this latter
difference would disappear if our observation of the conditions of the
phenomenon were complete.
Within mechanics there is a class of phenomena, the extent of which
has recently been measured by chaos theory, that casts doubt on the
practical possibility of a complete observation. There are those phe-
nomena that are sensitive to the initial conditions.26 A very small
difference in these conditions and imperceptible to the observer pro-
vokes with the divergence of the resultant trajectories our inability to
predict them. The indiscernibility of observables makes one think of
the paradox of the twins and of the skeptical solution, or rather reduc-
tion, of the Master Argument. 27
These doubts, nevertheless, have no bearing on law (1). From our
state of ignorance they draw unexpected consequences in classical me-
chanics about the predictability of phenomena. They in no way modify
either the idea of a physical state or the idea of the observable. The so-
lution they can bring to the Master Argument remains external to the
aporia itself. Rather than invalidating it, they render its significance
unclear. This uncertainty, in turn, justifies a maxim of indifference.
What difference does the invalidity of axiom (C) make if in situations
of instability I generally remain incapable of distinguishing a possible
that will be realized from one that will not?
But suppose now the physical world to be bound by the laws of
quantum mechanics and no longer obeying (1) but obeying (2). The
26
P. Berge, Y. Pomeau, Ch. Vidal, L'ordre dans le chaos, Paris, Hermann, 1984
(passim, and p. 287).
27
See above, p. 219; P. Suppes, Probabilistic Metaphysics, Oxford, Blackwell, 1984,
p. 129 (where the author remarks that "the matter is sometimes discussed as if
the whole problem resides in the initial conditions, but this is not the case": in
Laplace's doctrine moreover the explicit concept of a superior intelligence was called
into question, in dealing with the three-body problem, before the implicit concept
of a superior observer was.
 CONTINGENCY AND NATURE / 263

existence of the intercrossed term in (2) doesn't come from a modifi-

cation of the probability calculus. It results from the fact that it is
no longer probabilities that are added but probability amplitudes. It
attests that during its flight the particle is in state of superposition.28
It invalidates Diodorus' argument in changing the sense of its premises.
On the classical interpretation axiom (A) signified both the neces-
sity of the past and the impossibility of realizing the possible in the
past. Obviously definition (2) of the probability of the transition (f\i)
in no way changes the relation of the probability to the time parame-
ter: the impossibility of realizing the possible in the past is respected
and there is nothing to hope for, for the invalidation of (A), in the idea
of eternal return. Nevertheless, since the intercrossed term of interfer-
ence in (2) would disappear if we were to observe through which hole
the particle passed, it is impossible to say for a particle detected at x
whether it passed through hole A or hole B. Letting t' stand for the
time of the particle's passage through the holes, whereas formerly we
could say that it either passed through A at t' with the probability
1 and then its probability of passing through B at t' was nil, or vice
versa, that retrodiction now ceases to be true. (A) is thus falsified in
the sense in which it proclaims the necessity of the past and in which
the poet said that God could not make it so that what had been done
had not been done.
Axiom (B) subordinates the diachronic possible to its contraction.
If we compel nature to choose then between two dichotomous possibles,
M/vpt and MAT ~ Pt, the positive and negative diachronic possibles
will have to contract into Mtlpt1 and M<2 ~ pt2 with N < t\ < t and
N < t-2 < t. But in virtue of (NH) contradiction will be avoided only
if t\ ^ t%. Consequently, two contrary synchronic possibles cannot be
superposed in time. In quantum mechanics, by contrast, where condi-
tional necessity no longer has its former sense, the state of synchronic
superposition of two contrary amplitudes of probability is entirely le-
gitimate so long as we do not measure through which hole the projectile
has passed, which would destroy superposition. At the moment of pas-
sage, t', we can say that there is a non-null amplitude of probability
that the particle is passing through hole A and a non-null amplitude
of probability as well that it is passing through hole B. All that is
required is that there be no observation that comes to transform the
28
We obtain the expression of this state in projecting the initial state \i > on the
pair of intermediate states \ua > and \ub >:
\i >= \Ua >< Ua\i > +Ub >< Ub\i >= Ca\Ua > +Cb\Ub > .
264 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

state of superposition into a state proper. Thus axiom (B) turns out
to be truistically confirmed and even reinforced since the synchronic
contraction extends to all moments preceding that of the reduction.
Axiom (C) is verified by a possible that is not realized. But since
the term of interference precludes our observing through which hole the
particle has passed, it is impossible to say which of the two contrary
probability amplitudes has been realized, so that there ceases to be an
exclusion between axiom (C) and the conjunction of axioms (B and
NH) and, furthermore, axiom (C) is strictly speaking undecidable.
What of axiom (NH)? It is verified when there is reduction of the
wave-particle, that is to say, when the process of measurement selects
one particular value of the observable. If such a reduction takes place
at the moment the particle passes through one of the holes, the in-
terference effect is destroyed for detection. If it takes place only at
the moment of detection, the interference effect is produced. In other
words, nature chooses the statistical mix or superposition accordingly
as we observe or do not observe where the particle passes. To decree
along with classical mechanics that the state of a physical system is
composed of a set of physical magnitudes and that these magnitudes
are all simultaneously observable is to choose law (1). To choose law
(2) is to define the general state of a physical system by superposition:
such a state is not observable. An observable is associated with the
action of an operator on it and it is that action that produces a proper
state.
What dismantles the Master Argument's premises is the new dis-
tinction in the history of modal notions between a probability and a
probability amplitude. Classical physics was content with the oppo-
sition 'This particle passes through A' versus 'This particle has the
probability IT of passing through A'. This opposition has nothing to do
with ontology: it incorporates what is due to our ignorance into the
determination of natural phenomena. Instead of attributing a prop-
erty or magnitude to a physical system, we attribute it a disposition or
propensity to have that property or magnitude. Probability measures
that disposition or propensity that belongs to the system in act. A
probability amplitude is something altogether different. We can com-
pare it to an embryonic probability as the inventors of the infinitesimal
calculus compared the "moment" of motion to an embryonic motion
that an integration would bring to a state of "whole" motion. But
the comparison limps. For the probability amplitude, which is gener-
ally a complex quantity, does not figure among the elements of reality.
To obtain a probability we must multiply two conjugated probability
amplitudes. This means that, when we attribute that amplitude to
 CONTINGENCY AND NATURE / 265

a system, it is attributed neither as an actual property or magnitude

nor as an actual disposition or propensity to having such property or
magnitude, but as a purely virtual disposition or propensity to having
it. The second-order potentiality, as it were, thus put into play is no
longer the measure of an ignorance that might have some chance of
being only provisional. It is physical. It describes nature.
BIBLIOGRAPHY

Albert of Saxony, Perutilis Logica Magistri Alberti de Saxonia, ed.

Petrus Aurelius, Sanatus, Venice, 1522.
Alexander of Aphrodisias, In Aristotelis Analyticorum Priorum librum
I Commentarium, GAG II, P. 1, ed. Wallies, Berlin, 1883.
-Praeter Commentaria Scripta Minora, Quaestiones, De Fato, De
Mixtione, GAG Suppl. 2, P. 2, ed. Bruns, Berlin, 1982. (Numbers
in parentheses refer to "Traite du Destin et du Libre Pouvoir",
in Nourrisson, Essai sur Alexandre d'Aphrodise suivi du Traite....,
Didier, Paris, 1870.)
Ammonius, In Aristotelis de Interpretation Commentarius, GAG IV,
P. 5, ed. Busse, Berlin, 1897. (Numbers in parentheses refer to
the Latin translation of William of Moerbeke: G. Verbeke, Am-
monius, Commentaire sur le Peri Hermeneias d'Aristote, Louvain-
Paris, 1961, Corpus latinum Commentariorum in Aristotelem Grae-
cum, t. I.)
- In Aristotelis Analyticorum Priorum librum I Commentarium,
GAG IV, P. 6, ed. Wallies, Berlin, 1899.
St. Anselm, Ein neues unvollendetes Werk des hi. Anselm von Can-
terbury, ed. F.S. Schmitt. (Beitrage zur Geschichte der Philosophic
und Theologie des Mittelalters, 33/3, Miinster, Aschendorff, 1936,
24.16.)
- Opera omnia, ed. F.S. Schmitt, Edinburgh, Nelson, 1961,
photomechanical reproduction by F. Fromann (G. Holzboog),
Stuttgart, 1968.
Aristotle, Aristotle's Metaphysics, Text and Commentary by W.D.
Ross, Oxford, Clarendon Press, 1924, 2 vol. .
- The Works of Aristotle, transl. into English by Edghill, ed. Ross,
vol. I, Oxford University Press, 1928.
- Aristotle's Prior and Posterior Analytics, W.D. Ross, Oxford,

267
268 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Clarendon Press, 1949

- Aristotelis Opera, ex. rec. I. Bekkeri, ed. O. Gigon, Berlin, de
Gruyter, 1960, 2 vol.
- De la generation des animaux, texte et traduction P. Louis, Paris,
Les Belles Lettres, 1961.
- Categoriae et Liber de Interpretation, ed. L. Minio-Paluello, Ox-
ford, 1949, 3° ed. 1981.
- Du Ciel, texte etabli et traduit par Paul Moraux, Paris, Les Belles
Lettres, 1965.
- Metaphysique, traduction avec commentaires, nouvelle edition,
par J. Tricot, 2 vol., Paris, Vrin, 1953.
- Physique, texte etabli et traduit par H. Carteron, 2 vol., Paris,
Les Belles Lettres, 1931.
Arnim, J. von, Stoicorum veterum fragmenta (S. V.F.), Lipsiae, I
(1905), II (1903), III (1903), IV (1924).
Atticus, Fragments, texte etabli et traduit par E. des Places, Paris, Les
Belles Lettres, 1977.
St. Augustine, St. Augustini opera, Paris, Muguet, 1695, 8 vol.
Bailey, C., Epicurus, Oxford, Clarendon Press, 1926.
Baudry, L. Lexique philosophique de Guillaume d'Ockham, etudes des
notions fondamentales, Paris, Lethielleux, 1958.
Becker, A., "Bestreitet Aristoteles die Giiltigkeit des "tertium non
datur" fur Zukunftsaussagen", Actes du Congres international de
philosophic scientifique, VI, Paris, 1936, pp. 69-74.
Becker, O., "Zur Rekonstruction des "Kyrieuon Logos" des Diodoros
Kronos", Erkenntnis und Verantwortung (Festschrift fur Theodor
Litt, Diisseldorf, 1961).
Berge, P., Pomeau, Y., Vidal, Ch., L'ordre dans le chaos, Hermann,
Paris, 1984.
Bergson, H., La pensee et le raouvant, essais et conferences, Paris,
Alcan, 1934.
Blanche, R., "Sur 1'interpretation du Kyrieuon Logos", Revue
philosophique de la France et de I'Etranger, n° 155, 1965, pp. 133-
149.
Boethius, Commentaria in Librum Aristotelis Peri Hermeneias, Se-
cunda editio, Leipzig, ed. Meiser, 1877-1880.
Bollack, J., Bollack, M., Wismann, H., La lettre d'Epicure, Paris, Edi-
tions de Minuit, 1971.
Borel, E., Elements de la Theorie des Probabilites, Paris, Albin-Michel,
1950.
Boudot, M., Logique inductive et probabilite, Paris, A. Colin, 1972.
- "Temps, necessite et predetermination", Les Etudes philosophi-
 BIBLIOGRAPHY / 269

ques, n° 4, oct-dec., 1973, pp. 435-473.

Boursin, J.-L., Caussat, P., Autopsie du hasard, Paris, Bordas, 1979.
Boyance, P., "Epicure et M. Sartre", Revue philosophique, 143, 1953,
pp. 426-431.
Brehier, E., La theorie des Incorporels dans I'Ancien Stoicisme, 2° ed.
Paris, 1928.
- Chrysippe et I'Ancien Stoicisme, Paris, P.U.F., 1910, 2° ed. 1951.
- La philosophie de Plotin, Paris, Vrin, 1928, 2° ed. 1961.
- Les Stoiciens, textes traduits par Brehier, E., edites sous la di-
rection de Schuhl, P.-M., Paris, Gallimard, 1962.
Brochard, V., Etudes de Philosophie ancienne et de Philosophie mod-
erne, Paris, Vrin, introduction de Delbos, V., 1954.
Bunge, M., "Possibility and Probability", in Harper and Hoocker (ed.),
Foundations of Probability Theory, Statistical Inference and Statis-
tical Theories of Science, Dordrecht, Reidel, vol. Ill, 1976.
Buridan, J., Johannes Buridani Sophismata, Consequentiae Buridani,
Denidel, A., de la Barre, N., Baligault, F., Paris, 1495/6.
- Quaestiones super libros quattuor de Caelo et Mundo, ed. Moody,
E. A., Cambridge, Massachussetts, Mediaeval Academy of America,
1942.
Burks, A.W., "The logic of causal propositions", Mind, n° 5, 60, 1951,
pp. 363-382.
Celluprica, V., // capitolo 9 del "De Interpretation" di Aristotele,
Rassegna di studi 1930-1973, Societa editrice II Mulino, Bologna,
1977.
Chalcidius, Timaeus a Calcidio translatus commentarioque instructus,
ed. Waszink, J.-H., (Plato Latinus, t. IV), Leyde, 1962.
Cherniss, H., Aristotle's Criticism of Plato and the Academy, New
York, Russel and Russel, 1962.
- Plutarch's moralia, t. XIII, Part. I, II, Cambridge-London, Loeb,
1976.
Chevalier, J., La notion du necessaire chez Aristote et chez ses
predecesseurs particulierement chez Platon, Paris, Alcan, 1915.
Cicero, Academica, ed. J.S. Reid, London, 1885.
- De Fato, ed. Plasberg-Ax, 1938.
- De natura deorum, ed. A.S. Pease, Cambridge, 1955.
- De divinatione, ed. A.S. Pease, Darmstadt, 1963.
Cornford, P.M., Plato's Theory of Knowledge', London, Routledge,
1935, 2° ed. 1960.
Cournot, A., Essai sur les fondements de nos connaissances et sur les
caracteres de la critique philosophique, ed. J.-C1. Pariente, Oeuvres,
Paris, Vrin, t. II, 1975.
270 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Crombie, J.M., Examination of Plato's doctrine, vol. II, Plato on

Knowledge and Reality, London, 1963.
Cumont, F., Astrology and Religion among the Greeks and Romans,
New York, London, Putnam, 1912.
- La theologie solaire du paganisme romain, Memoires presentes
a 1'Academie des Inscriptions et Belles-Lettres, Paris, Imprimerie
Nationale, t. XII, 2° partie, 1909.
- Les religions orientales dans le paganisme romain, Paris, Geuth-
ner, 4° ed. 1929.
Descartes, PL, Oeuvres et Correspondance, publiees par Adam et Tan-
nery, Paris, Le Cerf, 1897-1913 (referred to as A.T.), 13 vol.
Diels, H., Kranz, W., Die Fragments der Vorsokratiker, 3 vol., Berlin,
Weidmann, 1959.
Diogenes Laertius, Vitae Philosophorum, ed. G. Cobert, Paris, Firmin-
Didot, 1878.
Doring, K., Die Megariker, Kommentierte Sammlung der Testimonien,
Amsterdam, Gruner, B.R., 1972.
Duhem, P., Le systeme du monde, Paris, Hermann, t. I, 1913 (1° ed.),
1951 (2° ed.); t. II, 1914.
Duns Scotus, loannis Duns Scoti Opera Omnia, editio nova... a P.P.
Franziscanis... recognita, Parisiis apud L. Vives, 1891-1895.
Effler, R., Duns Scotus and the Physical Approach to God, Studies in
Philosophy and the History of Philosophy, vol. 3, John Duns Scotus,
ed. Ryan and Bonansea, Washington, the Catholic University of
America Press, 1965, pp. 170-190.
Epictetus, Epicteti dissertationes, ed. Schenkl, collect. Teubner,
Leipzig, 1° ed., 1898, Minor, 2° ed., 1916, Maior.
Epicurus (see Bailey)
Eusebius, Werke, Achter Band, Die Praeparatio Evangelica, Herausg.
v. K. Mras, I-II, Berlin, 1954-1956.
Faust, A., Der Moglichkeitsgedanke, Systemgeschichtliche Unter-
suchungen, 2 vol., Heidelberg, C. Winter, 1931, 1932.
F011esdal, D., "Quantification into causal contexts", Boston Studies in
the Philosophy of Science, vol. II, New York, Humanities Press Inc.,
1965, pp. 263-271.
- "A modal theoretic approach to causal logic", Det Kgl. Norske Vi-
denskabers Selskabs Skriften, 1966, Nr. 2, Trondheim, 1966, pp. 1-
13.
Van Fraassen, B.C., "Presuppositions, Superevaluations and Free
Logic", in The logical Way of Doing Things, ed. K. Lambert, New
Haven, Yale University Press, 1969, pp. 67-91.
 BIBLIOGRAPHY / 271

Fraenkel, H., Wege und Formen Fruhgriechischen Denkens, Miinchen,

C.H. Beck, 2° ed. 1960.
Frede, D., Aristoteles und die "Seeschlacht", Gottingen, Van den Hoeck
und Ruprecht, 1968.
Frede, M., Die stoische Logik, Gottingen, Van den Hoeck und Ruprecht,
1974.
Fritz, Kurt von, C.R. of Schuhl, P.-M., "Le Dominateur et les possi-
bles", Gnomon, 34, 1962, pp. 138-152.
- "Frede, Aristoteles und die "Seeschlacht", Pseudepigrapha,
Geneve-Vandoeuvres, Fondation Hardt, 1972, pp. 241-250.
Furley, D., Two Studies in the Greek Atomists, Princeton, Princeton
University Press, 1967.
- "Variations on themes from Empedocles in Lucretius "Poem"",
University of London, Institute for Classical Studies, Bulletin n°
17, 1970, pp. 55-64.
Gassendi, P., Disquisitio Metaphysica, texte traduit et annote par Ro-
chot, P., Paris, Vrin, 1962.
Aulus Gellius, Les Nuits Attiques, ed. et trad. Marache, R., Paris, Les
Belles-Lettres, 1978, 2 tomes.
Genest, J.-F., "La liberte de Dieu a 1'egard du passe selon Pierre
Damien et Thomas Bradwardine", Annuaire de la 5° section de
I'Ecole Pratique des Hautes Etudes, Paris, tome 85, 1976-1977,
pp. 391-393.
Geyer, F., Ueberweg, F., Grundriss der Geschichte der Philosophic,
II, Die patristische und scholastische Philosophie, Darmstadt, Wis-
senschaftliche Buchgesellschaft, 1958.
Gilson, E., Jean Duns Scot, Introduction a ses propositions fondamen-
tales, Paris, Vrin, 1952.
Gilson, E., Bohner, P., Christliche Philosophie, 2te Aufl., Paderborn,
Schoningh, 1954.
Goldschmidt, V., Platonisme et pensee contemporaine, Paris, Aubier,
1970.
- Le systeme stoicien et I'idee du temps, Paris, Vrin, 3° ed., 1977.
- La doctrine d'Epicure et le droit, Paris, Vrin, 1977.
- "La theorie epicurienne du droit", Archives de philosophic du
droit, tome 26, L'utile et le juste, pp. 73-91.
Granger, G.-G., La theorie aristotelicienne de la science, Paris, Aubier,
1976.
Gueroult, M., L'evolution et la structure de la Doctrine de la science
chez Fichte, Paris, Les Belles Lettres, 1930, vol. I & II.
- Descartes selon I'ordre des raisons, vol. I: L'Ame et Dieu; vol. II:
L'Ame et le Corps, Paris, Aubier, 1953.
272 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

- Malebranche, vol. I: La vision en Dieu, Paris, Aubier, 1955; vol. II:

Les cinq abimes de la Providence, Paris, Aubier, 1959: a) L'ordre et
1'occasionnalisme; vol. Ill: les cinq abimes de la Providence, Paris,
Aubier, 1959: b) La nature et la grace.
- Spinoza, vol. I: Dieu, Paris, Aubier, 1968; vol. II: L'Ame, Paris,
Aubier, 1974.
- Etudes sur Descartes, Spinoza, Malebranche et Leibniz,
Hildesheim, New York, Georg Olms, 1970.
Hamelin, O., "La logique des Stoiciens", Annie Philosophique, Paris,
101, p. 26.
- Essais sur les elements principaux de la representation, Alcan,
Paris, 2° ed. by Darbon 1925.
- Sur le De Fato, public et annote par M. Conche, Paris, Editions
de Megare, 1978.
Hintikka, J., Time and Necessity, Oxford, Clarendon Press, 1973.
Hughes, C.E., Cresswell, M.J., An Introduction to Modal Logic, Lon-
don, Methuen, 1972, 2° ed.
Hume, D., A Treatise of Human Nature, London, ed. Green and Grose,
new edition, 1890.
- Traite de la nature humaine, oeuvres philosophiques choisies, tra-
duction M. David, Paris, Alcan, 1930.
- Dialogues sur la religion naturelle, traduction M. David, Paris,
Pauvert, 1964.
Kant, I., Kant's Gesammelte Schriften, hrsg. von der Akademie der
Wissenschaften zu Gottingen, Band I-XXXV, Berlin, Reimer, 1910-
1917. Berlin und Leipzig, W. de Gruyter, 1910-1955/1966.
- Critique de la raison pure, edition et traduction Tremesaygues-
Pacaud, Paris, P.U.F., 1944 (1967).
- Critique de la raison pratique, traduction Picavet, Paris, P.U.F.,
1943 (1976).
- Critique du jugement, traduction Gibelin, Paris, Vrin, 1946.
Kneale, M. & W., The Development of Logic, Oxford, Clarendon Press,
1962.
Lambert, K., "On Logic and Existence", Notre Dame Journal of For-
mal Logic, vol. 6 (1965), pp. 135-141, n° 2, april.
Lambert K., Scharle, T., "A translation theorem for two systems of free
logic", Logique et analyse, 10° annee, 39-40 (dec. 1967), pp. 328-
341.
Laplace, P.S., Essai philosophique sur les probability's (1813), Paris,
Gauthier Villars, 1921, 2 vol.
Lejewski, V.C., "Zu Lesniewskis Ontologie", Ratio, 1957-1958, pp. 50-
77.
 BIBLIOGRAPHY / 273

Leibniz, G.W., Die Philosophischen Schriften..., ed. Gerhardt, C.I.,

Berlin, Weidmannsche Buchhandlung, 1875, 1890, 7 vol.
- Nouveaux essais sur I'entendement humain, Introduction, Note
et Appendice par Lachelier, H., Paris, Hachette, 1886.
- Essais de theodicee suivis de La monadologie, Preface et Notes de
Jalabert, J., Paris, Aubier, 1962.
- Oeuvres, editees par Prenant, L., avec introduction, tables et
commentaires, Paris, Aubier, 1972.
Lindgren, A.W., McElrath, G.W., Introduction to Probability and
Statistics, London, MacMillan, 1969.
Longpre, E., La philosophic de B.J. Duns Scot, Paris, 1924.
Lucretius, De natura rerum (commentaire exegetique et critique), 3
vol., Ernout, A., Robin, A., Paris, Les Belles Lettres, 1925-26-28.
- De la nature, tomes I et II, livres I a VI, traduction Ernout, A.,
Robin, A., Paris, Les Belles Lettres, 1947.
Lukasiewicz, J., Aristotle's Syllogistic from the standpoint of modern
formal logic, (1° edition, 1951), Oxford, Clarendon Press, (2° edi-
tion, 1957) (french translation by Mrs. Zaslawsky, Paris, Armand
Colin, 1972).
- "On determinism", Polish Logic, 1920-1939, edited by Storrs Mc-
Call, Oxford, Clarendon Press, 1969.
Macrobius, Commentaire sur le songe de Scipion, translation Stahl,
W.H., New York, London, Columbia University Press, 1952.
Maier, A., Zwei Grundprobleme der scholastischen Naturphilosophie,
Das Problem der intensiven Grosse, Die Impetustheorie, 2. Aufl,
Roma, Edizioni di storia e letteratura, 1951.
Marcus Aurelius, Pensees, texte etabli et traduit par Trannoy, A.I.,
preface de Puech, A., Paris, Les Belles Lettres, 1925, 2° edition,
1964.
Mates, B., Stoic Logic, Berkeley, Los Angeles, University of California
Press, 1961.
Mignucci, M., "Sur la logique modale des Stoiciens", Les Stoiciens et
leur logique, ouvrage collectif, Paris, Vrin, 1978, pp. 317-346.
- "f2q erci TO iroXu et Necessaire dans la Conception Aristotelicienne
de la Science", Aristotle on Science: the Posterior Analytics, Studia
Aristotelica, 9, ed. Antenore, Padova, 1981, pp. 173-203.
Moody, E.A., Truth and Consequences in Medieval Logic, Amsterdam,
North-Holland Publishing Co., 1953.
Moreau, J., Le sens du platonisme, Paris, Les Belles Lettres, 1967.
Mugler, Ch., "Sur quelques particularites de I'atomisme ancien", Re-
vue de Philologie, N° 79, 1953, pp. 141-174.
- La physique de Platon, Paris, Klincksieck, 1960.
274 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

- Platan et la recherche mathematique de son epoque, van

Berkhoven, Naarden, 1969 (Reprint of the P.H. Heitz edition pub-
lished at Strasbourg in 1948).
Nemesius, De natura hominis, graece et latine, Matthaei, Halae
Magdeb., Gebauer, 1802.
Ockham, W., The Tractatus de "Praedestinatione et de Praescientia
Dei et de Futuris Contingentibus", ed. Boehner, P., Franciscan In-
stitute Publications, n° 2, St. Bonaventure, New York, 1945.
- Predestination, God's Foreknowledge and Future Contingents,
translated by Adams and Kretzmann, New York, Appleton Cen-
tury Crofts, 1969.
Oesterle, J.T., On Interpretation (Aristotle), Commentary by St.
Thomas and Cajetan, Milwaukee, Marquette University Press,
1962.
Origen, Contra Celsum, Origenes, Werke, Band 1-2, ed. Koetschau,
P., Leipzig, 1899-1913, 3 vol.
Pauly, A., Pauly's Real-Encydopddie der classischen Alterthumswis-
senschaft, Stuttgart, Metzler, 1893.
Philoponus, In Analytica priora commentaria, ed. Wallies, CAG XIII
2, Berlin, 1905.
Plato, Oeuvres completes, tomes I-XIII, Paris, Les Belles Lettres, 1920-
1964.
Poincare, H., La Science et I'Hypothese, Paris Flammarion, 1968.
- La valeur de la Science, Paris, Flammarion, 1970.
Popper, K., La logique de la decouverte scientifique, traduction Rutten-
Devaux, Paris, Payot, 1973.
Plotinus, Les Enneades de Plotin, par Bouillet, M.-N., Paris, Hachette,
1857, 3 vol.
- Plotini Opera, ed. Henry, P., Schwyzer, H.-R., Paris, Desclee de
Brouwer, Bruxelles, L'Edition Universelle, 1951, 3 vol.
- Enneades, ed. et trad, par Brehier, E., Paris, Les Belles Lettres,
1964, 7 vol.
Plutarch, Plutarqui Moralia, Leipzig, Teubner, tomes I-VIII, 1953-
1974.
- Plutarch's Moralia, tome XIII, Parts I & II, english edition and
translation by Cherniss, H., Cambridge London, Loeb, 1976, 2 vol.
Prior, A.N., Time and Modality, Oxford, Clarendon Press, 1957.
- Formal Logic, Oxford, Clarendon Press, 2nd edition, 1962.
- Past, Present and Future, Oxford, Clarendon Press, 1967.
Quine, W.v., Word and Object, Cambridge, Massachusetts, M.I.T.
Press and John Wiley and sons, 1960.
 BIBLIOGRAPHY / 275

- The Ways of Paradox and other Essays, New York, Random

House, 1966.
Rescher, N., Urquhart, A., Temporal Logic, Wien, New York, Springer,
1971.
Robinson, R., "Plato's consciousness of fallacy", Mind, 1942.
Robin, L., La theorie platonicienne des idees et des nombres d'apres
Aristote, etude historique et critique, Paris, Alcan, 1908.
- Etudes sur la signification et la place de la physique dans la
philosophic de Platan, Paris, Alcan, 1919.
- Pyrrhon et le Scepticisme grec, Paris, P.U.F., 1944.
- Platon, Paris, Alcan, 1935, nouvelle edition mise a jour et
completee par Schuhl, P.-M., 1968.
Rodis-Lewis, G., Epicure et son ecole, Paris, Gallimard, 1975.
Russell, L.B., The principle of Mathematics, London, Allen-Unwin,
1956.
- Logic and Knowledge, ed. Marsch, R.C., London, Allen-Unwin,
1956.
Sartre, J.-P., Situations HI, Paris, Gallimard, 1949.
Sedley, D., "Diodorus Cronus and Hellenistic philosophy", Proceedings
of the Cambridge Philosophical Society, 203, New Series, 23, 1977,
pp. 74-120.
Sextus Empiricus, Opera, English translation by Rev. R.G. Bury, Lon-
don, Heinemann, W.; Cambridge, Massachusetts, Harvard Univer-
sity Press, 1967-1971, 4 vol.
Schuhl, P.-M., Le Dominateur et les possibles, Paris, P.U.F., 1960.
Simplicius, In Aristotelis Physicorum libros quattuor posteriores com-
mentaria, CAG X, Berlin, Hermannus Diels, 1895.
- In Aristotelis de Caelo commentaria, CAG VII, Berlin, Heiberg,
1894.
- In Aristotelis Categorias commentarium, CAG VIII, Berlin, Kalb-
fleisch, 1907.
Stobaeus, Anthologii libri duo priores sen Eclogae physicae et ethicae,
Berlin, Wachsmuth, 1884.
Storrs, Me Call ed., Polish Logic 1920-1939, Oxford, Clarendon Press,
1967.
Suppes, P., Probabilistic Metaphysics, Oxford, Blackwell, 1984.
Tannery, P., Memoires scientifiques, publ. par Heiberg & Zeuthen,
Toulouse, Privat, Paris, Gauthier Villars, 1912-1950, 17 vol.
Tarski, A., McKinsey, J.O., "Some theorems about Sentential calculi of
Lewis and Heyting", Journal of Symbolic Logic, XIII, 1948, pp. 13-
14.
276 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Tarski, A., Logics, semantics, metamathematics, papers from 1923-

1938, trad. Woodger, J.H., Oxford, Clarendon Press, 1956.
Tatianus, Tatiani oratio ad graecos, ed. Schwartz, Heinrich, 1888.
Taylor, A.E., Plato, the Man and his Work, London, Methuen, 2nd
ed. 1969 (1926).
Themistius, In libros Aristotelis de Caelo Paraphrasis hebraice et la-
tine, CAG V, Berlin, Landauer, 1902.
St. Thomas, Expositio Libri Peryermenias, in Opera Omnia, ed. Leon-
ina, t. 1*1, Roma-Paris, 1989 (see also Oesterle).
- In Aristotelis libros de Caelo et Mundo, de Generations et Cor-
ruptions, Meteorologicorum Exposito, ed. Spiazzi, Turin-Rome, Ma-
rietti, 1952.
- In duodecim libros Metaphysicorum Aristotelis Expositio, ed.
Cathala-Spiazzi, Turin-Rome, Marietti, 1964.
- In octo libros Physicorum Aristotelis Expositio, ed. Maggiolo,
Turin-Rome, Marietti, 1965.
Vlastos, G., "Minimal Parts in Epicurean Atomism", Isis, vol. 56, part
2, n° 184, 1965, pp. 121-147.
- Platonic Studies, Princeton, Princeton University Press, 1973.
Vuillemin, J., Physique et metaphysiques kantiennes, Paris, P.U.F.,
1955.
- Mathematiques et metaphysiques chez Descartes, Paris, P.U.F.,
1960.
- De la logique a la theologie, Cinq etudes sur Aristote, Paris, Flam-
marion, 1967.
- Le Dieu d'Anselme et les apparences de la raison, Paris, Aubier,
1971.
- "Kant aujourd'hui", Actes du Congres d'Ottawa sur Kant, octo-
bre 1974, Ottawa University Press, 1975, pp. 17-35.
- "L'argument dominateur", Revue de metaphysique et de morale,
1979, pp. 225-273.
- "La theorie kantienne des modalites", Akten des 5. Int. Kant
Kongresses Mainz 4-8. April 1981, Teil II, pp. 149-167, Bonn,
Bouvier, 1981.
- "Le carre chrysippeen des modalites", Dialectica, vol. 37, fasc. 4,
1983, pp. 235-247.
- "Les formes fondamentales de la predication: un essai de classi-
fication", Recherches sur la philosophic et le langage, Langage et
Philosophic des sciences, Cahier n° 4, DREUG, Grenoble, Paris,
Vrin, 1984, pp. 9-30.
- What are philosophical systems? Cambridge University Press,
1986.
 BIBLIOGRAPHY / 277

Wajsberg, M., "Untersuchungen iiber den Aussagenkalkiil von A. Heyt-

ing", Wiadomosci matematyczne, vol. 46, 1938, pp. 45-101.
Williams, B.C., "The Sea Fight Tomorrow", Structure, Method and
Meaning, ed. Henle, P., New York, 1951.
Windelband, W., Heimsoeth, H., Lehrbuch der Geschichte der Philoso-
phic, Herausg. v. Heimsoeth, Tubingen, JCB Mohr, 1948.
Wright, G.H. von, An Essay in Modal Logic, Amsterdam, North Hol-
land Publ. Co, 1951.
- "The Master Argument of Diodorus", Essays in Honour of Jaakko
Hintikka (ed. Saarinen, Hilpinen, Niiniluoto and Hintikka, Dor-
drecht, Reidel), 1979, pp. 297-307.
- "Time, Truth and Necessity", Intention and intentionality, Es-
says in Honour of Anscombe, G.E.M., ed. by Diamond, C., and
Teichmen, J., Brighton, The Harvester Press, 1979, pp. 237-250.
- "Diachronic and Synchronic Modalities", Theorema, vol. IX/3-4,
1979, pp. 231-245.
Zeller, E., "Ueber den xupieuwv des Megarikers Diodorus", Kleine
Schriften, I, 1910.
Index of quotations of ancient and
mediaeval texts

Albert of Saxony 153 (289-290): 144nl9,

Logica n20
II, ch. 10: 216n25
St. Anselm
Alexander Aphrod. Proslogion
In Anal. Pr. t. I, VII: 192n77
177, 25-33: HOnlO
180: 98n53 Aristoteles
181,13: 98n54 De Int.
183,34-184,10: 57n27 5, 17a12: 158
De Fato 7, 17b28-33: 137n2
IX 175-176 (208-209): 9, 18a28-19b4: 133-148
163n58 13, 22a36-23a16: 32nl7
X 177 (211): 163n55, n56 Anal. Pr.
XII 181 (223): 163n58 I 23, 41a23-30: 19n6
XXII 191-192 (260-261): Phys.
163n59 I 3, 186a24-25: 152n31
XXII 192,21-24 (262-263): I 7, 191a7: 152n32
163n57 I 8, 191*27-32: 152n32
I 8, 191b12-17: 152n32
Ammonius II 5,196b34-197a4: 161n52
In De Int. II 6, 197b14-15: 161n52
130-131 (251-252): 64n43, II 6, 197b18-21: 161n52
141nl2 II 8,198b29-32: 162n54
138-139 (264-265): 138n4 III 1, 200b32: 230nl9
140 (267-268): 141nl4 III 1, 201b7-13: 231n20
141 (269): 142nl5 III 2, 201b29: 231n21
148 (274-275): 142nl6 IV 10, 218a6: 152n32
279
280 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

VI 1, 231b21-232a17: T 8, 1012b2-4: 140nll

53nl5 A 3, 1025a14-34: 162n53
VI 10, 240b10-241a6: E 2, 1027a20-26: 162n53
53nl5 E 4, 1027b25-1028a2:
VIII 5, 257a31-257b13: 160n46
232n23 0 1, 1046al-2: 152n32
VIII 5, 257b9-10: 231n22 0 2, 1046a36: 31nl7,
De Caelo 144n20
I 12, 281*6-7: 19n6 0 3, 1046b29: 144n20
I 12, 281a15: 18n3 0 3, 1046b29-32: 62n38
I 12, 281bl: 152n30 0 3, 1047a20-29: 26nl4
I 12, 281b10-20: 22 0 3, 1047a24-25: 245n2
I 12, 281b15-20: 152n30, 0 4, 1047b3-14: 152n31
227 0 4, 1047b8-9: 151n29
I 12, 282a22-25: 152n30 0 5, 1048a16-21: 127,
I 12, 283a24-29: 152n30 246n3
I 12, 283a27: 152n31 0 10, 1051b2-5: 159n45
I 12, 283b6-17: 15-33 0 10, 1051b30-1052all:
De Gen. et Corr. 160n47
II 11, 337b3: 138n4 0 10, 1051b35-1052a2:
De anima 160n48
I 4, 408a34-408b30: 1 1, 1053a33: 143nl8
232n25 K 8, 1064b30-1065a14:
Problemata 214n21
XVII 3, 916a18-39: 97n47 K 8, 1064b32-1065a3:
XXX 1, 954a25: 144nl8 161n52
XXX 14, 957a32-35: K 8, 1065a6-21: 149
143nl8 K 8, 1065a14-17: 147n23
Metaph. K 8, 1065a16: 95n45
A 8, 989a30-989b21: K 8, 1065a21-26: 150n27
137nlO A 1, 1069a30-1069b2
T 3,1005b24: 139n7 A 2, 1069b14-34: 152n32
T 7, 1011b25: 158 Eth. Nic.
T 7, 1011b25-29: 139n9 VI 2, 1139a13-14: 26
159 VI 2, 1139b5-ll: 27, 79
T 7, 1011b26-29: 157n42
T 7, 1011b29-30, 36: St. Augustine
139nlO De Civitate Dei
T 7, 1012a2-5: 156-157 XIII 17: 228n8
T 7, 1012a25: 139n7
T 7, 1012a26: 139n8 Atticus
T 8: 140 Fragments
 INDEX OF QUOTATIONS OF ANCIENT AND MEDIAEVAL TEXTS / 281

801a-804b: 227n6 IX 18: 212

IX 19: 212
Boethius IX 19: 167n3, 210nlO, 212
In Arist. De Int. IX 20: 211nl6
234,9-12: 44n2, 44n3, X-XII: 207-213
115n20 X 20-21:185n54, 208n3
234,10-235,9: 57n29
X 21-22: 208n3
412,8-21: 57n29
X 22: 169n3
Buridan X 23: 169n2
Consequentiae XI 24-25: 209n6
I, c.l: 216n24 XI 26: 209n7
Quaestiones super libros De XII 27: 10nl3, 209n8
Caelo XII28:209n9, 210nll
118: 77nl3 XIII 30: 116n24
119: 215n23 XIV 32-33: 210-211
Sophismata XV 33: 215n22
IV, 13: 218n35 XVI 37: 170n6
XIX 43: 118n26
Cicero XX 46: 174n23
Academica Priora (L. II) XX 46-48: 169n3
XXIII 74: 220n39 De Natura Deorum (L. II)
XXIV 77: 222n43 223n46 IX 24: 101n61
XXV 79: 223n47 XI 31: 101n62
XXV 81: 222n42 XIV: 118n30
XXVI 83: 222n41 XV 40: 101n63
XXVI 84: 223n45
XXXIII: 118n30
XXIX 93: 112nl7 XLVI 118: 98n50
XXX 97: 169nl 170n4
XXXI 99: 223n48 Diodoros Kronos (Boring,
XXXI 100: 212nl7 1972)
XXXIII 105: 221n40 fragm. 116-120, p. 129sq.:
XLI 126: 101n60 10nl7
De Fato
VI 12: 57n28, 62n36, Diogenes Laertius
63n41, 109 Vitae Philosophorum
VII 13: 31, 57n28 VII (75); 175,45-176,1:
VII13-14: 63n40 115nl9,119
VII 14: 106-107
VIII 15: 106-107 Duns Scotus, Johannes
VIII 16: 109 De rerum principio
IX 17: 44nl, 62n36, 64n42, IV, 297-299; q. 3, a. 3,
212, 213nl8 sect.l: 85n35
282 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

IV, 305; q. 4, a. 1, sect. 1: XXIV, 125-126; IV, d. 11,

85n34 q. 3, n. 22: 237n35
IV, 307; q. 4, a. 1, sect. 3: XXIV, 519; IV, d. 43, q. 3,
85n34 n. 18: 236n32
IV, 315-317; q. 4, a. 2,
sect. 5: 88n40 Epictetus
Opus Oxoniense (= In IV Li- Dissertationes
bros Sententiarum) II 19, 1-4: l-2nl
X, 628-629; I, d. 39, q. un.: II 19, 10: 4n2
82n29
X, 629-630;!, d. 39, q. un.: Epicurus
86n36 Treatise against the Megari-
X, 630-631;!, d. 39, q. un.: ans
89n41 Bailey, 159: 185
X, 637; I, d. 39, q. un.: Treatise on Fate
89n41 Bailey, 159: 185
X, 641; I, d. 39, q. un.: Letter to Pythocles
89n41 97: 183n44
XI, 523; II, d. 2, q. 10, n. 98:184n51
1: 237n38 Letter to Herodotus: 174n23
XII, 558; II, d. 12, q. 1, n. 40: 179n37
11: 236n34 41, 45b:186n61
XIII, 197-198; II, d. 25, q. 47: 178n35
un., n. 2: 239n44 50, 50-51: 175n25, 176n30
XIII, 199-201; II, d. 25, q. 57: 184n60
un., n.2: 83n30 77: 181n48
XIII, 201; II, d. 25, q. un., 79-80: 181n43
n. 7: 240n47 Letter to Menoeceus: 174n3
XIII, 207-208; II, d. 25, 133-134: 185
q. un., n. 12: 237n38, Principal Doctrines
238n40, 239n46 XI: 183n49
XIII, 208; II, d. 25, q. un., XII: 176n27
n. 13: 238n41 XXIV: 176n28
XIII, 208-209; II, d. 25, q. XXXVII-XXXVIII:
un., n. 14: 239n42 182n42
XIII, 221; II, d. 25, q. un., Fragments
n. 22: 239n46 LV: 185n58
Opus Parisiense (= Re- Vita Epicuri
portatorum Parisiensium 27-28: 185n56
Libri IV) 34: 175n25, 176n29
XXIII, 6; II, d. 12, q. 1, n.
10: 236n34 Eusebius
 INDEX OF QUOTATIONS OF ANCIENT AND MEDIAEVAL TEXTS / 283

Praeparatio Evangelica: 98- Philo (the Megarian)

99 Pragm. 135-137 (Boring, 41-
42): 128-129
Aulus Gellius
Nodes Atticae Philoponus
VI2:118n27 In Anal. Pr.
XII 1-3: 12n22 169, 17-21: 57n28
Lucretius Plato
De Rerum Natura Nom.
II, 252-293: 184n53 VII, 818 A7-E2: 226n3
II, 1047-1089: 186n63 IX, 898 C: 233n27
V, 526-533: 183n44 Farm.
129 D-E: 233n28
Macrobius Politic.
In Somnium Scipionis 273 A-C: 227n7
II, 13: 228n9 Soph.
Marcus Aurelius 254 D-F: 233n28
Thoughts Tim.
IX 41: 184n53 32 C: 225nl
38: 229nl4
Nemesius 40 E - 41 C: 225-226n2
De Natura Hominis 48 A: 226n4
38:101n58
Plotinus
Ockham Enn.
Tract, de Praedest. I, 8, §7, 4-7: 226n4
Q. 1, A 4: 79n23, 80n26 III, 7, §6, 21-26: 229nl2
Q. 1, C 5-6: 72n4, 73n5 III, 7, §6, 43-48: 229nl3
Q. 1, M 12-13: 71n3 III, 7, §7-10: 228nll
Q. 3, A 32: 82n28 III, 7, §11, 27-33: 230nl6
Q. 3, A 32-33: 82n29 III, 7, §11, 59-60: 229nl5
Q. 3, A 33:84n31 III, 7, §12, 11-14: 230nl7
Q. 3, B 33: 84n32, 85n35, III, 7, §12, 15-17: 230nl8
87n38 III, 8, §5, 10-14: 234n29
Q. 3, C 34-35: 89n41 III, 8, §10: 234n29
Q. 3, C 35: 88n39 VI, 2, §7: 234n29
Q. 5, A 38: 73n7 VI, 2, §16: 234n29
VI, 3, §23, 1-12: 234n29
Origenes VI, 3, §27, 28-35: 234n29
Contra Celsum (SVF)
V 20; II, n. 626: 100n56, Plutarch
102n66 Moralia (L. XIII)
284 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Plat. Quaest. 195,31-196,24: 57n28,

VIII, 1007 D: 226n3 129n51
De Comrn. Notit. adv. Stoic.
14, 10065 E - 1066 A: Stobaeus
120n29 Eclog.
De Stoic. Repugn. I, 17, 3 (p. 153, 7 W):
46, 1055 D-E: 57n28 103n68
Tatianus
Sextus Empiricus
Adv. Graecos
Adv. Math.
V, 14 (p. 4, 2 sq.): 101n57
I, 213: 177n32
I, 309-312: 152n32 Themistius
VIII, 11-13: 46n7 In De Caelo
VIII, 113: 52nlO 86,4-29 (ad I 12, 283b3-
VIII, 115: 52nll 12): 17n2
VIII, 132-139: 46n7
VIII, 254-256: 126n48 St. Thomas
VIII, 330-333: 52-53nl2 S. Th.
X,48:55n21 I, q. 14, a. 8, ad 3: 143nl8
X,86: 53nl5 I, q. 80, a. 2, resp.: 232n24
X, 87-88: 55n22 In De Int.
X, 91: 53nl6 ad 9, 18a28-33 (I 13,
X, 97: 54nl7 1-103: Editio Leonina,
X, 98: 54nl8 t. I* 1, p. 65-67): 137-
X, 99-100: 54nl9 138n3, n5, n6
X, 112-117: 66n45, 66n46 In Phys.
X, 142-143: 55n22 ad III 2, 201b29 (p. 546,
X, 347-349: 55n24 n. 2305 Marietti): 231n21
In De Caelo
Pyrr. Hyp.
ad I 12, 283b8 (p. 138 Ma-
II, 110-111: 53nl3, 62n37
rietti): 153n29
II, 111-112: 107n3
In Metaph.
II, 242: 55n22
ad K 8, 1065a14-17 (L. XI,
II, 245: 55n22
1. VIII, n. 2282 Marietti):
III, 71: 55n22
147n23
Simplicius
In De Caelo
335-336 (ad I 12, 283b6):
28nl6
In Phys.
886: 100n52
In Cat.
Index of Proper Names

Abelard, 33 216, 218, 220, 223, 225-

Academics, 222, 223 228, 230, 232, 233, 235,
The Academy, 133, 220 238, 242-245, 249
Adams, 74 Arius Didymus, 103
Ae'tius, 98 von Arnim, J. (see STOICS), 4,
Agathon, 26 98, 100-103, 121
Albert the Great, 238 Atticus, 227
Albert of Saxony, 216-218, 238 St. Augustine, 227, 228, 235,
Alcmeon, 97 240
Alexander of Aphrodisias, 57, Marcus Aurelius, 184
100, 110, 111, 113-115, Avicenna, 85, 235
121, 145, 163, 165
Ammonius, 64, 138, 141, 142, Bailey, C. (see EPICURUS,
144, 145 notes)
Anaxagoras, 139, 140, 174, 175 Barcan-Marcus R., 23, 218
Angstl, H., ix Baudry, L., 88
St. Anselm, 192
Bayle, P., 32, 64, 169, 172
Antiochus, 169
Becker, O., 10
Antipater, 3, 4, 102
Bekker, I., 147
Arcesilas, 220
Berge, P., 262
Aristarchus of Samos, 101
Bergson, H., 148, 149, 246
Aristotle, vii, 7, 8, 10, 12, 14-
16, 18, 19, 21, 22, 24-34, Bernoulli, J., 250
36, 47, 53, 55, 58, 60, 62, Blanche, 11, 45
74, 79, 86, 89, 91, 93- Bochvar, 158
97,99,102,110,121,127, Boehner,-P., 91
128, 133-167, 169, 170, Boethius, 44, 57, 58, 115, 139,
171, 174, 178, 179, 185, 144, 147
193, 208, 210-212, 214, Bollack, J., 186
285
286 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Bollack, M. (see BOLLACK, 207, 208, 212-214, 220-

J.) 223
Bonitz, H., 147 Cleanthes, 3, 43, 69, 91-96, 98,
Borel, E., 247, 248, 254, 255 99, 101, 102, 106, 113,
Boudot, M., 6, 10-13, 45, 51, 119, 122, 133, 172, 244
61, 71-75, 81, 144 Cornford, P.M., 233
Boursin, J.-L., 251, 254, 255 Cournot, A., 150, 162
Boyance, P., 183 Cratippus, 144
Brehier, E., 4, 44, 64, 64, 98, Cresswell, M.J., 19, 76-78, 107,
100, 101, 106, 109, 112, 109, 110, 121, 141, 157,
116, 169, 170, 174, 184, 218
221-223, 234 Cumont, P., 101, 102
Bull, R.A., 121 Cynics, 144
Buridan, 76, 77, 207, 215-219
Burks, A.M., 121 Damien, P., 96
Bury, 53-55, 62 Darwin, Ch., 201
Democritus, 174, 183, 184, 186
Cajetan, 31, 32 Descartes, R., 170, 188, 189,
Carneades, 91, 92, 115, 133, 191-193, 195, 196, 198-
170, 185, 195, 207-216, 203
219-223 Dialecticians, 13
Caussat, P. (see BOURSIN, J.- Dicaearchus, 144
L.) Diels, H., 103
Celluprica, V., 142 Diodorus, vii, 3-6, 9-13, 24, 29,
Chalcidius, 10 31-33, 35, 36, 43-48, 51-
Chaldeans (see BABYLONI- 55, 57-59, 61-71, 75, 79,
ANS), 102, 106, 111 80, 82, 93, 94, 96, 102,
Chanut, 191 107, 108, 111-113, 121,
Cherniss, H., 14, 15, 118, 152, 123, 128, 133, 142, 146,
220, 226, 232, 233, 237, 152-154, 163, 167, 172,
239 185, 197, 205, 212-214,
Chevalier, J., 151 218, 220, 225, 244-247,
Chrysippus, 3, 7, 11, 12, 23, 249-251, 254-261, 263
43, 63, 98-102, 105-116, Diogenes the Babylonian, 98
118-126, 128, 133, 172, Diogenes Laertius, 92, 115, 119
185, 207-211, 216, 244, Diogenes of Oenoanda, 184
245, 254-256 Doring, Kl., 4, 10, 12, 53-55,
Cicero, 10, 33, 44, 57, 62-64, 57, 62, 66, 128, 129, 152,
68, 98, 101,106,109,112, 153
116, 118, 119, 121, 133, Duhem, P., 98
169, 170, 172, 174, 185, Duns Scotus, 74, 82, 83, 85, 86,
 INDEX OF PROPER NAMES / 287

88, 89, 92, 225, 228, 235, Godfrey of Fontaines, 83, 239
238-240 Goldschmidt, V., 4, 98, 100,
126, 176, 177, 181, 182
Edghill, 137, 141 Granger, G.-G., 12, 18
Effler, R., 232, 238 Gregory of Rimini, 90
Eleatics, 55, 153 Gueroult, M., 192, 193, 198-
Elisabeth (Princess), 202, 203 200
Empedocles, 162, 199
Epictetus, vii, viii, 3, 4, 6, 8,13, Haack, S., 157, 158, 161
16, 22, 29, 43, 62, 92, 96, Hamelin, O., 118,121,128,129,
125, 133 151, 155, 223
Epicureans, 47, 133, 144, 170, Heimsoeth, H., 183, 184
171, 183, 187, 222, 223, Heinzmann, G., 161
230 Heracliteans, 98, 139
Epicurus, 7, 133, 140, 150, 155, Heraclitus, 98, 102, 139
169, 170, 172, 174-180, Hermite, Ch., 188
182-189, 192, 195, 198, Herodotus (EPICURUS' corre-
199, 201, 207-209, 212, spondant), 174-176, 178,
222, 228, 261 179, 183, 186
Euclid (of MEGARA), 67, 153 Herophilus, 55
Eusebius, 98 Heyting, A., 173
Hintikka, J., 14, 144, 152
Faust, A., 228
Hobbes, 33
Festus, 10
Hughes, C.E. (see CRESS-
Fichte, J.G., 198
WELL, M.J.)
F011esdal, D., 121
Hume, D., 109
van Fraassen, B., 160, 161
Fraenkel, H., 10 Jalabert, J. (see LEIBNIZ,
Frede, D., 141, 150 notes)
Frede, M., 63, 108, 109, 111, James, W., 246
115, 121
Frette, 31 Kant, I., 170, 186-192, 194-
von Fritz, K., 5, 95, 155 198, 203-205
Furley, D., 175, 177, 178, 185, Kelsen, H., 182
199 Kleene, S.C., 174
Kneale, W. and M., 6, 12, 44,
Galileo, 190 59, 111, 151,217
Gassendi, 192, 195, 201 Krantz, W. (see DIELS, H.)
Aulus Gellius, 12, 118 Kretzmann (see ADAMS)
Gerhardt, C.J. (See LEIBNIZ,
notes) Lambert, J.H, 157, 161, 189,
Gilson, E., 83-86,232, 235, 236, 219
239, 240 Laplace, K., 246, 261, 262
288 / NECESSITY OR CONTINGENCY: THE MASTER ARGUMENT

Leibniz, 20, 31-33, 64, 69, 73, Nicole, P., 201

86, 92-96, 118-120, 154, Nicomachus, 26, 30, 93
155, 165, 168, 170, 172,
185, 193-197, 199-204 William of Ockham, 13, 69-76,
Lejewski, C., 219 78-82, 84-92, 96, 108
Lesniewski, S., 219 Oesterle, G.T., 31, 32, 138
Lewis, 107 Origen, 98, 102
Lindgren, B.W., 247, 252, 253 Pacius, 137
Lucretius, 183, 184, 186, 199 Panaetius, 98
Lukasiewicz, J., 91, 145, 150, Pauly, A., 144
155-158, 165, 171, 172, Peripatetics, 44, 98, 144, 230
174 Philo of Megara, 52, 107, 108,
Lyceum (see PERIPATETICS) 128, 129
Lychet, 90 Philoponus, 57
Plato, 15, 36, 92, 127, 133, 153,
McElrath, G.W. (see LIND-
220, 225-228, 232, 233,
GREN, B.W.)
235-237, 240
Macrobius, 228
Platonic, 15, 33-35, 48, 150,
Maier, A., 238
220, 226, 227, 230, 234,
Malebranche, 200
235, 244
Mates, B., 51-53, 58, 59
Plotinus, 226, 228, 229, 233-
Megarians (school
235
ofMEGARA), 10,12,13,
Plutarch, 57, 118, 121,226
55, 62, 64-67, 141, 142,
Poincare, H., 255, 261
144, 152-154, 185, 213
Poisson, S., 247, 249, 252, 254,
Meinong, 217
256
Menoeceus (EPICURUS' cor-
Pomeau, Y. (see BERGE, P.)
respondant), 174, 185
Popper, K., 180, 181
Mersenne, Father M., 193
Porch (see STOICS)
Mesland, 192
Prior, A.N., 8-13, 19, 70-72,
Mignucci, M., 44, 110, 115, 122
74, 79, 80, 91, 105, 108,
Minio Paluello, L., 139, 147
112, 120-122, 145, 156,
William of Moerbecke, 138,
157, 171, 173, 214, 216,
144, 147
218
Moody, E.A., 216-218
Pythagoras, 98
de Morgan, A., 171-173
Pythocles (EPICURUS' corre-
Morus, Th., 192 spondant), 183, 184
Mosconi, J., 37
Mugler, Ch., 175, 227, 235 Quine, W.V., 155,170, 213, 214
Nausiphanes, 183 Regnier, M., 98
Nemesius, 101 Rescher, N., 6, 56
 INDEX OF PROPER NAMES / 289

Robin, L., 64, 208, 211 Tricot, J., 147, 149, 157
Rodis-Lewis, G., 176, 177, 183,
184 Urquhart, A., 6
Ross, D., 139, 140, 147, 149, Valla, L., 20
150, 158, 232 Verbeke, G. (see AMMONIUS)
Russell, L.B., 51, 217 Vidal, Ch. (see BERGE, P.)
Ryle, G., 112 Vives (see DUNS SCOTUS)
Vlastos, G., 175
Sands, M. (see FEYNMANN,
Vuillemin, J., 12, 67, 115, 122,
R.)
189, 199
Sartre, J.P., 183
Scharle, T., 219 Weyl, H., 140
Schuhl, P.M., 5, 95 Windelband, W. (see HEIM-
Sedley, D., 10, 13, 51, 53, 58, SOETH, H.)
66,67, 112,115, 142, 174, Wisemann, H. (see BOLLACK,
220 J.)
Sextus Empiricus, 47, 52-55, von Wright, G.H., 5, 6, 51, 78,
62,66,107,108,114,126. 148
176, 177, 180, 183, 188 Wyclef, 33
William of Shyreswood, 217
Simplicius, 28, 57, 98, 100, 129 Zeller, E., 4-6
Socrates, 220, 236 Zenon of Citium, 98, 99, 101,
Spencer, H., 201 119, 222
Spinoza, B., 203 Zenon of Elea, 10
Stagirite (see ARISTOTLE) Zenon of Tarsus, 99
Stobaeus, 102, 103
Stoics, 6, 10, 32, 33, 44, 47,
67,98-102,117,129,150,
163, 165, 169, 197, 210,
215, 222, 223, 229, 230,
255
Storrs McCall, 171, 174
Suppes, P., 262

Tannery, P., 101

Tarski, A., 156, 157, 159, 160
Tatianus, 101
Taylor, A.E., 227
Themistius, 17
St. Thomas, 31, 32, 83, 138,
143, 144, 147, 149, 151,
231, 232, 238, 239