North American Philosophical Publications

Time, Truth, and Knowledge in Ancient Greek Philosophy

Author(s): Jaakko Hintikka
Reviewed work(s):
Source: American Philosophical Quarterly, Vol. 4, No. 1 (Jan., 1967), pp. 1-14
Published by: University of Illinois Press on behalf of the North American Philosophical Publications
Stable URL: http://www.jstor.org/stable/20009222 .
Accessed: 17/12/2012 10:27

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact [email protected].

University of Illinois Press and North American Philosophical Publications are collaborating with JSTOR to
digitize, preserve and extend access to American Philosophical Quarterly.

http://www.jstor.org

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
American Philosophical Quarterly
Volume 4, Number i, January 1967

I. TIME, TRUTH, AND KNOWLEDGE IN

ANCIENT GREEK PHILOSOPHY
JAAKKO HINTIKKA

I
have thought that they can perceive in the back?
ground of the Greek mind with what we actually
IN this essay I shall discuss a tacit presupposition, know about Greek thinkers or of other facets of the
or a group of presuppositions, which seems to Greek civilization.3 Hence the largely justified
lurk behind certain doctrines of Aristotle's and qualms of professional philosophers about impu?
which seems to have been rather widespread in tations of implicit general presuppositions to the
ancient Greece. Greeks.
A generalization concerning such widespread It seems to me, nevertheless, that a closer study
tacit presuppositions has of necessity something of the most
articulate and systematic Greek
self-defeating about it. If such a generalization is philosophers may serve a purpose here. Such a
correct, the presuppositions
it postulates
were
philosopher is much more likely to make explicit
shared by the great majority of philosophers and of some of the presuppositions he shares with his
ordinary people within a culture. If so, there was countrymen than the majority of these. He may
little occasion for anyone to challenge these pre? even have to rely on these presuppositions in his
suppositions, to discuss them, or even to bring philosophical arguments. A careful study of the
them out to the open. In such circumstances, not
general presuppositions of an individual Greek
very much direct evidence is likely to be available philosopher may therefore throw some light on the
to show the existence of these presuppositions. implicit conceptual presuppositions of the ancient
This does not go to show that broad generaliza? Greeks in general.
tions concerning
more or less unconscious ways of Weare primarily concerned here with certain
thinking within this or that culture are without general features of Aristotle's philosophical think?
philosophical interest. In fact, they seem to be ing. Some of the assumptions he makes appear to
worth a
great deal more attention than
professional have parallels in other Greek philosophers, and
philosophers have devoted to them of late. The hence occasion the question whether there is some?

difficulty Imentioned perhaps explains part of this thing in the common background of all these
lack of interest, however. In most cases, it is not
philosophers to which these assumptions might be
very difficult to put forward intriguing suggestions related and which might partly explain them. I
the features of people's ways of shall make some these lines to?
concerning general suggestions along
thinking in different cultures or at different periods ward the end of the paper.
of intellectual history. The speculative philosophy
of history from Hegel onward bristles with such
II
However, it is much more
proposals. usually
difficultto substantiate them. Often it is rather The group of presuppositions we deal with here
difficult to connect the more or less implicit is connected with the notion of time.2 Many pre?
Weltanschauung which some philosophers of history suppositions in this group seem to stem from a

1These are illustrated of genuine

difficulties by the scantiness evidence for those fashionable contrasts between the ancient
Greek and the ancient Hebrew Weltanschauungen which have recently been effectively criticized by James Barr in The Semantics
of Biblical Language (Oxford, Clarendon Press, 1961). The theoretical problems involved in inferences from linguistic to cultural
data are also discussed, inter alia, by Joseph Greenberg in "Concerning Inferences from Linguistic to Nonlinguistic Data,"
Language in Culture, ed. by H. Hoijer (Chicago, University of Chicago Press, 1956); reprinted in Psycholinguistics, ed. by S.
Saporta (New York; Holt, Rinehart and Winston; 1961) and by Max Black in "Linguistic Relativity," The Philosophical Review,
vol. 68 (1959), pp. 228-238 ; reprinted inMax Black, Models andMetaphors (Ithaca, Cornell University Press, 1962), pp. 244-257.
2 to the category
The importance of people's attitudes of time as an indication of their value orientation is brought out
clearly by F. R. Kluckhohn and F. L. Strodtbeck in Varieties in Value Orientation (Evanston, Illinois; Row and Peterson; 1961).
A I

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 2 AMERICAN PHILOSOPHICAL QUARTERLY

characteristic tendency which permeates many asked to

give
an
example of an sentence,
arbitrary
different parts of Aristotelian thought. It is safer to he might have chosen something like "Socrates is
speak of a tendency here than of an assumption, awake" or "Socrates is
walking"?and these we may
for apparently Aristotle does not consciously express equally well by "Socrates is now awake"
choose this way of thought in preference to and "Socrates is walking at the present moment."

explicitly formulated alternatives. Rather, he takes If Aristotle's tacit assumption is expressed in this
this mode of thought as the only natural however, we are already
one, way, siding with the
without ever becoming quite clear of the alterna? moderns against him to some extent. What we have
tives it might have and certainly without ever said already expresses to some extent the spon?
articulating the alternatives. If we nevertheless taneous reaction of almost all modern logicians
want to formulate this tendency as an explicit and philosophers to the sentences of the kind we
we may say that for Aristotle the are discussing. A modern
assumption logician is likely to avoid
typical sentences used in expressing human the use (and the mention) of such sentences as
knowledge or opinion are not among those Quine much as possible.4 They are usually thought of by
calls eternal sentences
(or,
even among standing him as incomplete or indefinite sentences whose
sentences) but among those Quine calls occasion "meaning"
or "content" depends
on the circum?
sentences.z That is to say, they
are not sentences to stances in which they are uttered or otherwise
which we assent or from which we dissent once and propounded. Modern philosophers generally prefer
for all. They
are sentences to which we can sub? not to deal with such temporally indefinite sen?
scribe or with which we must disagree on the basis tences as "Socrates is awake" or "Socrates is
of some feature or features of the occasion on which
walking"; they prefer to discuss and to use instead
they are uttered (or written). In particular, the sentences obtained from these by somehow speci?
sentences Aristotle is apt to have in mind are fying the time to which they refer independently
temporally indefinite, they depend on the time of their of the moment of their utterance. The laws of logic
utterance.3a They may be said to be relative to the are formulated with only or mainly sentences of the
moment at which they are propounded. This latter sort in view, and other procedures
are some?
relation may be implicit, but it may also be made times thought of as being somehow fallacious.
"
explicit by the occurrence of such token-reflexive" It is not my purpose here to take sides for or
as "now" or "at the moment" the Aristotelian as contrasted
expressions present against assumptions
in the sentence in question. (Among these we have with the modern ones. It is important, however, to
to count also such as or them to see how differ. How would a
expressions "yesterday" compare they
"tomorrow" where another moment or
period of modern thinker argue for his view that the "con?
time is specified by reference to the present tent" or "meaning" of a temporally indefinite
moment.) In a sense they can be fully understood sentence (say, "It is now raining") varies? One
only if we know what moment this "now" is, i.e., way of doing so might be as follows. Suppose this
when the sentence in was uttered or is sentence is uttered on two different
question occasions, say
thought of as being uttered. If Aristotle had been yesterday and today. Then the facts that make
3W. Word and Object (New York and London; the Technology Press of M.I.T.,
V. O. Quine, and John Wiley and Sons;
i96o)> ?9 and ?40.
3a For a of the contrast between definite and indefinite statements see N.
development temporally Rescher, "On the
Logic of Chronological Propositions," Mind, vol. 75 (1966), pp. 75-96 (see especially pp. 76-78).
4 Instances are far too numerous for me to give more than a sprinkling of examples here: Bertrand
An Inquiry into
Russell,
Meaning and Truth (London, Allen and Unwin, 1940), p. 113; A. J. Ayer, Philosophical Essays (London, Macmillan, 1954),
pp. 186-187; Nelson Goodman, The Structure of Appearance (Cambridge, Mass.; Harvard University Press, 1951), p. 297;
Donald C. Williams, "The Sea Fight Tomorrow," in Structure, Method and Meaning, ed. by P. Henle (New York, Liberal Arts
Press, 1951), pp. 282-306, especially p. 287. Further examples will be given in the course of the discussion.
My list might give the idea that only fairly recent philosophers of an analytic bend of mind favor eternal sentences over
sentences. as Appendix A to A. N. Prior's Time and Modality
occasion This impression would be incorrect, (Oxford, Clarendon
Press, 1957) convincingly shows. In fact, in this respect most of our logical tastes go back to the seventeenth century at least.
An early version of the doctrine that eternal sentences are superior to occasion sentences is especially interesting in that it was
developed in conscious opposition to Aristotle. For Aristotle, the time-relation in a sentence is carried by the verb. This doctrine
is closely related to the fact that a typical Aristotelian sentence refers to the moment at which it is uttered or written, for it
was the verb that carried the assertoric element in a sentence and was thought of as creating the judgment expressed by the
sentence. In contradistinction to this view, it was already held by the authors of the Port-Royal Logic that time-reference is not
logically speaking a part of the verb. For them, the temporal aspect is, as it were, part of the subject-matter and not created
by creating the judgment.

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 TIME, TRUTH, AND KNOWLEDGE IN ANCIENT GREEK PHILOSOPHY 3

this sentence true or false are different in the two These quotations show that Aristotle saw no
cases. Yesterday it referred to yesterday's weather; difficulty in combining the two assumptions which
today it refers to today's weather. On one day the to a typical modern thinker are likely to seem
sentence is verified or falsified independently of its incompatible, viz., the assumption that the truth
verification or falsification on the other. It may be value of a temporally indefinite sentence changes
true on one
day and false on the other. All these with time, and the assumption that the sentence in
are taken to show that the two utterances of nevertheless one and the
things question may express
the sentence in cannot carry the same same content or or, as Aristotle
question proposition puts it,
meaning. Although the sentence in the grammatical one and the same belief or opinion (doxa) on the
sense of the word is one and the same, its content different occasions on which it is uttered or other?
or, as it is often said, the proposition it expresses on wise propounded. Apparently Aristotle did not
the two occasions is not the same.5 Hence if we find anything strange or awkward about his
want to have a
satisfactory correspondence between reconciliation of the two assumptions. Later, we
our thought and our language, between the logical shall examine some reasons why he felt this way.
and the grammatical form, we
really must use a Some comments are in order here. First of all,
different form of words on the two occasions. the authenticity of the Categories has sometimes
Whether the points I just made are correct or been challenged. Hence the relevance of the
not, Aristotle would not have accepted them. passages we just cited is perhaps not beyond doubt.
Aristotle would apparently have accepted the However, even if this work is not by Aristotle
doctrine that the sentence "It is raining" is made himself, it reproduces views current in the Lyceum
true or false by different sets of facts accordingly at a very early date. Hence its testimony can be
as it is uttered today or yesterday. However, he trusted provided that there are parallel statements
would not have been worried about the conse? in undisputably genuine works of Aristotle's. This
quence that one and the same sentence may be is in fact the case. Exactly the same point that was
true at one time and false at another. He would made in the Categories recurs in a slightly briefer
have rejected the notion of a proposition and would form inMetaphysics IX, 10, 1051hl3 ff. Hence it is
have stuck instead to the actual thoughts of the surely Aristotle's own view. Many of the passages
people who uttered the sentence on the two occa? that will be quoted in the sequel constitute evi?
sions. When doing so, he would have been willing dence to the same effect.
to argue that the thought expressed by the sentence It is also interesting to observe that the view put
today and yesterday is one and the same. And all forward in the Categories is not due to Aristotle's
this he not only would have been willing to say; desire to rule out recalcitrant facts or to enhance
he as much as said so quite explicitly: the architectonic neatness of his system. On the

contrary, it would have suited the argument of the

same Categories much better if the author could have
For the statement (logos) seems to be both true
and false. for example, that the statement taken the modern view that the truth-value of
Suppose,
that somebody is sitting is true; after he has got up this and sentences never
opinions full-fledged changes.
statement will be false. Similarly with beliefs. Suppose For the point made there is that substances are the
you believe truly that somebody is sitting; after he has
only entities that can take contrary attributes at
got up you will believe falsely if you hold the same different moments of time and nevertheless remain
belief about him. (Categories 5, 4324-29.)6 one and the same 5, 4a 1
numerically (Cat. off.).
Statements and beliefs, on the other hand, them?
To this doctrine the changing truth-values of
selves remain completely unchangeable in every way;
opinions and of sentences constituted an un?
it is because the actual thing changes that the contrary
comes pleasant counter-example. The
to belong to them. For the statement that some? counter-example
is sitting remains the it is because of a
was eventually ruled out by means of a rather
body same;
in the actual that it comes to be true at manoeuver. are the
change thing unsatisfactory (Substances only
one time and false at another. with beliefs. entities that assume contrary attributes because
Similarly
(Ibid., 4a34~4b2.) they themselves change; opinions and sentences
5
See,e.g., G. E. Moore, "A Defense of Common Sense," Philosophical Papers (London, Allen and Unwin, 1959), p. 35.
6The
translation is taken from J. L. Ackrill, Aristotle*s "Categoriae" and "De Interpretationen (Oxford, Clarendon Press, 1963).
In quoting other works of the Aristotelian Corpus, I shall normally use the familiar Oxford translation, edited by W. D. Ross,
although departures from it will not always be explicitly indicated. In quoting Plato, I shall mostly follow the translations in
the Loeb Classical Library.

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 AMERICAN PHILOSOPHICAL QUARTERLY
4

do this only because the facts they refer to change.) way. That such a tendency really existed is
The fact that the modern view was not even con?
betrayed by the fact that Aristotle frequently uses
sidered by the author of the Categories although it temporally indefinite sentences as putative
would have served his purpose perfectly illustrates examples of syllogistic premisses, contrary to his
the hold of the contrary view of him. own
explanations. In at least one passage he
We may also register the fact that the statements indicates explicitly that his example is of this sort.8
we quoted from the Categories are completely A study of Aristotle's usage also tends to support
categorical.
A sentence or a belief remains "com? our view of his notion of truth. He often speaks of
pletely unchangeable in every way" although the what was or will be or would be true to say at some
facts it refers to change. Furthermore, nothing is particular time. The following are cases in point:
said of beliefs (opinions) and sentences which
never their Hence For when nothing was separated out, evidently
change truth-values.
perhaps
could be truly asserted of the substance that then
the author is clearly thinking in terms of temporally nothing
existed (Met. I, 8, 98906-7).
indefinite sentences. Because of the parallelisms
But there are two senses of the expression "the
between the Categories and other parts of the
primary when in which something has changed." On
Aristotelian Corpus, we may infer that the same the one hand it may mean the primary con?
"when"
was the case with the Stagirite. the of the process of
taining completion change (the
I conclude, then, that Aristotle saw no obvious moment when it is true to say "it has . . .
changed")
difficulties in the assumption that a temporally (P^.VI,5,236a7-9).9
indefinite sentence one and the same
expresses
care with
thought or opinion on the different occasions of its Notice also the which Aristotle
utterance. My main suggestion is that he tended to habitually qualifies his statements of the law of
take temporally indefinite sentences as paradigms contradiction: one and the same sentence cannot

of all informative sentences. This led him, among be true and false at the same time, he says.10
other things, to define some of his key notions so Locutions of this sort are by no means infallible
as to be applicable only or primarily to temporally evidence. Such locutions could conceivably be
indefinite sentences or to
"opinions" correspond? used by a philosopher who is not subscribing to the
ing
to such sentences. assumptions I have imputed to Aristotle. In con?
nection with his explicit pronouncements on the
Ill relation of truth to time, however, they constitute
The fact that Aristotle preferred temporally useful circumstantial evidence. Their frequency
indefinite sentences is not belied by the fact that in shows how deeply ingrained the ways of thinking
his syllogistic theory, and in his theory of scientific were which we have found in Aristotle.
method which was built on his syllogistic theory, There are certain passages in Aristotle's works

Aristotle frequently says in so many words that a which may seem to call for a different interpreta?
in all the different
general premiss has to take
tion. Some passages seem to that in addi?
suggest
individuals of a certain sort, no matter whether tion to temporally indefinite sentences Aristotle
exist now or at some other moment of time. sometimes considered sentences some
they containing
For instance, "All men are mortal" is about all kind of more definite time-reference. The evidence
men, present, past, or future.7 That Aristotle had is not conclusive, however, but points predomi?
to stress this feature of the syllogistic premisses nantly in our direction. When Aristotle says that a
was a ten?
equally well indicate that there
sentence includes a he
may time-reference, normally

dency among his audience and perhaps even in means simply that its verb is in the present, past, or
himself to understand these premisses in a different future tense.11 Such a sentence normally implies
a

7
Analytica Priora I, 15, 34^17-18; Analytica Posteriora I, 4, 73a28-2g, and 8, 75021-36.
8At An. Pr. a syllogistic
I, 10, 3ob37~38 Aristotle considers conclusion which is not necessarily true, but is true "so long as"
the are true. These premisses thus cannot refer to individuals existing at all the different times, for if they did,
(ews) premisses
their truth-value could not change. An equally revealing example occurs at 3108-10, where Aristotle uses the terms "sleeping"
and "waking" in a counter-example to a syllogism. These are as clear-cut examples as one might wish of terms which apply to an
individual at a time only. Hence they contradict Aristotle's own admonitions as to how syllogistic premisses ought to be con?
strued.
9 Similar statements are found, e.g., in Met., IV, 7, ioi2a27~28, Met, V, 30, 1025314-15, Physics, VI, 8, 239328-29.
10 De Int. 10, 2oai6-i8.
E.g., Met. IV, 3, ioo5big-20, 23-32; 6, ionbi5-i8; Top. II, 7, 113322-23;
11 See De Int. De An. Ill, 6, 430330 if.
3, i6b6-i8;

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 TIME, TRUTH, AND KNOWLEDGE IN ANCIENT GREEK PHILOSOPHY 5

reference to the moment at which it is uttered. told by the Stoics, because "we want to know who
There are also passages in which Aristotle has in [writes]." Nevertheless, a sentence like "Dion is
mind some kind of closer specification of the time walking" is said to express a complete lekton, in
to which a sentence refers. But a second look at spite of the fact that it leaves room for the ana?
these passages shows that he is thinking of a speci? logous question: "When is it that Dion is
fication (of the time of the occurrence of an event) walking?"15
in terms of "the measurable stretch of time from From this it followed that the Stoics spoke
now onwards to or . . . from that on to now" of in the truth-value of a sentence
that, freely changes
(Phys. IV, 13, 222a24~28). For instance, in Parva and also (more properly) in the truth-value of
Natur alia (De Memoria 2) 452030 if. Aristotle dis? lekta. As is brought out very clearly by William
cusses the difference between "exact" and "in? and Martha Kneale, a lekton could change its
exact" estimate of time in connection with truth-value and even cease to exist.16 In his list of
memory.12 One of his examples is "the day before the different senses of a?rjd'fjs in the Stoics, Benson
yesterday," which is a specification of time in Mates distinguishes the use of this notion in
relation to the present day. In fact, the context connection with propositions (Sense I) from its
makes it quite clear that what Aristotle is thinking use in connection with sentences that can change
of in this whole passage are exact and inexact their truth-value or, as Mates calls them, propo
specifications of the length of time which separates sitional functions with a time-variable (Sense II).17
the remembered event from the present moment The distinction does not seem to be motivated,
e.g., Hence we have an however. In some of Mates's own of
(cf., 45208 ff.). again fact, examples
instance of a temporally indefinite time-reference. Sense I are easily seen to involve sentences (or
lekta) with changing truth-values. Cases in point
are found in Diogenes Laertios, Vitae VII, 66,
IV
where the temporally indefinite sentence "It is
The same predilection for temporally indefinite day"
occurs as an
example,
as well as in Sextus
sentences is found in other ancient philosophers, Against the Logicians
II, 10-13.18 The
Empiricus,
although not always in as explicit a form as in latter passage is not unambiguous by itself; how?
Aristotle. Virtually all the examples of singular ever, its import is brought out when Sextus later
sentences which were used by the Stoics as returns to the same topic (op. cit. 85, 88-89), using
examples and which are preserved to us seem to be as an example the same temporally indefinite
temporally indefinite.13 What is more important, sentence as
Diogenes.
such temporally indefinite sentences are put for? One is also reminded here of the famous Stoic
ward by the Stoics as examples of sentences which controversies concerning the conditions of the
are to a
taken express complete lekton. These validity of implications.19 This whole controversy
complete assertoric lekta or in short axiomata of the is couched in terms of sentences (or lekta) with
Stoics are in many reminiscent of the truth-values. For Diodorus
respects changing instance,
"propositions" which many modern philosophers Cronus held that "if p, then #" is true if and only if
as
postulate meanings of eternal assertoric sen? q is true whenever p is true.20 If p and q were
tences.14 However, axiomata differ from proposi? temporally definite sentences, this would reduce to
tions in that they are temporally indefinite in the our own truth-table definition of material
implica?
same way as occasion sentences. tion. However, the Diodorean doctrine is known
By saying "writes,"
one does not a we are to have been directed
yet express complete lekton, against Philo's definition

121 am indebted to Professor John W. Lenz for to this passage.

13 calling my attention
Cf., e.g., Benson Mates, Stoic Logic (Berkely and Los Angeles, 1961, originally published as vol. 26 of the University of
California Publications in Philosophy) which contains a full discussion of the Stoic of a number
logic and translations of
sources. The closest approximations to genuine from this type of example that I am aware of are found in Cicero
exceptions
(De Fato, ix, 19, and xiii, 30).
14
For examples of this sort, see, e.g., Mates, op. cit., pp. 96, 113, 118, 121, 123, etc.
15 See
Diogenes Laertios, Vitae VII, 63; Mates, op. cit., p. 16.
16William
and Martha The Development
17
Kneale, of Logic (Oxford, Clarendon Press, 1962), pp. 144-146, 153-155.
Mates, op. cit., p. 132.
18
See the Loeb Classical Library edition of Sextus, vol. 2, pp. 244-247.
19 See
Mates, op. cit., pp. 42-51 ;Kneale and Kneale, op. cit., pp. 128-138.
20
See, e.g., Sextus Empiricus, Against theMathematicians, VIII, 112 ff.

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 6 AMERICAN PHILOSOPHICAL QUARTERLY

which is essentially just our own truth-table important word yiyvuoK , especially in Homer,
definition. Hence Diodorus is clearly presupposing and to other verbs, too, as
pointed
out by Snell.25
that/? and q are temporally indefinite sentences. Similar observations have been made often. W.
Mates explains the Diodorean definition in G. Runciman sums up his patient examination of
terms of quantification
over a time-variable.21 the relevant aspects of Plato's Theaetetus as follows :
This is justified and illuminating, provided that we "The general impression left by the Theaetetus is
realize that there is no trace whatever of such a that Plato continued to think of knowledge as a
treatment in the Stoics themselves.22 sort of mental seeing or touching."26 This impres?
sion is not changed by Runciman's scrutiny of the
Sophist: "Although Plato says that all statements
V
are either true or false and that all judgments are
One the most
of interesting things about the merely unspoken statements, he does not thereby
we are with is that it helps us . . . commit himself to any modification of what we
assumption dealing
to understand some of the most characteristic have seen to be his earlier position on the nature
features of the Greek epistemology. Since these and objects of knowledge."27
features are found not only in Aristotle but also in For our purposes it is especially relevant that
many other Greek philosophers, the existence of there was a marked
tendency to conceive of the
this connection suggests that we are really dealing highest forms of knowledge as being somehow
with a common tendency of many Greek thinkers. analogous to immediate observation as distin?
The most important feature I have in mind is guished from mere hearsay. In short, the highest
the widespread Greek doctrine that we can have form of knowledge was thought of as being com?
genuine knowledge only of what is eternal or at the parable to that of an eyewitness. In the introduc?
very least forever changeless.23 This doctrine tion to his edition of Plato's Meno, R. S. Bluck says
becomes very natural if we consider it as the out? that "the inferiority of op67j Sofa to ?morqfjLrj as
come of two tendencies: (i) A tendency to think of a state of awareness of the a priori is analogous to
temporally indefinite sentences as typical vehicles the inferiority of second-hand information about
of communication; (2) A tendency to think of empirical matters to the certainty of one who has
knowledge in terms of some sort of direct acquaint? learnt from personal experience."28
It is also
ance with the objects of knowledge, e.g., in terms instructive to note that the kind of universal
of seeing or of witnessing them. knowledge which Plato ascribes to the soul in his
I cannot here document the second tendency
as famous doctrine of recollection is explained by
fully as it deserves. That there is something here Plato (whether metaphorically or not is not at issue
:
worth being documented is already shown by the here) as being due to earlier personal experience
facts of the Greek language. One of the common
The soul, then, as being immortal, and having been
Greek ways to claim that I know was to use the
born times, and having seen all things that
again many
verb otSa which literally taken amounts to saying whether in this world or in the world below, has
exist,
that I have seen the thing in question. If we are to
of them all. 81
knowledge (Meno c.)
believe Bruno Snell, this was not a mere piece of
etymology but a fact the speakers of the language As pointed out by Snell, there is a striking
were aware of. According to him, "in the Greek example of this way of thinking in Homer.29 When
language we can frequently discern that the verb he appeals to his omniscient Muses to help him, he
elbivai means, to know on the basis of one's own does not their omniscience as a conse?
represent
observation."24 The same also to the quence of superhuman intelligence or of more
applies

21
Mates, op. cit., p. 45.
22
Cf. P. T. Geach's review of Mates in The Philosophical Review, vol. 64 (1955), pp. 143-145.
23This is one of the most of both Plato and Aristotle, and is found in other ancient
tendency striking characteristics philo?
sophers as well. Its role and background in ancient Greek thought does not seem to have been systematically studied, however.
24 Bruno vol.
Snell, "Ausdr?cke f?r den Begriff des Wissens in der vorplatonischen Philosophie," Philologische Untersuchungen,
29 (Berlin, 1924), p. 25.
25Bruno Harvard
Snell, The Discovery of theMind (Cambridge, Mass.; University Press, 1953), p. 13.
26
W. G. Runciman, Plato's Later Epistemology (Cambridge, England; Cambridge University Press, 1962), p. 52.
27
Ibid., p. 121 ; cf. p. 125.
28R. S. Bluck
(ed.), Plato's Meno (Cambridge, England; Cambridge University Press, 1961), p. 33.
29
Snell, The Discovery of theMind (op. cit.), ch. 7, especially p. 136.

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 TRUTH, AND KNOWLEDGE IN ANCIENT GREEK PHILOSOPHY 7
TIME,

insight into the laws that govern the events than knowledge changes, it may have changed in the
the humans possess. In Homer's own words, interval between his seeing it and his making the
statement. If the snow sometimes melts from Mount
Tell me now, Muses that dwell in the palace of
Olympus, then the fact that I have seen the snow
Olympus?
For you are you are at hand and know there does not go to show that there is snow there
goddesses,
all things,
now. Only if the snow never melts does the "I have
But we hear only a rumour and know nothing? seen it" assertion amount to
knowledge concerning
Who were the captains and lords of the Danaans.30 the state of affairs. the two tenden?
present Hence,
cies I mentioned made it very natural for the
As Snell puts it, "the goddesses are superior to
man for the simple reason that they are always at Greeks to adopt the view that there can be genuine

hand, and have seen everything, and know it knowledge only of what is unchangeable (and
now. . . ." In seen perhaps also of what is being perceived at the
fact, their having it was for the
Greeks almost a prerequisite for their knowing it. present moment).
more than this old This is in fact the way in which Aristotle argues
Perhaps significant testimony is
the fact that almost exactly the same view is for his doctrine that we have knowledge in the full
sense of the word only of what is eternal or forever
echoed by Plato and Aristotle:
:
unchangeable
You both admit, to begin with, that the gods perceive,
we
see, and hear everything, that nothing within the
Now what scientific knowledge (imaTf^fXTj)is, if
are to speak exactly and not follow mere similarities, is
compass of sense or falls
knowledge [therefore?]
from what follows. We all suppose that what we
outside their tr. A. plain
cognizance (The Laws X, 901 d; E.
know is not even capable of being otherwise; of
Taylor; cf. Aristotle, Poetics 15, 1454^2-6, and
things capable of being otherwise we do not know,
Empedocles fr. 129).
when they have passed outside our observation,
whether they exist or not. Therefore the object of
VI scientific is of necessity. Therefore it is
knowledge
eternal. . . . Ethics VI,
What happens now if this idea of genuine knowl? (Nichomachean 3, U39b20-23.)

edge as an eyewitness's knowledge is applied to the It is worthwhile to notice Aristotle's locution

kind of knowledge that can be expressed by means "We all suppose." It shows that Aristotle thought
of temporally indefinite sentences ?One case seems that what he was saying was not a peculiarity of his
to be clear: We know the things which are at the but rather a
among the Greeks.
commonplace
present moment within our
sphere of perception. A concurrent reason why itwas easy for Aristotle
But what about things which are not under our (and for the other Greeks) to accept the doctrine
present observation? Take, for instance, the piece that there can be knowledge only of what is
of putative knowledge expressed by the sentence eternally the same is implicit on his view that
"There is snow on Mount as uttered sentences like "Socrates sits" one and the
Olympus" by express
somebody who is not within the sight of the famous same
opinion every time
they
are asserted. It is
mountain. On what conditions is a man right who natural to say that an opinion of this sort cannot
claims to know what this sentence asserts? The amount to real knowledge if it is sometimes false.
Greeks had, if I am right, a tendency to take this For "false knowledge" struck the Greeks, as it
question as being equivalent to the question: strikes us as a misnomer.31 Hence
today, opinions
When does his claim, "I have seen it," amount to a which correspond to temporally indefinite sen?
conclusive evidence that the things
are now as he tences can constitute knowledge only if they are
says they are ? The obvious answer is :Only if the always true, i.e., only if they pertain to facts which
thing in question never changes. Only on this never change. And if opinions of this kind are
condition does it follow from his earlier observa? thought of as typical, then one may be inclined to
tion that things are still as they were at the time of that we can have knowledge
say generally only of
the observation. If the purported object of his what is indestructible and unchangeable.32

30
Homer, Iliad, beginning of the Catalogue of Ships.
31
That it seemed a misnomer to Plato may be gathered from Gorgias 454d, from Republic 476e, and from Theatetus 152c and
186e. Gf. also Parmenides, fr. 2 (Diels), lines 7-8.
32This is to all intents and purposes the point of Aristotle's remarks in the Categories 7, 7D27-30: "Destruction of the knowable
carries knowledge to destruction. . . .For if there is not a knowable there is not knowledge?there will no longer be anything
for knowledge to be of... ."

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 S AMERICAN PHILOSOPHICAL QUARTERLY

If we keep in mind that for Aristotle what is Plato accepted his [Socrates'] teaching, but held that

always necessary and that for him what is con? the problem [of a universal definition] applied not to
sensible but to entities of another kind?for
tingent sometimes will fail to be,33 we can see that things
this reason, that the common definition could not be a
essentially this point is made by Aristotle in
definition of any sensible thing, as they were always
Posterior Analytics I, 33, 88b3i~34:
changing (Met. I, 6, 98704-8).
. . . that which is necessary cannot be otherwise; but
A few lines earlier, Aristotle alleges that Plato
there are which, true and real,
propositions though the Heraclitean doctrine "that all
are also of being otherwise. it is not accepted
capable Obviously
. . . sensible things are ever in a state of flux and that
knowledge that is concerned with these.
there is no knowledge about them." Whether
these attributions are correct or not, similar
In many pronouncements of Aristotle's, both
juxtapositions of doctrines are found in Plato's
these reasons appear intertwined. The following
From our modern of view, we be
seems to be especially worth quoting : writing. point may
puzzled and surprised by the facility with which
For this reason, there is neither definition nor Plato connects in the Theaetetus the view that
also,
demonstration about individual sensible "Knowledge is perception" and the doctrine of
substances,
because they have matter whose nature is such that Cratylus that all things are constantly changing.34
they are capable both of being and of not being; for This connection is introduced at 152 d, and it is
which reason all the individual instances of them are
frequently made use of in the argument; witness
destructible. If then demonstration is of necessary such passages as the following:
truths and definition is a scientific process, and if, just
as knowledge cannot be sometimes knowledge and sometimes If all things are forever in motion, every answer to any
ignorance, but the state which varies thus is opinion, so
question whatsoever is equally correct (Theaetetus
too demonstration and definition cannot vary thus,
183 a).
but it
is opinion that deals with that which can be
otherwise than it is, clearly there can neither be
should?how could?the constant and
definition nor demonstration about sensible indi? Why
universal change postulated by Cratylus make true
viduals. For perishing things are obscure... when they have
passed from our perception; and though the formulae
statements impossible? Clearly only if the state?
remain in the soul unchanged, there will no be ments in question primarily pertain to the moment
longer
either definition or demonstration (Met. VII, 15, of their utterance, and to other moments of time
my in so far as remain constant.35
I03gb27-i040a5; italics). only things
The notion of a sentence with a truth
changing
This may be compared with An. Post. I, 6, value may perhaps
serve to
explain Plato's strange

7^33 if., where a similar point is made. Here we doctrine of degrees of truth which Runciman finds
have one of the reasons for Aristotle, so A sentence cannot at any
obviously why, puzzling.36 given
there could not be any genuine knowledge of moment be truer than another true sentence, but it

sensible particulars, but only of uni versais. In may be truer than another in the sense of being
De Anima III, 3, 428b8-o, we similarly read: "But true more often than the latter is. Thus the different
true opinion only becomes false when the fact degrees of truth are in effect different degrees of
changes unnoticed." unchangeability in the objects these truths are
about. This connection is explicitly made by Plato
in the Philebus (58 a-59 c).
VII
If my interpretation of De Interpretatione 9 is
Similar considerations seem to have been opera? correct, Aristotle there in fact uses the expression
tive in Plato, too. In fact, Aristotle attributes "true more
/xaAAov aA^^s meaning often."Z7

exactly the same mode of argument to Plato : Another form of the same idea is that an opinion
33 See
my paper, "Necessity, Universality and Time in Aristotle," Ajatus vol. 20 (1957), pp. 65-90.
34 For reasons other than the one we are interested in which Plato may have had for connecting the two doctrines, see
Norman Gulley, Plato*s Theory of Knowledge (London, Methuen, 1962), pp. 78-80.
35 that if all things were at rest, "the same things will always be true and false," and
Aristotle, too, assumes that, conversely,
if all things were in motion, "nothing will be true and everything will be false" {Met. IV, 8, ioi2b24~28).
36
Runciman, op. cit., pp. 124-125.
37 See "The Once and Future Sea Fight: Aristotle's Discussion of Future Contingents in De
Jaakko Hintikka, Interpretatione
IX," The Philosophical Review, vol. 73 (1964), pp. 461-492, especially p. 486.

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 TIME, TRUTH, AND KNOWLEDGE IN ANCIENT GREEK PHILOSOPHY 9

(doxa) is the better the more permanent it is. Thus If even the superiority of the divine knowledge was
in the Meno (89 c) Socrates says of a certain essentially based on the greater share of perceptual
opinion : evidence that the gods possess, what more could
there possibly be to knowledge than perception ?
Yes, but not only a moment ago must it seem correct,
but now also and hereafter, if it is to be at all sound.

In the Republic (see 430 a) belief or opinion is in IX

the same spirit compared to "a dye, a dye designed Our observations also into an
may put appro?
to be as 'fast' as possible." The importance of the an argument which Plato's
priate perspective
immutability of the Forms for Plato also becomes commentators have found puzzling and even
intelligible from this point of view. One of the mistaken. In Theaetetus 201 a-c Plato lets Socrates
most important roles of the Forms was just to the suggestion that knowledge can be
disprove
provide absolutely immutable objects of knowledge defined as true opinion a contrast
by drawing
and thereby secure the possibility of genuine between the correct opinion of a jury which is
knowledge.38 giving the right verdict and the knowledge of the
From this point of view we can also appreciate event which is possessed by an eyewitness. This
the firm connection which there was in Plato's
may seem puzzling because the objects of which an
mind between, on one hand, the distinction to a crime has knowledge are not the
eyewitness
between knowledge and true belief and, on the kinds of things of which we can according to
other hand, the distinction between Forms and Plato's doctrines have
(or Aristotle's) explicit
sensible particulars:
genuine knowledge; they are not immutable
Forms of which alone we can according to Plato
own verdict, is this. If intelligence and true
My then, have knowledge in the full sense of the word.
belief are two different kinds, then these things? Now the idea that an eyewitness's knowledge is a
Forms that we cannot perceive but think of?
only case was seen
paradigm of genuine knowledge to be
certainly exist in themselves; but if, as some hold, true
belief in no way from
one of the motives that led to the doctrine that we
differs intelligence, then all
can have knowledge in the full sense of the word
things we perceive through the bodily senses must be
taken as the most certain reality (Timaeus 51 d). only of what is eternal. This suffices to explain
Plato's use of an
apparent eyewitness's knowledge
as an example of real knowledge without jumping
to the conclusion that we could according to
VIII
Plato have genuine knowledge of sensible reality.
In we can see that there is a close con? The underlying reason why Plato could use the
general,
nection two
example he mentions
between different and even is closely connected with the
apparently
contradictory preoccupations of Plato and Aristotle. fascination which the idea that "Knowledge is
On one hand, they paid a great deal of attention had for him. Of course Plato ended
perception" up
to the idea that "Knowledge is perception," rejecting this idea, but not before conceiving the
whether in the last analysis accepted it or not.
they highest form of knowledge in analogy to percep?
On the other hand, they were attracted by the tion, as a kind of "mental seeing or touching," to
idea that we can have knowledge only of what use Runciman's expression.
never We can now see that these two One could put the point as follows What
: Plato
changes.
types of knowledge were both assigned a privi? wants to establish in Theaetetus 201 a-c is a distinc?
leged position by the Greek tendencies to conceive tion between knowledge and true belief. For this
of knowledge as a kind of immediate awareness it suffices for him to show that genuine
purpose
and to think in terms of temporally indefinite is related to true belief in the same way
knowledge
sentences. The two are two as an is related to
preoccupations really eyewitness's "knowledge"
sides of one and the same coin. the last two being
"knowledge" by hearsay,
A reference to our quotation from Homer obviously different from one another. Cornford's
perhaps helps us to appreciate the appeal of the view that we are here given only an "analogous
idea that "Knowledge is perception" to the Greeks. contrast" has much more to recommend itself than

38Cf.
Parmenides I35b-c and Cratylus 43gd~440c.

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 10 AMERICAN PHILOSOPHICAL QUARTERLY

it has recently been given credit for.39 It is true that altogether. De Interpretatione 3, i6b6-i8 and io?
Plato does speak of knowledge when he discusses 19hl 1-18 as well as De Anima III, 6, 43oa3o ff. are
the example, but he seems to restrict the scope of cases in point. However, in other passages the
the
example
in so many words to "matters which tenseless sense
is clearly presupposed (see for
one can know only by having seen them and in no instance, Topics V, 3, 13^5-18).
other way." The connection between the notions of time and
knowledge which we have pointed out enable us to
see that the difference between the two senses was
X
much smaller for Aristotle than we might other?
The idea of "changing truth" which we have wise expect. For even if a present-tense statement
found in Greek philosophers has to be distinguished appears to refer exclusively to the moment at
from the modern idea of the historical relativity of which it ismade, it can embody real knowledge (as
truth. A historical relativist is apt to argue for an we saw) only if itwould have been true to make the
absence of any absolute criteria of truth, which same statement at any other moment of time. In so

results in an impossibility of obtaining truths that far as

present-tense statements
(without explicit
are not liable to be given up when the criteria are temporal specifications or explicit restrictions to
changed. What an ancient philosopher like the present moment)
are taken to represent claims
Aristotle had in mind is almost exactly the oppo? to knowledge, they thus automatically assume a
site. He was not concerned with changes in our "tenseless" or
"conjunctive"
sense.

criteria of truth, but changes in the objects our The features of Aristotle's thinking which we
truths are about. He was not concerned with have noted are related inmore than one way to his
changes in the opinions we have about reality, but other logical, semantical, and psychological doc?
with changes in the reality itself. Aristotle did not trines. The idea that a temporally indefinite
think that the discovery of truth is usually very sentence may express one and the same
opinion
difficult; the difficulty was, rather, that all the truths or belief when uttered at different moments of time
concerning changing things had to be discovered is encouraged by Aristotle's idea that "spoken
(as it were) all over
again
at each new moment. sounds are symbols of affections in the soul" (De
Int. 9, It is obvious that the sentence, "It
i6a3~4).
is raining," as uttered by me today, is made true
XI
or false by a set of facts different from those that
Our observations concerning the notion of verified or falsified my utterance yesterday, "It is
knowledge which naturally results from the raining." But it is very natural to say that in some

Aristotelian presuppositions put these presuppo? sense the state of mind or attitude toward my
sitions themselves into a new perspective. We have environment that is expressed by the two utter?
so far spoken almost as if a temporally indefinite ances is the same. The facts which yesterday's
sentence in the present tense would for Aristotle utterance refers to are referred to today by the

normally refer merely to the moment at which it is sentence, "It was

raining yesterday."
But the
uttered. We can see now that this is not the "state of mind" which this utterance appears to
only
way in which Aristotle was willing to understand express seems to be entirely different from that
such sentences and also he could assume a utterance,
why expressed by yesterday's present-tense
different interpretation without any compunctions. "It is raining." For instance, the former involves a

The alternative is to understand present-tense thought about a temporal difference between the
sentences in a "tenseless" or
perhaps better time the sentence refers to and the time of the
sense. In this sense, a sentence will utterance of the sentence, whereas the latter pre?
"conjunctive"
mean that things are always in the way the sentence supposes no awareness of such a difference.40
states them to be. A number of passages in the Hence the idea that spoken words are symbols for
Aristotelian Corpus seem to rule out such a sense unspoken thoughts encourages the idea that one
39F. M. Plato*s Theory of Knowledge and Kegan cf. Runciman,
Cornford, (London, Routledge Paul, 1935), PP- 141-142; op.
cit., pp. 37-38. It may be instructive to observe that in the passage I quoted from Aristotle's Categories in n. 32 above he commits
an inconsistency somewhat similar to the one Plato seems to me to commit here. Saying that the "destruction of the knowable
carries knowledge to destruction" is from an Aristotelian point of view a solecism, for of things destructible we cannot have

genuine scientific knowledge (Aristotle's word is here episteme) in the first place.
40This is borne out by Aristotle's comments in De Memoria', One is here reminded of Russell's
see, e.g., 449^24-30. para?
"
doxical statement that 'present' and 'past' are primarily psychological terms . . ." (op. cit., p. 113).

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 AND KNOWLEDGE IN ANCIENT GREEK PHILOSOPHY II
TIME, TRUTH,

and the same temporally indefinite form of words the concepts of time and truth a reflection of the
expresses one and the same belief or opinion at the same attitude. It is as if philosophers were so
different times when it is uttered. absorbed in the present moment that they tended
to think in terms of sentences that contained a
reference to the present moment and which
XII
therefore dealt primarily with the present state of
It is tempting to express this point by saying that affairs. It is not easy, however, to find direct
Aristotle presupposed a principle of individuation evidence for (or against) this suggestion.
for propositions different from ours. This way of There are in any case other general features of
scene relevant to our sub?
putting my point turns out to need qualifications, the Greek intellectual
but it is nevertheless useful for many purposes. For ject. Whether or not there was a direct connection
instance, we can now see that one of my earlier between the general Greek attitude toward time
explanations has to be qualified. The typical and their philosophers' ways of handling this
sentences were said to there is the fact that the Greeks were not
considered by Aristotle notion,
contain, explicitly or implicitly, a token-reflexive very successful with their timekeeping?much less
expression like "now" or "at the present moment." successful than some earlier civilizations, not to
This makes the sentences in question in a sense speak of the Romans. Different cities could have
token-reflexive from our modern of view. different calendars to such an extent that the year
point
The proposition such a sentence expresses is might begin at different times and that the thir?
different on different occasions. However, the very teenth month that periodically had to be added to
same sentences were not token-reflexive for Aristotle the year was added at different times.42 Combined
in this sense, for the belief or opinion they express with a general neglect of public timekeeping, it is
at different times was for him one and the same. no wonder that the failure of chronology some?
Notice also that a defender of our modern view times amounted to a public scandal, as shown by
cannot argue that Aristotle handles temporally Aristophanes in the Clouds. Because of these failures
indefinite sentences incorrectly because they of chronology, there simply was no handy way for
express a different meaning or thought (Frege's the Greeks to take the course modern philosophers
Der Gedanke) in different contexts. Aristotle's pro? generally assume to be the only satisfactory one,
cedure implies that we have to individuate thoughts viz., to replace all references to the indefinite
or opinions expressed by temporally indefinite sen? "now" in a sentence by references to some chronol?
tences in the same way as these sentences themselves. ogy independent of the moment at which the
sentence is uttered. What point would there have
been in replacing the sentence "It is raining in
XIII
Athens today" by a sentence in which the day in
It may be suspected that the peculiarities we question is specified by
a reference to a calendar
have discussed are connected with the general of the form "On such-and-such a
day of such-and
attitudes the Greeks had toward time. It has been such a year it is raining in Athens" if the day and
suggested that the Greeks "lived in the present perhaps
even the year were different in another
moment" to a extent than the members of to the moment of utterance must
larger city ? A reference
other cultures.41 It might seem to see in have been as useful in general as a reference to a
tempting
the Aristotelian way of handling the relations of badly kept calendar.43
41 i (New York,
Cf., e.g., the hyperbolic statements of Oswald Spengler's in The Decline of theWest, vol. 1926), p. 131. "Clas?
sical man's existence?Euclidean, relationless, point-like?was wholly contained in the instant."
42 See A
History of Technology, vol. 3, ed. by C. Singer, E. J. Holmyard, A. R. Hall, and T. I. Williams (Oxford, Clarendon
Press, 1954-1958)? P- 569
48The connection
between the problems of time-measurement and the logic of tenses cuts deeper than modern philosophers
sometimes realize. As Y. Bar-Hillel points out in his useful paper, "Indexical Expressions," Mind, vol. 63 (1954), pp. 359-379?
the problem of converting temporally indefinite sentences into definite ones is not without its presuppositions. To express "The
sun is now shining" in temporally definite terms, I must know what time it is (at the place I am talking about). What comes
up in the ancient Greek world is primarily the practical difficulty of setting up a useful frame of temporal references. But when
we consider statements made at spatially distant points, even difficulties in principle start cropping up. These are of the well
known relativistic variety. If we want to build up a tense-logic which applies to more than one world-line (as is necessary it we,
for example, want to study quantified tense-logic) and which is compatible with physical reality, we cannot accept the usual
simple tense-logic with a linear structure but must instead have one which has the structure of Lewis' S4. On this point,
see my paper "The Modes of Modality," Acta Philosophica Fennica, vol. 16 (1963), pp. 65-81, p. 76.
especially

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 12 AMERICAN PHILOSOPHICAL QUARTERLY

It is difficult not to see in this neglect of a words in question are actually uttered. Now this
systematic calendar a
symptom of a
general lack situation supplies what ismissing from a temporally
of interest in timekeeping and in chronology. It is indefinite sentence itself; it enables us to know
perhaps not impossible to find some independent what the moment of time actually is to which the
evidence of the same attitude. For instance, Her? spoken word "now" refers. A logician eliminates
mann Fr?nkel's the indefiniteness of temporally indefinite sen?
study of the early Greek ideas of
time has led him to state bluntly that in the Iliad tences by projecting himself, as it were, to the
"There is virtually no interest in chronology, audience of someone actually uttering the sentence
neither in absolute chronology nor in relative in question. In a written text such words as "now"
one."44 He does not find much more interest in and "at the present moment" are indeterminate in
later writers, either. This squares rather well with a sense in which they are not indeterminate as a
the early Greek ways of expressing temporal part of actual speech. In a written culture the
relations that Eric Havelock has recently em? replacement of temporally indefinite sentences by
phasized. He points out that events were located in definite ones is more important than in an oral
time not with reference to any absolute chronology culture?to the extent that it might appear to its
but rather by reference to each other: "The basic philosophers
as the only "correct" course.

grammatical expression which would symbolise The primacy of the spoken word in relation to
the link of event to event would be simply the the written word is not merely an' explanatory
'and next'."45 hypothesis of a historian; it occurs as an explicit
phrase
doctrine both in Plato and in Aristotle. Plato's
of the written word as a mere aid to one's
XIV rejection
memory and as being dead and helpless as com?
It seems to me, however, that there is another pared with the spoken word is well-known and
closely related feature in Aristotle's background need not be elaborated here.46 Aristotle expresses
which was more
important
or at least easier to himself in different terms but equally clearly;
document. This is the fact that in some obvious according to him "written marks are symbols of
though elusive sense the Greek culture was largely spoken
sounds" in the same way as the latter were

based on the spoken and not on the written word. In in turn symbols of "affections in the soul."47

philosophical literature this oral character of the Eric Havelock has emphasized the same feature
Greek thought is shown, for example, by the in the background of Plato as I am now stressing
of as a method of in the common of the of
importance dialogue presenting background philosophers

philosophical ideas. One is also reminded of the the Socratic school.48 The use he makes of this
origins of logic in the technique of oral argumenta? idea is different from mine, however. Havelock
tion. Every reader of Aristotle's Topics knows to considers Plato's philosophy as an expression of a
what extent he was concerned with transition from a oral tradition to a con?
extraordinary poetic
the tricks and pitfalls of verbal exchange. culture based on written records. This
ceptual
From this emphasis on the spoken word it approach is appealing, but even if the uses Have
follows that the reasons for replacing temporally lock makes of it are fully justified it does not
indefinite sentences by temporally definite ones exclude the possibility that many traces of a
which modern have were to some reliance on the spoken word persisted in Plato's
philosophers
extent absent. If the spoken word is primary in and Aristotle's thinking. If I am right, the heavy
relation to the written word, one is apt to think of reliance on indefinite sentences as a
temporally
and discuss logical and semantical matters from medium of knowledge and of opinion is such a
the vantage point of some situation in which the trace.49

44Hermann
Fr?nkel, Wege und Formen fr?hgriechischen Denkens, 2d ed. (Munich, C. H. Beck, i960), p. 2. Cf. Eric Havelock,
notes 22 and 27.
Preface toPlato (Cambridge, Mass.; Harvard University Press, 1963), pp. 192-193,
45
Havelock, ibid., p. 180.
46
Cf., e.g., Phaedrus 275d-276a. A full documentation is given in Paul Friedlander, Platon, vol. 1 (3d ed., Berlin, Walter de

Gruyter & Co., 1964), ch. 5, especially pp. 116-121.

47De
Interpretation 1, i6a4~6.
48
Havelock, op. cit., passim.
49 In matter at hand, pace Havelock, Plato is an arch-conservative
fact, I would go further and say that in the particular
who in his logical semantical doctrines relies entirely on the spoken word. The references given in n. 46 above show how deeply
Plato felt about this matter.

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 TIME, TRUTH, AND KNOWLEDGE IN ANCIENT GREEK PHILOSOPHY 13

It is also remarkable that in some unacknowl? word is addressed to someone else, for a man
surely
edged sense the spoken word was for Plato and understands his own words if they are explicit
Aristotle even logically prior to the thoughts it enough for others to understand them. Hence the
expresses. Plato explained the nature of thinking grammatical identity of a spoken sentence (form of
by calling it "the inward dialogue carried on by words) easily becomes a criterion of the identity of
the mind with itself without spoken sound."50 This the corresponding thoughts,
too. Thus a transi?
was not merely a metaphor, for Plato felt free to tion from the spoken word to thoughts it expresses
carry out arguments and considerations in terms of does not seem to necessitate any changes in the
spoken
sentences and then to transfer the results so
principle of individuation involved.
as to apply to the corresponding thoughts as well. The conceptual primacy of the spoken word over
This is in fact the strategy of Plato's interesting and thinking is related to the absence of a full-fledged
important discussion at the end of the Sophist. notion of a proposition among the Greeks. By and
Aristotle does not formulate the logical primacy large, logical and semantical matters were dis?
of the spoken word as compared with thinking as cussed by Greek philosophers in terms of actual
an explicit doctrine. However, he sometimes does utterances rather than in terms of the thoughts or
exactly the same as Plato, that is to say, he too beliefs expressed. It is true that in the opening
sometimes bases his views of what we can say of paragraph of De Interpretatione which we discussed
thoughts on what can be said of their the order of relative importance seems to be the
people's
words. The passages quoted above from the opposite: Aristotle says there that spoken sounds
Categories 5 are instructive cases in point.51 It may are symbols of affections of the soul. However,
be observed that the word which Aristotle uses in Aristotle seems to have rather limited purposes in
these passages and which was translated by mind in this passage. As Ackrill points out, what
Ackrill as ''statement," viz., logos, often refers to the reference to the "affections in the soul" is
spoken words and sentences.
primarily intended to elucidate is the fact that
It is perhaps not irrelevant to mention that both although spoken and written sounds are different
Aeschylus (Agamemnon 1. 276) and Homer occa? in different languages, what the words express may
sionally referred to people's thoughts as "wingless" be "the same for all men," as Aristotle himself
or
unspoken words. says.52 As Ackrill also points out, "the notion that
utterances are symbols of affections in the soul does
not have a decisive influence on the rest of De
XV Interpretationen Like other Greek philosophers,
Aristotle thought of logical and semantical matters
This primacy of concepts applying to spoken largely in terms of the spoken language.
words and sentences in relation to
concepts refer? The same point is applied to Plato by F. M.
ring to thinking must have encouraged the idea Cornford:53
that one and the same temporally indefinite sen?
. . . maintain that there is a false
tence expresses one and the same Logicians might
thought (opinion, . . . which has a I
belief) even when it is uttered on different occa? "proposition" meaning, though
cannot believe it. With that we are not concerned, but
sions. If a thought is, logically speaking, nothing
only with judgments and statements that can be
but a statement addressed by the "speaker" to
actually made and believed by some rational being.
himself without spoken sounds, the question Plato never discusses that no one
"propositions"
whether a statement a
specifies complete thought propounds.
becomes tantamount to the whether the
question
speaker to add something
has to this statement Similar remarks can be made of the Stoic con?
when he tacitly addresses it to himself in order to cepts of a lekton and of an axioma, closely though
make it fully understood by the "hearer." The they approximate our notion of a proposition in
obvious answer is that no expansion is needed as many other respects. Thus Mrs. Kneale writes as

compared with occasions when the same form of follows:54

50 See Theaetetus
igoa and Sophist 263e.
51
Categories 5, 4a24~2g, 4a34~b2.
52
Ackrill, op. cit., pp. 113-114.
53
Cornford, op. cit., p. 113.
54
W. Kneale and M. Kneale, op. cit., p. 156.

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions
 AMERICAN PHILOSOPHICAL QUARTERLY
14

As a previously quoted passage has shown, the Stoics identical actual "now" that is postulated in the
wished to insist that an axioma which is to be described
Physics.
as true or false must somehow be present when it is so

described;... in one place Sextus clearly assumes that

lekta exist only when they are expressed or meant. XVI
If there is anything to the contention that the
As far as Aristotle is concerned, it would not be Greeks "lived in the present" to a larger extent
strictly true to say he never considers propositions than the members of some other tradition, the
that no one
propounds.
He comes very close to
justification (and perhaps also the import) of this
envisaging in De. Int. 9, i8b36. It is no acci?
one suggestion has to be spelled out in terms of such
dent, however, that this passage is, if I am right, concrete facts as those that are known about the

exactly the passage in which Aristotle came ways of Greek philosophers with the concepts of
closest to considering
statements which refer to a time and truth. Whether or not the facts we have
unique singular event and which therefore need an pointed out go very far toward establishing any
objective chronology.55 The other similar cases broad generalizations, in any case they show us
that there are in Aristotle of unasserted proposi? how the different facets of the Greek way of think?
tions are much less clear, and probably explain? ing are related to each other. The suggestion that
able as mere
fa?ons de parler. the Greeks were immersed in the present moment
Physics we find a neat meta?
In the Aristotelian more deeply than we are easily provokes a reply
physical projection of the semantical idea we have which points to the preoccupation of the Greek
discussed. According to the latter, the belief philosophers with the eternal and the immutable.
expressed by a sentence containing the word We have seen, however, that this preoccupation is
"now" remains ordinarily
one and the same. In closely connected with the reliance of these
the Physics, Aristotle argues that the "now" is in philosophers on temporally indefinite sentences.
the actual sense of the word always the same. It is Far from being
a
counter-example to the
temporal?
what "holds the time together" and makes it ity of Greek thought, this pursuit of the eternal is
continuous. (See Physics IV, 13, 222a 1o ff., and more of a manifestation of this very temporality?
There seems to be an interest? or rather an to compensate for it.
11, 2igbio-22oa4.) perhaps attempt
was
ing analogy between the way different moments of Plato's and Aristotle's ideal of knowledge
time are actualized by becoming in turn identified knowledge of eternal truths just because the
with the eternally identical "now," and the way in vehicles by means of which truths were thought of
which the content of a known "now"-statement by them as being expressed tended to make all
becomes relevant to the world at the different other kinds of truths ephemeral. In a sense I am
moments of time by becoming utterable at the thus led to agreement with the view of R. G.
moment in question. Collingwood that "the Greek pursuit of the eternal
It is thus literally the case that the word "now" was as eager as it was, precisely because the Greeks
was not token-reflexive for Aristotle, for each themselves had an unusually vivid sense of the
uttered "now" referred to the same forever
actually temporal."56

University ofHelsinki
and

Stanford University Received November 4, ig6^

55 "The Once and Future Sea Fight,"

Hintikka, especially pp. 465-466.
56
R. G. Collingwood, The Idea ofHistory (Oxford, Clarendon Press, 1946), p. 22.

This content downloaded on Mon, 17 Dec 2012 10:27:10 AM

All use subject to JSTOR Terms and Conditions