Time, Reality &

Experience
ROYAL INSTITUTE OF PHILOSOPHY SUPPLEMENT: 50

EDITED BY

Craig Callender

CAMBRIDGE
UNIVERSITY PRESS
PUBLISHED BY THE PRESS SYNDICATE OF THE UNIVERSITY OF CAMBRIDGE
The Pitt Building, Trumpington Street, Cambridge, CB2 1RP,
United Kingdom

CAMBRIDGE UNIVERSITY PRESS

The Edinburgh Building, Cambridge CB2 2RU, United Kingdom
40 West 20th Street, New York, NY 10011-4211, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia

© The Royal Institute of Philosophy and the contributors 2002

Printed in the United Kingdom at the University Press, Cambridge

Typeset by Michael Heath Ltd, Reigate, Surrey

A catalogue record for this book is available

from the British Library

Library of Congress Cataloguing-in-Publication Data

Time, reality & experience/edited by Craig Callender.

p. cm.—(Royal Institute of Philosophy supplement, ISSN
1358-2461; 50)
Includes bibliographical references and index
ISBN 0-521-52967 0 (pbk.)
1. Time—Philosophy. I. Title: Time, reality, and experience. II.
Callender, Craig, 1968. III. Series.
BD638.T566 2002
115—dc21 2002067445

ISBN 0 521 52967 0 paperback

ISSN 1358-2461
McTaggart and the Truth about Time
HEATHER DYKE

1. Introduction
McTaggart famously argued that time is unreal. Today, almost no
one agrees with his conclusion.' But his argument remains the locus
classicus for both the A-theory and the B-theory of time. I want to
show how McTaggart's argument provided the impetus for both of
these opposing views of the nature of time. I will also present and
defend what I take to be the correct view of the nature of time.
McTaggart begins by noting that, when we think about when, in
the temporal order of things, an event is located, there are two ways
in which we can do this. On the one hand, we can locate an event as
in either the past, the present, or the future. Once we have desig-
nated an event as occurring, say, three days ago, then every other
event temporally related to that event will have some determinate
location in either the past, the present, or the future. McTaggart
called the series of events ordered in this way the A-series. But we
can also locate events in time without reference to the past, present
or future. We can locate events as temporally related to each other.
We say that an event is earlier than, later than, or simultaneous with
some other event. We can use these relations to order every event in
a temporal series. McTaggart called the series of events generated
in this way the B-series.
This claim of McTaggart's is an uncontroversial one about the
ways in which, as a matter of fact, we think about the temporal loca-
tions and ordering of events in time. The A-series and the B-series
are just two different ways of ordering the very same events and
moments. For instance, the Great Exhibition of 1851 occupies an
A-series location: it is 149 years in the past. It is also located in the
B-series. It is, for example, 63 years earlier than the outbreak of
World War I, which implies nothing about its location in the past,
present, or future. By drawing this distinction between the A-series
and the B-series, McTaggart has simply drawn our attention to the
fact that we can represent the temporal ordering of events in these
two different ways. But in the light of this distinction, genuinely
substantial metaphysical questions arise: is one of these two ways of

' Sprigge (1992) is an exception.

137
Heather Dyke

representing the temporal ordering of events more fundamental

than the other? Does one of them truly represent the nature of
time?
One characteristic of the A-series, that the B-series lacks, is that
events don't keep the same A-series position for very long. If an
event is present, then very soon it will be past. An event that is
already past is gradually becoming more past. The B-series, on the
other hand, is what we might call a static ordering of events. If an
event occurs two days earlier than another event, then those two
events are forever related to each other in that way. So, the notion of
the A-series involves what we might call A-series change, which has
also been called the flow of time, or temporal becoming.
Having made this distinction between the A-series and the B-
series, McTaggart proceeds to present his argument for the unreal-
ity of time. It consists of two theses: a positive and a negative the-
sis. The positive thesis is that, if time exists at all, it must involve an
A-series. His argument for this depends on the claim that there
could not be change unless the events and moments of time formed
an A-series as well as a B-series. So, the A-series is essential for
there to be change, and change is essential for there to be time. His
negative thesis is that the notion of the A-series is self-contradictory,
so it cannot be part of reality. The conclusion that McTaggart
draws is that, since the A-series must exist if there is to be time, but
the A-series cannot exist because it is a self-contradictory notion,
time itself does not exist.
In general, philosophers have accepted one of McTaggart's
theses and rejected the other. So, while they recognize that his
argument is valid, they have thought it unsound. However, they
have disagreed over which thesis to accept and which to reject. The
A-theorists agree with his positive thesis, that the A-series is
essential for the existence of time. A-theorists think that a descrip-
tion of time that does not make reference to the A-series is an
incomplete description of temporal reality. Consequently, A-
theorists reject McTaggart's negative thesis, that the notion of the
A-series is self-contradictory. B-theorists, on the other hand, tend to
accept McTaggart's negative thesis. The notion of the A-series is
indeed self-contradictory, so the A-series cannot be part of reality.
But they reject his positive thesis. They think that time can exist
without its constituents forming an A-series. In particular, they
argue that change is possible without the elements of time occupy-
ing A-series locations. So, the A-series is self-contradictory, but
since it is not essential to time, time itself is real, but consists only
of a B-series.
138
 McTaggart and the Truth about Time

2. Why I reject McTaggart's positive thesis

McTaggart thinks that change is of the essence of time. There is a

sense in which we all think this, since we all think that time is the
dimension of change. Change occurs when something possesses
incompatible properties at different times: a tree is fully clothed
with leaves, and then bare, and then fully clothed once more. But
McTaggart means something more than this. For him, time itself
exhibits change. Times, and the events that occur at them, change
from being future to being present to being past. When McTaggart
claims that time is the dimension of change, he means that it is the
dimension of A-series change.
Why does McTaggart think that the existence of change requires
the existence of an A-series? He argues that if time consisted only
of a B-series, change would not be possible. If all there is to time is
B-series facts about the temporal relations between events, then
there cannot be change, according to McTaggart, because B-series
facts never change. Facts about the B-series relations between
events are fixed; they do not change. The only way in which the
characteristics of an event can change is if it changes from being
future to being present to being past. McTaggart's charge against
the B-theory can be put another way. If there is only a B-series so
that all events are equally real, no matter when they occur, and no
event ever changes its B-series location, then nothing really changes.
Reality is a fixed and unchanging entity.
McTaggart's objections to a B-series account of change are, I
submit, question-begging. He argues that nothing about a B-series
ever changes, so the B-series cannot accommodate change.
However, he assumes, for the sake of his argument, that change
means A-series change. It may be true that the B-series itself never
changes, but that doesn't mean that the constituents of a B-series
cannot undergo change. It may be true, to use McTaggart's exam-
ple, that if a poker is hot at one time and cool at a later time, noth-
ing about those facts ever changes, but it doesn't follow that those
facts do not constitute a change in the poker. McTaggart is assum-
ing that the paradigm subjects of change are events. It is events that
change from future to present to past. But a proponent of B-series
change need not accept this assumption. She can argue instead that
the paradigm subjects of change are objects. It is objects that
change by having incompatible properties at different times.
To put my objection in another way, McTaggart's argument
establishes nothing more than that without an A-series there cannot
be A-series change. A B-theorist can accept this, because for her,
139
Heather Dyke

there is no A-series, and there is no A-series change. McTaggart's

conclusion is a conditional. He claims to have established that if
there is time, then there must be A-series change. But all that he has
really established is that if there is an A-series, then there must be
A-series change. This conditional is acceptable to a B-theorist, since
for her it is true because both antecedent and consequent are false.
If the existence of time depends on the existence of A-series
change, then it would indeed follow that without an A-series there
could not be time. But all McTaggart has established is that the
existence of the A-series depends on the existence of A-series
change.

3. Why I accept McTaggart's negative thesis

McTaggart's argument that the notion of the A-series involves a

contradiction is deceptively simple, and strangely uncompelling on
a first reading. His premises are that the A-series positions are
incompatible, and that if the A-series exists, and with it A-series
change, then every event occupies every A-series position. It follows
from these premises that the A-series does not exist. The obvious
response, as McTaggart notes, is that no event satisfies all of the
incompatible A-series predicates at the same time, but only succes-
sively, and there is no contradiction in anything satisfying incom-
patible predicates at different times.
The obvious response, however, doesn't work. It says that noth-
ing is ever past, present and future at once, but only at different
times. There are two ways in which we can cash out this response.
Are the different times at which an event is past, present and future,
different times in the A-series or in the B-series? Taking the second
option first, the response now goes as follows: of course nothing can
be future, present and past. But something can be future at one
time, <i, present at a later time, t2, and past at a still later time, £3.
This way of understanding the obvious response does indeed avoid
the contradiction. But it is unacceptable to anyone wishing to retain
a genuine A-series in her ontology. To say that an event, E, is future
at tj, present at t2, and past at £3, is to say no more than that E occurs
at t2, which is later than t± and earlier than t3. The A-series claims
collapse into B-series claims. By anchoring the possession of incom-
patible properties to different times in the B-series, the A-series,
and A-series change, have fallen out of the picture. Qualifying the
A-series claims in this way yields B-series claims, which do not
change their truth-value as things change their A-series position.
140
 McTaggart and the Truth about Time

All that is described by these qualified claims is a fixed and

unchanging B-series.
So in order to avoid the contradiction, and retain A-series facts
and change, a defender of the A-series must relativize the posses-
sion of the incompatible A-series predicates to different times in the
A-series. Now the response goes as follows: of course nothing can
be future, present and past. But something can be present now,
while it was future and will be past. This move also succeeds in
removing the contradiction, but it does so by introducing a set of
second level temporal predicates, and while some of these are com-
patible with each other, there are some that are not. But if the A-
series, and A-series change, are real, then every event possesses
every second level temporal predicate, even the incompatible ones.
So the contradiction has not been removed, merely shifted up to
these second level temporal predicates.
What has happened is this. By saying that an event is present, was
future, and will be past, we have described things as they are now.
But because reality undergoes A-series change, things have not
always been as they are now, and they won't remain as they are now.
In order to incorporate A-series change into our description of A-
.series-involving temporal reality, we must recognize that the same
event also will be future, and was past. But these second-level tem-
poral predicates are incompatible with the ones we used to avoid the
contradiction in the first place. So, relativizing the possession of
incompatible A-series predicates to different times in the A-series
cannot eliminate the contradiction.

4. An alternative expression of McTaggarts paradox

I often find that people are initially resistant to McTaggart's rea-
soning in establishing his negative thesis. I therefore wish to
unearth the contradiction that he identified in a different way.
Recall McTaggart's A-series. The properties of events by which
they are ordered in the A-series are the properties of being past,
being present and being future.2 If we suppose that events really are
ordered according to these A-series characteristics, then we must
also admit that they change their A-series characteristics over time.
An event, like the Sydney Olympics, was once in the remote future,
and was recently in the near future. It is now in the present, will
soon be in the recent past, and will gradually recede into the more
2
There are also finer gradations of A-series locations such as being three
weeks ago, being this week, and being two minutes hence.
141
Heather Dyke

remote past. The question for McTaggart, and for us, is: does time,
in reality, exhibit these A-series characteristics? Do events really
possess the characteristics of being in the past, present or future,
and do they really change in respect of them over time?
Let's suppose that events really do possess these characteristics.
In that case, we can plausibly suppose that they are properties in
some sense. Indeed, this is a common way of ascribing an ontolog-
ical status to these characteristics, by those who think they are real.3
How do the properties of pastness, presentness and futurity differ
from each other? One thing that we can say is that they have differ-
ent extensions. The property of pastness applies to all those things
that are earlier than the present moment. Presentness applies to all
those things that are occurring simultaneously with the present
moment. Futurity applies to all those things that occur later than
the present moment. But now notice that I have presented a picture
of temporal reality that is only accurate for a moment. We can dis-
tinguish between past, present and future, in terms of their exten-
sions, but by doing so, we leave out the other feature of the A-series:
the continual change from future to present to past that everything
undergoes.
So we must try to distinguish between past, present and future in
a way that accommodates A-series change. But accommodating A-
series change removes our means of distinguishing between past,
present and future. Because everything successively possesses every
A-series property, it follows that the extensions of the properties of
pastness, presentness and futurity are all exactly the same. They all
apply to everything. And it is not simply that these properties have
the same extensions as a matter of mere contingent fact. If A-series
change occurs, then they necessarily have the same extensions, rein-
forcing the conclusion that there is no genuine distinction between
them.
One could object that the extensions of these properties are not
identical if there is a first or a last moment of time. A first moment
of time is never future, and a last moment of time is never past. But
this does not avoid the co-extensiveness objection. In that picture,
the property of presentness is co-extensive with the property of
being either past or future, or alternatively, of being non-present.
My conclusion stands, as being present, and being non-present are
necessarily co-extensive if A-series change occurs.
Here, then, is McTaggart's paradox in my terms. We can only
distinguish between the properties of pastness, presentness and
3
See for example, Smith (1993).
142
 McTaggart and the Truth about Time

futurity at some moment of time. But this yields a static 'snap-shot'

picture of tensed time, a picture that is patently false, because
everything is constantly changing its A-series property. But as soon
as we try to incorporate A-series change into the picture, we lose our
means of distinguishing between the A-properties. The distinction
between pastness, presentness and futurity collapses because every-
thing successively possesses them all.
To put my point another way, to suppose that the A-series is real
requires commitment to two theses. Firstly, one must hold that
there is a real, observer-independent distinction between past, pre-
sent and future. Secondly, one must hold that different distributions
of past, present and future obtain at different times. But it seems
that one cannot hold both of these theses. Marking the objective
distinction between past, present and future requires leaving A-
series change out of one's account because one can only distinguish
between past, present and future at a particular moment of time.
Holding the second thesis, that the distribution of pastness, pre-
sentness and futurity changes from moment to moment, involves
relinquishing our grip on the first thesis, that there is an objective
distinction between past, present and future. As the distribution
between A-properties changes the distinction between them col-
lapses, since they all apply to everything. The entire account thus
collapses under the weight of this contradiction.
It follows that time cannot be such that its constituents form an
A-series. To suppose that the A-series is real is to suppose that time
has these two features: an absolute distinction between past, present
and future, and a continual change in respect of this distinction that
the constituents of time undergo. But time cannot possess both of
these features, so the A-series is not real.

5. The A-theory's options

If, as I have argued, McTaggart's attack on the reality of the A-

series succeeds, what options remain? One option is simply to deny
that times and their contents form an A-series at all. There is no
objective, observer-independent distinction between past, present
and future; nothing really changes from being future to being pre-
sent to being past. Taking this line involves explaining why, if there
is no past, present, and future, we are misled by our experience into
thinking that there is. But for many this sort of response will be
unsatisfactory. What is needed, they argue, is not an error theory of
our possession of A-series concepts, but an account of them that
143
Heather Dyke

does not collapse in the face of McTaggart's paradox. I think there

are two potentially viable options for those sympathetic to the A-
theory, which I will briefly outline.
McTaggart himself suggests the first option when he says 'It is
never true that [an event] is present, past, and future. It is present,
will be past, and has been future.' (McTaggart (1927) 21). For
McTaggart the explanation cannot stop here, since it merely intro-
duces more complex tenses than the three simple ones, and because
A-series change is continually occurring, every event has every
complex tense, just as it has every simple tense, and some of them
are incompatible. So, as we have seen, for McTaggart this line of
response cannot avoid the contradiction. But an A-theorist could
take issue with McTaggart's claim that the explanation cannot stop
at this point. Take any event, E, that is happening now. We can say
of E that it is present, was future, and will be past, and in saying this
we do not contradict ourselves. What we have described is the pre-
sent state of affairs. E has the property of being present. It also has
the past tense property of being future, and the future tense prop-
erty of being past. Provided the A-theorist is willing to concede that
the present state of affairs is all that there is, she can avoid
McTaggart's paradox.
There are some A-theorists, presentists, who are willing to make
this concession,4 and I grant that it does offer a way out of
McTaggart's paradox, but whether it can constitute a viable meta-
physics of temporal reality is another question. For many, commit-
ment to the unreality of past and future will be too high a price to
pay for avoidance of McTaggart's paradox. Those presentists who
are willing to pay it must still show us that their picture of the world
is coherent, and coheres with our experience. And it is not obvious
that they can do this. The presentist's response to McTaggart effec-
tively involves denying that A-series change takes place. Certainly
the presentist can talk about events that will be present, and events
that will, in a week's time, be two weeks past, and this way of talk-
ing gives the impression that A-series change is consistent with the
presentist picture. But all these expressions really convey is that
every event is located somewhere in the A-series, and that, were a
different moment present, they would be located elsewhere in the
A-series. Presentism, it seems to me, cannot accommodate the
change in A-series positions that events and times undergo, for as
soon as it attempts to do so, it falls right back into McTaggart's
paradox.
4
For example, Prior (1970), Bigelow (1996), Craig (1998), and Hinchliff
(1996).
144
 McTaggart and the Truth about Time

The second A-theoretic response to McTaggart, that seems

viable at first sight, is suggested by Dummett (1960). Dummett
argues that what McTaggart's argument really shows is that there
cannot be a complete description of reality independent of some
perspective. According to Dummett, McTaggart implicitly assumes
that there can be such a description. When that assumption is com-
bined with his thesis that the temporal cannot be completely
described without the use of A-series expressions, the contradiction
quickly follows. If time is real, then the complete description of
reality contains incompatible facts, viz., for any event E, E is past,
present and future. McTaggart concludes that, since the complete
description of reality cannot contain incompatible facts, time is not
real. Dummett concludes instead that the false premise is the one
that says that there can in principle be a complete description of
reality. So time is real, but reality only contains some of the incom-
patible temporal facts.
If there can be no complete, observer-independent description of
temporal reality, then one of two possibilities follows. Either tem-
poral reality consists of two domains: that which we can consistently
describe and that to which we can in principle have no epistemic
access. If this is the right interpretation of Dummett, then the
burden of proof lies squarely with him. Why should we think that
there is any more to temporal reality than that to which we have
epistemic access? Alternatively, we can interpret Dummett as argu-
ing that the maximal consistent description of temporal reality that
can be given from a particular temporal perspective describes all
that there is. This alternative reduces to presentism. We can give a
complete description of the A-series location of every event given a
particular temporal perspective, and this would constitute a com-
plete description of present fact. If present fact is all there is, then
presentism is true. But if this is the right interpretation of
Dummett, then he faces the same problem that I outlined above for
the presentist.

6. Moving on from McTaggart

I think that time itself exists, but that the A-series doesn't. There
are no characteristics of pastness, presentness or futurity. There is
no flow of time. Nothing really changes from future to present to
past. But all of our temporal experience seems to suggest that there
is an A-series. How come we seem to be deceived by our experience
on such a massive scale? In what follows I will present an account
145
Heather Dyke

of our temporal experience that appeals only to the existence of a B-

series.
The feature of our experience that is most suggestive of the exis-
tence of an A-series is that we talk about events as if they were
located somewhere in the past, present or future. We say, for exam-
ple, 'It's nearly 5 o'clock', which suggests that 5 o'clock is located in
the proximate future, and will soon be present. We say 'World War
II ended 55 years ago', which suggests that the end of World War
II is located 55 years in the past. And when we say things like this,
what we say is determinately either true or false. I will call sentences
like this, which appear to locate events or times somewhere in the A-
series, A-sentences. It is undeniable that many A-sentences are true
when they are uttered, but what makes them true, if not the fact that
a certain event or time is located somewhere in the A-series?
According to the B-theory of time, the fact that makes a sentence
like 'World War II ended 55 years ago' true is the fact that the event
that it is about (the end of World War II) is 55 years earlier than the
utterance of the sentence.5 All events stand in fixed and unchanging
B-series relations to each other. Utterances of sentences are events
like any other, so they stand in temporal relations to other events. In
particular, they stand in temporal relations to the events that they
are about. An A-sentence that appears to locate an event somewhere
in the A-series will be true if and only if that event and the utter-
ance of the A-sentence itself stand in the requisite temporal relation
to each other. An A-sentence that locates an event in the present is
true if and only if the utterance of the A-sentence and the event
occur at the same time as each other. An A-sentence that locates an
event in the future is true if and only if the event occurs after the
utterance of the A-sentence. An A-sentence that locates an event in
the past is true if and only if the event occurs before the utterance of
the A-sentence. Facts about the temporal relations that obtain
between events and utterances about them are sufficient to account
for the truth of every true A-sentence.
The B-theory thus treats time in a way that is similar to our
treatment of space. When I say that 'London is here' I am not
attributing to London the property of being here. What makes my
utterance true, if it is true, is that the utterance occurs in the same

5
There are two different B-theoretic accounts of the facts that make A-
sentences true: the date version and the token-reflexive version. I argue
elsewhere (Dyke, forthcoming) that the token-reflexive version is prefer-
able. For the sake of simplicity, I only discuss the token-reflexive account
here.
146
 McTaggart and the Truth about Time

place as London. Similarly, when I say that 'it is now Autumn', I am

not attributing to Autumn the property of being present. What
makes my utterance true, if it is true, is that the utterance occurs at
the same time as Autumn.
A-sentences appear to change their truth-value over time. The
sentence 'The train is now arriving' is true at some times and false
at other times. According to the A-theory of time, the fact that A-
sentences change their truth-value over time reflects the fact that
events and states of affairs are continually changing their location in
the A-series. The reason why the sentence 'The train is now arriv-
ing' is sometimes true and sometimes false, is because the fact that
the train is now arriving is only a present fact at some times, but not
at others. It is only when it is a present fact that the sentence is true.
All this is denied by the B-theory of time.
The B-theory invokes the distinction between sentence-types and
sentence-tokens. A sentence-type has a 'changing' truth-value if
and only if some of its tokens are true and others false. Two tokens
of the sentence-type 'The train is now arriving' might have differ-
ent truth-values, but the truth-values they have are fixed and
unchanging. The truth-value that any token of this type has
depends on when it is produced. So, the claim that A-sentences
change their truth-value over time is wrong. The fact of the matter
is that some tensed sentence-types have some true and some false
tokens. This gives the impression that the sentence-type itself is a
determinate object with a changing truth-value, but sentence-
tokens are the proper bearers of truth and falsity, and their truth-
values are fixed and unchanging.

7. An objection to the token-reflexive version of the B-theory

The B-theory provides a token-reflexive account of the truth con-

ditions of A-sentences. According to this account, a token of an A-
sentence is true if and only if the event the A-sentence is about
stands in the appropriate temporal relation to the token of the A-
sentence itself. For example:
For any token u of 'The train arrived 2 hours ago' u is true if and
only if the train's arrival is 2 hours earlier than u.
The token-reflexive version is so called because the token itself con-
stitutes part of its own truth conditions. It intuitively delivers the
right truth conditions for tokens of A-sentences, but it has been
criticized on the grounds that there are some circumstances where
147
Heather Dyke

it delivers the wrong truth conditions.6 For example, William Lane

Craig (1996) argues that that 'The New B-theory can give no coher-
ent account of the truth conditions of tensed sentences which are
not tokened.' (Craig (1996) 18) He asks what the truth conditions
are for a sentence like 'There are no sentence-tokens now.' Any
token of this A-sentence-type would be false, because if the token
existed, the time at which it existed would not be a time of which it
is true to say that there are no tokens. However, it also seems to be
the case that there are some times of which it is true to say that there
are no tokens then. The point of this objection is that the sentence-
type 'There are no sentence-tokens' can express something true
even though no true token of it can ever be produced.
The force of this criticism stems from the intuition that truth, or
what is true of the world, does not depend on what anybody hap-
pens to say. But the token-reflexive version seems to imply that
truth depends on true tokens being produced. This criticism has
been articulated in some depth by Quentin Smith (1993), so I shall
address his statement of it. He says:
'If a normal A-sentence is used on some occasion to express
something true, what the A-sentence expressed on that occasion
would have been true then even if it had not been expressed.'
(Smith (1993) 83)
Smith is appealing to an intuition that we have about the concept of
truth. The intuition is that the way the world is does not depend on
there being utterances expressing that the world is that way. Smith
thinks that the token-reflexive theory is committed to the denial of
this intuition because it gives truth conditions that can only be ful-
filled when sentence-tokens are produced. I shall argue that the
token-reflexive theory is not defeated by this objection.
Suppose an event occurs, and lasts for a certain amount of time.
A forest fire starts at tx and burns itself out by t2. During that peri-
od of time no one utters any sentence that expresses that the forest
is now burning. Because the forest actually burns during this peri-
od of time, our intuition is that if someone had uttered such a sen-
tence it would have been true. But how can the token-reflexive
theory cohere with this intuition? I would explain it by putting for-
ward the following counterfactual: between t\ and t2, if someone
had uttered a token of the sentence-type 'The forest is now burn-
ing' , that token would have been true. The reason why it would have
been true is that its tenseless token-reflexive truth conditions would
6
See, for example, Smith (1993), Craig (1996) and Mellor (1998).
148
 McTaggart and the Truth about Time

have been satisfied. The truth conditions for a token of this sen-
tence-type are:
Any token u of 'The forest is now burning' is true if and only if
the burning of the forest is simultaneous with u.
In order for these truth conditions to be satisfied two events must
occur simultaneously: the burning of the forest and the production
of a token of the sentence-type. Between t\ and t2 the forest burns
so if, during that period of time, a token of the sentence-type is pro-
duced, its truth conditions would ipso facto be satisfied. However, if
no such token is produced, the forest still burns during that period
of time, but there is no token the truth or falsity of which we have
to account for.
The project of providing truth conditions for A-sentence-tokens
has both semantic and ontological significance.7 On the one hand, it
specifies what the world must be like in order for those tokens to be
true. This is its ontological function. If the truth conditions of A-
sentences only require the existence of B-facts, then that shows that
A-facts are not needed to account for the truth of A-sentence-
tokens. It also explicates how the truth or falsity of a sentence-token
depends^on what its semantic constituents mean when produced in
a given_gpntext. This is its semantic function. If the project is suc-
cessful it will show that the world need not be an A-world to
account for the fact that we sometimes utter true and meaningful A-
sentence-tokens, and it will also explain why the true A-sentence-
tokens we utter are true. The provision of truth conditions makes
perspicuous both the relationship between truth and reality, and
that between truth and meaning.
The concept of truth is connected both to meaning and to reality.
We might even say that it is ambiguous in that it has two distinct
domains of application. Linguistic entities are capable of being true
or false, and the world is that which makes true or false our utter-
ances about it. Consider the difference between the predicate 'true'
and the operator 'It is true that'.8 The predicate 'true' applies to lin-
guistic entities. It is sentence-tokens that can correctly be described
as true or false. However, if we prefix a sentence with 'It is true that'
we are making a claim about the world, not about the sentence. I can
describe the sentence 'The forest is now burning' as true or false.
But if I say 'It is true that the forest is now burning' I am making a
7
Davidson (1986) recognizes both kinds of significance when he says
'The truth of an utterance depends on just two things: what the words as
spoken mean, and how the world is arranged.' (Davidson (1986) 309)
8
I am grateful to Colin Cheyne for suggesting this explanation to me.
149
Heather Dyke

claim about what the world is like; I am describing reality, not a sen-
tence about it.
It is important to be clear, when expressing one's intuitions about
truth, whether those intuitions are about the connection between
truth and meaning or that between truth and reality. It is the con-
nection between truth and reality that generates our intuition that
truth is independent of the production of any sentence-tokens. The
world is the way it is independently of what anyone happens to say
about it. This is the intuition that Smith thinks the token-reflexive
theory is unable to explain. His example constitutes a sentence-
type, a token of which would have been true if it had been uttered
at a certain time, but no such token was uttered. He argues that the
token-reflexive theory cannot account for our intuition that this
sentence expresses a truth whether or not a token of it is produced.
But the token-reflexive theory can account for this intuition, simply
by upholding the distinction between the ontological and the
semantic aspects of truth. We have an intuition that the sentence
'The forest is now burning' is true between tx and t2. The intuition
can be explained by appealing just to the ontological aspect of truth.
Between t^ and t2 it is true that the forest is burning, but if no sen-
tence-token is produced, there is nothing to which we can ascribe
the predicate 'true'. But reality remains the same whether or not
sentences about it are produced.
Lastly, to return to the original challenge, I must explain how I
would assign truth conditions to the sentence-type 'There are no
sentence-tokens now' of which there can be no true tokens, even
though what it expresses can be true. The general truth-conditional
formula for this sentence-type is:
Any token, u, of 'There are no sentence-tokens now' is true if and
only if u occurs at a time at which there are no sentence-tokens.
Now, consider some arbitrary token of that sentence-type. The time
at which it is produced cannot be a time at which there are no sen-
tence-tokens, so no true token of that sentence-type can be uttered.
And the token-reflexive analysis explains why this is the case.
However, if we turn now to the ontological aspect of truth, we can
see that reality can be such that there are times at which it is devoid
of sentence-tokens. That is all that is meant by the claim that there
are times at which this sentence, or what it expresses, is true. It is,
indeed, misleading to say that there are times at which this sentence
is true. It is misleading in two ways. Firstly, it is ambiguous between
sentence-types and sentence-tokens. As I have argued, it is sen-
tence-tokens, not sentence-types that have truth-values, and there
150
 McTaggart and the Truth about Time

can be no times at which a token of this sentence can be true.

Secondly, it equivocates between the semantic and the ontological
aspects of truth. There are times at which it is true that there are no
sentence-tokens, but there are no tokens of this sentence-type that
can be described as true.
I have only dealt with one objection to the token-reflexive version
of the B-theory of time, but it is, I believe, one of the most com-
pelling. Those who reject this theory, very often do so on the
grounds that it cannot account for the truth of unuttered proposi-
tions.9 By restricting the application of the predicate 'true' to sen-
tence-tokens, and by upholding the distinction between the seman-
tic and the ontological aspects of truth, the problem for the token-
reflexive theory evaporates. Furthermore, I hope to have deflected
the criticism that the token-reflexive theory has the unacceptable
consequence that the way the world is depends on what human
beings happen to say about it.

8. Conclusion

My hope in presenting this paper has been to guide you all on a

journey, starting from McTaggart's argument for the unreality of
time, and finishing up with what I believe to be the truth about
time. I think McTaggart was right to argue that our world cannot
be an A-world: it cannot be a world in which anything is really past,
present or future. But I think he was wrong to argue that our world
has to be an A-world if time itself is to be a part of it. Our world is
a B-world, in spite of the fact that we talk and think as if it were an
A-world. Indeed, the fact that we live in a B-world can provide the
best explanation for the truth of our true A-sentences.

References

Bigelow, J. 1996. 'Presentism and Properties', in James E. Tomberlin (ed.)

Philosophical Perspectives 10, Metaphysics, (Oxford: Blackwell
Publishers Ltd, 1996) 35-52.
Craig, W. L. 1996. 'Tense and the New B-theory of Language', Philosophy
71, 5-26.
1998. 'McTaggart's Paradox and the Problem of Temporary
Intrinsics', Analysis 58, 122-27.
9
Mellor (1998), for example, rejects the token-reflexive theory in favour
of the date theory for just this reason.
151
Heather Dyke

Davidson, D. 'A Coherence Theory of Truth and Knowledge', in Ernest

LePore (ed.) Truth and Interpretation: Perspectives on the Philosophy of
Donald Davidson (Oxford: Basil Blackwell, 1986) 307-19.
Dummett, M. 1960. 'A Defense of McTaggart's Proof of the Unreality of
Time', Philosophical Review 69, 497-504.
Dyke, H. (forthcoming, 2002) 'Tokens, Dates and Tenseless Truth
Conditions', Synthese.
Hinchliff, M. 1996. 'The Puzzle of Change' in James E. Tomberlin (ed.)
Philosophical Perspectives 10, Metaphysics, (Oxford: Blackwell
Publishers Ltd, 1996) 119-36.
McTaggart, J. M. E. The Nature of Existence, Vol II, (Cambridge:
Cambridge University Press, 1927)
Mellor, D. H. Real Time II, (London: Routledge, 1998)
Prior, A. N. 1970. 'The Notion of the Present', Studium Generale 23,
245-48.
Smith, Q. Language and Time, (New York: Oxford University Press, 1993)
Sprigge, T. L. S. 1992. 'The Unreality of Time', Proceedings of the
Aristotelian Society 92, 1-19.

152