PHIL 15-12

Philosophical Perspectives, 15, Metaphysics, 2001

Mc TAGGART ON TIME

Judith Jarvis Thomson

MIT

1.

McTaggart’s argument for the conclusion that time does not exist is noto-
riously hard to understand.1 C. D. Broad says that when properly interpreted,
its main part can be seen to be “a philosophical ‘howler’.” 2 Others see things
in it that they regard as true and important, or if not true, then anyway impor-
tant. But I have not seen any interpretation of it that seems to me to get it
exactly right. And I think that it pays to get it right: there are lessons to be
learned from consideration of what goes on in it.
By way of reminder, McTaggart’s argument has two parts. The first part
aims at the conclusion that time does not exist unless the A series exists. The
second part aims at the conclusion that the A series does not exist. It follows
that time does not exist.

2.

What is the A series? We do better to begin with what McTaggart calls the
B series.
Actually, we do best to begin with what McTaggart should be taken to
mean when he talks about a series’ existing or not existing. He does not say
what he means, so we have to supply an interpretation.
McTaggart gives us descriptions of two series.3 Both descriptions have the
following form: the series such that (i) its members are the things that have
property P, and (ii) one precedes another in the series just in case the one has
the two-place relation Q to the other. It is clear enough that when he asks, rhe-
torically, whether a series so described exists, what he is asking is whether there
is a series that satisfies the description. I suggest that we should take him to
think that there is a series that satisfies the description if and only if there are
things that have P, and for every pair of things that have P, one has Q to the
other. I suggest that we follow him in this usage.
230 / Judith Jarvis Thomson

Consider the Happy Lion Series Description: the series such that (i) its
members are the lions in my back yard, and (ii) one precedes another in the
series just in case the one is happier than the other. Is there a series that satis-
fies that description? No. That is because there are no lions in my back yard.
Consider the Happy Squirrel Series Description: the series such that (i) its mem-
bers are the squirrels in my back yard, and (ii) one precedes another in the
series just in case the one is happier than the other. Is there a series that satis-
fies that description? Well, there are squirrels in my back yard. Then there is a
series that satisfies the description just in case for every pair of squirrels in my
back yard, one is happier than the other.
It is worth noticing that opting for this account of the conditions under
which there is a series that satisfies a description of the kind we are consider-
ing is compatible with opting for a (by now) familiar account of the conditions
under which a series S is identical with a series S9. According to that account,
S is identical with S9 just in case they have the same members, and their order
in S is the same as their order in S9. The fact that these accounts are compatible
emerges as follows.
Suppose that there are only two squirrels in my back yard, Alfred and Bert,
and that Alfred is happier than Bert. Then there is a series that satisfies the Happy
Squirrel Series Description. Call it the Happy Squirrel Series. Consider now the
Fat Squirrel Series Description: the series such that (i) its members are the squir-
rels in my back yard, and (ii) one precedes another in the series just in case the
one is fatter than the other. Suppose that Alfred is not merely happier than Bert
but fatter than Bert. Then there is a series that satisfies the Fat Squirrel Series
Description. Call it the Fat Squirrel Series. What is the relation between the
Happy Squirrel Series and the Fat Squirrel Series? We are entirely free to say—as
the familiar account of the identity conditions for series directs us to say—that
they are one and the same series. Indeed, that they are both identical with the
ordered pair ^Alfred, Bert&. Three series-descriptions but only one series.
We must of course allow that the following might have been the case: Al-
fred is happier than Bert but no fatter or thinner than Bert. If that had been the
case, then there would all the same have been a series satisfying the Happy
Squirrel Series Description; so on any view, the Happy Squirrel Series would
have existed. But there would have been no series satisfying the Fat Squirrel
Series Description; would the Fat Squirrel Series have existed? We are free to
say that the Fat Squirrel Series would have existed, for we are free to say that it
simply would not in that case have satisfied the Fat Squirrel Series Descrip-
tion. We are free to say, more generally, that unless P is an essential property of
the things that have it, and Q such that if x has it to y then x necessarily has it
to y, a series that in fact satisfies a description of the kind we are considering
might have failed to do so.
Would McTaggart have accepted this (by now) familiar identity conditions
for series? There is only one place at which it will matter whether he would
have accepted it; we will get to it in section 5 below.
 McTaggart on Time / 231

3.

Let us now turn to the B series. What are its members? McTaggart intro-
duces us to the B and A series together: he says they are series of what he calls
“positions in time”—I take these to be time points or chunks.
But he goes straightway on to say:

The contents of any position in time form an event. The varied simultaneous con-
tents of a single position are, of course, a plurality of events. But, like any other
substance, they form a group, and this group is a compound substance. And a com-
pound substance consisting of simultaneous events may properly be spoken of as
itself an event. (10)

Presumably the compounding he has in mind is mereological fusion. And from

here on, he throughout writes as if he meant us to suppose that the members of
the B and A series are events. For example, he goes straightway on to say:
“[t]he first question which we must consider is whether it is essential to the
reality of time that its events should form an A series as well as a B series.”
(11) Let us say that an event is an E event if and only if it is the mereological
fusion of all the events that occur at and only at a given time point. And let us
say, then, that the members of both the B and A series are the E events.
What fixes which E events precede which in the B series? McTaggart tells
us about the B series: “the terms have to be such that, of any two of them,
either the first is earlier than the second, or the second is earlier than the first.”
(10) So we may take it that the following is McTaggart’s description of the B
series:

(B Series Description) the series such that (i) its members are the E events,
and (ii) one precedes another in the series just in case the one is earlier
than the other.

This may well strike us as puzzling, however. It is not, of course, transpar-

ent what it is or would be for time to not exist, but it is intuitively very plausi-
ble to think that if there are events, some earlier than others, then time does
exist.4 So if there is a series that satisfies the B Series Description, and it is the
B series, then doesn’t the existence of the B series suffice for the existence of
time? But if so, then what are we to make of the first part of McTaggart’s
argument? Why does the existence of time require the existence of the A series
if the existence of the B series would suffice?
McTaggart does not make explicit what his answer is. There are two re-
plies available to him.
He can reply, first, that that intuitively very plausible idea is mistaken. Thus
he can say that the existence of events, some earlier than others, does not suf-
fice for the existence of time. He can say that what the existence of time re-
232 / Judith Jarvis Thomson

quires is that there be time points. I am sure that many people do so use “Time
exists” that it is true if and only if there are such entities as time points.
But it is clear that McTaggart would not make this first reply. If he had
said that the B series is a series of events and the A series a series of time
points, then perhaps it would be right to think he would make this reply. But as
I said, he does not say this. I have pointed to reason to believe he intends the
members of both series to be events.
Moreover, there are many places in the text that show that he thinks the
existence of the B series really would suffice for the existence of time. For
example, having proved to his satisfaction that the existence of the A series is
required for the existence of time, McTaggart says that the B series

cannot exist except as temporal, since earlier and later, which are the relations which
connect its terms, are clearly time-relations. So it follows that there can be no B
series when there is no A series, since without an A series there is no time.” (13) 5

So I suggest that we should attribute to him the second response available

to him, namely that the existence of the B series would suffice for the exis-
tence of time. He can also say that the existence of the A series is required for
the existence of time: he need merely add that the existence of the B series
would suffice for the existence of the A series. We will return to this idea in
section 5 below.

4.

The members of the A series, like the members of the B series, are the E
events. Under what condition does one member precede another in the A se-
ries? There is only one passage that explicitly answers this question, namely:
the A series “runs from the far past through the near past to the present, and
then from the present through the near future to the far future” (10). Let us
help ourselves to the following complex two-place relation:

x has the Temporal Perspective Relation to y just in case

(1) x is past, and either y is nearer past than x is, or y is present, or y
is future, or
(2) x is present, and y is future, or
(3) x is future, and y is farther future than x is.

Then we can say that the A series is the series that satisfies the following
description:

(A Series Description) the series such that (i) its members are the E events,
and (ii) one precedes another in the series just in case the one has the Tem-
poral Perspective Relation to the other.
 McTaggart on Time / 233

5.

We now have before us characterizations of McTaggart’s B and A series:

if there is a series that satisfies the B Series Description, it is the B series, and
if there is a series that satisfies the A Series Description, it is the A series.
Whether there are series that satisfy these descriptions is a matter we will turn
to in the following section. We might well want to ask first, however, how
they are related to each other—supposing that they do exist.
Suppose they do exist. They have the same members: each has the E events
as its members. Is the order of the events different in the two series? Surely it
is obvious that an event x has the Temporal Perspective Relation to y just in
case it is earlier than y. (McTaggart himself gives no hint of thinking other-
wise.) Then the order of the events in the A series is the same as the order of
the events in the B series. So if we opt for the familiar account of the identity
conditions for series that I mentioned in section 2—namely that series S is iden-
tical with series S9 just in case they have the same members, and their order in
S is the same as their order in S9—then we are committed to supposing that the
A series just is the B series.
I asked, rhetorically, in section 2, whether McTaggart would have ac-
cepted this identity condition for series. We should notice now that if he
did accept it, he could easily explain why the existence of the B series
would suffice for the existence of the A series: the existence of the B
series would suffice for the existence of the A series because the A series
just is the B series.
On the other hand, he plainly thinks that the A series is not identical with
the B series; so he did not accept this identity condition for series. So what
should we take him to think marks the one series off from the other?6 Perhaps
that the relation being earlier than and the Temporal Perspective Relation are
different relations, and therefore that the series whose order is fixed by the one
is different from the series whose order is fixed by the other.
But it is easy to see that this won’t do. Suppose there is a series of all E
events such that one precedes another in the series just in case the one has the
Temporal Perspective Relation to the other; call every such series an alpha se-
ries. Suppose also that there is a series of all E events such that one precedes
another in the series just in case the one is earlier than the other; call every
such series a beta series. Let us grant McTaggart that the Temporal Perspective
Relation is not the same as the earlier than relation. Still, an event x is earlier
than an event y just in case x has the Temporal Perspective Relation to y.7 It
follows that every beta series is an alpha series. Therefore McTaggart cannot
have that every beta series is different from every alpha series in that their or-
dering relations are different.
So we really had better suppose that if there is such a thing as the A series,
it is the B series. This supposition is not going to turn out to have been unfair
to McTaggart, since, as we will see, his argument for the nonexistence of time
234 / Judith Jarvis Thomson

does not anywhere rely on there being a difference between the A series and
the B series.

6.

Let us turn now to the first part of the argument; it aims at the conclusion
that time exists only if the A series exists.
McTaggart says that time exists just in case some things change, and he
argues that things change only if the A series exists.
Is McTaggart right to say that time exists just in case some things change?
No doubt there is room to argue that he is not. Let us bypass that question: let
us suppose that McTaggart is right to say this.
His argument to the effect that things change only if the A series exists is
another matter, and needs a closer look.
What exactly is it for a thing to change? McTaggart invites us to agree that
a thing changes just in case it has a property (“quality”, “characteristic”) at one
time and lacks it at another.8 We may well think this intuitively plausible.
But McTaggart would have us agree that it is not so easy to supply exam-
ples. Let us suppose that a certain poker was not hot until a certain Monday,
hot on that Monday, and not hot thereafter. We might think that our supposi-
tion, together with McTaggart’s account of change, entitles us to conclude that
the poker changes. Not so, McTaggart says. He says:

this makes no change in the qualities of the poker. It is always a quality of that
poker that it is one which is hot on that particular Monday. And it is always a
quality of that poker that it is one which is not hot at any other time. Both these
qualities are true of it at any time—the time when it is hot and the time when it is
cold. And therefore it seems to be erroneous to say that there is any change in the
poker. (15)

McTaggart takes it that for the poker to be hot on and only on that Monday is
for the following to be the case:

(S1 ) The poker has the two-place relation being hot-at to that Monday and
lacks it to all other days.

But he says that the truth of (S1 ) does not suffice for the poker to change: after
all, the poker always has that relation to that Monday, and always lacks it to all
other days.
We might wonder why McTaggart does not take it that for the poker to be
hot on and only on that Monday is for the following to be the case:

(S2 ) The poker has, on that Monday, the property being hot, and lacks it
on all other days.

If that is true, then his account of change yields that the poker does change.
 McTaggart on Time / 235

I suggest that we must take McTaggart to think that there is no such prop-
erty as being hot. Let us call a property a temporary property just in case it is
possible that a thing has it at one time and lacks it at another. McTaggart thinks
that things change just in case they have a property at one time and lack it at
another. Thus things change only if there are temporary properties. I am sure
McTaggart would say that if there were such a property as being hot, it would
be a temporary property, and the truth of (S2 ) would suffice for the poker to
change. What I am suggesting is that we must take him to think that being hot
is not a temporary property because there is no such property; and (S2 ) is there-
fore false.
Which, on McTaggart’s view, are the temporary properties? He says:

The fact that it [the poker] is hot at one point in a series and cold at other points
cannot give change, if neither of these facts change—and neither of them does.
Nor does any other fact about the poker change, unless its presentness, pastness, or
futurity change. (15)

He had just above said the same of events, thus that the death of Queen Anne,
for example, is “in every respect but one” unchanging—the one respect in which
it does change is this: it was far future, then nearer future, then present, then
past, then further past. “Such characteristics as these are the only characteris-
tics which can change. And, therefore, if there is any change, it must be looked
for in the A series, and in the A series alone.” (13)
He clearly thinks that the only temporary properties are being present, be-
ing past, and being future. (Perhaps we should we add: being far future, being
far past, and so on. For brevity, I ignore them.) It follows that a thing changes
just in case it has one of these three properties at one time, and lacks it at
another.
Which are the things that might be thought to have one or other of the
three at a time? Events, and only events.9 So a thing changes just in case it is
an event and has one of the three at a time, and lacks it at another. Consider the
event (state of affairs?) that consists in the poker’s being hot.10 If there is such
an event, then, given our original supposition, it has the temporary property
being present on the Monday, and not on other days. So it is not the truth of
(S1 ) or (S2 ) that would suffice for there to be change; it is rather the truth of

(S3 ) The poker’s being hot has, on that Monday, the property being present,
and lacks it on all other days

that would. We need to keep in mind, however, that the truth of (S3 ) would not
suffice for change in the poker; what it would suffice for is only change in the
event that consists in the poker’s being hot.
We have not quite reached McTaggart’s conclusion. We have reached the
conclusion that a thing changes just in case it is an event that has one of the
236 / Judith Jarvis Thomson

three temporary properties at a time, and lacks it at another. I take it he would

say that if any event has one of the three temporary properties at a time, then
all events do, including the E events. I take it he would say that if any E event
has one of the three properties at a time, then it has the others at other times.
From here it is only a short step to the conclusion that there is change just in
case for every pair of E events, one has the Temporal Perspective Relation to
the other—and from here an even shorter step to McTaggart’s conclusion that
there is change only if the A series exists.
Should we accept this argument? Let us begin with temporary properties.
Are there any temporary properties? Suppose that “F(x)” stands for a prop-
erty, and that “A” refers to something, and that in saying

F(A)

you would be predicating that property of A. It is a very natural idea that if you
do say “F(A)” you assert a proposition, and moreover, that the proposition you
assert if you say it at one time is the same as the proposition you assert if
you say it at any other time.11 For example, suppose that “Hot 3AM,11/2/00 (x)”
stands for the property being hot-at-3AM,11/2/00, and that “the poker” refers
to McTaggart’s poker, and that in saying

Hot 3AM,11/2/00 (the poker)

you would be predicating that property of McTaggart’s poker. It is very natural

to think that if you do say “Hot 3AM,11/2/00 (the poker)” you assert a proposition,
and moreover, that the proposition you assert if you say it at one time is the
same as the proposition you assert if you say it at any other time. That is surely
very plausible.
Now suppose that there is such a property as being hot. And suppose that
“Hot(x)” stands for it, and that “the poker” refers to McTaggart’s poker, and
that in saying

Hot(the poker)

you would be predicating that property of McTaggart’s poker. According to the

very natural idea, if you do say “Hot(the poker)” you assert a proposition, and
moreover, the proposition you assert if you say it at one time is the same as the
proposition you assert if you say it at any other time. But the property being
hot—if there is such a property—is a temporary property. That is, it is possible
that a thing has the property at one time and lacks the property at another time.
Therefore it may be that the proposition you assert when you say “Hot(the
poker)” at one time is true and the proposition you assert when when you say it
at another time is false. It follows that the proposition you assert at the one
 McTaggart on Time / 237

time is not the same as the proposition you assert at the other time, and the
very natural idea is therefore false.
A devotee of the very natural idea therefore concludes that there is no such
property as being hot. And so similarly for all temporary properties: there aren’t
any. Similar reasoning yields that there are no temporary relations. For exam-
ple, while there is such a non-temporary three-place relation as “x is taller than
y at t”, there is no such temporary two-place relation as “x is taller than y”.
It may have been something like this very natural idea that led McTaggart
to think that there is no such temporary property as being hot, and to say, there-
fore, that for the poker to be hot on a particular Monday is not for the poker to
have, on that Monday, the (temporary) property being hot, but rather for it to
have the (non-temporary) two-place relation being hot-at to that Monday.
But perhaps not. For if that very natural idea was what led him to think
that there is no such temporary property as being hot, then it is puzzling that he
should have failed to notice that the same holds of the (putative) temporary
properties being present, being past, and being future. Surely one who says

Present(the poker’s being hot)

at different times no more asserts the same proposition on both occasions than
does one who says “Hot(the poker)” at a time at which the poker is hot and
then again at a later time at which the poker is no longer hot. Thus if the very
natural idea had been what led him to think that there is no such temporary
property as being hot, then it is hard to see why he rejected being hot out of
hand but took being present seriously.
We do not have to accept the very natural idea. There are alternatives on
offer in the literature.12 But there remains this to be said for McTaggart: any-
one who wants to have that there are temporary properties and relations has to
suppose that a satisfactory alternative is available or can be found.
There also remains this to be said against McTaggart: there is no good
reason to distinguish among (putative) temporary properties as he does. If there
are such temporary properties as being present, being past, and being future,
then there is no good reason to reject such temporary properties as being hot.
Should we also say that if there are such temporary properties as being
hot, then there is no good reason to reject such temporary properties as being
present, being past, and being future? We might think that whether or not there
are such temporary properties as being hot, there anyway are no such tempo-
rary properties as being present, being past, and being future. Or we might not.
Let us not stop over the questions that arise here: let us grant McTaggart his
favored temporary properties, and reserve criticism for the use he makes of
them. If it helps, we can suppose the following. For an event to be present is
for it to be occurring, or going on. For an event to be past is for it to be over,
done, completed. For an event to be future is for it to be going to occur or go
on.13
238 / Judith Jarvis Thomson

To return now to McTaggart on change. He invites us to accept two theses.

The first, (I), says that a thing changes just in case it has a property at one time
and lacks it at another. This yields that the truth of

(S1 ) The poker has the two-place relation being hot-at to that Monday and
lacks it to all other days

does not suffice for the poker to change, for the poker always has that relation
to that Monday and always lacks it to all other days.
The second thesis, (II), says that the only temporary properties are being
present, being past, and being future. This yields that

(S2 ) The poker has, on that Monday, the property being hot, and lacks it
on all other days

is false. However (II) leaves open that

(S3 ) The poker’s being hot has, on that Monday, the property being present,
and lacks it on all other days

may be true.
McTaggart therefore invites us to agree that there is change only if the
likes of (S3 ) are true.
I have suggested that if there are such temporary properties as being present,
being past, and being future, then there is no good reason to reject such tempo-
rary properties as being hot. We should therefore reject thesis (II).
We should also reject thesis (I). Suppose you accept the very natural idea
that I described, and therefore conclude that there are no temporary properties.
Are you really committed to thinking that nothing changes? That is hard to
believe. (The very natural idea is an idea about what a person asserts by saying
certain sentences. How could it plausibly be thought to yield the metaphysical
conclusion that nothing changes?) In particular, you might well say that the
truth of (S1 ) does suffice for change—indeed, change in the poker. You might
well say, more strongly, that for there to be change just is for the likes of (S1 )
to be true. It is not in the least obvious why this should be thought a mistake.
A further alternative is available, which some would say is preferable.
McTaggart believes that there is no such temporary property as being hot be-
cause there is no such property as being hot. Other philosophers would agree
with McTaggart that there is no such temporary property as being hot, but for a
different reason: on their view, there is such a property as being hot but it is
non-temporary. David Lewis is an example.14 On his view, it cannot be the
case that a thing has being hot at one time and not at another. He therefore
concludes that we should accept the metaphysic of temporal parts. In particu-
 McTaggart on Time / 239

lar, that we should say that for the poker to be hot on and only on that Monday
is for

(S4 ) The poker’s that-Monday temporal part has the (non-temporary) prop-
erty being hot, and the poker has temporal parts prior to its that-
Monday part, and temporal parts after its that-Monday part, and all
of them lack the (non-temporary) property being hot

to be true, and that the truth of (S4 ) suffices for change—once again, change in
the poker. Lewis would say that for there to be change just is for the likes of
(S4 ) to be true.
I find this idea unsatisfactory for a variety of reasons, one of which I will
draw attention to in section 9. Meanwhile, however, it supplies us with a sec-
ond way of understanding change that is incompatible with McTaggart’s thesis
(I) that a thing changes just in case it has a property at one time and lacks it at
another.
In sum, the first part of McTaggart’s argument does not succeed.

7.

The second part of McTaggart’s argument aims at showing that the A se-
ries does not exist. If there is a series that satisfies

(A Series Description) the series such that (i) its members are the E events,
and (ii) one precedes another just in case the one has the Temporal Per-
spective Relation to the other,

where

x has the Temporal Perspective Relation to y just in case

(1) x is past, and either y is nearer past than x is, or y is present, or y
is future, or
(2) x is present, and y is future, or
(3) x is future, and y is further future than x is,

then it is the A series. Following McTaggart, we are supposing that the A series
exists if and only if there are E events, and for every pair of them, one has the
Temporal Perspective Relation to the other. McTaggart now argues for the con-
clusion that nothing can have the property being present, or the property being
past, or the property being future. If that argument succeeds, then nothing has
the Temporal Perspective Relation to anything else, and the A series does not
exist.
He gives more than one argument for his conclusion. I bypass the first
since I am not at all sure I have understood it. McTaggart himself says that the
240 / Judith Jarvis Thomson

second points to “a more positive difficulty” facing the idea that x might have
one or other of the three properties we are concerned with, so let us turn to it.
The argument is a reductio. Let us assume

(1) M is past.

McTaggart says:

If M is past, it has been present and future. If it is future, it will be present and
past. If it is present, it has been future and will be past. Thus all three characteris-
tics belong to each event. (20)

So if (1) is true, so also is

(1*) (a) M is past, and

(b) M is present, and
(c) M is future.

But (1*), he says, is self-contradictory.

He knows perfectly well that his opponent has an answer. McTaggart says:

It is never true, the answer will run, that M is present, past, and future. It is present,
will be past, and has been future. Or it is past, and has been future and present, or
again is future, and will be present and past. The characteristics are only incompat-
ible when they are simultaneous, and there is no contradiction to this in the fact
that each term has all of them successively. (21)

Thus (the answer runs): “what follows from (1) is not (1*) but only

(2) (a) M is past, and

(b) M was present, and
(c) M was future.

And unlike (1*), (2) is not self-contradictory.”

McTaggart asks:

But what is meant by “has been” and “will be”? And what is meant by “is,” when,
as here, it is used with a temporal meaning, and not simply for predication? (21)

His answer is:

When we say that X has been Y, we are asserting X to be Y at a moment of past

time. When we say that X will be Y, we are asserting X to be Y at a moment of
 McTaggart on Time / 241

future time. When we say that X is Y (in the temporal sense of “is”), we are assert-
ing X to be Y at a moment of present time. (21)

Now “M is (temporal sense) past” is of the form “X is (temporal sense) Y.” So

(2)(a) says “M is past at a moment of present time.” More precisely, it says “M
is past at a t such that t is present.” But all three characteristics belong to each
moment of time just as all three belong to each event; thus if t is present it is
also past and future. So if (2) is true, so also is

(2*) (a) M is past at a t such that t is present, and t is past and t is future,
and
(b) M...,

which is self-contradictory.
We can see how it goes on. McTaggart’s opponent now says “No, no. What
follows from (2) is not (2*), but only

(3) (a) M is past at a t such that t is (temporal sense) present and t will be
past and t was future, and
(b) M...,

which is not self-contradictory.” Aha, says McTaggart, the second clause of

(3)(a), namely “t is (temporal sense) present”, is of the form “X is (temporal
sense) Y”, so what it says is “t is present at a moment of present time.” More
precisely, what it says is “t is present at a t9 such that t9 is present.” But all
three characteristics belong to each time t9; thus if t9 is present it is also past
and future. So if (3) is true, so also is

(3*) (a) M is past at a t such that t is present at a t9 such that t9 is present

and t9 is past and t9 is future, and
(b) M...,

which is self-contradictory.
As I said, we can see how it goes on. McTaggart now says:

Such an infinity is vicious. The attribution of the characteristics past, present, and
future to the terms of any series leads to a contradiction, unless it is specified that
they have them successively. This means, as we have seen, that they have them in
relation to terms specified as past, present, and future. These again, to avoid a like
contradiction, must in turn be specified as past, present, and future. And, since this
continues infinitely, the first set of terms never escapes from contradiction at all.
(22)

His idea is this. If (1) is true, so also is (1*), which is self-contradictory. We

try to escape from contradiction by declaring that what follows from (1) is not
242 / Judith Jarvis Thomson

(1*) but only (2). But (2) is tensed; properly interpreted, it can be seen to
yield (2*), which is self-contradictory. We try to escape from contradiction by
declaring that what follows from (2) is not (2*) but only (3). But (3) is tensed;
properly interpreted, it can be seen... .And so on. So we never escape from
contradiction.
We must therefore reject our assumption

(1) M is past;

that is, it cannot be the case that M is past. Similar reasoning yields that it
cannot be the case that M is present or that M is future. So no event can have
any of the temporary properties being past, being present, and being future. It
follows that the A series does not exist.
What is crucial to this argument is obviously McTaggart’s idea that if an
event M, or a time t, is past then it is also present and future. His reason for
thinking this lurks in a passage I quoted above, namely

But what is meant by “has been” and “will be”? And what is meant by “is,” when,
as here, it is used with a temporal meaning, and not simply for predication? (21)

When we say “M is past”, we may be using “is” with a temporal meaning. That
is, we may mean “M is now past”. But we may instead be using “is” simply for
predication. That is, it may be that what we assert is what we would have as-
serted if we had instead said “ Past(M)”, in saying which we would simply be
predicating pastness of M.
How did he get from here to his idea that if an event M, or a time t, is past
then it is also present and future? I suggest that his route went as follows.
In saying “ Past(M)” we are simply predicating pastness of M. But that
“simply” should be properly understood. In saying “ Past(M)” we do assert a
proposition. The role of that “simply” is simply to indicate that in saying
“ Past(M)” we are not asserting that M is past at the time of our speaking. We
are not asserting anything about when M is past.
It follows that the proposition we assert if we say “ Past(M)” at one time is
the same as the proposition we assert if we say it at any other time.
[What has surfaced here is what I called the very natural idea in the pre-
ceding section.15 You might have thought that McTaggart would therefore con-
clude that there is no such temporary property as being past, and a fortiori, that
there is no such thing as predicating ‘it’ of M. For don’t we assert a false prop-
osition if we say “ Past(M)” at a time before M is past but a true one if we say
it at a time when M is past? (Compare the argument I set out to show that there
is no such temporary property as being hot.) Note that there being no such
temporary property as being past—and also no such temporary properties as
being present and being future—would have been good enough for his pur-
 McTaggart on Time / 243

poses: no event would have the Temporal Perspective Relation to any other,
and it would follow that the A series does not exist. But McTaggart does not
conclude that there is no such temporary property as being past. What he does
instead is to deny that we assert a false proposition if we say “ Past(M)” at a
time before M is past but a true one if we say it at a time when M is past. To
return to McTaggart:]
Suppose that M is at some time past. Then the proposition we assert if we
say “ Past(M)” is true—true whether or not M is past at the time of our saying
“ Past(M)”. For remember: in saying that sentence, we are not asserting any-
thing about when M is past.16
But if M is at some time past, it is also at some time present and at some
time future. So the propositions we assert if we say “ Present(M)” and “Fu-
ture(M)” are also true. Destination reached: if an event M is past then it is also
present and future. Similarly for any time t.
It pays to notice that if McTaggart’s reductio succeeded, a similar argu-
ment would succeed in showing that nothing that is ever hot and ever cold can
be either hot or cold. That is, while I began with “M is past,” I could as well
have begun with “McTaggart’s poker is hot”, for a route analogous to the one I
suggest that he took to “If an event M is past then it is also present and future”
would have taken us to “If McTaggart’s poker is hot then it is also cold.” It
might have been thought (and I think that some people have thought) that what
started the regress was the fact that in saying “M is past,” we predicate a mys-
terious property. Not so. Or not so unless a property’s being temporary itself
suffices for it to be mysterious.
In any case, the combination of ideas at work in McTaggart won’t do. If P
is a temporary property then there is no such thing as ‘simply predicating’ it of
a thing. And in sum, the second part of McTaggart’s argument is no more suc-
cessful than the first.

8.

Michael Dummett says that those who think McTaggart’s argument for the
nonexistence of time is a “trivial sophism”—as some people do—have failed
to see what is really going on in it.17 Dummett asks the good question why
McTaggart did not go on to supply an analogous argument for the nonexistence
of space. Dummett himself thinks that there is an important difference between
time and space that emerges when we contrast McTaggart’s argument against
time with its analogue against space.
What Dummett focuses on in McTaggart’s argument is its first part. (He
says it contains “the heart of” McTaggart’s argument for the nonexistence of
time.) As I interpreted it in section 6, McTaggart there argues that there is no
change unless things acquire and lose the temporary properties being present,
being past, and being future. Dummett interprets it differently. According to
Dummett, McTaggart there argues for the following conclusion:
244 / Judith Jarvis Thomson

a description of events as taking place in time is impossible unless temporally token-

reflexive expressions enter into it, that is, unless the description is given by some-
one who is himself in that time. (354)

Let us call this the Time Thesis. According to Dummett, McTaggart does not
argue for the analogous Space Thesis since McTaggart thinks it is not true.
What is the analogous Space Thesis? Dummett does not spell it out, but we
may take it to say:

a description of physical objects as occupying places in space is impossi-

ble unless spatially token-reflexive expressions enter into it, that is, unless
the description is given by someone who is himself in that space.

According to Dummett, it is because McTaggart thinks the Time Thesis is true

and the Space Thesis is not true that McTaggart goes on to argue that time does
not exist and does not argue in the same way that space does not exist.
Why does thinking the Time Thesis true lead McTaggart to think that time
does not exist? According to Dummett, McTaggart is implicitly appealing to a
prejudice that lies very deep in many people, namely that “there must be a
complete description of reality; more properly, that of anything which is real,
there must be a complete—that is, observer-independent—description.” (356)
Thus presumably a description that token-reflexive expressions do not enter
into.
Dummett himself thinks that the Time Thesis is true and the Space Thesis
false. Moreover, he says that he is himself very strongly inclined to accept the
prejudice he says McTaggart is appealing to. But he thinks that McTaggart is
mistaken in concluding that time does not exist, and that this shows we must
abandon the prejudice.
Dummett says that in any case, it is this deep issue that McTaggart’s argu-
ment for the nonexistence of time draws our attention to, and that McTaggart’s
argument is therefore quite certainly not a trivial sophism.
Let us bypass the question whether Dummett has interpreted McTaggart
correctly. What is more interesting is the question whether the views Dummett
attributes to McTaggart are correct.
I think we do suppose that something like what Dummett calls the preju-
dice is true—that, as Dummett also puts it, “reality must be something of which
there exists in principle” a description free of token-reflexives, or that, as he
also puts it, “[t]he description of what is really there, as it really is, must be
independent of any particular point of view.” Surely it is not enough for there
to really be a cylindrical gray rock on Cambridge Common that it looks to me,
from where I now stand and in the prevailing light, that there is one there.18
Let us set the prejudice aside for the time being and return to the Time
Thesis and the Space Thesis. You might well think it long over-due that I say
something about what Dummett means by them, since that is hardly transpar-
 McTaggart on Time / 245

ent. To find out what he means we need to look at the reasons he gives for
thinking the one true and the other false.
Dummett begins by inviting us to suppose a person is in process of observ-
ing a sequence of events. Dummett then says:

Then even if he knows both what he has observed and what he is going to observe,
he cannot give a complete description of his observations without the use of tem-
porally token-reflexive expressions. He can give a complete narration of the se-
quence of events, but there would remain to be answered the question, ‘And which
of these events is happening now?’ (354)

This suggests that what Dummett means by the Time Thesis is: you cannot
give a “complete description” of a series of events that you are in process of
observing unless it includes “Event M is happening now”.
What does he mean by the Space Thesis? He thinks it is false, and we can
see what he means by it when we see why he thinks it is false:

By contrast, the use of spatially token-reflexive expressions is not essential to the

description of objects as being in a space. That is, I can describe an arrangement of
objects in space although I do not myself have any position in space. An example
would be the space of my visual field. In that space there is no here or there, no
near or far: I am not in that space. We can, I think, conceive, on the strength of this
analogy, of a being who could perceive objects in our three-dimensional physical
space although he occupied no position in that space. He would have no use for
any spatially token-reflexive expressions in giving a description of the physical
universe, ...19 (354)

This suggests that what he means by the Space Thesis is what was to have been
expected, namely that you cannot give a “complete description” of an array of
physical objects that you are observing unless it includes “Object O is here”.
The Space Thesis, then, is the spatial analogue of the Time Thesis, and Dum-
mett thinks the Space Thesis false and the Time Thesis true.
I have flagged Dummett’s expression “complete description” in my state-
ment of the two theses for the obvious reason: it is not at all clear what Dum-
mett can mean by it. What could it come to for there to be a complete description
of a thing?
There was another place at which Dummett used that expression. His first
statement of the prejudice was this: “there must be a complete description of
reality; more properly, that of anything which is real, there must be a complete—
that is, observer-independent—description.” There are two ways of interpret-
ing the final clause of that sentence. We can take the clause to say “there must
be a description that is complete in the sense of being observer-independent”.
But this way of understanding the expression will not do for Dummett’s pur-
poses, since there are observer-independent descriptions of events just as there
are observer-independent descriptions of physical objects. Alternatively, we can
246 / Judith Jarvis Thomson

take the clause to say “there must be a complete description that is observer-
independent”. But this is no help to us: it leaves us still confronted with the
expression “complete description”.
Perhaps we should give up on that expression, and so also on my interpre-
tations of the Time Thesis and the Space Thesis. Perhaps what Dummett has in
mind is not a difference between possibilities of description, but rather a dif-
ference that he might express in the following way:

(i) We can conceive of a being who could perceive objects in our three-
dimensional space although he occupied no position in that space,

but

(ii) We cannot conceive of a being who could perceive events in our one-
dimensional time although he occupied no position in that time.

I extracted (i) from the passage I quoted just above. Dummett went on to
say there that the being described in (i) “would have no use for any spatially
token-reflexive expressions in giving a description of the physical universe”;
thus he would have no use for such descriptions as “The poker is here” and
“The poker is nearer to me than the piano is.” Surely something stronger had
better be true: since the being described in (i) is not in the same space as the
poker and the piano, those descriptions would be at best false, and perhaps
nonsense.
This helps us understand (ii), for we can take it to say that if a being is
perceiving a sequence of events that includes event M, then it must be that he
speaks truly if he now says either “M took place before now,” or “M is taking
place now,” or “M will take place after now.” Never mind whether he must say
one or other of those things if he is now to give a complete description of the
sequence he is perceiving; it is enough for the difference Dummett is pointing
to—on this interpretation of him—that one or other of those descriptions must
be true.
It is plausible enough to suppose that (ii) is true. But I can see no good
reason to accept (i), and there is good reason to reject it. The being described
in (i) is described as perceiving objects in our three-dimensional space al-
though he occupies no position in that space. But to be perceiving physical
objects, say a poker and a piano, is to be perceiving them from a certain angle,
and thus from within the space they occupy.
Dummett had suggested that we should accept (i) on the strength of an
analogy with one’s visual field: in his visual field, he says, “there is no here or
there, no near or far: I am not in that space.” Well, I am not in among the
things that are in my visual field. (If I am in the dining room looking into the
living room, then I am by hypothesis not in among the things in the living
room.) But it is nevertheless true that some of the things in my visual field (as,
 McTaggart on Time / 247

for example, some of the things in my living room) are nearer to me than oth-
ers are.20
It is of course conceivable that there should be an array of physical objects
whose space I am not in since I do not exist. But so also is it conceivable that
there should be a sequence of events whose time I am not in since I do not
exist.
In sum, I have found no way of interpreting the views Dummett attributes
to McTaggart under which they could be thought to point to a metaphysically
significant disanalogy between time and space—or indeed, to any disanalogy
between them at all.

9.

There are plenty of disanalogies between temporal and spatial descrip-

tions, of course; it might pay to mention one that is familiar to all of us but that
is nevertheless of interest for present purposes.
What I have in mind is a certain difference between our use of “now”
and “here”. Suppose that there was, a few years ago, a series of powerful
earthquakes in and around Dedham. Standing in Dedham, I say to you “The
countryside is hilly here now.” Presumably I mean to be telling you that the
countryside is hilly in the place in space that I indicate (the Dedham area) at
the place in time that I indicate (the time of my speaking). So far so good.
Now suppose that standing in Dedham, I say to you “This tomato is soft here
now.” Do I mean to be telling you that the tomato is soft at the place in space
that I indicate (the Dedham area?) at the place in time that I indicate (the time
of my speaking)? What would it come to for the tomato to be soft in the
Dedham area?
When I say to you “This tomato is soft here now” I typically accompany
my remark with a gesture; and what I mean to be telling you is that the tomato
is now soft at the place, not in space but on the tomato, that I indicate. What I
mean to be telling you, alternatively put, is that a part of my tomato is now
soft, namely the part that I indicate.
The countryside has parts too, of course. So perhaps when I say to you
“The countryside is hilly here now,” I am not indicating a place in space (the
Dedham area), and what I mean to be telling you is not that the countryside is
now hilly at that place; perhaps I am instead indicating a part of the country-
side (its Dedham area part) and what I mean to be telling you is that it is hilly
now.21
There remains a disanalogy. The countryside has a Dedham area part; the
tomato has no Dedham area part.
Suppose that the tomato is now lying on the Dedham-Needham line, half
in Dedham, half in Needham. Then the tomato has a part that is now in Dedham.
But the countryside does not have a part that is just now in Dedham. Tomatoes
move around, the countryside does not. And the familiar difference between
248 / Judith Jarvis Thomson

our use of “now” and our use of “here” is due to the very fact that so many of
the physical objects that we are interested in do move around.
But how is this ‘moving around’ to be understood? Motion is a species of
change, so we are now brought back to a question about change that I left open
at the end of section 6.
Let us focus on motion.22 Suppose we carried our tomato from Chatham to
Dedham. Then our tomato moved from Chatham to Dedham. What marks it as
having moved from Chatham to Dedham?
A friend of the metaphysic of temporal parts (MTP) says this: the tomato
had a series of discrete but spatio-temporally continuous temporal parts, the
first of which was in Chatham, and the last in Dedham.
But he needs to do something more: he needs to convince us that none of
the members of the series were themselves tomatoes. For consider a series of
what according to MTP were discrete but spatio-temporally continuous five-
minute temporal chunks of our tomato. Suppose that what really happened was
this:

(Story) Each of the five-minute members of the series was itself a five-
minute, entirely new tomato—new skin, new tomato-flesh, new
seeds, new particles composing it.

If Story is true, then surely no tomato moved from Chatham to Dedham when
we thought that ours was doing so.
It won’t do for the friend of MTP to say, simply, “Well, if Story is true,
then your thought was false: your tomato did not move from Chatham to
Dedham.” For it had better not be left a chancy business whether or not a to-
mato (or anything else) ever moves from one place to another.
It also won’t do for the friend of MTP to say “No surely about it! Story is
true since every (cross-sectional) temporal part of a tomato is itself a tomato.
So your tomato is the fusion of a great many tomatoes. And it did move from
Chatham to Dedham since one of the tomatoes was in Chatham and a later one
in Dedham.” It is an odd idea that a tomato should be the fusion of many to-
matoes, and not one offered us by the typical friend of MTP. Not surprisingly.
No tomato moved through a short time unless at least some of the skin, the
tomato-flesh, the seeds, the particles it was composed of at the start still com-
pose it at the end. If Story is true, however, then the second five-minute tomato
shares no skin and so on with the first, and the third shares no skin and so on
with the second or the first, and... . So no tomato moved even through the fourth
to the sixth minute, much less through all the time it took us to get to Dedham.
The friend of MTP must therefore show that Story is false. We do not have
the burden of proving it true or even possible. He has the burden of proof, for
he is the one who told us that for our tomato to move from Chatham to Dedham
is for new things to be coming into existence throughout that time—new red,
bulgy, juicy things, which in all respects other than their short lifespan behave
 McTaggart on Time / 249

exactly like tomatoes. But I can see no way in which he can show that Story is
false.
Similarly for other species of change. A tomato might change from hard to
soft. If what followed a five-minute hard tomato was a new five-minute soft
tomato, then no tomato thereby changed from hard to soft. Similarly of course
for sheer endurance. If what followed one five-minute tomato is another, then
no tomato thereby lasted from the one time to the other.
If we reject MTP, we still have three of the four alternatives I drew atten-
tion to in section 6. According to McTaggart, only events change, so tomatoes
in particular do not. It hardly needs saying that that fact counts conclusively
against the alternative he offers us.
That leaves us with two of the four alternatives, each of which allows us to
have that the tomato changes. First, we can appeal to temporary properties. We
can say that the tomato changes because it has the temporary properties being
in Chatham and being hard, at one time and not at another. Second, we can
appeal to non-temporary relations. We can say that the tomato changes because
it has the non-temporary relations being in-Chatham-at, and being hard-at, to
one time and not to another. I suggest that the very fact that these two alterna-
tives allow us to have that the tomato changes counts in favor of them.

10.

McTaggart’s dark discussion of time is like a Rorschach blot. Some have

seen a frog in it. Others a prince. Still others a clash of armies by night. I have
thought it worth trying to be clear about the features of the blot itself.23

Notes

1. Until indicated otherwise, numbers in parentheses are page numbers in McTaggart

(1927).
2. See Broad (1938).
3. In fact he describes three, but I will discuss the third only briefly, in footnote 5.
4. G.E. Moore famously thought he had refuted McTaggart’s thesis that time does not
exist by drawing attention to the fact that he had breakfast before he had lunch.
5. Another passage appears later. Having proved to his satisfaction that time does not
exist, McTaggart says that there nevertheless exists something he calls the C series,
whose members are what we are really observing when we (mistakenly) think we
are observing a time-series. The C series, he tells us, includes “as terms everything
which appears to us as an event in time, ... . And although Mr. Russell’s time-series
(which is identical with our B series) has a one-to-one correspondence with the C
series, still the two series are very different. The terms of the B series are events,
and the terms of the C series are not.” (31) If he had thought that the B series could
exist even though time does not, he would have had no need for the C series.
And what an odd business that C series is! For what are we to suppose its mem-
bers are? What would an entity be like which “appears to us as an event in time” but
250 / Judith Jarvis Thomson

is not an event? And which members precede which? McTaggart deals with these
questions at length in later chapters; I will not try to summarize his answers.
6. Some commentators fix on a dark footnote (10–11) which suggests that we may
think of one of the two series as “sliding along” the other, and if this is true, then
the A and B series really are two series. But series don’t in fact slide. And McTag-
gart says in the opening sentence of the footnote that the idea that these series do is
(mere) metaphor.
7. We might well think that this is not merely true, but necessarily true. On some views,
however, even that is compatible with the idea that the Temporal Perspective Rela-
tion is not the same as the earlier than relation.
8. Peter Geach says that meeting this condition is sufficient for ‘Cambridge’ change
but not for real change. (For Socrates meets it if Theaetetus grows.) See Geach
(1969), pp. 71–2. McTaggart would have been untroubled by that objection. He
says: “if anything changes, then all other things change with it. For its change must
change some of their relations to it, and so their relational qualities. The fall of a
sand-castle on the English coast changes the nature of the Great Pyramid.” (11–12)
For McTaggart’s purposes here, a Cambridge change is as good as a real change.
9. This is not quite right, for McTaggart will later take seriously the idea that time
points have them. (See section 7 below.) But our taking that idea seriously now
would merely introduce unnecessary complication.
10. McTaggart does not distinguish between events and states of affairs. The poker’s
being hot is his own example of an event.
11. I think of this as a Fregean idea. My informant on Fregeana, Michael Glanzberg,
tells me that while Frege never explicitly asserted it, he would have accepted it
without a second thought.
12. I mention two, a stronger and a weaker. The stronger says that there is no such
thing as ( just) predicating a property of a thing. On this view, to predicate a prop-
erty (whether temporary or non-temporary) of a thing is always to assert that the
thing has it at a time. From this view, however, it follows that we do not know what
proposition is asserted by a person who says “F(A)”, “( ∃ x)Fx”, and “( ∀ x)Fx)”
unless we know what time he meant—the time of his speaking? or 3AM, 11/2/00?
Some people might therefore regard this alternative as too costly.
The weaker (and cheaper) alternative says only that to predicate a temporary
property of a thing is always to assert that the thing has it at a time.
Sally Haslanger draws attention to alternatives to the very natural idea in Has-
langer (1989).
13. Notice that sentences containing these predicates take tenses. We may say of an
event M, “M is now occurring” or “M did occur” or “M will occur”. So also “M is
now over” or “M was over” or “M will be over”. So also “M is now going to oc-
cur” or “M was going to occur” or “M will be going to occur”.
Many of McTaggart’s commentators take him to think that once we have helped
ourselves to the predicates “past,” “present,” and “future”, then we need nothing
further in the way of an account of tense—thus that those predicates could some-
how do double duty as tenses. [Michael Dummett is an example. See Dummett
(1960), reprinted in Dummett (1978).] They can’t, of course; no predicates can. And
I think we should agree that McTaggart does not think they can. McTaggart goes on
to give an account of tense in terms of those predicates, but it does not consist in
mere identification. We turn to it in section 7 below.
 McTaggart on Time / 251

14. Lewis believes that the property being hot is a non-temporary property because he
thinks it an unintelligible idea that it is a temporary property. See Lewis (1986),
pp. 203– 4.
15. C.D. Broad, whose interpretation of McTaggart seems to me closest to correct, does
not see the attractiveness of that very natural idea. See Broad (1938).
16. We are also not asserting that there is a time at which M is past. If in saying
“ Past(M)” we were asserting that there is a time at which M is past, then so simi-
larly for “ Present(M)” and “Future(M)”; and “ Past(M) and Present(M) and Fu-
ture(M)” would not be self-contradictory. McTaggart’s reductio would therefore fail
at its first step. I thank Selim Berker for drawing my attention to this point.
17. From here on, numbers in parentheses are page numbers in Dummett (1978).
18. I adapt this example from Dummett’s own:
I can make drawings of a rock from various angles, but if I am asked to say
what the real shape of the rock is, I can give a description of it as in three-
dimensional space which is independent of the angle from which it is looked
at. The description of what is really there, as it really is, must be independent
of any particular point of view. (356)
19. The final sentence of the passage reads in toto: “He would have no use for any
spatially token-reflexive expressions in giving a description of the physical uni-
verse, and yet that description might be a perfectly correct description of the ob-
jects of the universe as arranged in space.” I omitted its final clause in quoting the
passage because of the appearance in it of “correct description” rather than “com-
plete description”. The possibility of a (merely) correct token-reflexive free descrip-
tion of the physical objects in an array cannot be what he has in mind, since a
(merely) correct token-reflexive free description of the events in a sequence is en-
tirely possible too. What clearly matters to Dummett is complete token-reflexive
free description, which he thinks is possible for space but not for time. I go on to
comment on Dummett’s notion ‘complete description’ shortly.
20. Can it be that what Dummett intends for us to suppose is in his visual field is not,
as it might be, his living room furniture, but rather an array of sense-data? Cer-
tainly he is not in their space; but that is because they are not in any space.
21. On one view, that had better be what I mean to be telling you. The view emerges as
follows. Let us give the name temporally partial properties to what I have been
calling temporary properties. Let us call P a spatially partial property just in case it
is possible that a part of a thing has P and another part lacks P. Now as we know,
there are people who think that there are no temporally partial properties: thus that
while there is such a property as being hot, it is not temporally partial. They may
well say, and for the same reason, that while there is such a property as being hilly,
it is not spatially partial. It follows that when I say “The countryside is hilly here
now,” I speak truly only if what I mean is that a part of the countryside (the part I
indicate) is hilly now. Or indeed: I speak truly only if what I mean is that a certain
temporal part of that part of the countryside is hilly.
By contrast, there are people who welcome temporally partial properties. They
may be expected to welcome spatially partial properties too.
Still others (McTaggart, perhaps?) may be expected to think that there is no
such property as being hilly just as there is no such property as being hot, and that
for the countryside to be hilly now in the Dedham area is for the countryside to
252 / Judith Jarvis Thomson

have the three-place relation hilly-in-at to the Dedham area and the time at which I
write.
22. I thank Adam Elga for criticism of an earlier version of the following material on
motion.
23. I am grateful to Michael Glanzberg for helpful discussion along the way. I am also
grateful to Catherine Elgin for comments on an early draft, and to the members of
the seminar in metaphysics given by Ned Hall and me in fall, 2000, for comments
on a later draft.

References

Broad, C.D. (1938), Examination of McTaggart’s Philosophy, Volume II, Part I, Cambridge: Cam-
bridge University Press.
Dummett, M. (1960), “A Defense of McTaggart’s Proof of the Unreality of Time,” Philosophical
Review, vol. LXIX.
Dummett, M. (1978), Truth and other enigmas, Cambridge MA: Harvard University Press. [This
reprints Dummett (1960).]
Geach, P. (1969), God and the Soul, New York: Schocken Books.
Haslanger, S. (1989), “Endurance and Temporary Intrinsics,” Analysis, Vol. 49.
Lewis, D. (1986), On the Plurality of Worlds, Oxford: Basil Blackwell Ltd.
McTaggart, J.M.E. (1927), The Nature of Existence, Volume II, ed. C.D. Broad, Cambridge: Cam-
bridge University Press.