Personal Identity

Author(s): R. G. Swinburne
Reviewed work(s):
Source: Proceedings of the Aristotelian Society, New Series, Vol. 74 (1973 - 1974), pp. 231-247
Published by: Blackwell Publishing on behalf of The Aristotelian Society
Stable URL: http://www.jstor.org/stable/4544858 .
Accessed: 11/06/2012 06:56

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

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

The Aristotelian Society and Blackwell Publishing are collaborating with JSTOR to digitize, preserve and
extend access to Proceedings of the Aristotelian Society.

http://www.jstor.org
 XIII* PERSONAL IDENTITY
by R. G. Swinburne

A great deal has been written in the last few years on the
problem of personal identity. My excuse for adding to the
literature is my belief that much of the writing has produced
fundamentally misleading solutions to the problem, arising
from a confusionof two issues. There are two distinct questions
which can be asked about personal identity. The first is-what
does it mean to say that a person P2 at a time t2 is the same
person as a person PI at an earlier time t,? The second is-
what evidence can we have that a person P2 at t2 is the same
person as a person P, at t, (and how are different pieces of
evidence to be weighed against each other) ? These questions
are often confused in writing on this subject. Much of the
writing seems to me to be on the right lines if it is regarded
as attempting to provide an answer to the second question
but to be extremely misguided if it is regarded, as it often
appears to be, as attempting to provide an answer to the
first question.
Let us briefly survey the kinds of solution currently given
to the problem of personal identity, treating them for the
moment as answers to our first question. The first kind of
answer, which derives ultimately from that given by Hume
in the Treatise,is that personal identity is a matter of similarity
of memory' and character. P2 at t2 is the same person as P, at
tl if P2's memories include most of those of P,, and P2 behaves
in ways similar to P,. If there is a considerable temporal
interval between t, and t2, then it suffices,for P2 to be the same
person as P,, that there is a series of persons P. at times tm
intermediate between t, and t2, such that the memories of

* Meeting of the Aristotelian Society at 5/7, Tavistock Place, London,

W.C.I, on Monday, 13th May 1974, at 7.30 p.m.
' I use "memory" and cognate terms in the "weak" sense in which a
man's memories of what happened may or may not be correct. This is
contrasted with the "strong" sense in which necessarilyif a man remembers
p, thenp.
23I
232 R. G. SWINBURNE

each person include almost all those of any person slightly

earlierin the series, and each person is very similarin character
to any member of the series existing at a temporallyproximate
moment. A clause may be added to this solution to deal with
the case where two persons at t2, P2 and p2*, both satisfy
its criteria for being the same person as P1. It may state that
in such a case neither P2 nor P2* is the same person as Pl.
While this and other qualifications may be added, the central
idea of such a theory remains that personal identity is a
matter of similarity of memory and character.
The second kind of solution claims that personal identity
is a matter of bodily continuity. P2 is the same person as P1
if he has the same body as P1. A body B1 is the same body as a
body B2 if they are spatio-temporallycontinuous, in the sense
that they are connected by a continuous spatio-temporal
path at each point of which there exists a body somewhat
similar qualitatively (i.e., in appearance and construction)
to its neighbours. This second solution is often amended in
recent writing in a very important respect. Bodily continuity
is interpreted, not as continuity of all parts of the body, but
as continuity only of that part, i.e., the brain, which is re-
sponsible for a person's memory and character. On this view
P2 is the same person as P1 if he has the same brain as P1, which
may have been transplantedinto P2's body.
Neither of these solutions is especially plausible if taken in
isolation from a solution of the other kind-for well-known
reasons. The first one rules it out as logically impossible
that a man should lose his memory, something which we
ordinarily suppose to be possible. The second solution
rules out as logically impossible that a man should move from
one place to another without passing through intervening
space. Yet it seems far from obvious that it is logically
impossible that I should pass through a brick wall without
disturbing the bricks. For these and other reasons some
sort of compromise theory tends to be favoured. Such a
theory states how the various criteria to which I have referred
work together. Thus a compromise theory may claim that
satisfaction of either criterion by P2 at t2 makes P2 the same
person as P1 at t,, so long as there is no other person P2* at
t2 who satisfies the other criterion. The theory will then go
 PERSONAL IDENTITY 233

on to tell us what to say in the latter circumstances,e.g., that

neither person at t2 is the same person as P1 at t1, or that the
person who satisfies the bodily criterion is the same person as
P,; or the theory will provide some other solution. Or the
theory may provide some other detailed account of how such
criteriaas bodily continuity, similarityof memory and character
are to be understood and balanced against each other.
Now I see no objection to a theory of personal identity on
these lines if it is regardedas an answerto my second question-
what is the evidence that a person P2 at t2 is the same person
as a person P1 at t,? Without doubt continuity of their bodies
and especially of their brains is strong evidence that P2 and
P1 are the same person, and lack of such continuity is strong
evidence that they are not. Memory and characterare however
also relevant. If a woman turns up claiming to be Princess
Anastasia and can tell us details of independently verifiable
incidents in the Princess' life which would have been almost
impossible for her to find out unless she were the Princess,
that counts in favour of her being the Princess. Exactly how
the different criteria are to be weighed against each other I
do not propose to judge. What I am concerned to emphasise
is that such an account is an account of what we would be
justified in claiming about whether or not P1 and P2 are the
same person on the basis of different kinds of evidence.
However many recent philosophers seem to have wanted
to put forward accounts on the lines which I have sketched
above as accounts of what it is for P2 to be the same person as
P1, and it is in these terms that I sketched above the pattern
of such accounts. I wish to argue in this paper that such accounts
are not at all on the right lines. I will call a theory which
analyses personal identity as a matter of bodily continuity
and continuity of memory and character an empiricist theory
of personal identity.
I begin by drawing attention to the fact that an empiricist
theory of personal identity has as an inevitable consequence
that sometimes (it is logically possible) there will be no right
answer to the question whether two persons are the same, and
this will be so, whatever the detailed form of the empiricist
theory. On an empiricist theory there are clear cases where
P1 and P2 are the same person (I and the Swinburne whom
234 R. G. SWINBURNE

you saw yesterday) and clear cases where P1 and P2 are

different people (I and the Bloggs whom you saw yesterday),
but it is easy to imagine a whole host of cases where the empiri-
cist criteria give no clear verdict. This will be so whatever
the details of how the different criteria are to be weighed
against each other, simply because there will be cases where
each criterion separately gives no clear result. There will be
cases where P2 has fairly similar memories and fairly similar
character to P1, so that P2 lies on the border between satisfying
and not satisfying the criteria of memory and character for
being the same person as P1. The same applies to the criterionof
bodily continuity. How much of P2's body (or brain) has to
be continuouswith PI's in orderfor this criterionto be satisfied?
If P2 acquires only P1's left arm, clearly the criterion is not
satisfied, and if P2 acquires all Pi's body except the left arm,
clearly it is satisfied.But what about the case where P2 acquires
half of Pi's brain, or all his body except a quarter of the brain?
However you interpret the criterion of bodily continuity it is
easy to imagine circumstancesin which P2 lies on the border
between satisfying and not satisfying this criterion. Because
there are imaginable circumstanceswhere any criterion gives
no clear result, there are imaginable circumstanceswhen any
empiricist theory of personal identity would give no answer
as to whether P2 is the same person as an earlier P1. In such
circumstances,on an empiricist theory it is the case that not
merely are you or I unable to find out whether P2 and P1
are the same person, but there is no one right answer to the
question. The answer that they are the same is as near to the
truth as the answer that they are different. You can say which
you like. The situation is similar to that of some other cases
of identity such as the identity of a society or of an army.
(Hume compares the identity of persons to the identity of
"a republic or commonwealth".)2 When is army A2 at time
t2 the same army as army Al at tl? Clearly it is if A2 has
all the same soldiers in it as has A1. Clearly it is not if all the
soldiers of A1 mutinied at an intermediate time and were
replaced by new soldiers. But suppose that A2 has a quarter

2 David Hume, A Treatiseof HumanNature (first published I739), (ed.)

L. A. Selby-Bigge, Oxford, i888, p.26I
 PERSONAL IDENTITY 235

of the soldiers who were in A1 but that the other soldiers of

A1 have been gradually replaced by new soldiers. Are the two
armies the same? Our criteria of 'same army' give no definite
answer-you can say that the armies are the same or you
can say that they are different. There is no right answer.
Hume described the similar situation which he thought held
with regard to the identity of persons by saying that "the
identity, which we ascribe to the mind of man, is only a
fictitious one".3That is a misleadingway of putting the matter;
the identity or lack of it is in most cases real enough-it is
simply that in somecases there is no right answer as to whether
two persons are the same-according to an empiricist theory.
Having outlined this importantconsequenceof any empiricist
theory of personal identity, I now propose to argue against
all such theories. In order to do so it will be useful to classify
them in a way different from the familiar classificationwhich
I adopted earlier, into three classes. I will then proceed to
use a different pattern of argument against theories of each
class. First we have theories which allow for duplication. Such
a theory claims that P2 at t2 is the same person as P1 at t, if
P2 satisfies some criterion, where it is logically possible that
more than one person at t2 satisfy that criterion. An obvious
example of such a theory is of course the pure memory-and-
character theory, that P2 at t2 is the same person as P1 at t,
if and only if P2 has similar character and memories to P1.
Now, as Bernard Williams pointed out in I956,4 it is logically
possible that there be hundredsof personsat t2 who have similar
characters and memories to P1 at tl. If the fact that Charles'
memories and character in I956 are very similar to those of
Guy Fawkesin I604 makeshim the same person as Guy Fawkes,
then if Robert in I956 also has very similar character and
memoriesto Guy Fawkesin I604, he also would beGuy Fawkes.
Yet if P2 at t2 is the same person as P1 at tl, and so is P2* at
t2, then P2 and P2* are the same persons as each other. But
that is absurd-for of logical necessity the same person cannot

3Id. p.259. Hume goes on to say that the same applies to the identity
"which we ascribe to vegetables and animal bodies."
4 "Personal Identity and Individuation", Proceedingsof the Aristotelian
SocietyI956-7, 57, 229-52.
236 R. G. SWINBURNE

be (wholly) in two places at once. Any theory which allows

duplication allows as a logical possibility that two persons at
t2 in different places be the same person as an earlier person,
and that is clearly not a logical possibility. So any theory
which allows duplication must be rejected.
Among theories which do not allow duplication, we must
distinguish two further types. A theory of the second type of
empiricist theory of personal identity is simply a theory of the
first type with a clause added stating that a subsequent person
at t2 is not the same person as an earlier person if some other
person at t2 satisfies the criteria equally well. Thus the theory
could have the following form: P2 at t2 is the same person as
P1 at t1, if and only if P2 has roughly the same memories and
character as P1, and no other person at t2 has roughly the same
memories and character as P1. Another theory of this type, and
one which incorporates a criterion of bodily continuity, is the
following theory: P2 at t2 is the same person as P1 at t, if and
only if there is spatio-temporal continuity between some part
of Pl's brain and some part of P2's brain and there is not
spatio-temporal continuity between any part of Pi's brain
and any part of the brain of any person at t2 other than P2.
By this theory if we remove Pi's brain, destroy half of it and
successfully transplant the other half into P2's body, P2 will
be the same person as P1.
Now theories of the second type have a different implausible
consequence from the absurd consequence derived from
theories of the first type. This is that a man's identity (i.e.,
whether or not he is identical with a certain past person)
could in conceivable circumstances, as a matter of logical
necessity, depend on the success or failure of an operation to a
brain and body other than his own. Who I am could depend
on whether or not you exist. We can illustratethis with reference
to the theory sketched at the end of the last paragraph, by
telling a mad surgeon story.5 A surgeon removes P 's brain
and divides it in half. He successfully transplants the left
half into one body from which the existing brain has been

5 The value of such stories was brought out in the very important article
by Bernard Williams, to which this paper owes much, "The Self and the
Future", PhilosophicalReview,1970, 79, i 6i-i 8o.
 PERSONAL IDENTITY 237

removed. The half-brain takes. We will term the resulting

person P2. The surgeon also attempts to transplant Pl's right
half-brain into another body from which the existing brain
has been removed. Whether P2 will be P1 depends on whether
this second transplant takes. If it does not, he will be; if it
does, he will not, since there will then be two persons who
satisfy equally well the criterion for being the same person as
P1. Yet how can who I am depend on what happens to you?
A theory of personal identity which has this consequence
does not seem to be analysing our ordinary concept of personal
identity. Another absurd consequence of a theory of the second
type is that the way for a man to ensure his own survival is
to ensure the non-existence of future persons too similar to
himself. Suppose the mad surgeon had told P1 before the opera-
tion what he was intending to do, adding that while he felt
confident that the left half-brainwould transplant successfully,
he had some doubt whether the right half-brain would take in
the new body. P1 is unable to escape from the clutches of the
mad surgeon, but is nevertheless very anxious to survive the
operation. If the empiricist theory in question is correct there
is an obvious policy which will guarantee his survival. He can
bribe one of the nurses to ensure that the right half-brain
does nottransplant successfully.Yet it seems absurd to suppose
that as a matter of logic a man's survival can depend on the
non-existence of some person. A concept of personal identity
which has this consequence is not recognisably ours.
All theories other than theories of the second type, which
do not allow duplication, I will call theories of the third type.
A theory of the third type rules out duplication as logically
impossible but it does not do so by having a clause forbidding
duplication added to a theory of the first type. It is a more
natural consequence of the theory that no more than one
person at t2 is the same person as P1 at t,. The obvious example
of such a theory is the following: P2 at t2 is the same person as
P1 at tLif and only if the body of P2 is spatio-temporallycon-
tinuous with the body of P1 in such a way that any body on the
spatio-temporalchain joining them has almost all (e.g., go%)
the same brain matter as any body at a temporally proximate
earlier moment (e.g., within one second earlier) on the chain.
This chain cannot divide, for if a body is divided into two parts
238 R. G. SWINBURNE

(e.g., a part is cut off), then no more than one of the resulting
parts can have almost all the same matter as the previous
body. Yet theories like this have an essential arbitrariness.
Why go%? Perhaps "over half" would be more natural.
But then it would be a consequence of the theory, that a man
who lost half his brain at a blow would no longer exist whereas
a man could lose half his brain gradually over a few seconds
and yet continue to survive.
This essential arbitrariness about theories of the above
kind has the following consequence.Either the theory has rather
demanding conditions for survival (e.g., continuity of go% of
brain matter) or it has less demanding conditions (e.g., con-
tinuity of half the brain matter). In the former case (and
perhaps the latter case too) there seems nothing incoherent
in a man hoping to survive, even though the theory's survival
conditions are not met. In the latter case (and perhaps the
former case too), there seems nothing incoherent in a man
fearing that he will not survive, even though the theory's
survival conditions are met. To illustrate the former, let us
suppose that there is currentlyfashionablethe theory described
in the last paragraph, which I will call T, that P2 at t2 is the
same person as P1 at t, if and only if the body of P2 is spatio-
temporally continuous with the body of P1 in such a way that
any body on the spatio-temporal chain has go% of the same
brain matter as any body within the previous second on the
chain. A man A is told that he has a brain tumour and that
unless he is operated upon, he will suffer considerable, though
not unbearable, pain for the rest of his life. An operation would
need to remove at a stroke more than a tenth of the brain
in order to provide relief from pain to the person whose brain
it would be. A wants relief from pain but, more than that,
he wishes to continue to exist. Should he allow himself to be
operated on? The obvious thing to say here, to use Bernard
Williams' telling word,6 is that it would be a "risk"-the
fashionable theory T might or might not be right. In hoping
to survive A does not appear to be refusing to look logical
facts in the face. In that case it must be logically possible that
A survives and hence the claim of T to state a logical truth is

6 "The Self and the Future" p. I80.

 PERSONAL IDENTITY 239

mistaken. Similar considerations apply to similar empiricist

theories. Mere logic could hardly show you that hoping to
survive a serious brain operation (or fearing that you would
not survive it) would be hoping for (or fearing) something
the description of which made no sense. Even if you do think
that it is clear that a man does not survive if none of his brain
survives in a body, and that he does survive if all of it does,
you have to admit that what happens in intermediate
cases is unknown. Mere reflection of the meanings of words
would not appear to provide the answer. It appears to be an
empirical matter whether or not a man survives a large-scale
brain operation, yet not one which can be settled conclusively
by observationswhich we can make. The person P2 who has a
certain fraction of P1's brain will be expected to make some
memory claims which P1 would be expected to make and to
have somewhat the same character. But how much suffices to
make P2 the same person as P1? There seems no natural answer.
And yet a man may hope to survive even if he does lose some
of his memories and comes to react to circumstances in ways
other than he used to react. But how can we say for certain
when his hope has proved justified?
A furtherawkwardnessfor any such empiricisttheory is that,
as we have noted, there will be inevitably border-line cases
for its satisfaction. Suppose that T is right, and A is told that
he is to have removed from his brain at a strokea tumourwhich
consists of approximately io% of the matter of the brain, so
that the subsequent person will have approximately go% of
the brain of A. Will A survive the operation? The answer
given by T is that it is as true to say that he will as that he will
not. A is told this answer, and is told that the subsequent
person will then be tortured. Has he cause to fear? Presumably
less cause than if the person to be tortured were fully himself,
and more cause than if it were not at all himself. But how can
an intermediate reaction be justified? Each subsequent
person will either be tortured or not; no half-tortureswill be
laid on. An intermediate reaction would be justified if A did
not know who would be tortured, i.e., whether it would be
himself or someone else. But A has been told who will be
tortured, i.e., someone who is equally well described as A
or as not-A. How can any suffering affect A unless he suffers
240 R. G. SWINBURNE

it all or suffers part of it ?-and neither of these alternatives

is what is being suggested here.
The arguments of the previous paragraphs, and especially
the argument concerned with border-line cases, suggest an
essential difference in the survival conditions for persons
(and perhaps animals too) from the survival conditions for
inanimate objects. There is nothing puzzling about a future
car constructedfrom bits of my old car and other bits as well
being a borderline case for being my car. As cars do not have
feelings or hope to survive, the arguments of the previous
paragraph cannot be deployed to argue against this possibility.
But a conscious thing such as a person may wish to continue
to be conscious, and there seems no intermediate possibility
between a certain future consciousbeing being that person and
not being that person.
So much for objections to empiricist theories. What can we
put in their place? Wherein does the identity of persons
consist? The identity does not consist solely in the continuity
of one or more observable characteristics, for empiricist
theories took all these into account. (If they left one out, an
empiricist theory could easily be constructed which took
account of it, and, there is every reason to suppose, would be
found wanting for reasons of the kind which we have already
considered.) The only alternative is to say that personal
identity is something ultimate.7 It is unanalysable into con-
junctions or disjunctionsor other observableproperties.Bodily
continuity, continuity of memory and character, are however
the only evidence we have of its presence; it is observable only
by observing these. In general there is plenty of evidence,
normally overwhelming evidence, of bodily continuity,
memory and character, as to whether or not two persons are
the same, which gives very clear verdicts in the overwhelming
majority of cases. Yet while evidence of continuity of body,
memory, and characteris evidence of personalidentity, personal
identity, is not constituted by continuity of body, memory,

7 This was of course the view of Butler and, less explicitly, Reid. See
J. Butler Of PersonalIdentityin his The Analogyof Religion,London, 1902,
p.328; and T. Reid Essays on the IntellectualPowers of Man (ed. A. D.
Woozley), London, 1941, III.4.
 PERSONAL IDENTITY 24I

and character. Hence the evidence may on occasion mislead,

and two persons be the same, although the best evidence
which we have shows that they are not, and conversely. Also
on occasion the evidence of observable characteristics may
give no clear verdict as to whether P2 is the same person
as P1; but that does not mean that there is no clear answer to
this question, merely that we do not know and cannot even
make a reasonable guess at what it is. That evidence of
continuity of body, memory and character is evidence of
personal identity cannot of course be something establishedby
enumerative induction (since personal identity cannot be
observed apart from observations of continuity of body,
memory, and character). Although it might in theory be
established by some more complex form of inductive inference
that the former is evidence of the latter, this is in fact, I
suspect, either an analytic truth or some basic principle which
we assume to be intuitively obvious, like the basic principle
of induction itself that what has happened (described under
some simple description) invariably in the past is evidence
of what is going to happen.
One comes to understand the meaning of "same person",
not by being provided with a definition in terms of continuity
of body, character, and memory, but by being provided with
clear examples of pairs of personswho are and pairs of persons
who are not the same, and being shown the grounds on which
judgments about personal identity are made. By being shown
the evidence and clear cases where the evidence points one
way rather than the other, we come to have an understanding
of what is at stake. But there is no reason to suppose that the
understanding is simply an understanding of the evidence
(i.e., that we mean no more by personal identity than some
conjunction or disjunction of the kind of features which lead
us to make judgments ascribingit) and the arguments of this
paper count against that supposition.
Because of the difficulty of what to say in puzzle situations
about whether or not two persons are the same, one or two
recent writers, and notably Derek Parfit,8 have argued that
8 See, e.g., his "Personal Identity" PhilosophicalReview
I97i, 8o, 3-27
and "On 'The Importance of Self-Identity' " Journal of Philosophy1971,
68, 683-690.
242 R. G. SWINBURNE

our concept of personalidentity is a confusedone. The questions

which I have posed assume all-or-nothing answers, that P2
is or that P2 is not the same person as the earlier P1. Once we
assume that, to give one or other answerin the puzzle situations
seems unjustified, and that for some is worrying. So Parfit
wishes to treat personal identity as a matter of degree, and
instead of saying that two persons are or are not the same per-
son, to say rather that they are exactly the same or almost exact-
ly the same or pretty much the same or hardly at all the same.
This is indeed a very natural development of the empiricist
theory of personal identity. This is after all what we say of the
identity of inanimate things. We say that Italy of 1972 iS
exactly the same country as the Italy of 1952 (its boundaries,
constitution etc. are the same) but only pretty much the same
as the Italy of I942 (which had somewhat differentboundaries,
and a very different constitution) and hardly at all the same
as the Italy of I862. If we started to talk in this way about
people, then P2 would be the same person as P1 to the extent
to which bodily continuity, and similarity of memory and
character existed. Then it would be clear how to describe
the puzzle cases. The less and the less gradual the brain and
other continuity etc., the more P2 would be a different person
from P1.
However, the consequence of this way of talking is that we
have to say of a man P1 in his late fiftieswho does not remember
most of the events of what we would ordinarily call "his"
childhood, has a very different character from the character
which, we would ordinarily say, "he" had when young, and
has very few of the same brain cells as those which, we would
ordinarily say, "he" had fifty years before, that P1 is only
somewhat the same person as the boy he is ordinarily said to
have been. Now if we do say this we are clearly using the
word "person" in a different way from the normal way. For
on our normal understanding the boy who had developed
into a man in his late fifties without any brain operations,
sudden losses of memory or sudden changes of character, is
not more or less the same person as the boy was; he is the same
person simpliciter.That, given our normal concept of person,
is at least a very well evidenced judgment, if not an analytic
truth. If Parfit makes a different claim, he is introducing a
 PERSONAL IDENTITY 243

different concept of personal identity. For on our normal

concept of personal identity he is not saying the right thing.
But why should we scrap our normal concept of personal
identity which we find it natural to use in normal circum-
stances and even in abnormal circumstances (when a man
wonders whether or not he will survive an operation)? Parfit's
only reason for objecting to the normal concept seems to be
provided by a philosophical dogma. This is the verificationist
principle that a proposition has no factual meaning if no
evidence of observation can count for or against it." In certain
puzzle situations identified by a full description of the extent
of bodily continuity, similarity of memory and character
between persons P1 and P2 at different times, it is unclear
whether or not P1 and P2 are the same person, and yet no
further evidence of observation can count for or against the
claim that they are. Hence, the claim that in that situation
P1 and P2 are the same person is, given the verificationist
principle, a meaningless one. Hence, Parfit argues, the normal
concept of personal identity does not apply in such puzzle
situations. We want a concept which does apply there. Hence,
Parfit can argue, the need for a new concept.
However this argument of Parfit's must be adjudged worth-
less until he has produced arguments for the verificationist
principle. The verificationist principle seems to me false and
I have argued against it elsewhere.10

9 That Parfit uses the verificationist principle to establish his view can be
seen on p.s of his article. Here he considers a man's brain being split into
two and each half transplanted into another body. He rejects the sug-
gestion that one and only one of the resulting persons is the same person
as the original person on the grounds that "each half of my brain is exactly
similar, and so, to start with, is each resulting person. So how can I survive
as only one of the two people? What can make me one of them rather than
the other?" The assumption here is that if one of the resulting persons is
to be the original person he must differ from the other resulting person in
some observable respect, which would be evidence for his being the original
person; which assurmptionseems to derive from the principle that to be
factually meaningful a claim must be confirmable by observation-state-
ments, i.e., must be such that evidence of observation can count for or
against it.
10 "Confirmability and Factual Meaningfulness", Analysis I973, 33,
71-76. I have discussed an alternative account of factual meaningfulness
in my SenseandNonsensein Physicsand Theology,University of Keele Inaugural
Lecture, 1973.
244 R. G. SWINBURNE

The view which I have put forward is then that bodily

continuity, similarity of memory and character (and any
other observable characteristics) are evidence of but do not
constitute personal identity. This view is, I believe, sup-
ported by other thoughts which we have, things which we
regard as conceivable, about personal identity. Consider, to
begin with, the resurrection of the dead (whose dead bodies
have decayed). Most people, I suggest, uninfluenced by philo-
sophical theory, would allow this to be a logical possibility.
The affirmationthat there is life after death in another world
or reincarnation on Earth is widespread. Perhaps equally
widespread is the denial that these things happen. Yet most
who deny that these things happen seem to allow that it makes
sense to suppose that they do happen, while denying that in
fact they do. Now an empiricist theory which allows life
after death must claim that in such a case personal identity
is a matter of similarity of memory and character. A man
survives death if there exists after his death a man with similar
memory and character to his (subject, possibly, to the proviso
that there is no more than one such man). We saw earlier
that there are difficultiesin such theories. But here is a further
one. If it is logically possible that I should survive my death,
I have a coherent hope if I hope to do so. On an empiricist
theory, for me to hope for my resurrectionis for me to hope
for the future existence of a man with my memories and
character, that is, a man who will be able to remember the
things which happened to me and react to circumstances
somewhat as I do. But that's not at all what I hope for in
hoping for my resurrection.I don't hope that therebe a man of
that kind-I want it to be me. If it isn't to be me, then despite
my hope for my resurrection, I am probably relatively in-
different to whether or not a man rises with my character
and memories. And if I am to rise again, I probably shouldn't
mind all that much if I had lost many of my memories and
much of my bad character. What matters is that I rise. So
hoping for my resurrectionis not analysable as hoping for the
resurrection of a person in various ways like me. And so an
empiricist theory which says that it is false.
Another situation which an empiricist thory must rule out
as not logically possible is that I should be you and that you
 PERSONAL IDENTITY 245

should be me (in a certain sense). Many people wish they

were somebody else in the sense that they wish they were in
his shoes with his body, position and relationships,appearance,
memory and character. Perhaps my body is withered, my
own position and relationships are unsatisfactory, my looks
are ugly, my memories give me no joy, and I am profoundly
dissatisfied with my own character. You on the other hand
seem very satisfactoryin these ways. So I wish I were you (in
the above sense). Is the wish coherent? Am I, that is, wishing
for the existence of a logically possible state of affairs different
from the present state? Superficially, yes. Many have wished
wishes of this kind, and believed themselves to be wishing for
such a state. Yet if it is logically possible that I should have
your body, memory, character, etc., it is also logically possible
that you should have mine. And if the former state of affairs
differs from the present one, so does the latter one. And your
acquiring my body, memory and character seems to make a
further difference from the present state from the difference
which would be made by my acquiringyour body, memory and
character. But if I acquired your body, memory and character
and you mine (or if I always had the former and you the
latter) the world would be exactly the same as it is now in
respect of the bodily continuity, memory, character etc. of
persons." If the possibility that I be you and you be me in
the above senses can be coherently entertained, as appears
to be the case, then any empiricist theory of personal identity
is mistaken.
Before concluding, I should mention one further considera-
tion in support of my account of personal identity, though one
somewhat less compelling than the earlier ones. This is reached
11 Leibniz' comment on a situation of this kind seems mistaken. See his
Discourseon Metaphysics? 34 (in his Philosophical
PapersandLetters,a selection
translated and edited by Leroy E. Loemker, Vol. I, Chicago, 1956,
P.502):
Suppose some private man should suddenly become King of China,
but only on condition that he forgot what he had been, just as if he
were being born anew; would it not be the same practically, or so
far as discernible effects are concerned, if he were annihilated and
a king of China created at the same instant in his place? This
particular man has no reason whatever to desire this.
Even if the situation is the same "so far as discernible effects are con-
cerned", a man still has reason to desire one state rather than the other.
246 R. G. SWINBURNE

by reflecting on various moral judgments which we make and

seeing that the view of personal identity which I have ex-
pounded is implicit in them. What, for example, is wrong
with killing people? A naive utilitarian would say that killing
is wrong because the sum total of happy human experiences
would thereby be diminished. But that cannot be the answer
because, it would be generally felt, it would be wrong to kill
people even if we speedily produced new people to replace
the old ones, and even if we had good reason to believe that
the new people would be happier than the old ones would
have been. The correct answer surely is that killing is wrong
because existingpeople would be deprived of future experiences
(whereas in omitting to breed you are not depriving anyone
of anything, for there is nobody to deprive). But on the empiri-
cist theory of personal identity to say that existing people
would be deprived of future experiences is just to say there
would not be persons whose bodies are continuous with those
of the living, or who had similar memories and character to
the living, having experience. But then what's so important
about the persons, who are to have experiences, having bodies
continuous with personsnow living or having similar memories
and character? Surely in itself nothing at all. If instead there
are newly bred personswith healthierbodies, happier memories,
and more amiable charactersthan those murdered,why should
killing be wrong just because of the lack of bodily and other
continuity? In itself surely such continuity has no value. The
importance of continuity must therefore, it is plausible to
suppose, lie in the fact that bodily continuity, similarity of
memory and character, are evidence of something else (not
analysable in their terms)-personal identity. Personal identity
must be distinct from this continuity.
It can be argued that this view of personal identity underlies
many other moral judgments which we make-judgments,
for example, about the morality of punishment and the
goodness of forgiving. Such arguments do not prove that my
account of personal identity is correct, but only that it is
deeply embedded in moral views which intuitively we feel
to be correct. My main arguments are those given earlier.
If anyone wishes to oppose the account which I have given
of personal identity, it seems to me that he has two possible
 PERSONAL IDENTITY 247

lines of attack. First he may invoke some general philosophical

principle about meaning such as a verificationist principle
of meaning, and show that my account falls foul of it. But he
will need to prove his principle first,and in so far as the evidence
for his principle comes from what claims we would normally
admit to be meaningful, some of the examples of claims
adduced in this paper do, I suggest, count against any such
principle. Alternatively, he may attempt to refute my detailed
arguments by arguing that descriptions of situations which I
have claimed to be coherent (e.g., a man surviving the removal
of a large brain tumour with loss of memory and change of
character) are really incoherent, and that descriptions of
situations which I have claimed to be incoherent (e.g., my
surviving an operation as both of two persons) really are
coherent. My various arguments have proceeded either from a
claim that the description of some situation is coherent or
from a claim that the descriptionof some situationis incoherent,
to the claim that only an account such as mine can adequately
account for these facts. It is certainly possible that descriptions
which appear coherent really are not so, and conversely. But
the onus is on the philosophical critic, at any rate when he
claims the former, to prove his point. A superficial survey of
much religion, literature, and popular beliefs and hopes
would surely show that many people believe to be coherent
such descriptionsof situations as I have claimed to be coherent.
A philosophercannot just tell millions of people that when they
think they are talking sense, they are really talking nonsense;
he must argue his point. In default of a successful counter-
attack on one or other of the above lines, the case for rejecting
all empiricist theories of personal identity and adopting a
theory on the lines which I have outlined, is, I submit, over-
whelming.