The Plastic Self

Neurologists have demonstrated plasticity in the spatial sense of self, or body-image.  Not only are the ‘phantom limbs’ of amputees somewhat malleable, under special circumstances (and also remarkably intractable under others), but we can be induced to perceive a detached rubber hand, lying on a table, as our own hand, and even to feel that the table-top itself is part of our own body.  But how plastic, if at all, are our ideas of ourselves extended into the future and the past?

Published in 1970, Bernard Williams’ “The Self and the Future” is one of the seminal papers that gave rise to the contemporary philosophical debate on personal identity.   Williams presents two series of thought-experiments, which lead his own intuitions in opposite directions on the question whether a future person, described as having a certain relationship to his present self, would be himself or someone else.  One set of cases tends to persuade him that what matters in personal identity – the relation that makes us the same person over time – is psychological continuity.  The other set of cases makes him think that bodily continuity is more important.  Williams reports being left “not in the least clear” which is right.

Williams presents this as a philosophical problem.  I suggest it is more fruitfully regarded as an experimental result – a single data-point in a psychological experiment, with Williams as both subject and experimenter.  If repeated with a larger, and less contaminated, sample population, such an experiment could shed the light of empirical research on the question with which I opened this post.  I hope someone will undertake such a study, which could lead in interesting directions.

Cases Supporting the Psychological Criterion

Williams starts by presenting a series of cases which tended to persuade him that psychological continuity is the grounding relation of personal identity.  In these cases, the informational content of the brains of two persons, A and B, is read by some scanning device and stored in files.  A brain-programming device then transfers A’s information to B’s brain, replacing what’s there, and similarly replaces the information in A’s brain with that extracted from B.  When the subjects are awakened following the double information transfer, the ‘B-body-person’ (the person with the body that was B’s before the experiment) reports memories, thoughts and opinions previously expressed by A, and believes himself to be A; and the ‘A-body-person’ similarly believes himself to be B.  Both subjects remark on their changed appearance; they think of the experiment as an exchange of bodies.  Are they right?

To address this question, Williams steps back to just before the experiment, to a time when it is being planned.  A and B have full knowledge of the facts of what will happen, as described in my previous paragraph.  Williams sets up some incentives for the subjects to prefer being one of the survivors of the experiment rather than the other.  We may suppose there is nothing much to choose between the bodies of A and B: both are undamaged, normally endowed, and of presentable appearance.  The experimenter announces that one of the survivors will be rewarded with $100,000 and the other will be tortured.   He then asks the subjects to choose which of the survivors should be rewarded, and which punished.  Williams stipulates that the subjects are asked to choose ‘on selfish grounds’.

I don’t like this example much.  Physical torture, which philosophers are fond of invoking, tends to give thought experiments an air of unreality, perhaps because the philosophers themselves have never been in danger of torture.  Since the positive incentive is money, why not add balance to the experiment by making money a disincentive as well?  $100,000 in 1970 was worth about $600,000 today.   Suppose both subjects are reasonably well off to start with, each with about $1 million in net assets.  One condition of the experiment is that $600,000 will be transferred from the balance sheet of one to that of the other, leaving one comfortably set with $1.6 million, the other in sadly reduced circumstances with only $400,000 to his name.  Before the procedure, they are asked to choose which of the survivors will be enriched at the expense of the other.

In the first case of this series, Williams supposes that both subjects choose according to the psychological criterion of personal identity.

Suppose that A chooses that the B-body-person should get the pleasant treatment and the A-body-person the unpleasant treatment, and B chooses conversely (this might indicate that they thought that “changing bodies” was indeed a good description of the outcome.) (p 77)

Since the experimenter cannot fulfill both wishes, he picks B’s choice, rewarding the A-body-person (who psychologically resembles B), and punishing the B-body-person (who believes he is A).  When they wake after the procedure:

…the B-body person will not only complain of the unpleasant treatment as such, but will complain (since he has A’s memories) that that was not the outcome he chose, since he chose that the B-body-person should be well treated; and since A made his choice in selfish spirit, he may add that he precisely chose in that way because he did not want the unpleasant things to happen to him.  The A-body-person meanwhile will express satisfaction both at the receipt of the [$600,000], and also at the fact that the experimenter has chosen to act in the way that he, B, so wisely chose. (p 78)

Williams takes the experimental subjects’ affirmations as support for the psychological criterion.  “These facts make a strong case for saying that the experimenter has brought it about that B did in the outcome get what he wanted and A did not.”

Next, Williams supposes that A and B both choose according to the bodily criterion of personal identity:  A chooses that the A-body-person will be rewarded, while B favours the B-body-person.  Again, the experimenter resolves the conflict by rewarding the A-body-person.  The A-body-person is pleased with the outcome, although, since he has B’s memories, he will “reckon it good luck that the experimenter did not do what he recalls choosing.”  And the B-body-person will think he

has gotten what he chose, but not what he wanted.  So once more it looks as though [the A-body-person and the B-body-person] are, respectively, B and A, and that original choices of both A and B were unwise.  (p 79)

A third case, in which both B and A choose in favour of the A-body-person, supports the same conclusion on the same grounds.  And other cases Williams considers, in which physical or psychological differences between the two subjects are substituted for the financial incentives, all appear to support the view that

the only rational thing to do, confronted with such an experiment, would be to identify oneself with one’s memories, and so forth, and not with one’s body.  The philosophical arguments designed to show that bodily continuity was at least a necessary condition of personal identity would seem to be just mistaken. (p 80)

Common to all these cases is that the force of the argument for the psychological criterion stems from the credence we are inclined to give to the beliefs and attitudes of the experimental subjects, after the experiment.  We suppose they are not somehow mistaken.  Could they be mistaken?

Cases Supporting the Bodily Criterion

Williams launches into a second series of cases designed to make us doubt the conclusions drawn from the first set.   These cases also feature an experimenter with power over the fate of his experimental subjects.  But now there is just one main subject, Williams himself.  In the first case in this series, the experimenter

tells me I am going to be tortured tomorrow.  I am frightened, and look forward to tomorrow in great apprehension. (p 80)

The example works as well, I think, if the experimenter promises to appropriate Williams’ house and retirement savings, leaving him with the prospect of an impoverished life.  But let the torture example stand, since it figures centrally in Williams’ description of these cases.

The experimenter adds that he will induce amnesia, so that, when Williams is tortured, it will come as a surprise – he will not remember being told about the torture in advance.  “This certainly will not cheer me up,” says Williams, “since I know perfectly well that I can forget things.”  The fact of forgetting, or being made to forget, the announcement of torture will not eliminate the pain of the torture when it occurs.  The pain, therefore, remains something that it is appropriate for Williams to dread, now.  (And if you, like me, find the physical torture hard to take seriously, change the example to a Somali pirate holding you and your child until you arrange a payoff of 60% of your net worth, after which he will give you an injection to make you forget the whole experience, and set you adrift in a shipping lane where you will be rescued by a U.S. naval patrol boat.)   It seems reasonable for Williams to regard this prospect as a personal misfortune.

There is more.  The experimenter tells Williams that not only will he forget being told about the torture, he will forget everything.

This does not cheer me up, either, since I can readily conceive of being involved in an accident, for instance, as a result of which I wake up in a completely amnesiac state and also in great pain; that could certainly happen to me. (p 81)

All this is by way of setting up the first case in Williams’ series of cases designed to persuade the reader that personal identity is grounded in bodily continuity.  Note that this is actually three cases: first, the simple prospect of pain (or of loss of money); second, the prospect of pain (or loss) accompanied by forgetting the announcement of pain (or loss); and third, the prospect of pain (or loss) accompanied by forgetting everything about the past.  Forgetting everything is, of course, a major disruption of psychological continuity – sufficient, according to any plausible psychological criterion of personal identity, for identity to be lost.  According to the psychological criterion, it is unreasonable for Williams, before the amnesia-inducing drug is administered, to feel personal concern for the fate of the survivor of this operation.  It is rational for him to regard that survivor as a new (unfortunate) person, not himself.  For the proponent of the psychological criterion, the prospect of total amnesia is equivalent to the prospect of death.

By building up this case in three steps – torture, torture with selective forgetting, torture with blanket forgetting – Williams relies on our ordinary fears of future misfortunes carrying through to an extraordinary case where psychological continuity is completely lost – where the poor subject being waterboarded and mutilated cannot remember his name, the house he grew up in, or anything else about his past.   In contrast to the first set of cases, we cannot give credence to the experimental subject’s own convictions about who he is, because he does not have any such convictions.

If we are persuaded by this case, we have already abandoned the psychological criterion.  But Williams treats this as only the first, and least controversial, of a series of cases designed to support the bodily criterion of personal identity.

In case (ii), the fiendish experimenter promises not only to wipe out Williams’ memories, but to induce radical changes of character.  This does not make Williams feel better, or lessen his fear of the torture.

In case (iii), the experimenter says he will not only change Williams’ character, but replace his memories with false ones, that purport to be the personal memories of George IV, but in fact are just made up.  Williams regards this outcome as the prospect of insanity, added to the amnesia and torture featured in cases (i) and (ii).  The expectation of delusional pseudo-memories does not comfort him or lessen his fear of the torture.

In case (iv), the experimenter designs the character changes and pseudo-memories to faithfully emulate the character and memories of some real person, alive at the time.

Case (v) is like (iv), except that the character changes and pseudo-memories which the experimenter implants into Williams were captured from some real person by a brain-scanning process like that described in the first series of cases.  Williams comments:

Fear, surely, would still be the proper reaction: and not because one did not know what was going to happen, but because…one did know what was going to happen – torture, which one can indeed expect to happen to oneself, and to be preceded by certain mental derangements as well.

If this is right, the whole question seems now to be totally mysterious.  For what we have just been through is of course merely one side, differently represented, of the transaction which we considered before; and it represents it as a perfectly hateful prospect, while the previous considerations represented it as something one should rationally, perhaps even cheerfully, choose out of the options there presented. (p 81)

Two Descriptions

Williams notes two evident differences between the presentations of the first and second set of cases.  The third in this list is my own:

  1. In the second version, the dreadful outcome (torture) is described as happening to Williams himself.  The experimenter tells him outright, “You will be tortured.”  This biasses the interpretation from the start.
  2. Up to and including case (v), the story does not mention “the other man, except in the somewhat incidental role of being the provenance of the impressions of the past” that  Williams ends up with.
  3. In the second version, the story is built up gradually, so as to elicit agreement at each stage before being embellished with further details.  If the experimenter began by presenting case (v) all at once, the listener might be less inclined to leap to the conclusion that the recipient of torture would be himself.   It might also prompt him to question the fate of the other man.  If he received an answer which made this case resemble the first kind of case, he might begin to hope that he would escape the torture.

Williams takes one more step in the second series, adding case (vi), which just like (v), except that Williams’ brain too is scanned, and its information transferred to the brain of the person whose psychological states are destined for Williams’ body.

The reader will recognize that case (vi) in this series has all the relevant features of the cases in Williams’ first series, that seemed to support the psychological criterion.   Yet Williams intends us to draw a different, and opposite, conclusion.   Redescribing the second series of cases as happening to our original protagonists, A and B, with A taking over his own role, Williams has this to say about the transition from case (v) to (vi):

If A’s original fears could reach through the expected changes in (v), as they did in (iv) and (iii), then certainly they can reach through to (vi).  Indeed, from the point of view of A’s expectations and fears, there is less difference between (vi) and (v) than there is between (v) and (iv) or between (iv) and (iii). (p 85)

“From the point of view of A’s expectations and fears” is a crucial qualifier.  A huge difference between cases (v) and (vi) is that only the latter contains a plausible continuer of A – someone who claims to be A, knows things previously known only to A, possesses A’s psychological attributes – someone with whom he might plausibly identify.  But we are under the spell of the story of the second series of cases, and in that context, the new person introduced in case (v) seems irrelevant to A’s fate. What happens to the new person does not affect the A-body-person whom, in the previous cases, we have already decided is A.  The move from case (v) to case (vi) makes no difference to what happens to A’s body.  How could extrinsic events affecting a different body change what A can reasonably expect to experience?

In the present case there is absolutely no difference at all in what happens to [A], the only difference being in what happens to someone else.  If he can fear pain when (v) is predicted, why should he cease to when (vi) is?  (pp 85-86)

One reason is that there is a competing target for self-concern in case (vi), one that, if described differently, can be quite compelling, as we have seen.   Responding to Williams, Robert Nozick argued that the existence of this competing target actually makes all the difference.  He argued that because the B-body-person with A’s memories and personality is, in relevant respects, a closer continuer of A than the A-body-person, it follows that the B-body-person is in fact A.  Therefore it is rational for A, before the procedure, to feel prudential concern for the fate of the B-body-person.  According to this view, what such cases teach us is that being the same person is not an intrinsic relation.  Whether or not the A-body-person really is A at the end of the experiment, depends on whether or not someone else with a better claim to be A exists at that time.  Personal identity, on this view, is like monarchical succession.

Claims of Personal Identity Underdetermined by the Facts

But Nozick’s conclusion is stronger than the facts that support it.   This point can hardly be emphasized enough: any conclusion implying that I should expect to experience one person’s future rather than another’s makes a claim beyond observable facts.   A philosopher who cannot stomach the theory that events external to his own body can determine whose future is his own, need not do so; he can stick to his guns without being forced into contradictions or denial of demonstrable fact.

Williams’ paper is refreshing because he feels the force of arguments on both sides, acknowledging the pull of psychological and bodily criteria of personal identity.  The arguments, as we have seen, are just stories – different ways of describing the same underlying facts – with the power to persuade listeners to believe different interpretations and make different choices.

The persuasive force of these stories should be regarded not as philosophical insight but as psychological fact.  Williams’ cases demonstrate a certain amount of plasticity in the psychological phenomena of self-identification and self-concern.  The direction of this plastic flow appears to depend on the three factors outlined above: the (authoritative) use of identifying terms (“This will be done to you”),  the inclusion or omission of information about alternative targets of self-concern, and the order in which facts are presented.

Williams’ cases highlight the psychological facts that, in absence of a competitor, self-identification tends to follow the body; but when there are competing targets for self-concern, psychological continuity can trump bodily continuity.  These facts may help to illuminate the common human tendency (some would call it a weakness) to expect life following death.  Many of us seem strongly inclined to think of ourselves as continuing – to expect future experience, whatever happens.  If faced with competing candidates (which we almost never are, as yet, except in fiction or philosophers’ thought-experiments) we choose according to our beliefs about the nature of the substrate of continued personal existence.  For some, it is the body; for some, a soul (which resides in the body during biological life); for some sophisticates, it is a set of psychological attributes.  But when faced with the prospect of ordinary death, we lack evidence of any plausible candidates.  In that situation (which is everyone’s situation, the common lot of man) many of us invent a candidate.

Imagined afterlife is a topic to be explored on its own in a future post.  People entertain a variety of ideas on the subject.  They are often content for those ideas to be vague.  Belief does not require a high standard of evidence – many people believe in an afterlife on the shakiest of evidence, or none at all.

The fact that this plasticity exists, and the evidence that most human beings have a strong psychological motive to self-identify with someone or something – even something that is almost nothing – gives further grounds for doubt that we ever have real cause for self-concern.  We have no reason to fear or hope for our futures; we just do those things.

And since there is flexibility in the target of self-concern – it can go with body or with mind – we can sensibly ask whether self-concern could be stretched even further.  Could we feel about persons we customarily think of as “others” the same way we feel about ourselves?  If so, what conditions would make us feel that way?  How would the story have to be told?  These are psychological questions.

Undecideability and Conventional Decision

Williams supposes for the sake of argument that the psychological criterion of personal identity is used to decide case (vi), but that the bodily criterion should apply to case (i).  Looking backwards through the second series, considering cases from (vi) to (i), he can find no plausible place to “draw the line”.

Here someone will say: you must not insist on drawing a line – borderline cases are borderline cases, and you must not push our concepts beyond their limits.  But this well-known piece of advice, sensible as it is in many cases, seems in the present case to involve an extraordinary difficulty.  It may intellectually comfort observers of A’s situation; but what is A supposed to make of it?  To be told that a future situation is a borderline one for its being myself that is hurt, that it is conceptually undecideable whether it will be me or not, is something which, it seems, I can do nothing with; because, in particular, it seems to have no comprehensible representation in my expectations and the emotions that go with them.  (p 86)

Williams likewise rejects making an overtly conventional decision – just deciding where to draw the line.  We are trying to answer the substantive question of what we should expect in the future.  A substantive question cannot be settled by conventional decision.

Dwell for a moment on Williams’ description:  “no comprehensible representation in my expectations and the emotions that go with them.”  He cannot deal with there being no fact of the matter as to whether or not he will the one who experiences some future pain.  He believes there must be a fact of the matter.

Yet good empiricists, lacking evidence of such a fact, should assume there is none.  How are we to deal, psychologically, with that assumption?

How Can We Deal With This?

After considering the convincingness of the two presentations, and trying out conceptual undecideability, and conventional decideability, Williams is left

not in the least clear which option it would be wise to take if one were presented with them before the experiment.  I find that rather disturbing.  (p 90)

Disturbing, indeed, that there should be no answer to the question, “What is in my best interests?”

Williams concludes his paper by coming down very tentatively – in a “shaky way” – on the side of bodily continuity as the grounding relation for prudential self-concern.  To adopt it, he adds, “would be risky; that there is room for the notion of a risk here is itself a major feature of the problem.”

More plausibly, there is no real risk, because there is no fact of the matter as to whether the pain will be mine or not.  Again, how are we to deal with that, psychologically?

One way would be to take Jesus’ advice, and love our neighbours as ourselves.

I must insert a disclaimer: I am not a Christian.  I am, in fact, an atheist.  However, I do not think the fact that an idea is rooted in a religious tradition is grounds for rejecting it.  Religious thought, in many traditions, contains good ideas (as well as bad ones).

Jesus’ advice is on the subject of self-concern.  He is saying we should not withold from others the concern we invest in ourselves.  The facts are, there is pain and suffering in the world as well as joy and happiness.  There is no fact about which pain and joy will be mine and which will not.  That is a fiction – an illusion that evolved to motivate us to strive and compete.  The rational attitude – the one most compatible with an objective, scientific account of reality – is that it doesn’t matter whose pain and joy it is.

If we enter the realm of values, and assert that pain is bad, then everyone’s pain is bad.  If happiness is good, we should strive impartially for everyone’s happiness.

Try it.  The next time you pass a homeless person on a cold sidewalk, try expecting to feel her chill and hunger, even her resentful anger towards the better-heeled passersby, and the terrors engendered by ignorance and want.  Or alternatively, cultivate an indifference towards your own future equivalent to the indifference you are customarily inclined to feel towards the lives of the hundreds of people who pass you by on a busy downtown sidewalk.  These are good exercises, or at least interesting ones.

But they are not easy exercises.  The tendency to feel differential concern for a person one identifies as oneself is deeply ingrained – some would say ‘hard-wired’.  It is not easily overcome by a mere intellectual understanding that we have no more reason to dread our own suffering than that of others, as much reason to look forward to a stranger’s happiness as our own.   In me, as I suspect in you, self-concern is almost always present, a major determinant of how I behave in the world.  However, knowing that it is not rationally required gives me, I think, a little more freedom in how I respond to it.  I can decide to overrule my self-concern, and sometimes do.


I suspect that if Williams’ introspective, personal psychological experiment were repeated on a larger scale, results would demonstrate some plasticity in most, if not all, human beings’ self-concern for the future.   Introspection can still be a useful tool for psychological research, although care should be taken not to over-generalize from one’s own case, or one’s acquaintances.

It is particularly important not to lose sight of the extreme cases described in the neurological literature.  Extreme and pathological examples save us from the mistake of confusing normality with necessity.   Thomas Metzinger’s analysis of consciousness and the self is shaped by the requirement of including the entire known range of human behaviour within its explanatory scope.  Metzinger describes Cotard’s Syndrome, described in an earlier post, as “existence denial.”

In this case the first lesson to be drawn is this: You can be fully conscious and still truthfully describe the content of your own phenomenal selfexperience as “non-existence”. (Metzinger, 2003, p. 4)

In Cotard’s patients, self-concern seems to have disappeared entirely.  If so, that defines one limit of its plasticity.

Metzinger’s book, Being No One, is germane to our topic.  It is an impressive work which I hope to make the subject of future posts.

First, though, I want to examine another philosopher’s work on what I have come to think of as the anatomy of self-concern.  Raymond Martin’s book – titled, Self-Concern – an experiential approach to what matters in survival – examines the subject of self through a phenomenological lens, focussing more on how people actually allocate self-concern than on how they should.

Click here to return to the Phantom Self home page.


Martin, Raymond (1998), Self-Concern, Cambridge Studies in Philosophy

Metzinger, Thomas (2004),  Being No One

Metzinger, Thomas (2003), “Why Are Identity Disorders Interesting for Philosophers?” in T Schramme & J Thome (eds.), Philosophy and Psychiatry.

Nozick, Robert (1981) “Personal Identity Through Time”, in Martin and Barressi, ed. (2003) Personal Identity

Williams, Bernard (1970), “The Self and the Future”, in Martin and Barressi, ed. (2003) Personal Identity

One Response to “The Plastic Self”

  1. […] there is no reason in the end to discredit this kind of imagining.” [Johnston, 2010, 57/58]  I discussed such imaginings in a post on Bernard Williams’ “The Self and the Future,” explaining them as […]

Leave a Reply