Summary of Are We Bodies or Souls?
This book argues that human beings consist of two parts – body and soul, of which the soul is the essential part which makes us who we are. After defining several crucial words used for discussing this topic, it proceeds (in chapter 3) to discuss theories of personal identity — that is, theories which purport to analyse what it is for one person P2 at time T2 to be the same person as a person P1 at an earlier time T1. Almost all currently discussed theories are “complex theories”, in the sense that they analyse P2 being the same person as P1, (wholly or partly) in terms of P2 having enough of the same body, or brain, or memory and character, as P1. All such theories have insuperable difficulties; and the book argues that P2 being the same person as P1 consists in P2 having the same soul as P1.
Descartes gave an argument beginning from a premise “it is conceivable”, that is it involves no contradiction to suppose, “that I could exist without a body”, to reach the conclusion that having a soul is all that is necessary for anyone’s existence. Many contemporary philosophers object that no one knows to what “I” (as used by them) refers, and so Descartes is not entitled to that premise. This book claims that everyone can know to what they are referring by “I”, and so each of us can use Descartes’s argument to show that our soul is all that is necessary for our existence. Chapter 6 argues that not merely do brain events cause mental events (that is, events in our souls), but mental events cause other mental events and also brain events. Chapter 7 claims that although there might be a scientific explanation of the origin of a human soul from a foetus, and of the way in which that soul interacts with its brain, there could not be a scientific explanation of why the soul that evolves from a particular foetus is the particular soul it is. So there could not be a scientific explanation for why any of us exist.
Introduction to An Extract from Chapter 3 of Are We Bodies or Souls?
A strong complex theory of personal identity analyses P2 being the same person as P1 wholly in terms of P2 having enough of the same body or brain or memory and character, as P1. This book presents two objections to all such theories. The first objection is that any precise version of such a theory would involve a totally arbitrary criterion (e.g. requiring P2 to have 55%, rather than 54% of the brain matter of P1.) The second objection is that many such theories (e.g. theories which require P2 to have 45% of the brain matter of P1) could be satisfied by more than one future person, but they could not both be P1. The following extract from the book argues that attempts to meet these objections to strong complex theories all fail.
The only way for a strong complex theory to meet the objections that any precise version of such a theory would involve an arbitrary criterion (e.g. require the subsequent person to have 55 per cent of the brain matter of the original person) and often be satisﬁed by more than one candidate is by denying an assumption made in the previous discussion that it is always either true or false that some later person P₂ is identical with an earlier person P₁. Many of our concepts are such that for some object it is neither deﬁnitely true nor deﬁnitely false that the concept applies to them. When a surface is reddish-blue, it is neither deﬁnitely true that the surface is red nor deﬁnitely false that the surface is red. Harold Noonan has suggested that what we should say when some later person is by some set of criteria on the border between being and not being the same person as an earlier person is that this is an example of what he calls “semantic indecision”; that it is neither true nor false that that person is the same person as the earlier person. And so it would follow also, when there is more than one plausible candidate for being “the same person” as some earlier person, that it is neither true nor false of each candidate that he is the “same person” as the earlier person. Our concept of “same person” is not designed to ﬁt borderline cases, claims Noonan.
After all, that is what we do say when we are dealing with the sameness over time of ordinary inanimate physical objects. If you replace only one of the four legs of a table, it is true that the resultant table is the same table as the earlier table; if you replace the top of the table and two of its legs, it is false that the resultant table is the same table as the earlier table. But in an intermediate case, when you replace three of the legs but keep the table top and one leg, our concept of “same table” is not sharp enough to provide a clear answer to whether the resultant table is the same table as the earlier table; it is neither true nor false that the earlier table still exists.
But, despite the plausibility of this kind of answer to problems of the identity of inanimate physical objects, we can see its implausibility as an answer to the problem of personal identity, if we add a further detail to the earlier thought experiment. Suppose that Alexandra and two other unfortunate humans, Alex and Sandra, have been captured by a mad surgeon. The surgeon tells Alexandra that he is going to remove her cerebrum from her brain, divide it into the two hemi-spheres, put Alexandra’s left hemisphere into Alex’s skull from which her left hemisphere has been removed, and put Alexandra’s right hemisphere into Sandra’s skull from which her right hemisphere has been removed; the two hemispheres will then be connected to the other parts of the two brains of what were Alex and Sandra respectively. There will then at the end of the process, both the surgeon and Alexandra reasonably believe, be two conscious persons. The two resulting persons, they have reason to believe, will have equal degrees of physical and mental continuity (and connectedness) with the earlier Alexandra. The surgeon tells Alexandra that after the operation, he will kill one of these resulting persons; but he offers Alexandra the choice now of whether it is the person with her right hemisphere or the person with her left hemisphere who will be killed. Alexandra believes that the surgeon will do what she chooses, and she wishes to survive. How should she choose so as to have the best chance of surviving?
On a “semantic indecision” view, it won’t matter which choice Alexandra makes, since whatever she chooses, it will be neither true nor false that she will survive. But surely there is a truth about whether or not Alexandra will survive. Or at least there may be a truth that Alexandra will survive or that Alexandra will not survive the operation if she makes a certain choice; and yet the “semantic indecision” view has the consequence that whichever person is killed, that could not make any difference to the outcome of the operation. Yet it is surely (logically) possible, there is no contradiction in supposing, that, as with kidney or heart transplants, there is a truth about whether someone will survive an operation. It cannot be ruled out by some a priori philosophical theory that someone could survive an operation to remove some of their brain. When someone is about to undergo a brain operation of a more familiar kind, they normally hope to survive, and there seems to be nothing irrational in having that hope. How can this case be different, except in the respect that the doubt arises not from a doubt about whether the brain on which the operation is performed will ever again be the brain of a conscious person, but from a doubt about whether that brain will be the brain of the same person as the person whose it was before the operation?
The semantic indecision view holds that in almost all cases there is a truth about whether some future person will be me or not me; it is only in borderline cases that it is neither true nor false that a certain future person is the same person as a certain earlier person. But a semantic indecision view has to take a view about what are the boundaries of the neither-true-nor-false area. Thus, a semantic indecision version of a brain theory may hold that it is deﬁnitely true that P₂ is the same as P₁ iff P₂ has at least 55 per cent of P₁’s brain, and deﬁnitely false that P₂ is the same as P₁ iff P₂ has no more than 45 per cent of P₁’s brain, but it is neither true nor false that P₂ is P₁ iff P₂ has between 45 per cent and 55 per cent of P₁’s brain.
But this theory is now open to the arbitrariness objection — why take the neither-true-nor-false area as the area between 45 per cent and 55 per cent rather than the area between 40 per cent and 60 per cent? Any choice of the cut-off point as we take cases where P₂ is less and less continuous with P₁, where it becomes neither-true-nor-false that P₂ is the same as P₁, and any choice of the cut-off point as we take cases where P₂ becomes yet less and less continuous with P₁, where it ceases to be neither-true-nor-false that P₂ is the same as P₁ but to be deﬁnitely false that P₂ is the same as P₁, seems totally arbitrary. Hence it is natural for anyone sympathetic to the semantic indecision view to move to a more general view which I call the “partial identity” theory: that a future person P₂ at a time t₂ is less and less identical to a person P₁ at a time t₁ insofar as there is less and less physical and mental continuity (and perhaps connectedness) between them. On this view it is only if t₂ is immediately after t₁ and there have been virtually no changes at all in body or brain or a-memory or character between those times that P₂ is really, fully, identical with P₁.
Different writers have used different expressions instead of “partially identical” in order to state what is essentially a “partial identity” theory. Robert Nozick wrote about later persons being “continuers” (in physical and mental respects) of earlier persons; being the “best continuer” and so being identical to the earlier person is just having more of what makes for continuity. David Lewis wrote of personal identity as “a matter of degree” of a relation (of continuity and connectedness) which he calls the “R-relation”; Derek Parﬁt wrote of later persons being (to different degrees) “survivors” of the earlier person (to the extent to which they have physical and mental continuity and connectedness with them). (Note that Parﬁt’s sense of “survive” is not the normal sense of “continue to exist”, since in Parﬁt’s sense a person may have several different survivors at the same time.)
The theory common to all these writers is that a future person being really “identical” with a past person is just something like having a great degree (perhaps the maximum degree) of continuity and connectedness with that person, and a greater degree of continuity and connectedness with that person than with other persons. One consequence of a “partial identity” theory is, as these writers admit, that one’s concern for one’s own future happiness is just concern for the happiness of some future person who has a large degree of continuity (and connectedness) with oneself, and so it would only be rational to have a similar concern (even if one of lesser degree) for anyone who has any degree of continuity (and connectedness) at all with oneself. This is not claiming that altruism (concern for others) is rational, but claiming that being selﬁsh just is concern that people with much of one’s own body, brain, memories, and character should ﬂourish. Nozick, Lewis, Parﬁt, and most writers on personal identity who advocate a similar theory acknowledge that their theory seems initially implausible. But, they claim, that theory is the one to which reason leads.
There is, however, what I suggest is a conclusive objection to all “partial identity” versions of strong complex theories. All such theories which allow the possibility that P₂ at a time t₂ may be partly identical to an earlier P₁ allow the possibility that there can be at the same time t₂ another person P₂* who is also partly identical with P₁, although perhaps only to a lesser degree.
Now consider another version of the earlier thought experiment where Alexandra knows that each of her two cerebral hemispheres will be transplanted into the brain of one of two future persons, Alex and Sandra, replacing both of their hemi- spheres. Suppose this time that Alexandra knows that something very good will happen to Alex — for example, that Alex will win a million dollars in a lottery, and that something very bad will happen to Sandra — for example, that she will be kidnapped and tortured. If the operation has the result that the resulting Alex is partly identical to Alexandra (or “is a survivor” (in Parﬁt’s sense) or “continuer” of Alexandra (in Nozick’s sense), or is R-related to Alexandra (in Lewis’s sense)), then presumably Alexandra will have some good experience to which to look forward, but — since Alex is only partly identical to Alexandra — an experience which only a part of Alexandra will enjoy or which she will only partly enjoy or of which she will enjoy only a part. And if Sandra is partly identical (or whatever) to Alexandra, then presumably Alexandra will have some bad experience which she has reason to fear, but — since Sandra is only partly identical to Alexandra — something which only a part of Alexandra will ﬁnd unpleasant or which she will ﬁnd only partly unpleasant or of which she will ﬁnd only a part unpleasant. So Alexandra will rightly expect to have one or more future experiences which in one of these ways will be mixed partly pleasant and partly unpleasant experiences.
Now it certainly makes sense to talk of an inanimate physical object being “partly identical” to (and so “surviving as” or whatever) each of two subsequent physical objects, in the sense that each of the subsequent objects includes different parts of the original object. Both the table formed from three of the legs of a previous table and the table formed from one leg and the top of the original table could rightly be said to be “partly identical” with the earlier table. But although persons have parts, the parts don’t have experiences; the person has the experiences. When I have a visual sensation and an auditory sensation at the same time, it is not the case that one part of me has the visual sensation and another part of me has the auditory sensation. Maybe one part of my brain causes me to have the visual sensation and another part of my brain causes me to have the auditory sensation, but what they cause is co-experienced. Whoever is aware of the one sensation is aware of the other sensation, and that “whoever” is the person. So it makes no sense to suppose that “part” of Alexandra can have some experience, which another part does not have; and so it makes no sense to suppose that the partial identity of Alexandra with Alex and Sandra could consist in Alex already being a part of Alexandra and then subsequently having the pleasant experience, and Sandra already being a part of Alexandra and subsequently having the unpleasant experience.
But perhaps the partial identity of each of Alex and Sandra with Alexandra is not a matter of each being a part of Alexandra, but rather of Alexandra having an experience which she will only partly enjoy or of which she will enjoy only a part. Yet in the scenario which I have described, no one person has only partly enjoyable experiences, or experiences of which they enjoy only a part. Alex will have a fully enjoyable experience, not containing any unpleasant part; and Sandra will have a totally unpleasant experience, not containing any pleasant part. If the experiences of Alex and of Sandra were both in some sense also Alexandra’s experiences, then whoever has one experience should be to some degree aware of having the other experience. For if someone is not to any degree aware of some experience, that experience is not in any sense their experience. But in the suggested scenario Alex and Sandra are different persons from each other and have different lives, unaware of each other’s thoughts and feelings. I conclude that it does not make any sense to talk of a person being “partly identical” with an earlier person; and so that the theories of Nozick, Lewis, and Parﬁt claiming that the identity of a person with an earlier person is simply the maximum degree of a relation of which there are degrees are mistaken.
Derek Parﬁt imagined a spectrum of cases where a different proportion of the neurons of each brain is replaced by new neurons. In one brain 1 per cent of the neurons are replaced, in the next brain two per cent are replaced, and so on until we get to a brain where 99 per cent of the neurons are replaced. Parﬁt points out that the simple theory is committed to the claim that there is a precise point along this spectrum where the tiniest extra replacement of one neuron causes the existence of a different person. He writes that this claim “is hard to believe”. But at least it is comprehensible; it is easy to understand the claim that there is such a precise point. The trouble with the supposition of “partial identity” is that it is not comprehensible. I suppose that why Parﬁt ﬁnds the claim that there is such a precise point “hard to believe” is because he supposes that a small difference in a cause can only make a small difference in its effect.
But much recent science has shown that sometimes a small difference in a cause can make a very large difference in its effect. For example, chaos theory has taught us that a butterﬂy ﬂapping its wing in one part of the world can cause a hurricane in another part of the world a few days later. And the brain is a system in which very small changes can sometimes produce very big effects very quickly. A very small change in the amount of neurotransmitter chemical released at a synapse (the narrow space between adjacent neurons) or in the exact width of the synapse determines whether electric potential is transmitted from one neuron to the next neurone, so as to cause the latter neuron to “ﬁre”; and one neuron ﬁring can cause many other neurons to ﬁre. So there is no good reason to suppose that a very small change in the number of neurons replaced in Parﬁt’s spectrum cannot make all the difference to who the resulting person is.