Derek Parfit1 proposes that we separate the notions of identity and survival. This follows, he thinks, from the results of certain thought experiments—such as one put forth by David Wiggins—which seem to suggest that it is possible for A to survive into the future even if nobody will be identical to A at that time. This does not mean, however, that he places no conditions on survival. The conditions Parfit places on survival are possession of “character and apparent memories” (200). These conditions map fairly well onto a common use of the word ‘survive.’ They also suggest that many important questions can be decided without the use of the notion of identity, as was previously thought. However, many more questions are still left unanswered. Parfit’s conditions also lead to some quite counter-intuitive results which leads me to question the validity of Parfit’s new definition of survival. I will argue that for these reasons we should not think that one survives into the future if Parfit’s proposed conditions obtain.
Parfit takes these conditions over from those used in thought experiments by Sydney Shoemaker. He proposes that it makes sense to think of someone surviving after having his or her brain transplanted into a new body because “the resulting person has [the original person’s] character and apparent memories” (200). David Wiggins goes on to imagine a case where one person’s brain is split in two and each hemisphere is transplanted into a new body. Parfit thinks that in this case we must say that the person survives as two persons. This case leads Parfit to say, in order to avoid contradictions, that we should separate survival and identity since the two resulting persons are clearly not identical.
Parfit then redefines memory and other psychological relationships such that one need not be the same person as the one who originally experienced the thing remembered. This implies that survival is not transitive. If X remembers most of Y’s life and Y remembers most of Z’s life, X will not necessarily remember most of Z’s life. For Z to survive as X, X must have direct memories of Z’s life. This notion of survival already has some use in our common ways of speaking. We sometimes say that one can survive through his or her children. Parfit’s conditions would make sense of this by pointing to character traits shared by the parent and the child. And, in line with Parfit’s proposal, as generations go on we would say less and less that grandparents and so on survive through currently living persons.
Parfit’s conditions suggest, as he notes, that certain important questions, such as that of survival, can be solved without reliance on the notion of identity. But many other questions are left unanswered. It seems possible, at least in theory, to survive a process of division into several different persons. If someone enters into a contract and then divides who is responsible for upholding the contract? If I borrow money from someone and then she divides, to whom do I owe money? If someone commits a crime and then divides, who should we arrest? If a married woman divides should we arrest the husband of these women for practicing bigamy (illegal in all 50 states and many other countries)? It does not seem that these questions can be decided in any principled way without reference to the notion of identity.
These conditions of survival also leave unanswered more relevant questions. Parfit notes that these conditions on survival imply that survival is a relation of degree. Possessing half of A’s memories means that A half-survives. But, suppose someone is subject to brain damage such that only one character trait and one apparent memory remain. Should we say in such a case, that he survives? We might say that he only ‘survives’ in that his body is still alive, but that an otherwise unrelated person now inhabits the body. If Parfit says that he does not survive, then he will need to revise his conditions such that one needs to possess a certain degree of character and apparent memories in order to count as surviving. This would then need to be supported by an argument that tell us why that specific degree is the condition for survival rather than an arbitrarily chosen degree.
It might seem easier then just to stick with the current account and say that someone suffering brain damage such that only one character trait and one memory remain does survive the brain damage. But this leads to results even less satisfactory than before. Certain cases of source amnesia involve a subject who comes to believe, on the basis of hearing a story for example, that he or she has actually experienced the events of the story. Thus it is possible to remember experiences that actually happened to another person. Similarly, it is common for character traits to ‘rub off’ onto others with whom one has close contact. Imagine two friends, A and B, such that both of these scenarios obtain. B has at least one of A’s traits and at least one of A’s memories. Parfit’s conditions lead to the unintuitive result that A will have survived as B.
I think the severe brain damage case poses a real problem for Parfit. Assuming that in every case, a person either survives or doesn’t, either way Parfit could decide the survival of the brain damage patient seems unsatisfactory. Either Parfit needs to supplement his account by saying to what degree these psychological connectedness must obtain in order to count as survival, or his theory predicts that one can survive as another person without dying. We can even imagine, in the scenario with A and B, that B dies before A. This suggests that we do not judge survival on the basis of connectedness of psychological states.
1) Parfit presents these claims in his essay “Personal Identity.” Quotations refer to the essay as printed in Personal Identity. Ed. John Perry. University of California Press. 2008. Pages 199-223.