Konstantin Zarubin
Uppsala University, Sweden, spring 2018
Introduction
Derek Parfit’s Reasons and Persons is a modern classic. So many of the thought experiments Parfit takes up there have become standard fare in personal identity debates that it is hard to pick one. This is why I am happy that the choice has been made for me.
In this essay I will look at the thought experiment which Parfit calls My Division (pp. 254-261; all page references to Reasons and Persons are to the corrected 1987 version). In its basic outline, My Division follows a scenario discussed by Wiggins (1967). It plays a key role in Parfit’s argument that personal identity is not what matters in survival. The importance of My Division is stressed by Parfit himself: in a personal endnote, he confesses that he chose to “study philosophy almost entirely because [he] was enthralled by Wiggins’s imagined case” (p. 518).
In what follows I will first present Parfit’s set-up of the thought experiment and discuss how he uses it to argue that personal identity is not what matters. I will then try to show a potential weakness in one of Parfit’s premises, inherited from Wiggins’s original formulation of the case. I will use the example of episodic memory to show that assuming full psychological transfer in a My Division-type case can come at a price.
My Division in Reasons and Persons
Parfit asks us to imagine that he has two brothers, presumably of about the same biological age. One unlucky day Parfit suffers a fatal injury. Before he dies, his brain is taken out and split in two by some highly advanced neurosurgeons. The neurosurgeons then transplant one hemisphere into the body of one of Parfit’s brothers and the other hemisphere into the body of the other brother. (I can only assume that, for some other tragic reason, both brothers are braindead at the time of the surgery.)
We are asked to imagine that, before the split, the two hemispheres of Parfit’s brain “have the same full range of abilities” (p. 255). Moreover, the neurosurgeons manage to divide Parfit’s lower brain exactly in half and transplant it along with the hemispheres. In other words, once the surgery is over, there are two living, fully functioning human organisms – I will follow Parfit in calling them Lefty and Righty – who are both continuous with Parfit in equal measure. Each of them has the same share of Parfit’s psychology. Each has inherited the same amount of his neural tissue. Even their bodies are similar to Parfit’s now-defunct body to about the same degree. After all, both bodies were harvested from Parfit’s brothers.
It needs to be said at this point that Parfit, along with many others, distinguishes two broad kinds of views on personal identity. On the one hand, there are reductionist views. Reductionists believe that someone’s identity can be explained in terms of some physical or mental facts, such as the survival of the brain or a continuous chain of memories. There is no person or self that exists separately from those facts. By contrast, non-reductionists hold that personal identity can’t be reduced to such facts. Whether person A is identical with person B goes above and beyond any physical or psychological facts about A and B. Another tenet of non-reductionism is that personal identity is an all-or-nothing matter. Either A is the same person as B, or they are two different persons. Either A continues to exist, or she stops existing. There can be no intermediary stages.
Descartes’s view of the self is a textbook example of a non-reductionist approach to personal identity. A Cartesian believes that we are purely mental entities (whatever that might mean) that can exist with or without the body. Therefore, person A is identical to person B as long as the subjective experiences of A and B are had by the same mental entity.
Parfit, for his part, defends a thoroughly reductionist view of persons. He does not deny that persons exist or that we can make meaningful statements about them. However, Parfit argues, to say that there is a person is simply to say that there is a brain connected to a body and that this brain-body duo has mental states and carries out actions (pp. 251, 471). My Division is one of the cases that Parfit considers in his attempt to show that reductionism is more plausible than non-reductionism.
The rest of this section follows Parfit’s discussion, with some variations and additions. For a non-reductionist, Parfit notes, there is always a fact of the matter as to whether a particular person is me. If we pick any random person at some point in spacetime, it should be possible to say “Yes, this is me” or “No, that’s not me”. This seems intuitively attractive. After all, most of us want to survive, and it seems that we do not survive unless we can say “Yes, this is me” about some existing person. By the same token, it seems intuitive that we will be dead when we can no longer say “Yes, this is me” about any existing person.
But if that is correct, it should also be relatively easy to say what happens to Parfit in My Division. Recall that there are now two persons, Lefty and Righty, who each have exactly one half of Parfit’s brain and share Parfit’s psychology in equal measure. As Parfit points out (p. 256), there are four logically possible answers to the question “Which person is Parfit?” Either both of them are Parfit, or Lefty is Parfit, or Righty is Parfit, or neither of them is Parfit.
If neither of them is Parfit, it seems to follow that Parfit is now dead. It is, however, not clear why Parfit would be dead. To begin with, there are actual people with only one functioning brain hemisphere, and we do not consider them dead. Moreover, we can compare My Division with a more straightforward one-to-one transplant case. Suppose that Parfit only had one intact hemisphere after he suffered his fatal injury. That (left) hemisphere was transplanted into the body of his braindead brother. Now there is a person we call Lefty. Lefty thinks he is Parfit, says he is Parfit, writes books on utilitarianism and has 50% of Parfit’s brain. Are we willing to say that Parfit is dead in this case? If we are not, it is not clear what reasons we have to declare him dead in My Division. It feels odd to claim that one person (Parfit-Lefty) is dead simply because some other person (Righty) is alive. Indeed, such a claim would violate the intrisicality requirement, proposed by Bernard Williams and cited by Parfit with apparent approval (p. 267). According to the intrisicality requirement, whether Parfit is identical to Lefty – and therefore survives – must depend on some intrinsic relation between Parfit and Lefty. The existence of a Righty is not such a relation.
Let us consider our other options. If either Lefty is Parfit or Righty is Parfit, then there must be a reason why only one of them – and not the other – inherited Parfit’s personhood. What kind of reason could that be, given that Lefty and Righty have exactly the same degree of physical and mental continuity with Parfit? Perhaps one of them has a body that is in some way more similar to Parfit’s original body. For instance, Lefty’s body – but not Righty’s – might have a big Antarctica-shaped birthmark in exactly the same place. But this fact seems highly trivial. According to Williams’s non-triviality requirement, also cited by Parfit (p. 267), the question of personal identity is too important to be decided by a trivial fact.
Could there be a non-trivial difference between Lefty and Righty that might explain why one of them – but not the other – is Parfit? Perhaps we need to give up on reductionism after all and accept that a Cartesian soul can fit the bill. Suppose that Parfit’s essence was an immaterial Ego which stuck to the left hemisphere during the transplantation and now inhabits Lefty. This seems non-trivial enough to justify Lefty’s claim to being Parfit. Recall, however, that the two hemispheres were qualitatively identical. Why would Parfit’s Cartesian soul prefer the left one? It seems that we are back at square one, where we have to look for a trivial reason to explain the soul’s choice. As an alternative, we could argue that there was no reason at all. But that means that the soul’s choice was entirely random, and randomness in this matter seems even less acceptable than triviality.
Finally, what about Righty? He has no Cartesian soul, but he walks, talks and writes analytical philosophy just like Lefty. We might want to dismiss him as one of Chalmers’s philosophical zombies with no mental states, but we would hardly be justified in treating him as less of a person than Lefty. After all, we would have no way whatsoever to show he had no inner life – or indeed to establish which of the two, Lefty or Righty, had inherited Parfit’s Cartesian soul in the first place. It looks like non-reductionism remains a non-starter.
The only option that remains is to accept that Parfit is both Lefty and Righty at the same time.[1] However, it is not clear how that is possible. Personal identity is supposed to be numerical identity, not merely qualitative identity. By definition, a thing can only be numerically identical to itself and to nothing else (Stanford Encyclopedia of Philosophy). Parfit was one person; Lefty and Righty appear to be two. Therefore, Parfit can’t be identical to both of them at once.
There is a way around this conclusion. Perhaps Lefty and Righty only appear to be two persons when in fact they are still one person. In other words, perhaps we need to modify our concept of person by allowing two-body persons, three-body persons, and so on. However, such a move is very hard to square with how we normally think about persons. Suppose that a few weeks after the surgery Lefty runs into a charismatic Islamic theologian at a dinner party and has a life-transforming conversation. He converts to Islam and moves to Tajikistan, where he becomes a farmer. Ten years later, Righty is still an effete Oxford don writing a 2000-page follow-up to Reasons and Persons, while Lefty is a tough Tajik-speaking cattle-breeder with calloused hands and a sunburnt face. It seems absurd to regard the two as one person. In other words, allowing multibody persons may not be an option after all.
To sum up, none of the four possible answers to the question “Which person is Parfit?” in My Division seems tenable. This leads Parfit to conclude that the question is an empty one (p. 260). This, Parfit explains, does not necessarily mean that the question “Which person is Parfit?” has no answer at all. Instead, it means that it does not matter much which of the four answers we choose because they are simply four different descriptions of the same situation. We can know everything there is to know about the case without committing to any particular description.
Here Parfit seems to make a distinction between a correct description and a best description. As I understand his point, a correct description presents what there is to know about the case, while a best description simply offers a helpful perspective on the case. A correct description of what happens in My Division could look like this: first there is a brain-body called Parfit which does deeds and has mental states; then there are two such brain-bodies, called Lefty and Righty, both of which are psychologically and neurologically continuous with Parfit. There are no other deeper facts to know about this case; there is no metaphysical need to pose the question “Which person is Parfit?” A best description, on the other hand, comes in if we choose to pose that question. For Parfit, the most helpful description of My Division is that neither Lefty nor Righty is Parfit. But – and this is a crucial point – that does not mean that there was no survival. To be more precise, it does not mean that having your brain split and transplanted into two bodies is as bad as death.
Here then comes the punchline of My Division. According to Parfit, the thought experiment suggests that there is no need to talk about survival in terms of personal identity. Whether Parfit is the same person as either Lefty or Righty is an empty question. If it is an empty question, there must be something else that matters – something else that makes My Division less bad than death. For Parfit, this other factor is what he calls relation R: the relation of psychological continuity and connectedness. The reason why I would rather undergo a division than die is that relation R would hold between me and the two persons who would have my brain hemispheres after the transplantation.
Parfit’s conclusion seems convincing. Overall, I believe that Reasons and Persons makes a strong case for the relative unimportance of personal identity. I also believe that questioning the premises of thought experiments on empirical grounds is not always helpful or even meaningful. That said, a good philosophy paper is expected to engage with its subject critically. If I am to have an issue with anything at all in My Division, it will be the fact that Parfit’s thought experiment appears to rely on a highly unusual kind of brain. It is to this feature of My Division that I will turn in the next section.
A very special kind of brain
The view of persons that Parfit advocates in Reasons and Persons is both reductionist and naturalist. Persons are brains in bodies along with their actions and mental states. Their continued existence is the existence of the brains, bodies, actions and mental states involved. Of these four, mental states matter the most because they ensure psychological continuity and connectedness. Our intuitive reactions to imagined scenarios involving body-swapping or teleportation suggest that what we most care about are our memories, beliefs, desires, and other features of our mental lives. All these features are events in our brains; they have neural correlates.
It is a matter of ongoing research in neuroscience to what extent the neural correlates of various mental phenomena are localized in particular areas of the brain. However, there is good evidence that many of them are indeed highly localized. Lesions and injuries affecting particular parts of the cortex can cause loss of specific skills or memories as well as dramatic changes in character. It is true that brains, especially young brains, have a significant degree of plasticity; for instance, in aphasic patients lost skills can often be re-learned and taken over by intact parts of the cortex. But it is probably safe to assume that the brain does not lay down back-up neural pathways for every feature that is important for Parfit’s relation R. It does not make much evolutionary sense to duplicate every single neuronal structure as a matter of course. When we form a long-term verbal memory in the left half of the cortex, the same memory is not automatically copied to the right half.
In chapter 13, entitled “What does matter”, Parfit highlights the importance of episodic memory for psychological connectedness. Many of us believe that, all things considered, our lives so far have been worth living, and we “value highly our ability to remember many of our past experiences” (p. 301). For instance, I value greatly my memories of my deceased parents, of the summers I used to spend at the dacha as a child, of being 15 and lovesick on a slushy day in March many years ago, and so on. If I had to choose between having all those memories erased and being swiftly killed, it would not be much of a choice. Erasing those memories would destroy relation R, which has held my life together so far. I would cease to exist.
A key premise of My Division is that before the split the two hemispheres “have the same full range of abilities” (p. 255). Parfit does not explicitly mention the same range of memories,[2] but as we have seen, he does credit Wiggins, so it is perhaps worthwhile to quote Wiggins’s formulation of the thought experiment:
We are supposing that the transplanted persons, Brown I and Brown IT, claim to remember exactly the same things, that they are equally intelligent, and that they are equally at home in their new bodies. In this case, where cerebral material is actually transplanted we cannot simply disregard their (claimed) memories. For we understand far too well why they have these memories. (Wiggins, 1967, p. 53, italics in the original)
Parfit’s discussion of My Division (as well as the use he makes of Lefty and Righty in the subsequent chapters) strongly suggests to me that he shares this premise. To a significant degree, Lefty and Righty both enjoy psychological connectedness with pre-surgery Parfit because they have the same memories. Needless to say, this is entirely conceivable. At any rate, it is no less conceivable than body-swapping or teleportation to a Jupiter moon. However, this premise might have interesting consequences for what was conceivably the case before the split.
It is worth taking another look at the last sentence of the Wiggins quote above. At first glance, it seems almost trivially true. Surely we do understand why the post-split persons have the memories they do. They have them because their half-brains once belonged to the same person. But let us probe this a little further. How exactly is it possible that each of the two hemispheres ends up with exactly the same set of memories?
There are four possible explanations for this. For the sake of convenience, I will stick with the characters from My Division rather than switch to Wiggins’s Browns.
To begin with, it might be the case that Lefty and Righty do not, as a matter of fact, inherit a full set of Parfit’s memories. Instead, they only have the subset that was stored in both hemispheres at once. All the non-shared memories are somehow deleted during the transplantation. Another possible explanation is that every single memory Parfit has is stored in at least two copies. At least one copy is always in the left hemisphere; at least one other is in the right hemisphere. A third possibility is that Parfit’s memories before the split are completely unlocalized. Instead, they are somehow spread throughout the brain. A fourth and final possibility is that the hemisphere transplantation was combined with an even more science-fictiony procedure. First each hemisphere was scanned; then all memories that were missing from one hemisphere were copied to the other, and vice versa.
All four options have interesting implications for My Division. I will now consider each in turn.
(A) Lefty and Righty only inherit a subset of Parfit’s memories, namely those that were stored in both hemispheres at once. All other memories are deleted.
This is one of the two options that potentially involve a regular human brain. There is evidence – including evidence from split-brain patients with a cut corpus callosum – that the different hemispheres do contain a number of similar memories. Realistic or not, (A) is a poor fit for My Division because it compromises psychological connectedness. In (A), a lot of Parfit’s memories are irretrievably lost. Some of those memories may have been highly valuable to Parfit. They may have formed a core element of his personality, informing his goals and choices. But if that is the case, the question “Which person is Parfit?” seems to become less empty, provided that we link personal identity to psychological connectedness. The reply might be the same (“Neither of the two”), but this reply is no longer a best description. It begins to feel like a genuine answer.
(B) Parfit’s brain stores every single memory in at least two copies. Each hemisphere has at least one copy of every single memory.
At first glance, this explanation fits My Division quite well, but on closer examination it has an interesting consequence. If we accept (B), it follows that each hemisphere has a full set of qualitatively identical memories even before the split.[3] If we add to this “the same full range of abilities”, stipulated by Parfit, we get an organism that, for all intents and purposes, has two qualitatively identical brains with one stream of consciousness. Such an organism is certainly conceivable, even if spectacularly unlikely in evolutionary terms. However, it can be argued that the presence of two brains in Parfit’s body before the split turns the whole scenario on its head. Instead of one brain whose body is injured, we have two brains which are finally liberated from each other by a welcome surgical intervention. In a sense, this is a crude empirical version of David Lewis’s coinciding person-stages (Lewis, 1983). As such, it sits uneasily with Parfit’s agenda in My Division.
(C) Parfit’s memories are not localized. Every single memory is spread throughout the brain.
This explanation also seems compatible with My Division. However, just like (B), it becomes less attractive the more you probe it. It is possible to imagine a brain without any fixed neural correlates. Perhaps such a brain would work by generating some kind of mental cloud, where no particular mental event would ever be linked to any particular group of neurons. But there are at least two difficulties with this. Firstly, My Division requires that half a brain be just as good at supporting this cloud as an entire brain. Yet if each hemisphere is just as good at supporting the cloud as a whole brain, we run into the same problem as in (B) above: it appears that Parfit has two fully functional brains even before the split. Secondly – and perhaps more importantly – the idea of mental life that is not correlated with any particular part of the brain has a distinct non-reductionist flavour of a Cartesian Ego.
(D) Each hemisphere is scanned during the transplantation. All memories that are missing from one hemisphere are copied to the other, and vice versa.
Like (A), this scenario can conceivably involve a regular human brain, perhaps especially a young one. It also appears compatible with My Division. One problem I can see with (D) is that it makes My Division somewhat superfluous. Like Parfit’s teleportation cases, (D) involves scanning and recreating mental states from scratch on the basis of different physical material. If we accept (D), then My Division loses much of what makes it different from cases of the Branch-Line type (Parfit, pp. 200-201), where someone is replicated on Mars while the original remains alive on Earth.
In conclusion, it is time to revisit Williams’s claim that we know why the post-split persons in a My Division-type case have the memories they do. I have tried to show that accounting for this knowledge is less straightforward than might seem at first glance. Whichever conceivable explanation we choose for the equal transfer of memory across the split, any explanation appears to have significant consequences for the thought experiment. Though I have focused on episodic memories, the same reasoning can be extended to any phenomena of our inner life that have localized neural correlates (which might include most or even all of them). It might be worth bearing this in mind whenever we use My Division-type cases to argue for reductionism or for the relative unimportance of personal identity.[4]
Notes
[1] Parfit doesn’t discuss a scenario where Lefty and Righty take turns at being Parfit, presumably because it is only a minor variation on the One-or-the-Other option.
[2] Unless we include the abilities to retrieve specific memories in the full range of abilities.
[3] Here I ignore as uninteresting the possibility that one hemisphere might have more copies than the other.
[4] I originally planned to end the essay by considering a variation on My Division that would involve a more usual type of brain. My hope was to show that even a Regular Brain version of My Division could be used as evidence that personal identity is not what matters. I had to give up on this idea once I had got to 4000 words.
In case anyone cares to cite this essay
Zarubin, K. (2018). Parfit’s division: On split brains and personal identity. Unpublished manuscript. Retrieved from https://kostia.me/english-deutsch/patfits-division/
References
Noonan, H., & Curtis, B. (2017). Identity. In The Stanford Encyclopedia of Philosophy. (Spring 2017 Edition). Retrieved from https://plato.stanford.edu/archives/spr2017/entries/identity.
Lewis, D. (1983). Survival and Identity. Philosophical Papers, volume I. Retrieved from http://www.oxfordscholarship.com/view/10.1093/0195032047.001.0001/acprof-9780195032048-chapter-5.
Parfit, D. (1987). Reasons and persons. Oxford: Oxford University Press.
Wiggins, D. (1967). Identity and spatio-temporal continuity. Oxford: Basil Blackwell.
КОНСТАНТИН ЗАРУБИН 