Utilitarianism Wins Outright Part 44: The Replacement Argument

Against critics of utilitarian replacement

Some people aren’t very happy with the fact that utilitarianism says that there’s nothing intrinsically bad with replacing people with other equally happy people. Here, I will defend that conclusion.

Kate Sills @kate_sills

Did anyone else see this insane exchange? conversationswithtyler.com/episodes/sam-b…

To see one example of the challenge, Knutsson (2021, p.5) raises the following objection based on utilitarainism favoring replacing people.    

Suboptimal Paradise: The world has become a paradise with no suffering. Someone can kill everyone in this paradise and replace them with beings with (possibly only slightly) more well-being in total. Traditional utilitarianism implies that it would be right to do so.12

This does seem a little bit counterintuitive to some people.  However, it is entailed by several plausible principles.  

Objection 1: The Iteration Objection 

Suppose we continually iterate the process of destroying the world and replacing it with people with higher well-being.  If we accept 

Weak Replacement: There is some level of well-being that a world could have such that it would be worth replacing the world with people with that level of well-being.  

This seems really intuitive -- killing everyone to bring about some vast number of people who get, at each moment, more pleasure than has been experienced in human history would seem to be a good thing.  However, the person who finds the replaceabiltiy argument persuasive would have to deny this, if they accept the following.  

Iterative replacement: If replacing people with others with slightly higher well-being is wrong, doing this a vast number of times would be bad.  

Iterative replacement likewise seems plausible; it would be strange to hold that replacement is really bad -- but if you do it a bunch of times, then it stops being bad. 

If this were done an infinite number of times, that would seem to be infinitely bad, if we accept replacement is bad.  However, this then commits the defender of the replacement argument to the notion that horrifically torturing everyone to death would be less bad than an arbitrarily large amount of replacement -- this would be so no matter how extreme their pre-death suffering was.  

Objection 2: Death 

The person who holds the view that replacement is bad seems committed to some rather puzzling implications about death and birth.  The following principle is very plausible.  

Improvement over time: In a society where, after about 100 years people die, but new babies are born, where average welfare improves dramatically does not get worse over time. 

This principle is hard to deny.  If we accept that in our world average quality of life is steadily improving, it seems hard to deny that our world is getting better.  This is true despite the fact that, over time, people die and new people are born.  However, the person who rejects replacement seems committed to denying improvement over time.  After all, they think that if a person dies, but then a new person is born with a higher quality of life, that is actively bad.  

If a person dies, but then a new baby is born, the combination of those two things would not seem to be a bad thing.  However, the person who rejects replaceability would be committed to denying that.  

In fact, on this account, the world could be bad despite no one being badly off.  Suppose that we started with a world of people with low utility and then we replace them with other people with low utility.  On this account, despite no one suffering at all, this world would be bad to create, even though every person is rational to want to be in it.  This is not plausible. 

Objection 3: There is no personal identity 

Parfit (1984) has argued persuasively that there is no irreducible personal identity.  He explains the view as follows (p.212-213)

(9) Though persons exist, we could give a complete description of reality without claiming that persons exist. 

I call this the view that a complete description could be impersonal. 

This view may also seem to be self-contradictory. If persons exist, and a description of what exists fails to mention persons, how can this description be complete? 

A Reductionist could give the following reply. Suppose that an object has two names. This is true of the planet that is called both Venus and the Evening Star. In our description of what exists, we could claim that Venus exists. Our description could then be complete even though we do not claim that the Evening Star exists. We need not make this claim because, using its other name, we have already claimed that this object exists. 

A similar claim applies when some fact can be described in two ways. Some Reductionists accept (4), the claim that a person just is a particular brain and body, and a series of interrelated physical and mental events. If this is what a person is, we can describe this fact by claiming either 

(10) that there exists a particular brain and body, and a particular series of interrelated physical and mental events. 

or 

(11) that a particular person exists, 

If (10) and (11) are two ways of describing the same fact, a complete description need not make both claims. It would be enough to make claim (10). Though this person exists, a complete description need not claim that he exists, since this fact has already been reported in claim (10). 

Other Reductionists accept (5), the claim that a person is distinct from his brain and body, and his acts, thoughts, and other physical and mental events. On this version of Reductionism, claim (10) does not describe the very same fact that claim (11) describes. But claim (10) may imply claim (11). More cautiously, given our understanding of the concept of a person, if we know that (10) is true, we shall know that (11) is true. These Reductionists can say that, if our description of reality either states or implies, or enables us to know about, the existence of everything that exists, our description is complete. This claim is not as clearly true as the claim that a complete description need not give two descriptions of the same fact. But this claim seems plausible. If it is justified, and the Reductionist View is true, these Reductionists can completely describe reality without claiming that persons exist.11 

My claims about Reductionism draw distinctions that, in this abstract form, are hard to grasp. But there are other ways of discovering whether we are Reductionists in our view about some kind of thing. If we accept a Reductionist View, we shall believe that the identity of such a thing may be, in a quite unpuzzling way, indeterminate. If we do not believe this, we are probably Non-Reductionists about this kind of thing.

Consider, for example, clubs. Suppose that a certain club exists for several years, holding regular meetings. The meetings then cease. Some years later, some of the members of this club form a club with the same name, and the same rules. We ask: ‘Have these people reconvened the very same club? Or have they merely started up another club, which is exactly similar?’ There might be an answer to this question. The original club might have had a rule explaining how, after such a period of non-existence, it could be reconvened. Or it might have had a rule preventing this. But suppose that there is no such rule, and no legal facts, supporting either answer to our question. And suppose that the people involved, if they asked our question, would not give it an answer. There would then be no answer to our question. The claim ‘This is the same club’ would be neither true nor false.

Though there is no answer to our question, there may be nothing that we do not know. This is because the existence of a club is not separate from the existence of its members, acting together in certain ways. The continued existence of a club just involves its members having meetings, that are conducted according to the club's rules. If we know all the facts about how people held meetings, and about the club's rules, we know everything there is to know. This is why we would not be puzzled when we cannot answer the question, ‘Is this the very same club?’ We would not be puzzled because, even without answering this question, we can know everything about what happened. If this is true of some question, I call this question empty.

If there is no irreducible personal identity, then it seems that the replacibility argument loses its force.  After all, every moment of every day, I am being replaced by someone else.  I can only hope that the one replacing me has a higher level of well-being.  It does not seem that as time passes and I undergo both neural and physical changes, as cells die and are reborn, I become apprecibly worse off, or that something bad is happening.  

Providing an in-depth analysis of personal identity would be beyond the scope of this book.  Here, I shall provide only a brief rundown of some of the main arguments against personal identity being non-reducible.  

Spectrum Arguments 

Parfit (p.231) provides one argument that purports to show that there is not irreducible personal identity.  

Williams discusses a single case in which, after a few changes, there will be no psychological continuity. I shall discuss a spectrum, or range of cases, each of which is very similar to its neighbours. These cases involve all of the possible degrees of psychological connectedness. I call this the Psychological Spectrum. 

In the case at the far end, the surgeon would cause very many switches to be simultaneously flipped. This would cause there to be no psychological connections between me and the resulting person. This person would be wholly like Napoleon. 

In the cases at the near end, the surgeon would cause to be flipped only a few switches. If he flipped only the first switch, this would merely cause me to lose a few memories, and to have a few apparent memories that fit the life of Napoleon. If he flipped the first two switches, I would merely lose a few more memories, and have a few more of these new apparent memories. Only if he flipped all of the switches would I lose all my memories, and have a complete set of Napoleonic delusions. 

Similar claims are true about the changes in my character. Any particular switch would cause only a small change. Thus, if I am to be like Napoleon, I must become more bad-tempered, and must cease to be upset by the sight of people being killed. These would be the only changes produced by the flipping of the first two switches. 

In this revised version of the argument, which involves very many different cases, we must decide which are the cases in which I would survive. In the case at the near end, the surgeon does nothing. In the second case, I would merely lose a few memories, have a few delusions, and become more bad-tempered. It is clear that, in this case, I would survive. In the third case, the changes would be only slightly greater. And this is true of any two neighbouring cases in this range. It is hard to believe both that I would survive in one of these cases, and that, in the next case, I would cease to exist. Whether I continue to exist cannot be plausibly thought to depend on whether I would lose just a few more memories, and have a few more delusory memories, and have my character changed in some small way. If no such small change could cause me to cease to exist, I would continue to exist in all of these cases. I would continue to exist even in the case at the far end of this spectrum. In this case, between me now and the resulting person, there would be no psychological connections.  (1984)

The person who denies reductionism about personal identity -- who thinks the boundaries of personal identity are narrow and precise -- would have to think that there’s some robust fact of the matter about when you stop being yourself.  However, this seems totally implausible -- how could there be a precise point at which you shed your personal identity.  

Personal identity seems to require transitivity -- if I’m the same person as my four-year-old self, and my four-year-old self is the same person as my one-year-old self, then I must be the same person as my one-year-old self.  However, it seems like the following proposition is true 

Weak Insensitivity to Flips: For any person, you are at N flips, you are not a different person at N+1 flips.  

However, if we accept weak insensitivity to flips and transitivity we get the conclusion that even if you are fully replaced by Napoleon, you don’t become a different person.  

Each moment, cells in your brain are dying and being replaced.  You have very little in common with your four year old self.  I arguably have more in common with a person that’s a lot like me than I do with my four year old self.  Thus, it seems hard to imagine that there’s some continuous thread that unites me with my four year old self.  

Parfit (1984, p.236-237)  provides a similar, yet even more troubling objection to irreducible personal identity 

Consider another range of possible cases. These involve all of the possible variations in the degrees of both physical and psychological connectedness. This is the Combined Spectrum. 

At the near end of this spectrum is the normal case in which a future person would be fully continuous with me as I am now, both physically and psychologically. This person would be me in just the way that, in my actual life, it will be me who wakes up tomorrow. At the far end of this spectrum the resulting person would have no continuity with me as I am now, either physically or psychologically. In this case the scientists would destroy my brain and body, and then create, out of new organic matter, a perfect Replica of someone else. Let us suppose this person to be, not Napoleon, but Greta Garbo. We can suppose that, when Garbo was 30, a group of scientists recorded the states of all the cells in her brain and body. 

In the first case in this spectrum, at the near end, nothing would be done. In the second case, a few of the cells in my brain and body would be replaced. The new cells would not be exact duplicates. As a result, there would be somewhat less psychological connectedness between me and the person who wakes up. This person would not have all of my memories, and his character would be in one way unlike mine. He would have some apparent memories of Greta Garbo's life, and have one of Garbo's characteristics. Unlike me, he would enjoy acting. His body would also be in one way less like mine, and more like Garbo's. His eyes would be more like Garbo's eyes. Further along the spectrum, a larger percentage of my cells would be replaced, again with dissimilar cells. The resulting person would be in fewer ways psychologically connected with me, and in more ways connected with Garbo, as she was at the age of 30. And there would be similar changes in this person's body. Near the far end, most of my cells would be replaced with dissimilar cells. The person who wakes up would have only a few of the cells in my original brain and body, and between her and me there would be only a few psychological connections. She would have a few apparent memories that fit my past, and a few of my habits and desires. But in every other way she would be, both physically and psychologically, just like Greta Garbo.

These cases provide, I believe, a strong argument for the Reductionist View. The argument again assumes that our psychological features depend upon the states of our brains. Suppose that the cause of psychological continuity was not the continued existence of the brain, but the continued existence of a separately existing entity, like a Cartesian Ego. We could then claim that, if we carried out such operations, the results would not be as I have described them. We would find that, if we replaced much of someone's brain, even with dissimilar cells, the resulting person would be exactly like the original person. But there would be some critical percentage, or some critical part of the brain, whose replacement would utterly destroy psychological continuity. In one of the cases in this range, the carrier of continuity would cease either to exist, or to interact with this brain. The resulting person would be psychologically totally unlike the original person.

Split Brain Cases 

Parfit provides another argument that appeals to split brain cases in which parts of a brain is in two different skulls,

My Division. My body is fatally injured, as are the brains of my two brothers. My brain is divided, and each half is successfully transplanted into the body of one of my brothers. Each of the resulting people believes that he is me, seems to remember living my life, has my character, and is in every other way psychologically continuous with me. And he has a body that is very like mine.  (1984, p.254)

Later Pafit makes another, similar claim, criticizing alternative views, expanding on his earlier claim about split brain cases.  

It is natural to believe that our identity is what matters. Reconsider the Branch-Line Case, where I have talked to my Replica on Mars, and am about to die. Suppose we believe that I and my Replica are different people. It is then natural to assume that my prospect is almost as bad as ordinary death. In a few days, there will be no one living who will be me. It is natural to assume that this is what matters. In discussing My Division, I shall start by making this assumption. 

In this case, each half of my brain will be successfully transplanted into the very similar body of one of my two brothers. Both of the resulting people will be fully psychologically continuous with me, as I am now. What happens to me?

There are only four possibilities: (1) I do not survive; (2) I survive as one of the two people; (3) I survive as the other; (4) I survive as both.

The objection to (1) is this. I would survive if my brain was successfully transplanted. And people have in fact survived with half their brains destroyed. Given these facts, it seems clear that I would survive if half my brain was successfully transplanted, and the other half was destroyed. So how could I fail to survive if the other half was also successfully transplanted? How could a double success be a failure?

“Consider the next two possibilities. Perhaps one success is the maximum score. Perhaps I shall be one of the two resulting people. The objection here is that, in this case, each half of my brain is exactly similar, and so, to start with, is each resulting person. Given these facts, how can I survive as only one of the two people? What can make me one of them rather than the other?

These three possibilities cannot be dismissed as incoherent. We can understand them. But, while we assume that identity is what matters, (1) is not plausible. My Division would not be as bad as death. Nor are (2) and (3) plausible. There remains the fourth possibility: that I survive as both of the resulting people.

This possibility might be described in several ways. I might first claim: ‘What we have called “the two resulting people” are not two people. They are one person. I do survive this operation. Its effect is to give me two bodies, and a divided mind.’ 

This claim cannot be dismissed outright. As I argued, we ought to admit as possible that a person could have a divided mind. If this is possible, each half of my divided mind might control its own body. But though this description of the case cannot be rejected as inconceivable, it involves a great distortion in our concept of a person. In my imagined Physics Exam I claimed that this case involved only one person. There were two features of the case that made this plausible. The divided mind was soon reunited, and there was only one body. If a mind was permanently divided, and its halves developed in different ways, it would become less plausible to claim that the case involves only one person. (Remember the actual patient who complained that, when he embraced his wife, his left hand pushed her away.)

The case of complete division, where there are also two bodies, seems to be a long way over the borderline. After I have had this operation, the two ‘products’ each have all of the features of a person. They could live at opposite ends of the Earth. Suppose that they have poor memories, and that their appearance changes in different ways. After many years, they might meet again, and fail even to recognise each other. We might have to claim of such a pair, innocently playing tennis: ‘What you see out there is a single person, playing tennis with himself. In each half of his mind he mistakenly believes that he is playing tennis with someone else.’ If we are not yet Reductionists, we believe that there is one true answer to the question whether these two tennis-players are a single person. Given what we mean by ‘person’, the answer must be No. It cannot be true that what I believe to be a stranger, standing there behind the net, is in fact another part of myself.

Suppose we admit that the two ‘products’ are, as they seem to be, two different people. Could we still claim that I survive as both? There is another way in which we could. I might say: ‘I survive the operation as two different people. They can be different people, and yet be me, in the way in which the Pope's three crowns together form one crown.’ 41

This claim is also coherent. But it again greatly distorts the concept of a person. We are happy to agree that the Pope's three crowns, when put together, are a fourth crown. But it is hard to think of two people as, together, being a third person. Suppose the resulting people fight a duel. Are there three people fighting, one on each side, and one on both? And suppose one of the bullets kills. Are there two acts, one murder and one suicide? How many people are left alive? One or two? The composite third person has no separate mental life. It is hard to believe that there really would be such a third person. Instead of saying that the resulting people together constitute me—so that the pair is a trio—it is better to treat them as a pair, and describe their relation to me in a simpler way.  (1984, p.254)

Biases and Debunking 

There are lots of features that bias our assessment of the scenario, making our intuitions likely to be off about it.  As Greenberg (1983) notes, our moral intuitions are skewed to favor things that are good for us.  It’s no surprise that the existing wouldn’t want to be replaced by the non-existing -- after all, it would be worse for the existing.  Thus, the intuitions of those of us who exist would be expected to be skewed.  If we polled the non-existent who would replace us, they’d undoubtably have different intuitions. 

Additionally, we know that morality is affected greatly by empathy.  But it’s hard to empathize with the non-existent.  

Additionally, as we saw in the section providing an Objective List Theory of wantworthiness, we plausibly ought to have aims that conflict with replacement even if replacement is good, all else equal.  One ought to value their friends and loved ones and thus not want them to be replaced, even if their replacement would make the world marginally better.  

Additionally, our intuitions are affected by status quo bias.  As we’ve already seen, people are biased toward keeping things as they are.  By definition, replacing everyone is not keeping things as they are.  Thus, we’d expect people to be biased against it.

Conclusion

Thus, this objection, while at first seeming to have force, loses all of it when carefully examined. It, like many of the putative counterexamples to utilitarianism shows that when we carefully reflect, our reflective judgment converges with utilitarianism. Thus, in the arsenal of the critic of utilitarianism, the replacement argument should be replaced.