Doing the Best We Can: A Dialogue on the Nonidentity Problem, Harm and Future People

Two first-year undergraduates, Shelby and Rebecca, meet at the dining hall for dinner. There, they run into Jeremy, a second-year graduate student in philosophy who also happens to be a teaching assistant for the introductory ethics course Shelby and Rebecca are taking.

[An argument to the conclusion that meat-eating does not harm animals is introduced: the animals are already dead.]

Shelby: You’re only having salad for dinner? You’re so skinny – you can’t be on a diet!

Jeremy: No, I’m not trying to lose weight. But the ‘mystery stew’ they’re pushing tonight does not look at all appetizing to me. Anyway, I’m a vegan. I don’t believe in torturing animals.

Shelby: Hey, I don’t believe in torturing animals, either. But I don’t eat live animals. I only eat dead ones. You can’t torture something that’s dead. Ergo, there’s nothing wrong with eating meat.

Jeremy: (Blank stare.) What about you, Rebecca? Do you know what a ‘sophist’ is?

Rebecca: Well, I guess it has something to do with putting together arguments your philosophy T.A. doesn’t like very much.

Jeremy: You are both highly amusing. Truly funny. Have a good evening. (He turns toward a group of graduate students).

Jeremy: (Turning back.) No. I wouldn’t want your mystery stew to get cold. But I will ask the two of you to define ‘sophist’ for the entire section tomorrow morning. So: be prepared. Remember, the power to grade is the power to destroy. (Walks off.)

Jeremy’s Wednesday morning section meeting is in progress. Mark is still another first-year student.

[Utilitarianism is introduced; it is noted that, according to the utilitarian, the determination of whether an act creates the most happiness is to be made on the basis of how much happiness is created not just for human beings but for ‘all sentient creation’; reprisal of Shelby’s argument that meat-eating does not harm animals.]

Jeremy: Now, we’ve already seen that utilitarians, in general, think that agents ought to create the most good that they can. The moral permissibility of a given act is to be assessed, in other words, by determining whether the consequences of that act are as good, or better, than the consequences of any alternative act the agent might have performed instead.

Before we can actually apply this principle, we need to understand what ‘the good’ is – what makes one set of consequences better than another. Different philosophers have different visions of the good. But for purposes of class today we’re interested in John Stuart Mill. According to Mill, creating the most good that one can is a matter of creating the most happiness that one can. To be more specific: it’s a matter of creating the most ‘general happiness’ that one can. In other words, for Mill, morality requires not that we work toward increasing happiness for ourselves or the members of our own family or community but rather that we work to increase happiness impartially, for the ‘aggregate of all persons.’ Utilitarianism, Chapter IV, Paragraph 3.

My question for you guys is this. I’m a moral agent, trying to figure out the right thing to do. According to Mill, I, as a moral agent, should be thinking about creating happiness for ‘all persons.’ What does Mill mean by ‘all persons’ here? Mark, you tell us.

Mark: Well, you just said it. We ought to create happiness, according to Mill, for all persons, not just persons we know and love.

Jeremy: Let me ask the question another way. You remember Kant from two weeks ago, right? Now, Mill thinks that our obligations are in respect of ‘all persons.’ And Kant thinks that, strictly speaking, our obligations run only to those beings who themselves have the ability to think and reason and the capacity to conduct themselves in accordance with moral law. Are you saying, Mark, that the views of the great utilitarian, Mill, and the great deontologist, Kant, after all converge –on at least this one point?

Mark: I didn’t say that. Well, maybe I did. It depends on what you mean by ‘person.’ But Mill can’t really be saying in Chapter IV that we have obligations only to rational beings or only to individuals who themselves have the ability to act morally. I mean, you made a big deal, back in Chapter II, about how Mill thought happiness was something that we should try to create not just for people but for animals as well.

Jeremy: Not just for human beings but for non-human animals as well. In fact, for the ‘whole sentient creation.’ If you can experience pleasure and desire pleasure, and experience pain and desire its absence, then you matter, morally, for Mill. You are, for Mill, ‘on the list,’ and the agent has to think about creating happiness for you as well as for himself.[2]

Jeremy: O.K., for herself. Whereas for Kant – well, Kant doesn’t really think that non-human animals matter morally at all, except in the indirect sense that our treating non-human animals well keeps us in practice for treating those rational beings who really do count well.

So, Mark, let’s take it as a given that Mill isn’t contradicting himself. What, then, does he mean by ‘person’ by the time he gets to Chapter IV?

Mark: Well, ‘person’ will have to include non-human animals. Not all non-human animals. Not viruses, and probably not even insects. But the term could just mean any sentient creature – any animal, whether human or not, capable of feeling pleasure and pain.

Jeremy: Exactly! One other point. It’s not just the quantity of pleasure that Mill thinks is important. It’s the quality of that pleasure as well. The pleasure we take in Mozart gets weighted more heavily in Mill’s utilitarian metric than the swine’s pleasure in the slop that he is fed. Sorry, that she is fed.

Shelby: That he or she is fed. But, Jeremy, aren’t you being somewhat sophistic? I mean, Mill doesn’t actually say what you just said. Maybe the swine is capable of experiencing higher and lower pleasures as well, and maybe the swine’s higher pleasures count just as much as the pleasure that you say you take in Mozart.

Jeremy: Ah, Shelby, thank you for reminding me of that item of business! Tell us, what’s a ‘sophist’?

Shelby: I thought you’d never ask. A ‘sophist’ is someone who has a talent for cleverly making the lesser argument look like the better argument and the better argument look like the lesser.

Mark: There wasn’t really any sophistry. You were giving us one way of interpreting Mill’s idea that quality and quantity are both important in determining what act leads to the greatest happiness. It’s true there are other ways of interpreting Mill on that point. But your interpretation wasn’t, in itself, sophistic.

Rebecca: Right. You weren’t doing anything really tricky. You weren’t exactly blinding us with some clever logical or linguistic flourish to get us to buy into some ridiculous result that you’d like us to accept whatever its truth.

Jeremy: Gee, and I thought everything I said blinded with its brilliance. But thank you, Rebecca and Mark, for your dispensation. It is not my aim in life to be a sophist. So where were we? Putting aside exactly how interspecies comparisons between kinds of pleasure are to be made, we do seem to have reached a solid reading of Mill. For Mill, agents ought to create the most happiness they can for persons, where the term ‘happiness’ is understood to be determined by reference to both the quality and the quantity of the experienced pleasure and the term ‘person’ is interpreted to include at least many non-human animals. Oh – and by the way – to exclude at least some human animals.

The important point here is that, as Singer would put it, Mill is no ‘speciesist’: most human beings, though not all, as well as at least many non-human animals have moral status, on Mill’s view. Write that down.

Shelby: If persons, including some animals, have moral status, and if they have moral status because they can feel pleasure and pain, and if you can’t feel pleasure and pain once you are dead, then anything that is dead, including dead animals, has no moral status. And that means – well, you know what that means.

Jeremy: Everyone see Shelby’s idea? She’s trying to argue that at least one reason many people give for avoiding the consumption of meat and other animal products is in fact not entirely cogent. Your reason for not eating ham cannot be that you are concerned with the happiness levels of the dead pig. I mean, we vegans and vegetarians may still have other moral reasons for how we eat. We might not eat meat for the sake of our own health, or to save money that we can then give to OXFAM. But if Shelby’s right, we cannot cogently cite concern for the animal itself as a reason not to eat meat.

So, look, guys, we’ve got to stop, but think about Shelby’s argument – and why it fails – for Friday. And read Chapter V of Utilitarianism, where Mill explains why his utilitarianism is not completely at odds with principles of justice.

Just the students, back at the dorm, having demitasse in the living room after dinner.

[The claim is made that, while eating meat may not harm the animal one is then eating, it does harm, at least indirectly, future animals by virtue of the support and encouragement it gives to the industry of factory-farming.]

Mark: You know, Shelby, you seemed a little testy in class. You might want to watch that.

Shelby: Me, testy, Mark? You’re out of your mind. Jeremy may have been testy, but not me. All I’m doing is trying to figure out how these moral theories are supposed to apply to my life. I mean, the categorical imperative I never did get. I thought Kant was saying something like we ought to live by whatever rules – no, sorry, maxims – would lead to the best results for everyone assuming everyone followed them. But Jeremy said that was not it at all –– and I got a D on the paper. I never could figure out what those maxims were supposed to be. ‘Don’t lie.’ ‘Don’t kill.’ Aren’t there always going to be exceptions? Jeremy tried to explain that the specific maxims themselves were not really that important for Kant, and that the main test of morality was whether the agent does whatever he or she does rationally and out of duty. That just confused me more. Anyway, Mill makes more sense to me. He’s proposing one simple principle that we don’t have to read a lot of exceptions into – and that tells us to aim for something I, for one, think is really important. Happiness – the ‘greatest happiness for the greatest number.’ I was born to be a utilitarian.

Mark: Rebecca’s right. About that hamburger you had for dinner tonight. You obviously didn’t cause the cow that went into that hamburger any harm. But you did do your little bit to sustain one of the nastiest industries around – factory farming. Millions of animals live short and miserable lives and then die very nasty deaths every single year. You know. You’ve seen the PETA films. And there’s no good reason for it. It’s not like meat-eating represents some huge gain for the human race. Just the reverse. It’s an incredibly expensive and inefficient way to get the nutrients we need, and our hearts would be healthier and cholesterol levels lower if we just gave up meat.

Shelby: Don’t be a hypocrite. You ate a hamburger for dinner tonight, too, Mark.

Mark: Hey, I’m not even a utilitarian! I’m just trying to tell you what’s wrong with your argument.

Mark: You college women have the most glorified vocabularies. Look, I can’t sit around here talking all night about Mill. Major physics test tomorrow. Gotta go.

Rebecca: He’s completely forgotten I’m taking that course, too, and need to get to the library myself. See you tomorrow.

Professor Richardson and Jeremy discussing how the introductory ethics course is going in Richardson’s office.

[The nonidentity problem, or ‘NIP,’ is introduced as an objection against the claim that meat-eating harms future animals. According to the NIP, an act that brings a new person into existence does not harm that person even if it causes that person to suffer, so long as (1) the suffering is not too extreme and (2) the agent did not have a way to make things better for that person. Kavka’s slave child example is introduced. It is described how the ‘precariousness’ of existence makes it hard for agents to make things better for any particular future person, and why agents, if they do try to make things better, are likely instead simply to consign the one person to nonexistence and bring some ‘nonidentical’ person into existence instead.]

Richardson: So how do you think our students are doing, Jeremy? Are they getting it? Are they engaged? Do they like my lectures? Do they like my jokes?

Jeremy: It’s going well. Really. I’ve never T.A.’ed before, of course. But the students are talking a lot in the section meetings, and they seem to be understanding your lectures. I mean, they weren’t thrilled with their grades on the first paper. But you’d said the average grade should be in the C-/D+ range, so . . . .

Richardson: I said that? C-? D+?! I didn’t realize I had such high standards!

Jeremy: Oh. Well, uh, anyway. I thought you wanted, like, a C- average, so that’s the average the students in my section ended up with.

Jeremy: Yeah. Uh. A few. They argued that your other T.A.’s graded much more leniently than I did. To get a C- average, do you realize how many Ds and Fs I had to give out for each A or B? Maybe your other T.A.’s did grade more leniently. Maybe I misunderstood what you said.

Richardson: Who knows? I don’t remember what I said! But look – you’re the T.A. helping them achieve their fullest potential!

Jeremy: Hmmm. Well, anyway, at the last section meeting, we veered off into a discussion of the ethics of meat-eating.

Jeremy: I was actually a little surprised to find that some of them really are un-reconstructed meat-eaters. One first-year, in particular, this woman, keeps pushing the idea that you can’t harm animals by eating them since, by then, they’re already dead.

Jeremy: Yeah! Obviously, her eating meat doesn’t harm the one animal she happens to be eating at the time. But it hardly follows that it doesn’t harm others.

Richardson: Absolutely. Our moral calculation – whatever it may be – must take into account the plights of future animals as well as already-existing animals. Breeding cattle, or chickens, or whatever, for the purpose of factory farming, where their lives will be miserable and they may be doomed from the start to a horrifying death, all on the venal expectation that your student and others like her will buy the next sirloin steak and the next and the next – well, it’s wrong. And your student helps create and sustain that expectation, of course, with every purchase of meat and cheese and eggs that she makes.

Richardson: Of course it is. Except, of course, for the nonidentity problem. You don’t want to overlook that.

Jeremy: Hmmm. The nonidentity problem. Well, I read Reasons and Persons when I was an undergraduate. Wasn’t Parfit’s idea that when you do something that brings someone into an existence that is defective in some way you haven’t really harmed that person? Because it’s better to have that defective life, even if it involves suffering, than it is never to have existed at all?

Richardson: Well, that’s the basic idea. That’s part of it, anyway. It’s better to be slaughtered early in life and eaten than it is never to exist at all. But, before you can get to Parfit’s ‘no harm done’ result, there’s second point you need, too. It’s got to be the case that, for whatever reason, the agent did not have a way to confer on that same person some still better existence. I mean, if the agent could just snap his fingers and guarantee that the new person would exist and not suffer, then the agent harms the person when he fails to do that. At least, that’s my view. But then of course I’m a maximizing utilitarian. For me, harm’s not a matter of going around and making people absolutely miserable. Instead, harm is a matter of doing less for them than you can. Though no one agrees with me on this, of course, anymore! Everyone’s so Kantian these days!

Jeremy: So, we don’t have a way to confer a better existence on the animals who have already been slaughtered. But don’t we have a way to make things better for the ones who haven’t been slaughtered yet?

Richardson: Ah! You have to look at the bigger picture to get the full thrust of Parfit’s argument! The point is this: acts of meat-eating today, of course, cause future animals to suffer because they provide an incentive for factory farming, which in turn imposes a lot of suffering on animals. But those same acts of meat-eating also provide an incentive for the factory farmers to bring those same animals into existence to begin with. The very act that imposes some pain on the animal also confers on the animal the gift of life itself! Get rid of the one effect, and you get rid of the other effect as well! So – unless their lives are so miserable that they’d be better off never having existed at all – our eating meat today doesn’t, after all, harm the animals of tomorrow. If anything, it makes them better off than they would have been. Which would have been never having existed at all!

Meat-eating is, in its own way, a procreative act! As such, the choice to eat meat is not a bad thing for those future animals – a harm – but rather a profoundly good thing – a benefit.

In fact, the nonidentity problem shows us that many perfectly ordinary choices turn out to be, on close inspection, procreative in nature. It’s not just meat-eating. You know Kavka, yes?

Richardson: Ah – you should know Kavka! His slave child example is brilliant on this point! Consider a man and a woman who’d like to buy a yacht but aren’t interested in having children. At least not together. At least not now. Along comes a wealthy man who offers to give them the $50,000 they need to purchase their yacht – was that enough for a yacht, even in 1982? – in exchange for their producing a child who, at birth, would be turned over to the wealthy man as a slave. Now, think about this transaction from the child’s point of view. It’s a damned good thing that this rich guy has come along! Had there been no slave child contract, there’d be no child. And it’s better to exist as a slave than never to exist at all. Take away what you allege to be ‘harm,’ here, and all you manage to do is take away the child’s very existence!

Jeremy: But now wait a minute. You said before that it’s not enough just to produce some good for people. You said that when agents could do more for a given person, they harm that person when they instead do less. But surely ‘less’ is just what the couple has done in the slave child case. After all, they had the alternative of creating a lot more happiness for the child than they have. They could have simply not signed the contract and then produced the child. Maybe they wouldn’t have wanted to do that, and maybe they wouldn’t have done that, but the alternative existed for them. So haven’t they harmed the child?

Richardson: (Laughing.) Oh, so you think they could have ‘simply’ not signed the contract and produced that child? There’s nothing ‘simple’ about that at all! Think about it, Jeremy! You say the couple could have created still more happiness for the slave child they in fact have by not entering into the contract and yet still producing a child. What, exactly, do you think the chances are that they’d have managed to produce exactly that same child but now without the contract in place? What are the odds that they could have produced exactly that same child, but now as a non-slave? . . . . Actually – as your comment suggests – the non-slave part is easy. Obviously the couple they had the alternative of not entering into the contract and proceeding to produce some child or another. If they could have produced one child as a slave, they certainly could have produced some other child as a non-slave. The hard part is producing the very same child who now exists as a slave.

If this isn’t clear, just try to imagine that, contrary to fact, the couple had not entered into the contract and then tried with all their might to produce that very same child What are their chances of succeeding? Practically none at all, wouldn’t you say? ‘Simple’? I think not!

In the case of future persons, existence remains highly ‘precarious,’ as Kavka put it, until very late in the game. What he meant was that much has to go exactly as it did, in that long and detailed sequence of acts and events that ends in any one person’s coming into existence, for that one person ever to make it into existence at all. Had your grandparents never met, you would never have existed! Had your parents conceived a child one month earlier rather than one month later, you would never have existed! Had your conception taken place one hour, or even one minute, earlier than it did, or under even slightly different circumstances, you would never have existed! There are just so many gametes floating around out there! So many sperm cells vying to inseminate any given egg! A hundred million sperm a day produced on average by the healthy sexually active man! The least little variation in the causal sequence that ends with any one person’s coming into existence can so easily mean that a different sperm cell does the inseminating – and thus can so easily take that one person off-track for existence altogether! When we think about it, it’s amazing that any of us ever managed to make our way into existence at all!

But if even a minor variation in the causal sequence is very likely to take someone off-track for existence altogether, surely the couple’s not signing the contract would have taken the slave child off-track for existence. That’s a seismic shift! Had they not entered into the contract, the causal sequence would have been thrown off in a thousand ways! In the very best case, where the couple tries their mightiest to produce that child as a nonslave, all they’d have done is end up producing some distinct, ‘nonidentical’ child in place of the one, and the slave child who actually exists and suffers would never have made it into existence at all! Ha!

Richardson: Marvelous, isn’t it? Entering into the slave child contract doesn’t, after all, harm the slave child. But the same holds for the practice of meat-eating and factory farming and all those future animals you vegans are so worried about. Suppose we all abandon meat-eating, as you suggest, for the sake of the future animals who otherwise will be subjected to the miseries of factory farming. We stop eating meat and cheese and eggs, and we stop drinking milk, because we don’t want to be support an industry that does so little for the cows and chickens and pigs that it produces –

Richardson: – and instead the ‘tofu roast’ becomes our Sunday dinner staple. Now imagine that, contrary to fact, we went about trying to create a better alternative for at least some of those future animals that you so badly want to rescue from the miseries of factory farming. We could certainly collectively afford to pay the factory farmers – at least for awhile – to continue breeding and raising animals and then to care for those animals humanely and allow them to live out their own natural lives. Or we could buy live animals from those farmers – instead of dead meat – and then take care of those animals as our pets or at least let them live as pleasant ornaments for our front lawns or our woodsy compounds.

Jeremy: – I get it now. (Holds up hand.) We could work hard to produce animals who no doubt would be better off than the animals who will exist and suffer if things are left as they are and people continue to eat meat and create incentives for factory farming. But those better-off animals very, very likely would be ‘nonidentical’ to the animals we feel we ought to save from the system of factory farming. The chances of our managing to bring into existence even one of those future animals for are practically nil. So there really isn’t anything after all that we can do to confer a better existence on those very same ‘identical’ animals that will we know will suffer under a system of factory farming.

Richardson: Exactly! Think about it from the cow’s point of view: is it really better never to exist at all than it is to exist and suffer the degradations of the factory-farmed animal? Plausibly, it is not! But the option of existing and not suffering those degradations does not, in any real sense, exist for that cow. Even if agents tried to rescue her – by not eating meat now yet still working to produce that very same cow – they’d, very, very likely, fail. She’d simply never exist at all!

Richardson: Look, I’ve got to get back to my desk. Yale’s putting a lot of pressure on me to get this new book out. But I’ll take a stab putting this whole nonidentity thing together for the next common lecture. See you there then, right? This is fun stuff, yes? Right? Great stuff, isn’t it?

[Cases involving already-existing animals and future animals are distinguished. It is observed that, where the apparent victim already exists at the time the questionable act is performed, the nonidentity problem has no application. Notwithstanding the NIP, then, we can discern many avoidable harms imposed on already-existing animals. (The NIP’s claim that acts performed today cannot harm future animals – or future persons – is, for the moment, left unchallenged.)

The question is raised whether, in some cases involving great suffering, never having existed at all may be better for the person who suffers than existence is.]

Shelby: What a sophist Richardson is. I love that Boston accent. But there is something off about that nonidentity argument.

Jeremy: Why do you say that, Shelby? ‘That nonidentity argument’ seems to get you exactly the conclusion that you seemed to want – that meat-eating, on inspection, does not harm the animals but rather, if anything, benefits them. That result gives you a basis on which to argue that, after all, there’s nothing wrong with eating meat. But now you’re saying you’re not going for it?

Shelby: Well, I like the conclusion. I just think there’s something weird about the argument.

Mark: But, Shelby, all the really interesting and really good arguments make some interesting leap at some point along the way that surprises us. They take us from premises we think, on inspection, seem right to a conclusion that departs in an interesting way from the mainstream. I guess it says something about this place that mainstream thinking includes the idea that meat-eating is morally objectionable! Anyway, the ‘no harm done’ argument Richardson put forward seemed right up your alley!

Rebecca: Wait a sec, you guys. You don’t think that the fact that you happen to like a certain conclusion means that the argument that gets you there must be a good argument. Why on earth, then, do you expect Shelby to think that?

Jeremy: O.K., O.K. Let’s not get paranoid here. As it happens, I don’t like the conclusion, but I’m having a hard time figuring out what’s wrong with the argument. In fact, it seems like a pretty good argument to me. I truly hate to say this, but I’m beginning to think that protecting animals from harm is not a good reason not to eat meat. I mean, I have plenty of other reasons for being a vegan. But avoiding harm to the animals themselves does not seem to be one of them. The animals are dead – as Shelby is so fond of pointing out – when we eat them. And, as Richardson was arguing today, any future animals we might think we can rescue by not eating meat today effectively owe their life – their highly ‘precarious’ existence – to our practice of eating meat. “That they must be killed is the reason that they live.”[3] But never existing at all is hardly an improvement for them so long as they are treated in a minimally decent manner during their lives and then humanely, if prematurely, slaughtered. It’s got to better to have a flawed life than none at all. So our avoiding meat-eating isn’t going to benefit future animals, either. If anything, it’s going to harm them!

Shelby: Yes, he does. Sort of. According to Mill, it’s better for agents to create more happiness rather than less. But that means that you’re not exactly done when you’ve managed to bring a new human or cow or chicken into a not-so-horrible existence. If you can do more for a particular animal than that, then you harm him or her when you do less.

Jeremy: Of course! But, Shelby, you’re not appreciating the force of the nonidentity problem. As soon as you try to do any more for your future cow, by, for example, refusing to eat meat, all you end up doing is taking it off-track for existence altogether.

Shelby: Taking her off-track for existence altogether. I mean, it’s a cow . . . .

Jeremy: Whatever. The point is that, if you want to maximize happiness for any particular future cow, then eating meat today is probably just the way to make that happen. The moment you try to do more for your cow, all you end up doing is a lot less! Ergo, meat-eating is after all not a practice that harms the future animals we vegans say we care about so much.

Shelby: Your ‘ergo’ is not so clear, Jeremy. Beck, you explain what we talked about after dinner last night. I’m exhausted! I haven’t been getting any sleep at all lately! (Slumps in chair, puts feet up.)

Jeremy: Shelby, we don’t even want to know how you spend your nights. But you should try in the future to come to class with some reasonable ability to function. If nothing else, think of your parents who have to pay the huge tuition bills at this place.

Shelby: I’m functioning. You’re just haven’t noticed yet. And the tuition here is nothing to my parents, not that it’s any of your business.

Rebecca: Stop it, you two. Jeremy, just listen. This is what Shelby and I came up with last night. The choices we’re evaluating include the consumer’s choice to eat meat and the factory farmer’s choice to satisfy the consumer. We’ve already agreed that the choices made before the animal comes into existence, if made differently, will probably mean nonexistence for that animal. That’s Kavka’s ‘precariousness of existence.’ But what about the period after the cow has come into existence but before she’s slaughtered? Now, on any given morning during that period, the farmer makes a lot of choices that bear on the wellbeing of that cow. He (let’s suppose ‘he’) may choose, for example, to keep her in a dark and confined space rather than give her access to pastures for the purposes of grazing and socializing, or he may choose to remove her newborn calf from her side, or he may choose to load her on the truck and send her off to some gruesome meat-packing plant. If the farmer, one fine morning, ‘does more’ for that particular cow than any of those awful things, that cow’s existence is not going to be undone. That cow is not going to evaporate into thin air. The laws of physics as they apply to that cow are not going to spin out of control. Instead, that cow’s going to frolic, in the pasture, socializing with other cows, newborn calf at her side, etc., etc., etc. You see the point. As to those acts performed after existence is no longer ‘precarious,’ you no longer have a nonidentity problem.

Shelby: Precisely! You have an identity non-problem. The factory farmer might load the cow on the truck for the meat-packing plant. But, given that the farmer clearly has the alternative of doing better than that for that particular already-existing cow, he harms that cow when he loads her on the truck for the meat-packing plant.

Mark: Hmmm. This isn’t a bad point, Jeremy. I guess you could say the same thing about the slave child case. The nonidentity problem argues that the couple’s entering into the slave child contract does not harm the child. But, once the baby exists, the couple’s actually performing under the contract – that is, handing the baby over to the rich man so that he can have his slave – does harm the child. Because at that point, the child’s existence is no longer ‘precarious.’ The couple can do better for the child than do what the contract tells them today. When they do less, they harm the child. That’s so, even if there’s some penalty for their breach of the contract. Or imagine that the couple have a change of heart and ask a court to decide the question of whether they must comply with the slave child contract. Imagine that this happens after the child exists. Then, the court’s refusing to enforce the contract would hardly remove the child from existence.[4] So if the court does opt to enforce the contract and orders the child delivered into slavery, the court harms the child.

Jeremy: But now wait a minute. You guys have shifted the focus of the argument. We started off discussing whether the consumers – ordinary people like you and me – harm animals when they eat meat and thereby create incentives for factory farming. But you’ve now shifted the focus to the factory farmers and whether what they do harms the animals. And, Mark, you’ve shifted the focus from whether the couple’s entering into the slave child contract harms the child to the question of whether, after the child exists, the couple’s doing what the contract directs – or the court’s enforcing that contract – harms the child. Different agents, guys, performing different acts at different times – and you get different results on the question of harm. No one’s denying that.

Shelby: O.K., so let’s focus on meat-eating by me. The meat-eating that I do today. Surely it’s clear that a vast amount of misery that any particular beef cow experiences is a consequence of acts that are performed after that cow has already made it into existence – and, obviously, before that particular cow has been slaughtered. That means that those acts have nothing to do with bringing the cow into existence. The timing is all wrong for things to be otherwise; the cause can’t come after the effect. And that, in turn, means that, from the cow’s point of view, there’s nothing at all good to be said about those acts. They don’t confer a miserable existence on the cow because they don’t confer any existence on the cow. They just confer misery.

Jeremy: O.K., but what on earth does the meat-eating you do today have to do with all that misery? That’s my point!

Shelby: Well, that’s what you need to rethink. Consider some already-existing, not-yet-dead youngish cow. Call her ‘Bessie.’ Because Bessie is young, there is ample time for the meat-eating I do today to create some small incentive for the factory farmer to scrimp on Bessie’s care in order to maximize profits over a period of years and then in the end send that very same identical Bessie to an early and awful death. Now, you can’t cogently argue that my meat-eating helps Bessie by bringing her into existence. She already exists. So I just don’t see, Jeremy, how your nonidentity problem is going to get me – devoted meat-eater that I am – to the ‘no harm done’ result I want!

Jeremy: ‘Small’ does seem the operative word to describe the kind of incentive you see yourself creating here! Maybe your not eating meat will eventually have some bearing over the next several months or years on how the factory farmer in some distant state or country treats the cows he’s already bred and who are marked for eventual slaughter. But maybe not. In any event, what on earth do you think the factory farmer is going to do for your cow Bessie if you suddenly stop eating meat and if that choice does by some miracle have some effect on the farmer? Make her his pet? Obviously not. She’ll just find herself dead even sooner than if he had had the economic incentive to keep her alive. Be real, Shelby!

Shelby: The point, Jeremy, was one of principle. You can’t just assume – as you apparently have – that all the questionable things we do that relate in some way or another to cows bear on their coming into existence. It depends on the timing. The acts that take place after Bessie has come into existence can’t be defended on the grounds that they are, well, ‘existence-inducing.’ And, anyway, our options aren’t just eating meat and not eating meat. We can refrain from eating meat and still create some incentive for the farmer to keep Bessie alive and well-cared for. We, collectively, could buy, or kidnap, the damn cow – or pass a law! That’d be an incentive!

Rebecca: Anyway, why shouldn’t we think, not just about the ethics of meat-eating, but also about the ethics of factory farming? The acts of the factory farmers are hardly themselves beyond scrutiny.

Shelby: Right. I think my choices probably do, eventually at least, have some small effect on how particular cows are treated. You obviously disagree. But even you have to concede that factory farming clearly harms already-existing cows. I mean, the causal sequence between the agent’s choice and the effect on the cow is a really short one!

Mark: So you have a ‘direct’ harm imposed on the cow by the factory farmer, and an ‘indirect’ harm imposed on the same cow by the meat-eater who provides the factory farmer with an incentive to impose that ‘direct’ harm.

Shelby: Whatever. You have plenty of harm clearly done within the context of factory farming – by the farmers themselves and – I think – by the consumers whose dollars and pounds and yen drive the industry forward. Nice to think ‘no harm done,’ but I don’t see it myself.

Jeremy: Look, we have more to talk about. I’m not ready to rule out anyone’s view at this stage. But do something for me. Over the weekend, chart the nonidentity problem. The x-axis will reflect the causal sequence – what happens, and who exists, when. On the y-axis, identify the relevant agents. Then, try to figure out when the nonidentity problem kicks in to show ‘no harm done’ and when it’s silent. I think you’re going to see the following. There will be a very narrow class of cases in which meat-eating possibly harms animals and a broader class of cases in which factory farming harms animals. I’ll give you that. But what you won’t find are cases in which the meat-eating – or any other act you care to scrutinize – harms an animal who is not already in existence at the time that act is performed. But that’s the vast majority of cases: cases where acts of meat-eating today don’t really make already-existing cows suffer but rather bring future cows into existence and are really necessary for the next generation of cows to have any chance of coming into existence at all.

Rebecca: But if the class where harm is done isn’t so narrow that it’s empty, won’t that class still potentially have moral significance? A relatively small number of small harms is not no harm. You have a harm, and no good reason not to make things better for that animal than you have. Doesn’t that mean that meat-eating is morally impermissible – you know, wrong? So what that only a relatively few already-existing animals are harmed by the meat-eating that’s done on a given day rather than large numbers of future animals. If it’s wrong, it’s wrong.

Jeremy: No doubt that if it’s wrong, it’s wrong. But look, let’s stop here for today.

Shelby: And also we need to talk not just about the effect of my meat-eating but also about the effect of our meat-eating on particular already-existing cows. Collectively, what we as a society do in terms of meat-eating obviously does have an effect on those cows! Even if you think what I do as an individual won’t make any real difference to Bessie, you have to admit that what we do together most certainly will make a difference to Bessie!

Jeremy: Well, hmmm, OK, let’s just stop here for today. Chart the nonidentity problem in the way I suggested, and we’ll come back to this on Wednesday.

Shelby: And that’s not the only mistake Richardson makes, either. I mean, he just assumes that life for the factory-farmed animal is always better than never having existed at all. But I think in some cases it is surely better never to have existed at all than to exist and suffer. What makes Richardson so sure that the factory-farmed animal is not in exactly that boat? After all –

Jeremy: – Shelby, in about five seconds, you’re going to be in here talking to yourself. See you Wednesday!

Shelby: – the value of never existing to the person who doesn’t exist isn’t high but it’s not horribly low, either. I mean, it’s not negative . . . .

The next morning, discussion between Jeremy and Richardson in Richardson’s office.

[Reflections on the broad scope of the NIP, which can be construed to challenge the following ideas, among others: the widely-held view that agents today have all kinds of obligations with respect to future generations, including the obligation to provide for a reasonably high quality of life for their own future offspring as well as for remote generations; the related idea that we ought to avoid unduly depleting natural resources and ought not pollute the planet; the notion that descendants of individuals who have themselves been subjected to some great wrong, such as slavery, should be compensated in some way; and the so-called ‘person-affecting’ intuition that ‘what is bad must be bad for someone.’

The point is made that utilitarianism, because it rejects the person-affecting intuition, can find wrongdoing even in cases in which no harm has been done. The ‘total’ form of utilitarianism is introduced, and the ‘repugnant conclusion’ as an objection to totalism is noted.]

Richardson: Of course it is, Jeremy. Very big, in the sense that it is interesting and provocative.

Jeremy: But I mean it seems to show that we really just can’t harm future animals – or future persons. The more we try to do better for those individuals, the more likely we are to mess up the causal sequence that brings them into existence to begin with and keep them from ever coming into existence at all. My reason for avoiding meat-eating was to protect future animals from the harm of factory farming. But the nonidentity problem shows that that line of reasoning is completely mistaken. I would have thought that would-be parents needed to start taking care of their own future children before those children had even be conceived. I would have thought that people thinking about having children should stay away from serious drug use or toxic environments or, well, activity that puts them at high risk of HIV or whatever – that they should stay away from all those things that might lead to a bad outcome for the child once the child is conceived. But now I can’t see how that is so. I thought that the present generation owed the next generation, and the next and the next, clean air and water, and a fair share of available resources. But how can that be, when the environmental and population policies we adopt today are going to determine who exists tomorrow?

Richardson: Ah, ‘big’ in that sense. Yes, the scope of the nonidentity problem is quite broad.

Jeremy: Since so much of what they do will help to determine just who comes into existence in the future, the bottom line is that agents can do pretty much whatever they want so long as what they do does not harm any already-existing person.

Richardson: Well, some will argue that you can harm the dead. I don’t see that, myself.

Jeremy: Right. And the nonidentity problem shows that you can’t harm future persons either, because if you had tried to do any better for them than you have you’d never have produced them to begin with.

Richardson: Well, now, Jeremy, things are not quite that bad. You need to keep two points in mind.

First, the scope of the NIP isn’t really quite that broad. It’s true that what you do today and what others do to you, and in general that the future unfolds for you in one particular way rather than another, is likely to have a significant influence on the identities of any offspring you may find yourself parent of at some future date. The existence of any particular future child that you may produce thus is, as of the moment, highly ‘precarious.’ Correct?

Richardson: But what you do today, and what others do to you, and so on, won’t influence the identities of the offspring of people whose lives you do not in some way touch. Here’s an example. Madeleine and I – you know Maddy, yes? My wife? – haven’t had any children yet. Suppose that we decide that now would be a good time to try to have a child. Just suppose. But instead of repairing to the bedroom with Maddy, I go out into my back yard, where I spend a few hours digging an enormous hole where a swing set will one day be installed. And I bury in that hole – oh, I don’t know – some glass, or dioxin, or anthrax, or something awful. I return to the house and wash up, and we then conceive our first child.

You see, of course, where I’m going with this. Suppose that child – let’s call him Henry – in later years becomes infected with the anthrax I have buried in my back yard. That’s a terrible thing. But little Henry’s existence, despite the infection, will still be a quite tolerable one; Maddy would see to that! And suppose I’d tried to do better for Henry than that. Suppose I’d interrupted my zany little digging project before it had quite got off the ground. Suppose I’d repaired earlier rather than later to the bedroom with Maddy. That will change everything! The timing and conditions of conception will be completely off! Different gametes will combine. Same egg, very probably at least, but a different sperm. Different gametes mean that, in the end, a different child will be born. A child better off than, but still ‘nonidentical to, Henry. So I couldn’t really, after all, have done any better for Henry than I have! Too bad about the anthrax, but better to exist than not; and effectively that was Henry’s only way into existence! So, no harm done!

Jeremy: Right! Your burying the anthrax doesn’t harm Henry at all, but instead benefits him, by causing him to exist!

Richardson: Exactly! But now think about my next-door neighbors’ future child – my neighbors whom, let’s suppose, I avoid like the plague. Nothing I do in my own back yard will have any impact at all on the timing and conditions under which they conceive their own child.

Jeremy: So if their child happens to wander over to play on the swing set in your back yard and becomes infected with anthrax, then the nonidentity problem has no application. That child won’t ‘owe her very existence’ to your burying the anthrax. So we can say that that child, even though she did not exist at the time of your act, is ultimately harmed by what you’ve done.

Richardson: Precisely! And we can say that what I did was wrong in virtue of the fact that it harmed their child – though not, ironically, wrong in virtue of the fact that it harmed my own child.

Jeremy: But that seems like a – hmmm – tenuous way to secure obligations with respect to future people. I mean, what if you put a high, secure fence around your own back yard and keep out all the neighbors’ future children. Then, you’re off the hook, morally, despite the fact that you’ve infected your own child with anthrax?

Richardson: ‘Tenuous’! Good word! Yes, you’re correct that this criss-crossing device for securing obligations with respect to future persons operates in a fairly narrow context. It is, as you say, tenuous. In many cases, I do – frightening and unwanted though the thought may be – have at least a bit of influence over my neighbors’ day-to-day lives. Moreover, the bigger and the badder the act that I perform – or that we together perform – the greater the likelihood that that act will influence the identities of the future offspring of the most far-flung persons.

Jeremy: Right! That’s the force behind Parfit’s depletion example! We bring into existence an entire future population who suffers badly as a result of our own selfish choice to deplete the earth’s resources. But had we chosen conservation instead, all we would have ended up doing is bringing into existence an entirely distinct population, a population ‘nonidentical’ to the population that exists and suffers. A population that, though badly off in view of resource scarcity, ‘owes its very existence’ to our choice of depletion. Just think, in this connection, how the nonidentity problem applies to the case of global warming! If the nonidentity problem is correct, why should we waste one more second trying to figure out whether global warming is a reality or a liberal conspiracy theory – or any money at all in trying to stop it?

And don’t forget Sher’s slave trade example. How many African-Americans today who have ancestors who were slaves would ever have existed at all in the absence of the institution of slavery? They owe their very existence to slavery. They aren’t victims of slavery, but rather, if anything, its beneficiaries. So any claims to restitution that they might make, when you think about it carefully, are not, at base, cogent. And how many descendents of European Jews born after 1946 or so would have existed at all but for Hitler? Their parents and grandparents, depending on who they were and where they had the bad fortune to live, were clearly harmed by the atrocities. And their businesses and their estates may have been annihilated. But they themselves, ironically, owe their very lives to those same atrocities!

And think about how the nonidentity problem applies in the case where a new reproductive technology turns out to be riskier than anticipated. As long as the technology helps to bring the child into existence – and so long as the child’s existence is not so miserable that never having come into existence at all would have been better – it can’t be said to harm the child. Defense lawyers in medical malpractice actions must love this argument! No harm, no negligence, no malpractice – and no damage award to be paid to the plaintiff!

Richardson: Jeremy, take a breath. Calm yourself! We’ve just seen that the nonidentity problem isn’t quite as broad as you first suggested. Still, it’s true that the scope of the NIP is broad. What the argument shows us is that the cases in which our acts eventually do impose harm on future individuals are extremely rare. It’s just maddeningly difficult, in general, to discern harm to persons who don’t yet exist at the time the questionable act is performed.

But remember that I said that there were two important points to be made in this context. So far we’ve just made one. But now we need to attend to point two. You’re assuming, Jeremy, what Parfit called the ‘person-affecting intuition’ – the idea, as he put it, that ‘what is bad must be bad for someone.’ But – as Kavka’s case so cleverly illustrates – that fact is that bad acts need not be bad for anyone! We can put the point in our usual language of ‘harm’ and ‘permissibility.’ It just doesn’t follow, from the fact that no harm has been done, that the act itself is morally permissible!

You didn’t really think, do you, Jeremy, that we moral philosophers would rest content with the idea that it is morally permissible for the couple in Kavka’s slave child case to enter into the slave child contract? Or that it’s perfectly all right for the present generation to deplete the planet’s resources in ways that will inevitably lead to the serious suffering of future generations? Of course not! Rather, we moral philosophers have eagerly taken up the challenge of explaining how it can be that an act that harms no one can be wrong – how it can be, indeed, that an act that represents the very best agents can do for each existing and future person can be wrong!

Jeremy: I’m not following. You’re a utilitarian – you think that the agent’s obligation is to create the most happiness that that agent can create for persons. How can it be that an act that manages to do that for each and every person who ever exists be wrong? How can you, as a utilitarian, think that any plausible moral theory requires more of agents than that?

Richardson: Therein, Jeremy, lies the genius of totalism! Or averagism! Pluralism does a pretty good job with all this as well! Take your pick! All kinds of consequentialist approaches do a superb job of avoiding the NIP!

Jeremy: Wait . . . . I haven’t had your ethics seminar yet, or Matthews’ course on Rawls. You know, I’m only a second-year student. I probably shouldn’t even be a T.A. for this course. I’m just not following you. I know what totalism is: a form of utilitarianism that assesses acts on the basis of how much total happiness they produce. And I know that averagism assesses acts based on how much average happiness – how much happiness per person – they produce. But when agents have done the very best that anyone could have done for all people, including all future people – when they’ve maximized happiness for every person – won’t both totalism and averagism imply that what they have done is perfectly permissible?

Richardson: But again, Jeremy, you’re assuming that an act that is wrong must be wrong in virtue of the fact that it’s bad for, or decreases wellbeing for, or harms someone or another. You’re still taking the person-affecting intuition for granted.

Richardson: And you think that the agent’s obligation, under utilitarianism, is to ‘make people happy’?

Jeremy: Of course. The happier, the better. And if you are a totalist, you think the obligation is to maximize total happiness for people. If you’re an averagist, you think the obligation is to maximize average happiness for people.

Richardson: And whose wellbeing levels matter, for purposes of making those calculations?

Jeremy: Well, Mill thinks each agent has the obligation to create happiness for people. He didn’t actually talk about future persons – but he can’t have meant to say that they do not count.

Richardson: So we are to aggregate all those individual wellbeing levels – present and future – to figure out just which act it is that maximizes wellbeing. Whether we are totalists or averagists. According to Mill. At the same, an act will create varying individual wellbeing levels depending on the context in which it is performed. The same act may produce more or less wellbeing depending – among other things – on what other agents do, both at the time the one act is performed and at each later time.

Jeremy: I understand all that. I did take Professor Lampier’s Logic and Ethics course last term. I know that the utilitarian assessment of individual acts must be relativized to the distinct ‘possible worlds’ at which those acts are performed. You can think of distinct ‘possible worlds’ as different ways in which our own future might unfold, with each future, in all its gory detail, representing some precise resolution of all of the contingencies of life. For example, consider my act of – oh, I don’t know – running a stop sign this afternoon. And suppose that just prior to my act you’re standing at the side of the road waiting to cross. If the future unfolds in such a way that you are in the middle of the crosswalk at just the time I run the stop sign, my running the stop sign will cause your happiness level to fall. On the other hand, if the future unfolds in such a way that you haven’t yet started across, then my running the stop sign will not have that effect. We can’t consistently say that my running the stop sign both reduces and does not reduce your happiness level. But we can say that at one world my running the stop sign reduces your happiness level and at the other world it doesn’t.

Richardson: Fine, fine, all that is well and good. But what does a careful formulation of utilitarianism look like?

Jeremy: Simple. We can rank all the possible worlds that a given agent has the options of bringing about in terms of their ‘betterness.’ For the totalist, one world is better than another just in case the one world’s contains more aggregate happiness than the other, where aggregate happiness itself is just the summation of all the individual happiness levels for all the people who exist – and all those who will exist! – at a given world.

Richardson: Good. And, Jeremy, we might use the term ‘wellbeing’ rather than ‘happiness’ here. It may well seem right to us that agents ought to create the most good they can. But for purposes here we do not necessarily need to have a lengthy debate on what constitutes that good or whether that good necessarily consists of happiness. We can just use the term ‘wellbeing’ instead as a sort of placeholder – and understand that, in due course, we’ll need to figure out whether creating wellbeing for people is a matter of making them happy or something else entirely. Perhaps, for example, maximizing the good is a matter of maximizing people’s ‘capability,’ as Amartya Sen suggests, to achieve some variety of ‘functionings’ ranging from gaining nutrition and interacting socially to performing higher-level mathematics. Or perhaps it is something else entirely.

Jeremy: All right, fine. Suppose, then, we want to assess a particular act, ‘A,’ performed by a particular agent at a particular world, ‘w.’ If no worlds that exclude A contain more aggregate wellbeing than w does, then A, performed by that agent at w, is perfectly permissible. But suppose that there are worlds that exclude A and that are better than w – i.e., contain more aggregate wellbeing than w does. Then, A performed at w will be wrong if the agent’s performing some alternative act B instead of A is sufficient to insure that one of those better worlds would have obtained. Now, that principle just states a sufficient condition for A’s being wrong. There’s a controversy whether having the ability on one’s own to insure a better future also counts as a necessary condition for wrongdoing. . . .

Richardson: Yes, yes, yes, causal overdetermination, etc. But let’s not sink into that seductive swamp here? You’ve already said plenty to enable us to wind up the matter before us. So ‘aggregate wellbeing’ at a world refers just to – what?

Jeremy: Well, ‘aggregate wellbeing’ is just the summation of the overall, lifetime wellbeing levels for all persons who do or will exist.

Richardson: At that particular world, no? Because as the nonidentity problem makes so terribly plain to us, some people will exist at some possible worlds but not others. If all the same people existed in all possible worlds, there’d be no nonidentity problem.

Richardson: Ha, ha! Identity non-problems! I’ll have to remember that one! But to proceed. When we compare worlds to each other in respect of their betterness – that is, in respect of the aggregate wellbeing that each contains – do we find that one world will be better than another only if there is some person who exists in both worlds for whom the one world is better?

Jeremy: No. The test is simpler than that. All that matters is the bottom line – if the one world contains more aggregate wellbeing than the other, then it’s better than the other.

–Oh. I get it. For the totalist, the ‘betterness’ of one world relative to another has nothing to do with making any one person better off than that person is at that other world. It’s enough that there exists some person ‘nonidentical’ to the one person who exists in place of the one person but is better off than the one person.

Richardson: Precisely! Totalism takes an ‘impersonal’ form. It jettisons, in other words, the idea – the ‘person-affecting’ idea – that what is bad must be bad for someone. And that means that making each person as well off as you can isn’t enough to avoid wrongdoing. If, by bringing other ‘nonidentical’ persons into existence in place of those who suffer, you can increase aggregate wellbeing, then you do something wrong when you fail to do that.

Of course, we all start with the intuition that what is bad must be bad for someone. Even Parfit admits, or has on occasion admitted, that the intuition is deeply held. But he also thinks that cases like the slave child case and his own depletion example show that the intuition must be rejected. It’s a mistake! Yes, we are initially, intuitively, attracted to the idea that, for each person who does or will exist, agents ought, when they can, create more good for that person. They ought ‘make people happy’ in that sense. But the nonidentity problem shows that there is more to right and wrong than intuition suggests!

The genius of utilitarianism – totalism, for example; but averagism does the job just as well – is to show us exactly where intuition goes awry in this context!

Jeremy: Because ‘making people happy’ is not the only way to increase aggregate wellbeing. We can sometimes increase aggregate wellbeing by ‘making happy people’ – by bringing a new person, ‘nonidentical’ to and better off than the original, into existence in place of – or, I guess, in addition to – that original. I’ve seen the distinction between ‘making happy people’ and ‘making people happy’ before. But it didn’t mean anything to me at the time. I’m not even sure where I saw it.

Richardson: In Narveson, of course! But normative ethics has moved far beyond Narveson at this point! He brilliantly distinguished between the idea of creating more wellbeing for each existing and future person and the idea of creating more wellbeing simpliciter! But then – apparently without any real scrutiny – he put the former forward as the basic insight behind utilitarian thinking rather than the latter. And then just a few years later Kavka and Parfit blew the whole thing up for him with the nonidentity problem! I think Narveson is now a libertarian somewhere!

Jeremy: But wait a second. Doesn’t the focus on aggregate wellbeing mean that we can make things right by simply adding new people to the world rather than by making things better for the old ones? Can that really be right? Also, no one thinks we can justify making an existing child miserable on the grounds that doing so will somehow serve to bring still more miserable children into existence at some future date, even in the case where that choice increases aggregate wellbeing. I mean, you can’t justify the selling of your own children into slavery on the grounds that doing so will induce more would-be parents than ever before to have children whom they can then sell into slavery.

Richardson: Well, is it really so clear that you can’t? I guess some people find those sorts of results repugnant.[5] For my own part, I find such results on the whole at least palatable if not obvious. But you can’t expect any moral theory to handle every case! No theory is perfect!

Jeremy: I guess I need to think about all this. Professor Richardson, I really apologize. I mean, I should have known about the nonidentity problem and that there’s a big issue with claims about harming future persons. And I should have understood how easily totalism avoids the nonidentity problem. To tell you the truth, I’m not confident I should even be in graduate school in philosophy. I mean, how can I be teaching this stuff when I don’t know it myself? I’m considering a move over to the law school. They admitted me a couple of years ago when I was thinking about a joint degree. Maybe they’d give me another chance . . . .

Richardson: Jeremy, you’re doing a terrific job as a second-year grad student! Don’t expect so much of yourself! Just keep up the good work! Law school! Such nonsense. Think about it all you want. Just don’t do anything about it without talking to me first!

Midway through the next section meeting. A heretofore quiet student, Nikhil, becomes involved in the discussion.

[Some skepticism is expressed regarding whether the NIP and, in particular, Kavka’s ‘precariousness of existence’ point, successfully demonstrates ‘no harm done’ to future persons; the point is set aside for the moment. Implications totalism yields for the individual’s personal procreative choice whether or not to have a child, including the choice of abortion, are noted. Parallels are drawn between such choices and the repugnant conclusion. It is noted that totalism, averagism and certain other forms of utilitarianism are often considered vulnerable to objections based on equality, justice and fairness. Rawls’ idea that utilitarianism ‘fails to take into account the distinction between persons’ is introduced. The issue of whether Mill was himself a totalist – or, more precisely, an aggregationist – is noted in passing. Pluralism as an alternative form of consequentialism that has the power to address both the NIP and the repugnant conclusion is introduced.]

Mark: Totalism, then, avoids the nonidentity problem by assessing acts on the basis of how much happiness in the aggregate they create. The couple who enters into the slave child contract have not harmed the child they produce as a slave because they could not have produced that same child outside the contract. But under totalism the fact that the couple had the alternative of producing some distinct child as a non-slave is morally significant. After all, we can plausibly assume that the happiness the couple derive from their yacht, when put together with the happiness the rich man derives from his slaveholding, is outweighed by the happiness the distinct child derives from his or her existence as a nonslave. Totalism, then, implies that what the couple has done is wrong – that it is a bad act that is bad for no one. All this, by the way, is implied by averagism as well. What’s really important is that we focus, not on individual wellbeing, but rather on aggregate wellbeing.

Rebecca: You say, Mark, that the couple could not have produced the same child as a non-slave. For the record, Kavka points out that this is just a matter of probability. Note 15, p. 100. It’s simply very, very unlikely, he says, that the child the couple in fact produce as a slave would have come into existence as a nonslave, however hard the couple might have tried to make that happen for that child.

Mark: – but still, from the child’s own point of view, it’s better to exist as a slave than to take the very great chance of never coming into existence at all. So long, at least, as that existence is not so miserable that the child would have been better off never having existed at all.

Shelby: Right. You have to set aside the really horrible cases. But even when you do that, there’s something weird about that argument.

Shelby: But I was right last time. Right? I mean, I charted the nonidentity problem, like you suggested. It still seemed to me I was right. The factory farmers impose all sorts of harms on the animals that they could have avoided without putting the animals’ very existence at risk. Not slaughter them when they’re four; wait until they’re eight. Let them graze freely with the herd instead of keeping them in those little pens. Etc. And we, in turn, by eating meat – at, say, time t₁– influence the factory farmers to impose – at, say, t₂ – just those harms on animals who, as of t₁, had already made their way into existence. In that case, it makes no sense to say that the cow ‘owes her very existence’ to my eating meat.

But I’ve been thinking about your other category of cases, too – cases where the bad act is performed before the cow, the slave child, whatever, makes it into existence – cases where the act is performed at a time when the apparent victim’s existence is still ‘precarious.’ So: a premise of the argument in the slave child case is that, had the couple performed any seemingly better alternative act instead of the clearly wrong act they in fact performed, that same child very, very probably would never have come into existence at all. From there, we are supposed to infer that the clearly wrong act is better for that child than any seemingly better act would have been – and hence that it does not harm that child.

Jeremy: Shelby, I’m glad you’re such an enthusiastic and opinionated participant in our discussion. But, with all due respect, you’ve come in late again and you are now interrupting the discussion. I mean, you’ve high-jacked my section meeting. Anyway, I think Mark has just explained how that inference works. Surely, it’s better to suffer the flawed existence – within reason, as Mark points out – than it is to take the enormous chance of never existing at all. But, you got here late. Maybe you didn’t hear what he said.

Shelby: I heard. I just think the argument needs a closer look. It’s got this ‘post hoc ergo propter hoc’ flavor to it.

Jeremy: Look, I’ll be glad to talk to you about that, but not during class. We’ve already spent too much time on the nonidentity problem! We were just finishing up with the nonidentity problem, and we now need to move on to Chapter V of Utilitarianism. We need to talk about why totalism and principles of justice are not, according to Mill, really at odds. I mean, we really need to get through Chapter V today to have any hope of getting through the ten or twelve other topics the syllabus lists for this term. So – OK, Rebecca, I see you have something to say. I hope it has something to do with Chapter V.

Rebecca: I’m trying to get a sense of what this new theory, totalism, really comes to –

Jeremy: I’m not sure totalism counts as a ‘new theory,’ Rebecca. We really are just coming to a better understanding of what Mill has been saying all along.

Rebecca: Hmmm. Well, I’m trying to figure out what it means to say that we ought to do whatever it takes to increase aggregate wellbeing, including, if necessary, ‘making happy people.’ Think seriously about what those words mean. If the bottom line is aggregate wellbeing, then we must ‘make happy people’ even in cases in which doing so will require that we make already-existing people quite unhappy. Are you saying, then, that I have an obligation to be pregnant right now? I guess I could make that happen. It would, of course, be extremely deleterious to my happiness if it did. But – if I did have a child now – I know that I’d take good care of that child. He or she would have a happy life. The bottom line is that the amount of happiness that that child – that new person – would enjoy would more than counterbalance any loss in happiness for me. Doesn’t totalism therefore require me – and not just me but all of you guys as well – to start producing children, early and often?

Shelby: And doesn’t totalism imply that abortion is almost always wrong? Even very early abortions effectively are choices not to add a new person to the world. But won’t the addition of one more person almost always, even netted against the loss of wellbeing for others, increase aggregate wellbeing? But isn’t it just very implausible that very early abortions are morally problematic, even in the slightest way, so long as the abortion is something that’s good for the woman herself?

Nikhil: In a lot of cases where the abortion issue comes up, the choice not to have the abortion is the choice to add a ‘new’ person to a given scenario even at the expense of making some of the ‘old’ ones worse off. But so long as adding new people increases aggregate happiness, totalism says you have to do it.

Jeremy: Now I get it. Good. O.K., so, can you lay the repugnant conclusion out for us, and explain how it challenges totalism?

Nikhil: Sure. Suppose that you, as agent, see two buttons on a panel before you. And you have the choice as to which one to press, the green one or the blue one. And you know that if you press the green button a very large population of people all of whom will have lives well worth living will be produced. And if you press the blue button a vastly larger population all of whom have lives only barely worth living will be produced. You can press one button or the other but, exclusive ‘or,’ you can’t press both. Suppose, too, that when you take the summation of individual happiness levels at each of the two worlds, you find that aggregate happiness at the green-button world is less than that at the blue-button world. Assume you have no other choices. Totalism, then, implies that it is wrong to push the green button and that you ought to push the blue button. Parfit finds this conclusion ‘repugnant.’ As do I.

Rebecca: Me too. Abortion is really just a mini-repugnant conclusion. In some cases, at least. For example, consider the case where continuing the pregnancy and thereby bringing a new person into existence is bad for the woman and not particularly good for anyone else who does or will otherwise exist. But continuing the pregnancy will presumably be good for the future baby, at least eventually, and, in virtue of that fact, it will increase aggregate happiness. Forcing the woman not to have the very early abortion – forcing her not to have the abortion at a point where the fetus is not yet a person – is, I think, repugnant. And so is telling her what she’s done is wrong.

Jeremy: You know what, let’s not mix the repugnant conclusion up with the problem of abortion. In fact, we won’t be dealing with abortion at all in this course. But I’m glad we’ve taken the time to get the repugnant conclusion out on the table. You’ve all got the basic idea, I think. And we’ll be coming back to the repugnant conclusion later in the term, but just now we need to talk about utilitarianism and justice. We really need to finish up with Chapter V today.

Mark: Well, if you want to talk about justice, maybe you should talk more about the repugnant conclusion. I’ve started reading the Rawls assignment for next week, and isn’t the repugnant conclusion related to Rawls’ whole objection to utilitarianism? Utilitarianism fails, he says, ‘to take seriously the distinction between persons.’ Instead of considering what is good for each person who does or will exist, utilitarianism just focuses on the bottom line, treating people as a single, massed, squished-together being.

Jeremy: Well, that’s an attractive picture! Look, there’s something to what you’re saying. But we need to distinguish between the two sorts of problems. I think when Rawls said that utilitarianism fails to ‘take seriously the distinct between persons,’ he really meant to be referring to an array of inequality problems – cases where one person is left much worse off than another through no fault of their own and in a way that seems to us unjust. But that’s different from the repugnant conclusion. I mean, the justice objection against utilitarianism has been around forever, and the repugnant conclusion is a fairly new invention. Moreover, utilitarianism is well-positioned to addresses considerations of justice and equality. Both totalism and averagism can, for example, recognize and give an account of what is called the declining marginal value of resources. Both can easily explain why we ought to give the extra loaf of bread to the poor family rather than the well-off family: doing so will create greater average, and greater total, happiness altogether. The repugnant conclusion, in contrast, is much harder for the utilitarian to avoid.

Nikhil: Well, I’m not so sure, Jeremy, that that is true. There will still be certain badly unequal distributions of happiness that will turn out to be perfectly permissible under both totalism and averagism. If the total or average is high enough, it’s not going to matter that one person, or some relatively small number of persons, has been exploited in some really terrible way. It’s for that reason that totalism and averagism are so hopelessly outdated. Most contemporary utilitarians are not really totalists or averagists at this point but rather pluralists.

Jeremy: Nikhil, I see that you too are reading ahead. Can you explain what the idea behind pluralism is?

Nikhil: Well, utilitarianism is a mono-valued theory. Just one fact about the world has moral significance: whether pleasure, or happiness, or preference-satisfaction, or capability – or whatever you think makes life so precious to the one who lives – has been maximized. But the core intuition behind utilitarianism is that agents ought to create the most good that they can. Consider with that, we might well think that the good – the overall good now, not just one aspect of the good – itself comprises more than just whatever it is that makes life so precious to those who live. Perhaps, for example, equality, in addition to aggregate wellbeing, is part of the overall good. If we think that, then we might also think that betterness between worlds will be a matter not just of how much aggregate wellbeing each contains but also how much equality.

Now, going that far obviously helps us resolve certain equality problems. But it doesn’t really help us with the repugnant conclusion. After all, the vast population of persons whose lives are just barely worth living may well suffer equally. Inequality may not be their problem at all. To solve the repugnant conclusion, we therefore must take pluralism a step further. But why not? Why not say that the overall good comprises, not just aggregate wellbeing and equality, but something like ‘human flourishing’ as well? At a world where the population consists of a vast number of people with lives only barely worth living, there may be a lot of equality and a lot of aggregate wellbeing but not much in the way of human flourishing. There will likely be no Mozart, for example, and no one to appreciate Mozart if by some fluke he does exist. We thus have a basis for saying that the less populated world, where each person has a life well worth living, is better than the more populated world, where each person’s life is only barely worth living.

Jeremy: Good summary of pluralism, Nikhil. Look, we’re ten minutes over. We need to stop. Think about pluralism, and think about it in connection with Chapter V of Utilitarianism. Which we will turn to at the very beginning of our next section meeting!

Faculty club. Jeremy is lunching with Professors Richardson (Chris) and Lampier (Lamp), and Professor Rosen (Miriam) from the law school.

[Counterfactual, or ‘but for,’ test for harm considered and rejected. The suggestion is made that an act harms a person when no alternative act exists that would have been better for, or created more wellbeing for, that person. The discussion then shifts to the ‘betterness’ relation. It is argued that the determination that an act A is better than an act B for a given person p cannot cogently be based on a comparison between the actual value A creates for p is and the expected value B creates for p. The argument is then made that, when we switch to a comparison between expected values alone we do not, after all, obtain the result that the morally questionable act is better for the apparent victim than any other. The suggestion is made, in other words, that the particular type of NIP under consideration, rather than showing ‘no harm done,’ is based on a mistake regarding probabilities. Once that mistake is set aside, there is ample room for a finding of harm.]

Lampier: You’ve been talking about your slave child case, Chris, and the nonidentity problem for years. I just realized I had a question about it. The couple effectively sell their own future child into slavery by signing the slave child contract prior to conceiving that child. And signing the contract doesn’t harm the child – or so the nonidentity argument goes – because that child has no better alternative than existence as a slave.

Richardson: Right! Had it not been for the slave child contract, the couple would never have been interested in or willing to have any child at all. So that child would never have existed at all. And that would not have been better for the child than existing as a slave. If anything, it would have been worse.

Rosen: But wait. You’re saying that the agent’s act harms the victim only if, “but for’ – that is, in the absence of – that act, the victim would have been better off?

Richardson: Ah, Miriam. You’ve caught me. I was sloppy. I’m sure Jeremy already sees where I went wrong, but go ahead and explain your objection to the ‘but for’ test of harm.

Rosen: It’s a simple objection – one that all first year torts professors just love to trot out at the earliest opportunity for their students. But it’s on the basis of that objection that the law of negligence itself discarded decades ago the ‘but for,’ or counterfactual, test for harm in favor of a ‘substantial’ or ‘significant’ causal factor test.

Richardson: So an act causes harm to an individual just in case that act is a significant causal factor in that individual’s being made to suffer in some way? Begging the question has its advantages!

Rosen: So it does! Better to beg the question than to be wrong! So, Jeremy, perhaps you already know that the phenomenon of causal overdetermination creates a problem for the ‘but for’ test of harm. Suppose I shoot Chris in the arm. And suppose, too, I was so mad at Chris that had I not shot him in the arm I would have shot him in the heart. ‘But for’ my shooting Chris in the arm, then, Chris would have been left quite a bit worse off than he is. But that means that my not shooting him in the arm would have left him no better off than he is. Under the ‘but for’ test, my shooting Chris in the arm therefore does not harm him. Under a ‘but for’ approach, my shooting Chris in the arm, if anything, saves his life! But obviously I do harm Chris when I shoot him in the arm. No court in the land is going to make the mistake of finding otherwise when presented with such a case. And I sincerely doubt that very many philosophers are going to make that mistake, either.

Jeremy: But if the counterfactual test fails as a necessary condition for harm and the ‘significant causal factor’ test is question-begging, what test do we use for harm?

Rosen: Well, I think the right thing is for the court to consider whether the agent – a.k.a., if you’re a lawyer, the defendant – could have made things better for the victim than they in fact are. Is there something the defendant could have done, in other words, that would have made, or left, the victim better off than he or she in fact is?

Jeremy: But doesn’t that test imply that defendants can be held legally responsible not just for their acts but for their omissions? I mean, utilitarians generally think that omissions can be just as wrong as affirmative acts. Just as it’s wrong for me to push the small child into the swimming pool where he will almost certainly drown, it’s wrong for me not to rescue the child from the swimming pool when I stroll by and see that he’s in trouble. But I thought that the law was different – that we cannot be held legally responsible for our omissions but only for our affirmative acts.

Rosen: I know a lot of people have that idea about the law. But it’s not entirely accurate. Two points. First, the law of negligence has no trouble with the idea that failing to prevent a harm is itself a cause of harm. The law recognizes that an omission might cause as much harm as any affirmative act. But liability in negligence requires a finding of both harm and duty. And courts will often balk and refuse to find liability on grounds of duty. For most cases, the rule is that people have a duty to avoid acts that cause harm but no duty to perform acts that prevent harm. They have no ‘duty to rescue,’ in other words. Second point: that rule itself does not hold without exception. There are all kinds of ways that one can become legally liable for their omissions. You know that your doctor has the duty not to amputate the healthy leg rather than the one infected with gangrene. But your doctor also has the duty to test you for – oh, say – strep throat when you come to him or her with a certain set of symptoms. Your doctor then has a duty to treat that infection if the test comes back positive. If the doctor fails to test, or fails to prescribe, that failure to act – clearly, an omission – can be grounds for a lawsuit. In that case, it’s the special relationship between the doctor and the patient that gives rise to the duty against omissions.

Lampier: So can we get back to the problem I wanted to raise? I’m already on dessert here.

Richardson: Patience, Lamp, let me just first say what I should have said to begin with about the slave child case. We don’t suddenly find harm in that case the moment we abandon the defective ‘but for’ test of harm. It isn’t just that the couple would not have been interested or willing to have any child at all in the absence of the slave child contract. It isn’t just that they would not have done better for the child than they have. The point is rather that they really could not have done any better for the child than they have. There is no alternative act that they could have performed that has any real chance of bringing the child into existence as a nonslave. That’s Kavka’s point about the precariousness of existence! Just think about it! How, exactly, are the couple supposed to go about not signing the contract that enslaves their child and still producing that very same, identical child? The timing of conception, if nothing else, would have been different! Different timing, different gametes – or at least different sperm! Different sperm, different child! A better off child, to be sure, but a child ‘nonidentical’ to the one you are all so worried about nonetheless!

Lampier: Do I get to talk now? So, as you say, the point is really one about the probability of the child’s existing as a nonslave. It isn’t that it was somehow logically, or metaphysically, or even physically or biologically, impossible for that couple to produce that same child as a non-slave. Being a slave isn’t some essential feature, like being a person rather than a garage door, that the child could not have come into existence and not had. It’s just highly unlikely that the couple would have ended up conceiving the very same child as a nonslave, however hard they might have tried.

Richardson: Precisely! And – from the child’s own point of view – it’s surely better to exist as a slave than to take the very great chance of never having existed at all.

Lampier: So you want to say. But let me continue. First off, it will vastly simplify things if we can give our slave child a name. We can’t cogently name things that never exist – and when we pretend to do so, we run the risk of mixing things up. But here, by hypothesis, the slave child exists and suffers. So we can name her. Let her be ‘Molly.’

Then, your claim is that the couple’s entering into the slave child contract is better for Molly than any alternative act the couple might have performed instead. In other words: the couple’s entering into the slave child contract creates more wellbeing for Molly than do any other act available to the couple at the time in question. What I’m trying to figure out is how you get to that ‘betterness’ result. Let’s consider the possibilities. Maybe you’re comparing how much wellbeing Molly in fact has at the ‘actual’ world, where the couple’s entering into the contract has by hypothesis eventuated in Molly’s existence as a slave, against how much wellbeing Molly has at each other possible world – including those at which Molly exists yet the couple has refrained from performing the ‘bad’ act of entering into the slave child contract.

Lampier: Exactly. That can’t really be it. Consider the conjunctive act the couple in fact performs: the act of ‘entering into the slave child contract and producing a child.’ Call that conjunctive act A, for short. Now, at the actual world, the couple by hypothesis performs A and produces not just any child but rather Molly, who exists as a slave. But of course there are going to be worlds distinct from the actual world – there are going to be various possible worlds, in other words – where the couple refrains from entering the slave child contract but nonetheless produces not just some child or another but rather Molly. We just said, after all, that there’s no impossibility to Molly’s coming into existence as a nonslave.

Richardson: Well, sure. But the thrust of the nonidentity problem is not that Molly’s coming into existence as a nonslave is impossible but rather that it’s so highly improbable.

Richardson: Of course there are distinct possible worlds where the couple manages to bring Molly into existence as a nonslave! But the very instant the couple veer away from the act A they in fact perform and toward some alternative act that seems on the surface of things to be an act that is better for Molly, the chances that any of those distinct possible worlds will ultimately be realized drops to almost nothing. I just don’t see where you’re going here with all this, Lamp. You’re not telling us anything new.

Lampier: Bear with me. So you’re not establishing ‘betterness’ by comparing how Molly in fact fares at the actual world against how she fares at each possible world accessible to the couple at the relevant time. Cross that way of establishing ‘betterness’ off the list.

Richardson: Right! To get the nonidentity problem off the ground, we want to talk probability.

Lampier: So you want to talk about probability. For part of your betterness equation. But for the other part you seem to want to talk about actuality – how the future in fact unfolds. You want to determine that the couple’s performance of A is better for Molly than is any alternative by comparing how much wellbeing Molly in fact has at the actual world, where A has been performed and Molly in fact exists as a slave, against how much wellbeing Molly is likely to have, given that the couple does something other than A.

But can we do that? The standard way to bring probabilities to bear in assessing choice is to appeal to the concept of expected value. You know, one pill’s got a 90 percent chance of killing the patient and a 10 percent chance of fully curing the patient. The other’s a lot safer, but the cure rate is a lot lower. Under an expectational approach, we decide which pill it is better to give the patient by figuring out, first, the value of each possible outcome for the patient and, second, for each alternative act, the likelihood that that outcome will obtain, given the performance of that act. We then just sum up, for each act, the outcome values multiplied by outcome probabilities. That quantity represents the expected value of the particular act. What we don’t do, under any expectational approach I know of, is compare the actual value of giving the one pill to the patient, given how the future in fact happens to unfold for that patient, against the expected value for that patient of giving the other pill.

Richardson: Why not? Why can’t we do that – at least for those cases in which we do happen to know the actual value that the act has for our subject? And it is exactly that knowledge that we have, by hypothesis, in the slave child case! In that case, after all, a given future is already imagined to have unfolded; a given sequence of acts and events, culminating in Molly’s existence, has already taken place. So we know, in effect, what the couple’s choice of A does for Molly: it brings her into existence, albeit as a slave. But had they done anything other than A, Molly would have had only a very small chance of ever existing at all. Ergo, Molly’s actual wellbeing, given A, is greater than the expected wellbeing that any alternative to A would have created for her.

Lampier: Well, I think you may have a problem, if that’s really what you want to say about the slave child case. Actually I think you have two problems. The first is that the idea that betterness can be determined by a comparison between actual and expected values leads to contradiction. At the relevant time, the couple obviously had the alternative of trying their hardest to produce that same child as a nonslave. Sort of a weird alternative, to be sure, since the couple could not possibly know, at that time, that their performance of A would lead to Molly. But no matter. Call that alternative act – the act of refraining from entering into the slave child contract but proceeding to bring a child into existence anyway – B. And now suppose that we happen to be in one of those rare, but clearly possible, situations in which against all odds B, too, will, just like A, lead to Molly’s existence – but now, of course, as a nonslave. It then becomes plausible to say that the actual value for Molly, given B, is greater than the expected value for Molly, given A. But we’ve already shown that the actual value for Molly, given A, is greater than the expected value for Molly, given B. But if betterness can be determined on the basis of a comparison between actual and expected values, we can now infer both that B is better for Molly than A and that A is better for Molly than B. But surely ‘betterness’ is what we call an ‘anti-symmetrical’ relation. It’s like the ‘greater than’ relation on the natural numbers: if one number is greater than a second, then the second cannot also be greater than the first. Similarly, if one act is better for a person than a second, then that second cannot also better for that person than the first. So we now have a contradiction.

Richardson: Hmmm. Who’s to say ‘betterness’ is anti-symmetrical? Many consider ‘betterness’ intransitive. Perhaps it has other surprising logical properties as well.

Lampier: Hmmm. Well, I don’t think so. It’s true that there are people who think betterness is intransitive. But it doesn’t follow that they are right. In any event, anti-symmetry seems an even more basic property: if B is better for a person p than A is, how can A also be better for p than B is? If that’s right, then I doubt we have a clear enough concept of ‘betterness’ to get your argument off the ground to begin with.

But let me tell you my second problem. Shouldn’t the assessment that an act is morally permissible be based on what the agent is in a position to know, at least in theory, prior to performance?

Richardson: Well, yes, that’s my view. The moral analysis is most plausibly construed to provide the agent with a guide to action and choice. It must, accordingly, be couched as a prospective – not a retrospective – analysis. It can’t tell the agent the act is wrong after the fact when, at the critical time just prior to performance, all evidence is that that act constitutes the very best that the agent can possibly do. But the nonidentity problem is completely consistent with that point. Prior to performance, the agents can reason to the conclusion that any child, whether Molly or any other child, the act A happens to bring into existence will be one who has not been harmed by A. Agents can figure that out before they actually do anything.

Lampier: But doesn’t your commitment to a prospective account require you to agree that the determination of whether A represents the better act is to be made on the basis of what the agent knows about A and each alternative to A just prior to performance? And doesn’t that mean that, in coming to the conclusion that A is better-for-Molly than any alternative is, including, for example, B, we should basically screen out what happens to Molly after the agents perform A? We should screen out, in other words, the after-the-fact information that our after-the-fact hypothetical includes – namely, the information that Molly exists and hence enjoys a certain modest – modest because, after all, she’s a slave – level of wellbeing.

Richardson: Hmmm. I don’t really see any deep problem here, Lamp. We can easily deal with both your issues. We simply abandon any comparison between actual and expected values as a vehicle for figuring out betterness, and move instead to a consistently expectational approach.

Lampier: (Looking at Miriam.) I told you he’s a smart guy. That’s just what I was going to suggest.

Richardson: Well, I do think it’s plausible that a correct moral account will have the capacity in the usual case to tell agents – prospectively, when they’re still trying to decide what to do, and not just in hindsight – whether a given choice is morally permissible. So expected value seems the way to go.

Lampier: But now we have another problem. The performance of A does not guaranty that Molly will exist. But when we go to calculate the expected value of A, we have to take that fact into account – just as we do when we calculate the expected value of B. And that, in turn, means that your notion that A is somehow better for Molly than B is may not be right after all.

Richardson: Well, I suppose there is no guaranty that Molly will exist, given A. A lot of things have to go a certain way, along with A’s being performed, for it in the end to be Molly who comes into existence rather than some distinct person ‘nonidentical’ to Molly. But clearly A makes Molly’s existence considerably more likely than B does.

Richardson: Lamp, it’s the whole ‘precariousness of existence’ point. Any variation from the actual course of events is very, very likely to take Molly off-track for existence altogether. So the variation from A to B – which is an enormous variation, by the way – is very, very likely to take Molly off-track for existence as well. That means that, even though there’s no guaranty, A makes Molly’s existence significantly more likely than B does. And that, in turn, is going to mean that A creates more expected value for Molly than B does.

Lampier: Look, let’s just suppose that, if A makes Molly’s existence more likely than B does, then A creates more expected value for Molly than B does. Even though that’s not literally the case. My problem is that I’m not seeing how we can establish that A makes Molly’s existence more likely than B does. And – oh, hell, I’m late for class!

Lampier: I gotta go. But I think that once you raise the question it’s pretty obvious that there is no easy answer. Keep in mind that B represents the couple’s best attempt to bring the very child they will bring into existence when they do sign the contract and proceed to conceive some child or another – the child we’ve named ‘Molly’ – into existence as a nonslave. So it must be something like – oh, I don’t know – feigning to sign the slave contract and then having sex – actually having sex – on whatever schedule they anticipate, at that critical time just prior to performance, they will have sex once they sign the contract. Keep in mind, too, that prior to performance that anticipated schedule – the schedule the agents anticipate will unfold given that they will enter into the contract and produce some child or another – is itself very loose. For all the agents know prior to performance, conception, under D, might take place any time over a period of a week. And the exact timing will be critical in determining whether Molly, or someone else, is conceived. But precisely the same is true for E. The agents, prior to performance, have no clear idea what the precise timing of conception will be under E. Indeed, prior to performance, the timing of conception may not even exist as a determinate fact!

Look, all I’m really trying to say is that the ‘precariousness’ point cuts both ways: however unlikely Molly’s existence is under D, it’s just that unlikely under E. There’s nothing special about E vis-a-vis Molly’s existence. The two are basically independent events.

Rosen: The fact that Molly miraculously makes her way into existence after E is performed doesn’t imply, in other words, that E counts as any sort of special causal factor for Molly’s coming into existence.

Jeremy: I think it get it. As of that time t just prior to performance, there are a zillion ways the future might unfold, given that the couple decides not to enter into the contract but still produces a child. Only some of those ways eventuate in Molly. But as of t, there are equally a zillion ways the future might unfold, given that the couple does enter the contract. But again, only some of those ways eventuate in Molly. Looking at things prospectively from t, we therefore have no basis on which to think that it’s more likely Molly will come into existence if the couple does E than if they do D.

Richardson: Lamp, you’ve got to go to class. You’re already late. But we need to talk more. This is interesting stuff. If you’re right, we can after all discern harm to Molly in the slave child case. But even if we find harm in this one case, there are a whole slew of variations on the NIP that you’ll need to deal with as well!

Lampier: No way. The nonidentity problem is your baby, not mine. (Lampier looks at his watch and jumps up.)

Richardson: Mainly, cases where the child truly cannot exist and not suffer. That happens, for example, in cases where a baby is born with some chromosomal or genetic defect that we – at least at the moment – have no way of repairing.

Rosen: Depending on the severity of the condition, we call those ‘wrongful life’ or ‘wrongful disability’ cases over in the law school. They come up most often when a doctor or fails to inform a woman of her risk, given her own family history, her age and so on, of producing a disabled child or fails to recommend appropriate prenatal testing. Or when a lab misreads the amnio results and tells the woman the fetus is healthy when it’s not. Basically, by not letting the woman know there’s a problem, the defendant makes it impossible for her to make the right decision on getting pregnant to begin with or on abortion. Obviously, the doctor’s negligent omission doesn’t cause the disorder – and in fact seems not to harm, but if anything to benefit, the child. But you still feel that a wrong has been done – and you’d like there to be a cognizable cause of action.

Richardson: Exactly! Even if the slave child case evaporates as nonidentity problem, wrongful life and wrongful disability cases still seem to present scenarios in which the agent has done something ‘bad’ that is ‘bad for no one’ – something that is wrong but that harms no one.

Rosen: A wrong but no person harmed? No claimant? No plaintiff? No client? No contingency fees? No possibility of punitive damages? That’s a way of looking at things that we lawyers will want to avoid at all costs!

Richardson: Well that, my dear Miriam, is the nonidentity problem! But Lamp’s point about the slave child case is an interesting one! I’ve got to get back to my office to, well, transcribe our luncheon conservation! Perhaps Lamp and I can write a paper!

Rosen: You’re awfully cheerful for someone who may have to abandon some very deeply held views. Chris, you’ve been advocating the slave child case as an argument for totalism for years!

Richardson: My dear Miriam, as a philosopher I can’t afford to be clingy about my ideas! One always has to be willing to go back to the drawing board! But you must know, too, that one lunch does not a proof make. I’m serious about going back to my office and writing all this out. You just can’t believe how often it’s happened to me that something that sounds, well, beguiling has turned out to be mistaken in some way. You stare at something long enough, and it often starts to disintegrate. And I still feel that there’s something about the couple’s entering the contract that gives the slave child – gives Molly – a leg up on existence. I just do not think, at the moment at least, that the slave child argument really commits the fallacy of post hoc ergo propter hoc.

Rosen: Hmmm. I don’t know. After all, it’s part of the hypothetical that the future has already unfolded. It’s so easy to imagine, then, that one act leads to another, one event to another, in accordance with some internal drive, or logic, so that entering into the contract eventually unfolds into the existence of that particular child rather than any one of a vast army of other possible children. But expected value surely has to be calculated – surely Lamp is right on this point – just prior to performance. At least for lawyers, the whole point of the expectational calculation is to see if defendants – doctors, whatever – were negligent in taking the course of action they opted to take. But that means looking at things from the defendant’s point of view prior to performance. To the extent that we can see only in hindsight that the act they chose created some disaster, you’re outside of traditional negligence law.

This feeling you have that entering into the slave child agreement makes the child’s existence more likely – you may be importing the information you have about how the future in fact unfolds into your assessment of the chances that the particular child will exist, given that the couple enters into the contract. When we know something, it’s always hard to remember that we are supposed to pretend that we don’t. Or even worse, think as though we don’t. But to do the expectational calculation correctly, that’s just what you have to do.

Richardson: This is good stuff, isn’t it? I need some time to think. But let’s talk again soon. Monday? Lunch? I’ll call Lamp. And Jeremy, too. In the meantime, I managed to get Lamp off to his class but completely forgot I had one of my own. Let me go throw myself on the mercy of my students – via email, at least, since they’ve undoubtedly all wandered off by now!

[The issue of whether Mill is an aggregationist is revisited; a person-affecting form of consequentialism (‘personalism’) is mentioned in passing, in connection with Pareto principles and Rawls’ doctrine of leximin.]

Mark: The repugnant conclusion convinces me that Mill’s approach, after all, wasn’t much of an improvement on Kant.

Rebecca: But you don’t have to read Mill as a totalist. At least, nowhere does he say that what we ought to do is create the greatest total happiness – or even the greater average happiness – in the aggregate. Maybe, instead of an aggregationist, he’s a Paretian.

Mark: You means, he’s more of a continental philosopher than a British empiricist – he’s some sort of phenomenologist, an existentialist or something?

Jeremy: Rebecca, believe me. If you learn nothing else at this place, learn never assume people know stuff. You just cannot believe how much stuff people don’t know.

Rebecca: Fine. I like to talk. This I get from my economics course. The early twentieth century economist Vilfredo Pareto – Italian, not French – is famous for this nice, tidy little set of principles – the ‘Pareto’ principles. Their gist is that if agents can make one person better off without anyone else any worse off, then they are wrong when they fail to do that. It seems to me that his principles are perfectly consistent with Mill’s idea that we should create has much happiness as we can for people – but as individuals as opposed to in the aggregate.

Shelby: Yeah. Rebecca and I were trying to write this out last night. I need something to write on. Where’s my napkin? Oh! Here’s one!

Rebecca: Tsk, tsk, Jeremy. Be good. Let Shelby write it out for you. You’ll be interested in this. Really.

Shelby: That’s all right, Rebecca. I don’t really need to write it out. The idea is really just what Rebecca’s said. Instead of thinking of the beneficiaries of our choices as a single massed person, just think of them as the individuals that they are. Think in terms of maximizing happiness for each one.

Shelby: Right. But sometimes you as agent face conflicts, and find that you cannot maximize happiness for one person without harming someone else.

Shelby: But in the meantime you shift the quantifier over persons to the left to give it broad scope. Instead of saying that the agent’s obligation is to maximize wellbeing for all persons in the aggregate, we say that for all persons – each and every one of them – agents ought to maximize wellbeing for that person. And you make sure that the domain of the quantifier includes not just existing persons but future persons as well. You maximize, then, for each one –

Rebecca: – when you can. But sometimes there are tradeoffs. Sometimes to avoid imposing still more harm on one person, you have to not maximize wellbeing for – i.e., harm – another.

Mark: You guys are forgetting about the slave child case. The agents there have maximized happiness for the slave child, but they’ve still done something wrong. You’re also forgetting the minor detail that Mill commits himself to a principle of aggregation. The last footnote in Chapter V; the note on Spencer. Mill writes that ‘equal amounts of happiness are equally desirable, whether felt by the same or by different persons.’

Jeremy: Well, actually, guys, Richardson says he’s going to ‘shake the Etch-a-Sketch’ on the slave child case at the next common lecture. And I think Lamp’s going to be making a guest appearance.

Jeremy: Shelby, I don’t want to hear any complaints from you. You’re getting what you said you wanted. We’re going to be spending more time on the slave child case after all. Of course, even if we can manage to find harm in the slave child case, there are plenty of other types of nonidentity problems.

Rebecca: And, Mark, about that footnote. Mill may not be committing himself to aggregation at all there. He might just be making the uncontroversial point that a unit of happiness is just as desirable to a person who has more as to a person who has less. It’s not going to follow from that fact that giving the unit of happiness to the person who has more is morally indistinguishable from giving that unit to the person who has less, or that giving the unit of happiness to the person who has more rather than the person who has less is a morally permissible choice.

Jeremy: I don’t know, Rebecca. I’ll have to look back at the text. In any case, you two have an even bigger problem. Tradeoffs. I mean, figuring out what we ought to do would be a piece of cake – the whole field of normative ethics wouldn’t even exist, in fact – but for cases in which people’s needs and interests are in conflict. Isn’t, in other words, the tradeoff case basically the standard case? Your idea of Mill as a Paretian –

Rebecca: We call it ‘personalism’ – versus totalism, you know. Or averagism or pluralism.

Jeremy: Very cute. Your idea of Mill as a ‘personalist’ – as a person-affecting maximizer – seems to have a clear result only in the case where we really have no need for a clear result.

Shelby: Sure, if you don’t consider procreative choice – including abortion, just to mention one example – to be a case where we need a clear result. God, Rebecca, maybe they want to keep us permanently confused on anything having to do with procreation?

Shelby: Hey, Emily Post. I’m kidding. Jeremy’s a modern man. I trust him to be totally sensitive when it comes to understanding how important procreative issues are to women.

Shelby: I know. I am. But Jeremy, think about it. Procreative choice is hardly the only context where totalism and the other theories completely mess up. Almost any ‘extra person’ case where the choice is between making people happy and making happy people seems to create more of a challenge for all your little aggregative ‘isms’ than it does for a person-affecting approach.

Jeremy: O.K., so in effect what you’re done is to give an account of procreative choice, and ‘extra person’ cases in general, according to which there is no genuine tradeoff. By doing so, you get a quick route to a clear result.

Shelby: Right. You can’t harm someone who has never and will never exist at all.

Jeremy: At a world. In any morally significant way. Right. But now you need to deal with the case where, under your own view, you do have a genuine tradeoff.

Mark: Well, Rawls might say that you first do whatever you need to do, between two persons whose needs and interests are in conflict, to help the less well off.

Jeremy: That’s helpful, in the sense that it provides us with at least a partial rule. It’s what they call ‘leximin.’ On the other hand, though, just think about what it says in the extreme case. Imagine a world where the least well-off is someone – say, Billy – who could be helped by some tiny amount if some better off person sacrificed a vast quantity of wellbeing; and Billy could be helped by just a tiny bit more if some second better-off person were made to sacrifice a vast quantity of wellbeing; and so on.

Shelby: And on and on. I get the point. Now, in some cases, vast sacrifices like that will lead to lots of people just as bad off as Billy is. While our view doesn’t aggregate, there’s no reason why it can’t count numbers of badly off people. Between a choice that makes only Billy as badly off as Billy is at his worst, and a choice that makes a lot more people as badly off as that, the former is the correct choice. But not all cases will be so convenient. There will be more extreme cases where, in order to make the very badly off Billy just slightly better off, vast numbers of people will have to have their own wellbeing levels reduced drastically.

Rebecca: Jeremy, you’re imagining that transfers of vast wellbeing from large numbers of persons to poor Billy will do Billy some good but not much. You’re really imagining a case, then, where, for very little gain, wellbeing is simply destroyed.

Jeremy: Yeah. It seems to me leximin will, in some cases, prescribe just such a destruction.

Rebecca: But in those extreme cases, won’t there always be some other, more efficient way of helping Billy? I mean, if wellbeing is just the ‘capability to achieve functionings,’ as Sen suggests – well, wouldn’t there always be some way to increase Billy’s capability that did not involve such a whole-sale destruction of wellbeing?

Shelby: I think we have to go back to the drawing board, Rebecca – or, rather, the blackboard. At least, I’m not sure that I, at the moment, have all the relevant cases in mind.

Jeremy: Yeah, well. See what you can come up with. Better to have an incomplete theory, even a highly incomplete theory, than an obviously false –

Shelby: – My God! It’s eight o’clock! I’m sorry! I’ve got to get out of here! Haven’t gotten one bit of work done all day!

After the section meeting, which Shelby did not attend, Jeremy speaks to Rebecca.

[Conversation between Jeremy and Rebecca regarding Shelby’s behavior and attitude.]

Jeremy: You need to tell your friend to stop missing class. She wasn’t here today, and she missed Richardson’s lecture. At least, I didn’t see her there. And when she does bother to show up, she’s late. And she’s rude.

Jeremy: I will. I have. But listen, it doesn’t help her to rush in, issue weirdness edicts, and then just disappear for a week or so. It just alienates people.

Jeremy: I mean, you can’t just waltz in and trash something and waltz back out.

Rebecca: But Richardson himself seems now to be saying that there’s something weird about the slave child case.

Jeremy: Granted. Nonetheless . . . . Look, Rebecca, it really is beginning to seem to me that some of our best thinking goes on ‘below the radar.’ It’s just that we can’t trust it until we can articulate it and test it.

Jeremy: Well, there’s orderly articulation – and then there’s sabotaging the conversation.

Rebecca: Whatever. Anyway, there’s a lot I can’t articulate to you right now about Shelby.

Jeremy: Fine. I hope, then, to see both of you at Richardson’s lecture tomorrow.

[The student’s medical situation and need for a donor organ are revealed to the T.A.]

Richardson: So it seems that our mutual student, Shelby Ratliff, is a pretty sick young woman.

Jeremy: She’d been missing some section meetings but still very engaged with the discussions. If anything, all too engaged. Well, I guess she does look thin, and I guess pale. And she’s definitely frazzled. But they all look thin and frazzled. But I should have realized that something was wrong. One of her friends hinted around that there was a problem.

Richardson: Listen, Jeremy, not to worry. Shelby’s had superb medical care all along. That’s part of why she’s been missing some classes. No one, I can assure you, thinks there’s a thing you could or should have done differently. In fact, the reason I called you in was just to tell you she’d appreciate you dropping by the hospital this evening. That friend of hers – Rebecca? – is going over, too.

Richardson: Well, you see, there are some unusual factors, apparently. And there just aren’t enough organs in the donor pool to make it very likely that a good match can be found over the next few days. It’s ironic. Tommy Ratliff, her father, is about the richest man in Arkansas. Maybe in the whole south. He could probably buy Arkansas. I knew him quite well at Oxford years ago, where he was study for some kind of masters degree when I was finishing up the doctorate. Tommy met with some of the hospital’s board just yesterday. One of the board members suggested the possibility of trying to locate a donor on an international basis. The money’s there to stir up a considerable amount of interest.

Richardson: Well, it can be done very discretely. Very arm’s length. The problem is Shelby will not hear of it. Paying for donor organs seems problematic, to her. She thinks that a market in organs, harvested from accident victims, can’t be kept separate from a market in live donor organs. Or that, at least, she’ll have no way of insuring that the two really have been kept separate. She’s not going for it. She refuses to sign the consent. But, look, Jeremy, she’s asked to see you. Perhaps, well, you could talk to her about this. I know her family would appreciate it.

Richardson: Well, you do what you want. In any case, Jeremy, when you get to the hospital, go up to the cardiac intensive care unit on the third floor. You’ll find her there.

At Shelby’s bedside in cardiac intensive care unit are Jeremy, Rebecca and Mark.

[A discussion of the ethics of markets in donor organs – and in particular in markets for live donor organs – takes place at Shelby’s bedside. Among other things, the question of whether consequentialism requires, or permits, the more well-off to buy organs from the poor is debated.]

Rebecca: Yep. He didn’t know. I didn’t say a word. As per your instructions.

Shelby: Sit down! Try to be comfortable. Just ignore all the machinery. And the beeping. And nurses and etc. Mark has been a doll. He’s stayed with me all afternoon. We had to be little pushy to make that happen, but it did.

Mark: Look, I’m going to walk downstairs for some coffee. Can I bring you two anything?

Shelby: If you don’t do it, Mark, then I’ll just walk down and get it myself.

Shelby: Thank you, Jeremy, really for coming. I wanted to talk to you about a problem. A moral problem. And since I’m so sick, I know for once you’ll listen to me!

Jeremy: Shelby, I don’t anything about the specifics of your situation. But, well, your father talked to Richardson. Richardson told me that you refused to consider a heart transplant. Now, I spent the afternoon doing some research on heart transplants. They’re really getting better and better at these things. The five-year survival rate is well above fifty percent.

Shelby: It’s somewhat less if you’re missing the heart to transplant, Jeremy! And I am not going to let Daddy get involved in ‘finding’ one for me somewhere in India, or Mississippi.

Rebecca: Shelby, I agree with you that a profit-driven market in donor organs is not a wonderful thing. But it isn’t terrible, either. It would help you. And it might very well help some poor family whose just lost their husband and father – and who might well starve but for your father’s willingness to pay well for a donor heart. I mean, seriously, you know the kind of poverty that exists in some parts of India, and Africa.

Shelby: O.K. You want to talk about a donor heart? Daddy’s rich, it’s true. I’m sure he could find me a heart that was a perfect match – somewhere, somehow. No doubt about it. He’d pay $1000, $10,000, $100,000, whatever it took, to save my life.

Shelby: Mark, I was just saying that letting Daddy try to find a heart for me outside the U.S. donor organ system is not a good solution. For one thing, there’s absolutely no guaranty – in fact, the probabilities are on the other side – that the heart would come from some poor slob who’d just died in a car wreck or some other sort of accident.

Jeremy: You think, what, that some guy is going to shoot his wife, or his father, or something, and then claim there was an accident, all in the hope that he can make some money selling the heart on the black market?

Shelby: Well, I wasn’t actually thinking about murder. Though it’s a possibility. I was thinking about the lengths people will go to for their families. I mean, as Beck’s just pointed out, there’s a lot of poverty in India. And in Africa and elsewhere. I think a lot of people would willing to serve as live donors and, basically, sacrifice their own lives in exchange for $10,000 or so to keep their children from starving or even just to give them better lives. And I can promise you Daddy’s forgotten whatever ethics he learned at Oxford. He’d pay whatever it took – and justify the transaction on the grounds that the poor Indian has a right to sell his heart on the free market for whatever he can get, and that all we’re doing in taking that heart is respecting that right.

Mark: Well, I suppose there’s something to that. After all, the transaction is freely entered into, and it’s better for both parties to the transaction – better for both you and the Indian, or at least his family – that it go forward than that it not . . . .

Shelby: But since when is an act’s having those features sufficient for moral permissibility? I mean, it’s just not plausible that aggregate wellbeing is increased by having such a transaction go forward. It’d be far more efficient to let me die – and keep in mind that it’d be nice for me to die in peace, by the way? – and have Daddy just give the guy the $10,000 or whatever. The money would mean no more to him than the fingernail clippings his manicurist sweeps away. And if you take a person-affecting route, things get even worse. You then lose the wiggle room to speculate that my continued existence will produce, at the end of the day, at least as much or more aggregate wellbeing than the Indian’s. On a person-affecting approach, you have to address tradeoffs up front. But making the rich richer, and the poor poorer, is surely not the way we want to go on that. And I’ve already had this really great life.

Rebecca: But Shelby, you know very well your father’s not going to go out and write some poor guy who isn’t willing to donate his heart to you a check for $10,000!

Jeremy: But that’s beside the point, Rebecca. Shelby’s right: the morally relevant point is that there’s nothing that stops her father from doing just that. Consequentialism does not work by comparing the act you want to perform against whatever other act you would have performed instead. You have to consider all the alternatives the agent has at the relevant time. The obligation, then, is to perform that which creates the most good. That’s so, whether you’re a totalist, or a pluralist, or whatever.

Shelby: Just like in the slave child case, where it doesn’t matter that the couple would not have had any child had they not had the slave child. What matters is what they could have done.

Rebecca: But what about the poor guy whose children will starve? Are you saying he does something wrong if he agrees to serve as a live donor and chooses to sacrifice his own life for the sake of helping his children escape poverty?

Shelby: No, but it doesn’t follow that I do something right when I pay him to do so. Look, if you’re going to talk moral theory, you’ve got to talk moral theory in a careful and serious way. It’s not some verbal weapon that you wield as a pretext for getting what you want. Our aim in life is not to sophistry – right, Jeremy?

Anyway, I don’t really care. There are plusses and minuses to a heart transplant. And, Jeremy, you haven’t noticed, but I’m really smart. Being on heart-lung machine for literally hours can impair cognition and memory. And you have to take drugs forever to prevent tissue rejection. And you have to take other drugs to prevent clotting, which can cause you to bleed out overnight any time. Look – I signed the damned consent! But I wrote in plenty of provisos. They can’t get close to me with any heart that hasn’t come through all the proper channels. But I don’t really want a heart transplant. At least not that badly. Not particularly. I have had this incredible life. And, Jeremy, I strongly advise you to get one.

Nurse: She’s stable, for the moment. It’s an intermittent arrhythmia. Look, she needs to rest now. But she’s asked whether you can drop by again tomorrow.

Mark: You two go on. I’m just going to hang around the waiting room for awhile.

Rebecca: All right. But look. I’ll be back tomorrow morning, early. Call me, though, will you, if anything changes, all right? You have my cell.

Rebecca: I’ll come with you, Jeremy. But look, you need to stop agreeing with everything Shelby says.

Jeremy: And you need to stop arguing with her about a heart transplant. She not saying stupid things, and you should leave her alone.

[The question is discussed whether the student, whose heart failure seems to be genetically based, should agree to donate her eggs to her sister, who is infertile. While it is true in such a case that the child who may result from those defective eggs is perhaps not harmed – that child, after all, could not exist and not face the medical risk – the concern is raised that bringing that disabled child into existence will predictably result in burdens – harms – to others since substantial resources will need to be diverted to that child, given his or her special needs.]

Shelby: I washed it! I’m not going to lie in this bed with dirty hair. I told the nurse that if they didn’t unhook me from all their little sensors and shit that I’d just get into the shower with them on. I guess their choices were to tie me down or let me do what I wanted. And you know Daddy’d have a cow if they tried to tie me down. Hand me a dry towel, will you? And could I have the floor for just a moment? Without being shouted down? They’re going to be taking me back into surgery a little later this morning – just some procedure they say that have to run – so I don’t have much time.

Shelby: O.K. You thought my moral problem has to do with whether I should consent to a heart transplant, right? It doesn’t. I’m having another kind of organ donation problem – one that also happens to be a nonidentity problem. A really hard one! Look, I have already signed the consent to have most of my organs donated to whoever may need them. Maybe they’ll even be able to dissect my heart and learn something useful. It’s my ovarian tissue I’m worried about. And my sister, Jenna. She’s a lot older than I am, close to forty. We finally managed to get her married a couple of years ago but she hasn’t been able to conceive. To put it bluntly, she wants my eggs. But I don’t think I can give them to her. My germ line is clearly a mess. This heart defect I have is hereditary – my biological child will have a substantial chance of getting it. And I’ve really liked my life, but it’s been hell on my parents and Jenna for me to have been so sick, really, since I was a child. And my dying will be hard on them, too, though at this point they expect it. I just don’t want Jenna to have to go through that all over again with her own child. I mean, there are plenty of other potential egg donors out there!

Jeremy: So you think Jenna, instead of using your eggs, could perhaps arrange for donor eggs through a fertility clinic. And avoid any real risk of having a child with a serious genetic disorder.

Shelby: Exactly! Jenna, though, says that she wants the genetic connection. I mean, I’m sure she’d adore the offspring of one of those gorgeous Princeton women with their high SATs. But she thinks the family traits are worth something, too. She imagines having a child who reminds her of our mother, or our grandfather, or even me.

Rebecca: Well, this case is very different from the slave child case. Here, there’s absolutely no chance that a child produced from your eggs who does happen to be born with the heart anomaly could possibly exist without it. So Jenna’s having a child using your eggs is certainly not going to harm the child.

Jeremy: But you’re worried about Jenna, right? Is the defect one they can screen for at this point, Shelby? Is pre-implantation genetic diagnosis an option?

Shelby: I wondered about that, too. But apparently they haven’t located the exact gene sequence for my particular problem. And more than one gene may be involved. We are not even sure what the precise risk of transmission is.

Mark, so, what do you think? What does your Kantian gut tell you about this issue?

Mark: Hmmm. Well, I think I need to know more about your sister. Would she be a good mother, even if the baby is born with a defective heart?

Shelby: I have no doubts there. She’d be the best. And her husband is really not a bad person, either, though we’ve definitely had our moments. And, as I said, there are no other children.

Rebecca: It seems to me that in this case the only possible victim is the child. And since the child can’t otherwise exist, I think you should sign the consent. This would be your child, Shelby, and to bring this child into existence would be a wonderful thing. For Jenna. And for you. And for me, too.

Shelby: Rebecca, I think you’re forgetting your own motto: ‘an unexamined emotion is not worth having.’

Rebecca: (Pulls self together.) No, my heart and my head are together on this one.

Jeremy: Shelby, honestly, I don’t know. I understand why Jenna wants to use your eggs. She loves you, and will love the child. And the baby will, if he or she is anything like you at all, will be a terrific pain but also have a blast. Heart defect or not. On the other hand, I admire your realism. Maybe in the end this is one of those cases where you have to take a blind stab at what you think is right and hope for the best.

Shelby: I am worried about Jenna. But she’s a big girl. Maybe I should let her make the decision for herself. (Takes pen from Rebecca.) But it’s my decision, too. And the baby . . . . How do you do what you’re supposed to do for a sick baby and for all those other people you’re supposed to worry about as well? Whose interests to you set aside? The baby’s? Surely not. But it won’t do to set aside anyone else’s, either. (Puts pen down.)

I’m feeling a little woozy, and y’all don’t look so good either. Could someone please call the nurse?

Rebecca: What’s wrong? There were no alarms. But she’s so weak – she can hardly lift the pen. And so pale.

Nurse: She’s failing rapidly. And she’s taken off the sensors, so we’re not sure exactly what her status is. Look, she’s got a hard morning ahead of her. The best thing is for you to go home. Why don’t you check back with me later?

(Hurries off. Jeremy and students move toward waiting area for the cardiac unit. More activity in corridor. Then, huge stir at entrance to cardiac unit. Medical personnel rush in.)

Rebecca: Look – what do they have there? An ice chest? Wait a second – it’s got to be a heart. They have a heart for her!

Jeremy: God, she’ll never consent, not if there is any question her father has somehow procured this heart out of the country. She’s not going to have anything to do with a black market heart.

Attendant: This heart’s not black market. We’ve worked like hell to get this baby up here from Pittsburgh! Just arrived by helicopter! Bad car wreck early this morning. But one guy who was killed – this eighteen year old kid! – had signed a donor card and appears to be a match. The doctors won’t have told the patient yet that we have a potential heart – too much

uncertainty before the tests are completed; it’s just too much of a roller coaster for the patient – but they should have already started prepping her for the surgery.

[1] This dialogue owes much of its identity and certainly its coming into existence to Derek Parfit, Reasons and Persons (Oxford University Press, 1984), pp. 351-79 (‘The Nonidentity Problem’), and John Perry, A Dialogue on Personal Identity and Immortality (Hackett, 1978).

[2] The useful ‘on the list’ locution is Elizabeth Harman’s. ‘Can We Harm and Benefit in Creating?’ Meetings of the American Philosophical Association, Eastern Division (2003)

[3] Roger Scruton, ‘Eating your Friends: a carnivore's credo,’ University Center for Human Values, Princeton University, November 9, 2005 (quote approximate).

[4] Posner overlooks this point in his argument in favor of the court’s enforcement of contracts for surrogate motherhood. See Richard A. Posner, ‘The Ethics and Economics of Enforcing Contracts of Surrogate Motherhood,’ Journal of Contemporary Health Law and Policy 6, no. 21 (1989): 21-31.

[5] The ‘repugnant conclusion,’ like the nonidentity problem itself, has been made famous by Parfit. See Reasons and Persons, pp. 381-90.