I. Deprivation accounts and their counterintuitive consequences

In its most basic form, the deprivation account of the badness of death is the view that death is bad because, and to the extent that, it deprives a person of goods she would have gained had she not died when she did. The more net good death prevents the deceased from gaining, the worse death is for that person. To determine how bad one’s death is for the one who dies, the deprivationist needs some way to measure the net level of goodness of which death deprives its victims. Here I focus on Ben Bradley’s contextualist deprivationism because I think it the most plausible and well-developed account, although my argument is also applicable to other (viable) deprivation accounts. Bradley measures the amount of goodness of which death deprives its victims with the difference making principle, formulated as follows.

In other words, the value of an event depends on what would have happened in the nearest possible world where the event had not occurred. By looking at the most likely counterfactual alternative where an individual does not die, we can see exactly what death deprives its victims of and, consequently, measure how bad each person’s death is for them.Footnote ⁴

DMP appears promising. It nicely accounts for our intuitive judgements about standard cases of death. Consider, for instance, an otherwise healthy 10-year-old boy who contracts the Ebola virus and dies. One reason people are disposed to judge this child’s death to be bad for him is because he presumably would have lived a good life for many years had he not contracted Ebola. Compare this young boy’s death with that of a 90-year-old man. The young child’s death seems worse for the child than the elderly man’s death is for the elderly man. If the elderly man did not die of Ebola, he presumably would have died relatively soon after. Again, DMP gets the intuitively right result (i.e. the child’s death is worse for the child than the elderly man’s death is for the elderly man) and for the right reasons (i.e. the net good the child’s death deprives the child of is greater than the net good the elderly man is deprived of by his death).

DMP’s explanatory power also extends to common sense judgements about when death is good for the dying. Death is generally considered good for the dying in cases where the dying persons’ quality of life, were they to continue to live, would be worse than death. Imagine a case where a person in the final stages of amyotrophic lateral sclerosis (commonly known as Lou Gehrig’s disease) initiates physician assisted suicide. Intuitively, this person’s death is good for him because it prevents him from experiencing more of the overwhelming agony caused by amyotrophic lateral sclerosis. These considerations provide strong reason to accept DMP or some sufficiently close principle.Footnote ⁵ However, principles like DMP coupled with NBNL generate an absurd conclusion largely overlooked in the literature.Footnote ⁶ This consequence is best illustrated in the case of Sick Suzy.

As a result, according to deprivation accounts, Suzy’s death is no longer bad for her. In fact, the net value of Suzy’s death is positive since the nearest possible world in which she recovers is one where she subsequently endures many years of suffering worse than death. However, discovering Ned’s actions should bring no comfort to Suzy.

If this is not already intuitive, consider the following. Suppose that prior to acting, Ned bursts into Suzy’s room and says ‘I can’t secure a donor heart for you, but I can do the next best thing; hire someone to torture you for the remainder of your natural life if you get a donor heart.’ It would be quite odd were Suzy to reply ‘Yes, please do. That would be very comforting.’ After all, hiring the torturer wouldn’t make any difference to Suzy’s well-being. It wouldn’t affect her (actual or expected) total well-being or her level of well-being at any point in her life. Ned’s hiring of the torturer is counterfactually dependent upon Suzy not recovering. So, if she were going to get a heart transplant, Ned would not have hired the torturer. Moreover, whether or not Suzy would continue to live a good life were she to survive, she still could continue to live a good life were she to survive. Ned’s actions do not change that fact. Since hiring a torturer would not in any way affect Suzy’s well-being or prevent it from being possible (in any sense) for her to continue living a good life, Suzy should not be comforted by Ned’s actions. Generalizing from Sick Suzy, if DMP and NBNL are both true, then everyone could escape the evil of death by hiring counterfactual torturers. Everyone’s actual death could be turned into something that is good for them and consequently, no one should lament their own death. But, for reasons given above, this is wildly counterintuitive.

Inverse Sick Suzy cases can be constructed to illustrate the same point. Consider

Betty’s enemy has not given Betty any reason to change her attitude towards her death. Assuming that Suzy and Betty are similar in all other relevant respects, Suzy and Betty should lament their death in the same way and to the same degree. After all, we may suppose that each have the same expected lifespan, justified beliefs about the best possible lives they could lead, identical levels of well-being at any given time, same narrative structure and same total well-being.

One might try to resist these conclusions by appealing to a counterfactual comparative account of plural harm. Neil Feit develops the best account of plural harm, which can be formalized as follows.

According to QPH, although Suzy’s death is no longer bad for her, [Suzy’s death and Ned’s hiring a counterfactual torturer] is bad for her. This is the smallest plurality of events that make Suzy worse off than she otherwise would have been. This seems to be the correct result. At least, if both of these events did not occur, Suzy would be much better off. QPH identifies harms in a way that comports with our intuitions more frequently than alternative accounts. For this reason, among others, it is a very promising view.

Pairing deprivationism with QPH and NBNL would allow that Suzy still has something to lament. Given this, it may seem that QPH could be employed to resolve the tension between deprivationism and NBNL. This strategy, however, will not work. First, note that it doesn’t help in Betty and the Billionaire. Like DMP, QPH (correctly) entails that Betty’s enemy makes her death really bad for her. The smallest plurality of events that makes Betty worse off than she otherwise would have been is still her death. But, for reasons already given, Betty’s enemy does not give Betty any reason to change her attitudes toward her death even though she makes Betty’s death much worse for her. Second, I think a case in which death (as a result of a plural harm) preempts another death generates the same counterintuitive consequence that Sick Suzy did for DMP.Footnote ⁸ There are also multiple plural harms in such cases, which raises the question of which plural harms, if any, our attitudes should track. Answering these questions requires rejecting NBNL or so I argue. The lesson from Sick Suzy and Betty and the Billionaire cases, then, is that we must either reject (a) deprivation accounts or (b) NBNL.Footnote ⁹ In the next section, I argue in favor of the latter option.

II. Sick Suzy’s death is good for her

In this section I argue that, contrary to common sense, deprivation accounts get the correct results in Sick Suzy cases. A surprising consequence is that how bad one’s death is for the one who is dying and how much one should lament her death come apart. That is, NBNL is false. This opens up the possibility that we should be indifferent to deaths that are very bad for us and lament deaths that are good for us. Though DMP’s consequence that Suzy’s death is good for her initially seems counterintuitive, this prima facie implausibility can be overcome once we consider a revised Sick Suzy case. Before I give the revised case, a brief digression is necessary. This is because the correct judgment about the revised case appeals to a moral principle I call Prevent the Worst Death.

S’s obligation in PWD is generated from her moral reasons and, as is likely already assumed by the reader, PWD is true because ceteris paribus the fact that x’s death is significantly worse for x than y’s death is for y provides S with more moral reason to prevent x rather than y from dying. Before showing that PWD can be derived from a nearly incontrovertible moral principle, a few clarificatory points are in order. First, PWD can account for cases in which the process of dying is painful for one person, but not the other. As is standardly done in the death literature, I distinguish between the process of dying and death itself. Following Rosenbaum, we may understand dying as ‘the process whereby one comes to be dead or the process wherein certain causes operate to bring about one’s being dead’ (Reference Rosenbaum1986, 217). Death may be understood as the ‘time at which a person becomes dead, ’ which may be an instant in time or no time at all and being dead follows death (Reference Rosenbaum1986, 218). The crucial point is that dying is part of the persons’ life and may be experienced whereas death and being dead are necessarily non-experiential.

So, if dying is painful for x, but painless for y, then all things are not equal and PWD allows that x should be saved over y even though y’s death is substantially worse. The ceteris paribus qualifier renders PWD consistent with any normative ethical view on which there is a morally relevant distinction between an act and an omission. The ceteris paribus qualifier guarantees that PWD is remarkably weak since it makes no claims about what the other morally relevant facts are. Hence, it is consistent with any plausible form of consequentialism, virtue theory or deontology. It is also neutral with respect to the various accounts of the badness of death and so does not beg any important questions in the death literature.Footnote ¹⁰

Why accept PWD? It is entailed by a more general, quite plausible, moral principle.

PPCH initially appears to be an incontrovertible moral principle, but there is at least one small issue with this formulation. The prevention of a benefit and the infliction of physical pain are both harms according to DMPH. There is nothing in PPCH that provides a way to distinguish between the two. So, if preventing x from winning the lottery harms x significantly more than cutting y’s leg off and z can only prevent either x or y from being harmed, PPCH allows that z is obligated to help x. I am personally happy to accept this consequence, but I recognize that it is a controversial one. Fortunately, for the purposes of this paper, we can sidestep this issue by restricting PPCH such that it only concerns the same kinds of harm.Footnote ¹² Since death can only be a deprivational harm, I amend PPCH as follows.

It is hard to see what could lead one to reject PPCDH. It retains all of the virtues of PPCH and can accommodate DMPH without any controversial commitments. There is also good reason to accept PPCDH. Countless common sense moral judgements (at least implicitly) appeal to it. Here is a rather extreme example. Suppose you can either save one person from falling into a coma for a decade or another from falling into a coma for 10 min and there are no other morally relevant differences between the two options. You morally ought to save the person who would be in a coma for a decade and the best explanation for why appeals to PPCDH. If PPCDH is true, then so is PWD since PWD is simply a narrower version of PPCDH. PWD specifically concerns the deprivational harms brought about by death instead of deprivational harms in general.

Recall that PWD is supposed to help vindicate DMP, though it is probably not yet clear exactly how it is supposed to do that. The short answer is that one can see the explanatory power of the DMP once Sick Suzy is revised in such a way that (a) multiple people are dying and (b) S is in a position to save just one person from death. Consider

To whom ought the doctor to offer the heart transplant? The answer should be clear. It should be offered to Jenny because Suzy would be tortured for the remainder of her natural life if she receives it. Jenny, however, would go on to live a long happy life if she receives the heart transplant. Common sense judgements about Sick Suzy vs. Sick Jenny appeal to DMP coupled with PWD. At least, that is the best explanation of why it seems so clear that the doctor ought to save Jenny.Footnote ¹⁴ We have good reason then, to accept DMP’s consequence in Sick Suzy, which in the context of Sick Suzy vs. Sick Jenny, should no longer seem counter-intuitive. Jenny’s death is much worse for Jenny than Suzy’s death is for Suzy, but their deaths are equally lamentable. By stipulation, if Ned did not hire a torturer, Suzy and Jenny’s deaths would be both equally bad and lamentable. For reasons previously given, Ned’s hiring the torturer does not give Suzy any reason to change her attitude toward her death. So once Ned hires the torturer, he succeeds in making Suzy’s death good for her, but fails to make her death less lamentable than Jenny’s.

One might deny that Suzy’s death is worse for her than Jenny’s death is for Jenny, but maintain that the doctor is obligated to save Jenny. This option might appeal to contemporary EpicureansFootnote ¹⁵ and proponents of categorical desire views, Footnote ¹⁶ though it raises the question of why the doctor has an obligation to save Jenny. Proponents of DMP and PWD have a good answer, whereas it far from obvious that those who reject DMP do. Sick Suzy cases reveal that there is good reason to accept standard deprivation views (that appeal to DMP), so NBNL should be rejected. This surprising consequence raises a new question. Which facts about death should our attitudes track? I attempt to answer this question in the remaining sections.

III. Which attitudes should we have toward death?

Cases like Sick Suzy reveal a troubling complication for deprivation accounts. They are not easily amenable to intuitive assumptions about appropriate attitudes toward death. Rejecting NBNL severs the tie between harms and negative attitudes and leaves us with the puzzling question of how we should feel about death. If one shouldn’t lament her death when (and because) it is bad for her, when should she? What, if anything, makes death lamentable? Before sketching the beginning of a positive answer to this question, I consider and reject four salient possibilities. Each one has a nontrivial degree of prima facie plausibility. Nevertheless, each entails surprisingly radically counter-intuitive consequences and should be rejected.

Before I consider the four possibilities, a few clarificatory points are in order. First, I wish to reiterate that I am using lament in a generic sense to refer to negative attitudesFootnote ¹⁷ and agents can lament something to varying degrees. Second, attitudes may be more or less forceful. Roughly, the more forceful an attitude is, the more frequently the attitude is occurent and the greater emotional impact it has on the agent. I am here only concerned with the more forceful attitudes that an agent ought to have toward her death. Third, I am concerned with both the objective and subjective attitudes that an agent ought to have toward her death. Appropriate subjective attitudes refer to the attitudes that an agent ought to have relative to the evidence in her epistemic ken. This allows for the possibility that two agents whose deaths mirror one another in all of the relevant respects should have different attitudes toward their death because the evidence available to them differs. Appropriate objective attitudes refer to the attitudes that an agent ought to have if she were aware of all of the normatively relevant facts. In other words, objective attitudes pick out the attitudes that a fully rational omniscient agent would have.

The principles in the next two sections are formulated in terms of the subjective attitudes an agent ought to have. Everyone, for better or worse, is in a suboptimal epistemic state. Our attitudes towards our deaths are necessarily formed relative to the evidence available to us. Focusing solely on objective attitudes would then exclude an important practical consideration; I want to know how I should feel about my death, not just how an idealized version of myself would feel about my death. At the same time, I do not want to exclude objective attitudes. Each principle yields verdicts about appropriate objective attitudes too. To get the principles to generate verdicts about appropriate subjective attitudes, fill in the details about the agent’s relevant justified beliefs. To generate verdicts about appropriate objective attitudes, simply idealize the agent by imagining that the agent is omniscient.Footnote ¹⁸

I will now consider, and reject, four commonsense answers to the question of how agents ought to feel about their deaths. A natural thought is that death is lamentable to the extent that it deprives agents of a different possible life they could have led. Maybe the degree to which an agent should lament her death is dependent on some type of life she thinks it possible for her to have led were it not for her death. The first type of possibility I will consider is nomological possibility.

This proposal seems promising for a few reasons. First, it provides a clear explanation of why we generally think it more regrettable when a young person dies than when an elderly person does. Imagine, for instance, a 90-year old and a 19-year old are both struck and instantly killed by a bus. We might think it very likely that it was nomologically possible for the 19-year old to have lived much longer than the 90-year old and so, ceteris paribus, the 19-year olds’ death is more lamentable (for the 19-year old) than the 90-year olds’ death is (for the 90-year old). Second, on this formulation, one’s attitudes toward death track how things seemingly could have gone, not how they would have gone. This nicely handles overdetermination casesFootnote ¹⁹ and Sick Suzy cases. Although Ned successfully manages to prevent Suzy’s death from being bad for her, he doesn’t give her reason to change her attitude toward her death because (we might think) he did not prevent it from being nomologically possible for her to continue living a good life were she to receive a heart transplant. Perhaps that is what warrants Suzy’s negative attitude toward her death.

Nomological Possibility is too restricted to be accurate. Whether it can generate the intuitively correct verdicts in every case hinges upon whether we are justified in accepting determinism and the fixity of the past.Footnote ²⁰ If Suzy justifiably accepts both, she would believe that it is not nomologically possible for her to live longer than she will, in fact, live. NP would then entail that Suzy has no reason to lament her death. But this is too large a bullet to bite. Whether Suzy should lament her death does not hinge upon whether she is justified in accepting determinism and the fixity of the past.

The seriousness of the problem with NP is best illustrated by considering a revised thought experiment from Nagel’s ‘Death.’ Nagel asks us to imagine that ‘we were all inevitably going to die in agony – physical agony lasting six months, ’ before rhetorically asking if the inevitability of this agony would make it less unpleasant (Nagel Reference Nagel1970, 69). His idea seems to be that the degree of badness of this agony is not affected by its inevitability. We can revise Nagel’s though experiment a bit to concern the question of which attitude we should take toward the agony. Suppose that the laws of nature entail that everybody suffers six months of agony before they die. If we should lament being harmed iff we justifiably believe it is nomologically possible to have avoided said harm, it follows that we should not lament suffering six months of physical agony. But of course we should lament suffering six months of physical agony! The same reasons apply, mutatis mutandis, to the harm caused by death. So NP is false.

In response, one might suggest there is a relevant asymmetry between Nagel’s agony case and the case of death because agony is an intrinsic harm, while death is an extrinsic harm. Or, alternatively, agony is an experiential harm, while death is a non-experiential harm.Footnote ²¹ But these differences, in themselves, do not give us reason to treat the harms brought about by death and agony as relevantly different. We need some positive reason to think that these differences are relevant in some way to the attitudes we ought to have towards different kinds of harm. But we generally do not have asymmetric attitudes with respect to intrinsic and extrinsic harms or experiential and non-experiential harms. So there is no apparent reason to treat these harms differently.Footnote ²² In the absence of reason to think the cases are relevantly different, we should think they are relevantly similar. I conclude, then, that our attitudes toward agony and death should not be held hostage by debates over the truth of determinism and the fixity of the past.

NP fails, but an analogue view avoids this determinism problem. Consider

Metaphysical Possibility initially appears more promising, yet shifting the modality from nomological to metaphysical possibility incurs a unique set of problems. First, MP might entail that everyone’s death is equally lamentable. This would be the case if the evidence is such that the best metaphysically possible life anyone could live is an immortal life that contains an infinite amount of good.Footnote ²⁴ But this runs contrary to commonsense, as the aforementioned bus case illustrates. Ceteris paribus, the 19-year olds’ death is intuitively more lamentable for the 19-year-old than the 90-year olds’ death is for the 90-year-old. It is a cost to any view that has to deny this intuition.

Proponents of MP might attempt to salvage it by denying that it is metaphysically possible to live an immortal life with indefinitely extensive possible goods. Bernard Williams defends this view, arguing that immortality would necessarily be bad for humans because it would either result in the loss of one’s identity or extreme boredom.Footnote ²⁵ There are strong counterarguments to Williams, Footnote ²⁶ though his thesis could be weakened a bit and still allow MP to avoid this problem. Depending on individual persons’ essential psychological features, it might be the case that some could live an infinitely good immortal life while others could not.

I do not find any anti-immortality arguments compelling. However, this is orthogonal to the issue at hand, as MP is subject to two additional serious problems. First, it is unable to account for the role that expectations rightly play in determining our attitudes toward our death. The degree to which we ought to lament our death varies with respect to when we can reasonably expect to die. MP only concerns a particular type of life an agent could have lived, not at all tempered by the type of life it would be reasonable for the agent to expect to live. Second, MP seems to require agents to lament too much. There are countless metaphysically possible goods that do not seem to determine how lamentable one’s death is for them. For instance, the best metaphysically possible life I could lead might include Aladdin’s lamp or an ATM that distributes endless amounts of cash (to me, of course), but the lack of these goods in my life does not obviously warrant lament.

The next natural step changes the modality from metaphysical to logical possibility. Consider

Unfortunately, LP fares no better than MP and it might even be worse. Whether it is metaphysically possible to live an indefinitely good immortal life is an open question, but it is certainly logically possible to live such a life. As such, LP is committed to the claim that every person should lament her death to the same degree. Moreover, LP is also subject to the second and third problems that plague MP.

A tempting conclusion to draw from these considerations is that our attitudes toward death should not track how life could go in such a general sense. Perhaps they should instead track how life could go for the type of beings we are. This certainly has some intuitive appeal. For instance, people tend to regard the death of an 18-year dog to be much less lamentable for the dog than the death of an 18-year-old human is for the human, even in cases where each animal has the same cognitive capacity. A reasonable explanation is that people are inclined to make this judgment because the life expectancy of dogs is significantly shorter than the life expectancy of humans. In light of these considerations, one might propose the following.

LTP has the advantage of grounding our attitudes in something more realistic than metaphysical or logical possibility and manages to avoid the determinism problem to which NP succumbs.Footnote ²⁸ What is a reasonable life for me to expect to live, relative to the type of being I am, is not affected by whether I am justified in accepting determinism and the fixity of the past. It might be nomologically necessary that I die when I am 29, but my evidence relative to the type of being I am (e.g. human, person living in an affluent nation, etc.) is such that I can reasonably expect to live much longer.

In spite of these virtues, LTP is subject to one complication and one decisive objection. First, my formulation of LTP is misleadingly simple. Complications immediately arise upon attempting to categorize each different type of being in a single non-ad hoc way. Here are just a few of the most basic questions that this formula raises. Should types of beings be determined by the country in which one resides, the state, the city or something else? Should we categorize beings based on the medical technology available to the majority of people at a particular time? Or should we do it by what is available to persons in the same socioeconomic class or by what is affordable to each individual? Or should we do it some other way? It is far from clear that one could provide non-ad hoc answers to these questions or the countless ones like them. Nor is it desirable to allow multiple competing formulations of types of beings, for then LTP could generate an overabundance of conflicting attitudes that an agent ought to take towards her death.

I do not want to rule out the possibility that there is a principled way to pick out the relevant type of being. One strategy is to focus on individuals’ essential properties. All humans might be necessarily human, for instance.Footnote ²⁹ My life could have been very different. I seemingly could have been born into a wealthy family or have been born into extreme poverty. I might never have become a philosopher.Footnote ³⁰ But perhaps I am human in every possible world in which I exist. Whether I should lament my death might then depend on how well I fare relative to the life I could reasonably expect to live as a human. One worry with this response is that I might only be contingently human (as many in the transhumanism literature think). More generally, there might not be any essential properties.Footnote ³¹ Perhaps these worries could be overcome, however. The more important point is that even if this strategy avoids the first complication, LTP still succumbs to the following decisive objection.

LTP entails that those who do the best, relative to other beings like them, have no reason to lament their death. This is troubling. It strikes me as perfectly appropriate were Jeanne Calment (the oldest verified supercentenarian) to lament her particular death at age 122. Why? Because it was possible in some important sense for her to continue living a life worth living even if she could never reasonably expect to live such a life. To magnify the problem, consider the following example.

According to LTP, this humanoid should not lament her death. In fact, any humanoid who lives a sufficiently good life for more than a day should not lament her death, but that seems very counterintuitive. Perhaps, some will find this thought experiment too close to science fiction to have any clear intuitions about the case. However, these counterintuitive implications are revealed in analogue cases in the actual world. Duchenne muscular dystrophy is a disease that causes slow muscle degeneration and an early death. The average life span of a person with the disease is 25 years and almost no one lives to be in their thirties. If LTP is true, then Thomas Sulfaro should not lament his death since he is in his forties. Footnote ³² But that seems false. It is perfectly appropriate for Thomas Sulfaro to (forcefully) lament his death.Footnote ³³

As with NP, we can draw from Nagel’s agony case to illustrate another problem with LTP (Nagel Reference Nagel1970, 69). Broadening LTP to apply to all harms (including intrinsic and experiential) instead of just death, a person who only suffers five months of agony shouldn’t lament her agony at all. Again, this seems very counterintuitive. Suffering extreme agony for long periods of time (that is neither intrinsically nor extrinsically valuable) always warrants a forceful negative attitude.

These four proposals are each intuitively appealing in their own way and yet all have very counterintuitive implications as well. Certain considerations lead us to think that our attitudes toward death should be influenced by the life it is possible, in some sense, for the subject to live. Other considerations lead us to think that our attitudes toward death need to track something more realistic.Footnote ³⁴ I am not optimistic that either approach can succeed on its own. Instead, the lesson to draw from the virtues and vices of these proposals is that our attitudes toward death should be influenced both by realistic expectations about our death and the degree to which we miss out on the best (metaphysically) possible life we could have led. In the next section, I argue for the basics of such a hybrid view, though the full details remain to be worked out.

IV. A hybrid view

My primary focus in this paper has been negative. I aimed to show that standard deprivation accounts coupled with NBNL lead to absurd conclusions and that we should reject the latter. I then argued against a variety of alternative accounts of appropriate negative attitudes toward death. I now turn to my positive argument, which is more tentative than my negative argument. Working out a fully detailed positive account is beyond the scope of this paper. In this section, I argue for a sort of hybrid view, where our (forceful) attitudes toward death should be determined by analogue principles of MP and LTP.

One lesson some may draw from section III is that there may be no fact of the matter about how an agent ought to feel about her death full stop. Maybe the failure of the principles I considered help vindicate Bradley’s contextualism.Footnote ³⁵ I am sympathetic to this line of thought, though skeptical. It is certainly true that Suzy should, in some sense, be glad that her death prevents her from being tortured and subsequently murdered. On the supposition that she should have some attitude towards this fact, what else could it be? But, and this is a crucial point, this attitude should not be one that has any serious emotional impact on Suzy because it was never reasonable for Suzy to believe that she will be tortured and there is no chance that Suzy will, in fact, be tortured. Simultaneously, in some sense, Suzy should be saddened by the fact that her death will ensure that she could not possibly continue to live a good life. Yet, this contextualism cannot be the entire picture. Even if some form of contextualism is true, not all attitudes warranted by our death should be equally forceful. To see why, it will help to first consider a case that concerns two different deprivational harms that befall poor Unlucky Louie.

Let’s stipulate that possessing Aladdin’s lamp would improve Louie’s life significantly more than winning the lottery and so is a greater deprivation of goodness than the loss of the lottery ticket. But Unlucky Louie would rightly be regarded as irrational (indeed, crazy!) were his lack of Aladdin’s lamp to have a greater emotional impact on him than his stolen lottery ticket in any context in the actual world. This is so even if one thinks it perfectly appropriate (in certain contexts) for Louie to have a negative attitude of some kind towards the deprivational harm of not having Aladdin’s lamp (e.g. upon being primed to think about the fact that he does not have Aladdin’s lamp).

At the same time, which negative attitudes ought to be most forceful for Unlucky Louie is not solely determined by context. The lesson to draw from Unlucky Louie then, is that something roughly analogous LTP should partly determine which attitudes ought to affect the agent in a forceful way. Louie was justified in believing that he would soon be rich and so that good seemed realistically obtainable, whereas he was never justified in believing he would possess Aladdin’s Lamp, so that good was not. Ceteris paribus, being deprived of a good that was, at some time, a seemingly realistic prospect should affect the agent more than being deprived of a good that was not. Such a condition is intuitively compelling and would explain why we tend to think a seemingly healthy 25-year old who dies of a heart defect has a more lamentable death than someone who dies from old age, even if each person’s death precluded identical amounts of possible good life. I propose that the following principle best captures this idea.

EP has the virtue of being strictly compatible with the basic structure of contextualist accounts of the badness of death. It also accounts for the intuition that it is appropriate for Thomas Sulfaro to lament his death, though to a lesser degree than an otherwise seemingly healthy person who dies in his forties. More importantly, it avoids the problems with each of the principles discussed in section three. It is not held hostage to debates over determinism and the fixity of the past, does not entail that everyone’s death is equally lamentable, does not require agents to lament too much and allows that humanoids should lament their death in many cases.

Still, EP itself is insufficient. If it were the sole determining factor, it would entail that any fully informed person (i.e. one aware of the normatively relevant facts about her death) should not lament her death. This is an unacceptable consequence. We cannot prevent death from being lamentable simply by knowing everything about it. Keeping the symmetry with experiential and intrinsic harms also requires rejecting the EP as the sole relevant determiner of forceful attitudes toward death. EP doesn’t warrant negative attitudes for someone in Nagel’s thought experiment who suffers six months of agony (and no positive attitudes for someone who only suffers five months of agony). I propose that we can generate the intuitively correct results if we supplement EP with the following analogue of MP.

Having MP* tempered by EP gets around a major problem. It allows that Jeanne Calment should lament her death, but also explains why she should lament it less than a 25-year old who unexpectedly dies. So coupling EP with MP* seems to be a promising way to avoid the problems with NP, MP, LP and LTP, yet retain the virtues of each. That does not mean that this view is free of problems, however. Many details remain to be worked out. I have not explained the exact relationship between EP and MP*. Maybe we should have one attitude toward our death determined by these two principles or maybe each warrants a distinct incommensurable attitude. This view, described thus far, is also incomplete. It leaves open the possibility that we should have other forceful attitudes toward death. If we should lament not having the best metaphysically possible life, shouldn’t we be happy to not have the worst metaphysically possible life? I am inclined to think so, but recognize this is potentially counterintuitive. These issues will have to be worked out in another paper.

V. Conclusion

First, I sought to reveal a problem heretofore almost completely overlooked in the literature, viz. – that standard deprivation accounts cannot accommodate NBNL. This means that we either have to reject standard deprivation accounts or NBNL. I argued that we should reject the latter by first arguing for the truth of a moral principle I call Prevent the Worst Death and then showing that we need to appeal to this principle and standard deprivation accounts to get the correct result in Sick Suzy vs. Sick Jenny. After that, I explored the question of what, if anything, warrants forceful lamentation of death. I considered and rejected four prima facie plausible answers to that question before defending my positive account. I proposed that two principles (the Expectation Principle and Metaphysical Possibility*) each partly determine the forceful attitudes we should have towards death.