Philosophical Studies

, Volume 158, Issue 3, pp 507–513

Fischer on death and unexperienced evils


    • Philosophy DepartmentSyracuse University

DOI: 10.1007/s11098-010-9667-0

Cite this article as:
Bradley, B. Philos Stud (2012) 158: 507. doi:10.1007/s11098-010-9667-0

A simple Epicurean argument goes as follows. Dead people have no sensations, a fortiori no bad sensations. Nothing is bad for us unless it is, or causes, some bad sensation (the experience requirement). Thus, being dead is not bad; nor is the event of one’s death, since it does not cause one to be in a bad state.

This argument stinks. But why? John Fischer, like Thomas Nagel, Robert Nozick, Jeff McMahan, and many others, objects to the experience requirement. In several chapters of Our Stories Fischer argues that there can be things that are bad for us that do not involve bad sensations at all.1 Some examples: Nagel’s example of the man who is betrayed behind his back and never finds out about it, but lives happily to the end of his days (pp. 5, 37, 104); Nozick’s example of the person who is secretly videotaped in her bedroom and watched by people in Outer Mongolia, but never finds out or has any bad feelings as a result (pp. 5–6); McMahan’s example of the person whose daughter dies on a Himalayan adventure, unknown to him, just minutes before his own death (pp. 7–8, 106). In each case Fischer agrees there is a misfortune to the person. Each of these examples would be a counterexample to the experience requirement.

Here is a different kind of counterexample to the experience requirement. Suppose John has sent me some concert tickets in the mail. I don’t know about the concert or the tickets. If I did know, I would really want to go, and would enjoy going. Derk steals the tickets from my mailbox and I am never the wiser. The theft is a misfortune to me, but it does not cause me to have any bad sensations—it merely prevents me from having some good ones.

Notice a difference between these sorts of counterexample. The concert ticket example involves an event that is bad for me only because it deprives me of something good; it is instrumentally bad for me. The examples from Nagel, Nozick, and McMahan involve events that are bad for someone independent of what they bring about for that person or deprive that person from getting; they are events that are allegedly intrinsically bad for the person.

Now let us return to the Epicurean argument and add a bit of structure to it, to see how these two responses to the experience requirement result in different objections to the argument.
  1. 1.

    If death is bad for someone, then it is bad for that person instrumentally.

  2. 2.

    The only things that are instrumentally bad for someone are things that bring about something intrinsically bad for that person.

  3. 3.

    Death does not bring about anything intrinsically bad for its victim.

  4. 4.

    Therefore, death is not bad for us.

The deprivation theorist rejects premise 2.2 We may take the deprivation theorist to make the following claim about the badness of death:

D: Death is bad for its victim in virtue of preventing the victim from getting some intrinsic goods.

Death is like many other instrumentally bad things, such as illnesses and injuries, in preventing us from having the goods of life. It is unlike those instrumentally bad things in that one is not suffering when one is dead. A nice feature of this response to the Epicurean argument is that it is compatible with every theory of well-being, including hedonism.

Nagel, Nozick, McMahan and Fischer seem to target premise 3. They think there are things that are bad for us that are not bad sensations: undiscovered betrayals, unknown misfortunes of loved ones, secret rights violations, and the like. Not only are these not bad sensations, they need not bring about any bad sensations in order to be bad, nor prevent anything good from happening.3 They are intrinsically bad. Fischer claims (p. 5) to be defending the deprivation account, yet his defense involves arguing for the badness of events that are bad independent of whether they deprive the victim of anything.

It seems, then, that we have dialectical confusion. The view about death’s badness suggested by Fischer’s examples is not a deprivation view, but another view that, as far as I know, does not even have a name yet. How exciting! I hereby call the view the “more intrinsic badness” view:

MIB: Death is bad for its victim in virtue of bringing about something that is intrinsically bad for her.

Of course, the specific examples Fischer gives are not examples that we can appeal to in order to explain the badness of death; death does not bring about betrayal or the other unexperienced evils he discusses. How could death bring about something intrinsically bad for its victim, supposing that death marks the end of the victim’s existence? There are a few possible ways. According to desire fulfillment accounts of well-being, it is intrinsically bad to have a desire frustrated. Death can frustrate many desires, and so can bring about intrinsic evils (Feinberg 1993; Pitcher 1993; Luper 2009). According to achievement-based accounts, failure is intrinsically bad. If someone works hard on a project and her death prevents her from completing it, she fails, which is intrinsically bad for her. Fischer follows Velleman (1993) in claiming that well-being is at least partly determined by irreducibly global facts about narrative structure. A premature death can add some intrinsic badness to the story of one’s life; perhaps it may do so in one of the ways just discussed, by frustrating a desire or causing a goal to go unachieved.
Why does Fischer appeal to unexperienced intrinsic evils, rather than employing the more straightforward deprivationist reply to Epicurus? Recall the example of Derk and the tickets. According to the deprivationist account, Derk’s stealing the tickets is bad for me, even though it causes me no bad sensations. Those who defend the experience requirement make the following move: they say that Derk’s theft is bad for me only because it is possible for me to feel bad about the theft (Silverstein 1980; Rosenbaum 1993). If I were to find out about the theft, I would be angry! So there is a modified experience requirement that is satisfied in the ticket-theft case, but is not satisfied in the case of death:

ERII. Something is bad for someone only if it is possible for that person to feel bad about it. (p. 38)

The examples Nagel gives, like the betrayal case, aren’t obviously counterexamples to ERII either, since it is possible for someone to find out he has been betrayed. But Fischer notes that we can make them into counterexamples to ERII by adding a Frankfurt-style counterfactual intervener, whose presence ensures that the victim cannot find out about the betrayal, but whose presence seems to make no difference to whether any harm has occurred (pp. 6–7; 39–41; 105–107). I think Fischer is completely right about this. Furthermore, I think we can see that ERII is problematic without even appealing to any fancy examples. The problem I have in mind concerns the magnitude of harm. Suppose I never in fact find out about Derk’s theft, but if I did, my upset feelings would have a value for me of −10. How bad is Derk’s theft for me? Are we to determine the badness of the theft by looking at the disvalue for me of some non-actual negative feelings, so that the theft has value −10 for me, just as it would if I actually found out and had the bad feelings? That is absurd. In general, it seems like a mistake to say that some feature makes an event harmful, but plays no role in determining the extent of the harm.4 The mere possibility of being the object of bad feelings is required by ERII for an event to be harmful, but does not play a role in determining the extent of the harm. So ERII seems false for at least two reasons.
However, there is a third experience requirement that is more widely held:

ERIII: There are no unexperienced intrinsic goods or evils.

Fischer, Nagel, and the rest reject ERIII in the process of rejecting the other versions of the experience requirement. But surely we should not have to reject ERIII in order to reject those other versions, or to object to the Epicurean argument. (I hope not, because I think ERIII is true!) The fact that the deprivationist objection to the Epicurean argument is compatible with ERIII gives it a significant advantage over the Fischer/Nagel objection.

The Epicurean need not rely on an experience requirement in arguing against the badness of death, and he will not regard the deprivationist objection as convincing. He will appeal to another difference between death and ordinary deprivations: death marks the end of an individual’s existence. (Or so we may assume, to make the Epicurean argument strongest.) This fact, rather than an experience requirement, explains why death cannot be bad for its victim, though having my tickets stolen by Derk can be bad for me.

This Epicurean argument is based on the notion of badness at a time. Let us say that for an event to be bad for someone at a time, it must make that person worse off at that time than she would otherwise have been. Here is a way to formulate the argument (the Timing Argument):
  1. 1.

    If death is bad for us, it is bad for us at some time.

  2. 2.

    If death is bad for us at some time, then either it is bad for us before we die or after we die.

  3. 3.

    Death is not bad for us before we die.

  4. 4.

    Death is not bad for us after we die.

  5. 5.

    Therefore, death is not bad for us.


Death is unlike ordinary deprivations, the argument goes, because in the case of an ordinary deprivation, we can locate a time at which the victim is worse off. I am worse off as a result of Derk’s theft at those times at which I would have been enjoying the concert.

Several objections have been raised to the Timing Argument. Some, like Nagel (1979) and perhaps Feldman (1992), seem to argue that the evil of death is timeless, thereby rejecting premise 1. Others, like Feinberg (1993), Pitcher (1993), and Luper (2009), argue that death is bad in virtue of frustrating our desires, and that this badness is located at the time we have the desire; thus they reject premise 3. In my view, we should reject premise 4. Death is bad for us at times after we die, because at those times, just as in the case of ordinary deprivations, we are worse off than we would have been if we hadn’t died (Bradley 2009; Feit 2002; Grey 1999). Fischer agrees with me about this (p. 46). But there is an important difference between us.

On the view I favor, death is bad for someone at times after she dies in virtue of the fact that at some of those times, intrinsically good things would have been happening to her. She is worse off being dead than enjoying those intrinsic goods. This requires us to say that the dead have a well-being level of zero. Some have argued that this is incoherent, because it requires attributing properties to nonexistent things (e.g. Hershenov 2007). I have argued that it is coherent, and that in prudential evaluations we treat nonexistence as equivalent to existing with zero well-being (Bradley 2009, pp. 98–111). (Since Fischer agrees that the dead are harmed after they die, there is no reason to rehearse these arguments here.)

What is Fischer’s story about why death is bad after the victim dies? In arguing against ERII, Fischer describes the following case:

Imagine that your spouse and your best friend are on a space colony orbiting Mars, which is now on the opposite side of the Sun from Earth. Hence, it will take a few minutes for light waves to travel from Mars to the village on the Alaskan coast where you reside. They betray you. It turns out, however, that a gargantuan earthquake-induced tidal wave is going to kill you in fewer than the number of minutes it takes for light waves to travel from Mars to Earth. Here it is impossible (in a very strong sense) for you to experience something bad as a result of the betrayal. And yet it seems that you have been harmed… If you are harmed by the betrayal in the Mars case, why not also say that you can be harmed by your death, even though the death occurs after you have ceased to exist?… The time of the misfortune is the time during which you are dead. (pp. 45–46, emphasis in original)

Here we see that Fischer rejects premise 4 of the Timing Argument, as I do. Just as a betrayal can harm you after you die, so can death. But his reason for objecting to premise 4 is different from mine. My account is a pure deprivation account; death merely deprives one of intrinsic goodness, and the time of death’s badness is the time that goodness would have obtained. Fischer’s account is a “more intrinsic badness” account: one can “suffer” intrinsic evils such as betrayal at times after one has ceased to be. The time of the badness is the time of the intrinsically bad event, such as the betrayal.

The difference is important. First, note that according to Fischer’s account but not mine, one’s well-being level can rise and fall after one has died. According to the mere deprivation view, death and posthumous events are very different. Death deprives its victim of good things, and thereby makes a difference to how well things go for someone, but events occurring after death cannot make any such difference.5 According to MIB, posthumous events can make a difference to how well things go for the dead person, because that person’s well-being level can continue to rise and fall after death depending on what happens—e.g., depending on whether one’s spouse cheats with one’s best friend on Mars.

Others have also claimed that death and posthumous events are analogous (Feinberg 1993; Luper 2009). However, they have typically claimed that the harm of death is retroactive. Death frustrates desires, and desires are had only by the living, so death and posthumous events, when bad for someone, must be bad for the person at the time she had the relevant desires. Fischer’s view provides the more natural way to account for these harms. In non-death cases, we do not think that the frustration of a desire is bad for the desirer at the time of the desire; we do not, for example, think that a childhood is retroactively made worse when the child’s desires turn out to be unsatisfied in adulthood (Velleman 1993, p. 340).

But there is a glaring drawback to Fischer’s reply to the Timing Argument. It requires us to say that someone can have a positive or negative well-being level at a time at which she does not exist. This is hard to accept. If death brings about additional intrinsic badness for its victim, it seems more plausible to argue that the additional intrinsic badness is timeless—it makes the victim’s whole life worse than it would have been otherwise by reducing its irreducibly global value, without making it worse at any time.

While this might be a better way to develop Fischer’s view, it still leaves us unable to account for the full badness of death. Even if death is bad in virtue of diminishing the value of the victim’s life as a whole, e.g. by frustrating the goals the victim had while she was alive, it also seems bad in purely deprivational ways. Suppose Jennifer is extremely depressed; as a result, she has no goals, nor any desires for good things in life. Were she to die, her death would cause nothing intrinsically bad to happen to her, because it would not cause any desires to be frustrated nor any goals to be unachieved. But imagine that were she not to die, her depression would soon be cured, and she would begin to enjoy life. She would go on to form new desires that would be satisfied and set new goals that she would achieve. I think Jennifer’s death would be very bad for her. But it would not count as bad according to MIB, for nothing intrinsically bad happens as a result of her death. In other, more ordinary cases, in virtue of frustrating some goals and desires, death might be counted as bad by MIB, but not as bad as it seems to be. Death is one of the worst things that can happen to someone. Its badness cannot fully be captured by looking at the unexperienced intrinsic evils it causes, such as frustrated goals. The badness of being deprived of many years of the good things in life dwarfs whatever intrinsic badness death causes for its victim. What’s more, a great deal of that badness is episodic, rather than narrative or global badness. It is at least largely in virtue of depriving its victim of many particular good episodes that death is bad for that person.

So MIB is not a plausible competitor to the deprivation account. But of course it is possible that death could be bad in virtue of both the intrinsic goods it prevents and the intrinsic evils it causes. The point I wish to drive home is this: as long as we wish to say that part of the badness of death is purely deprivational, forward-looking, and episodic, as I think we should, we will face pressure to reply to the Epicurean worry that it is impossible to have a well-being level (even zero) after death. We cannot avoid this worry by appealing to unexperienced intrinsic evils.


Fischer 2009. All page references in the main text are to Our Stories unless otherwise noted.


See Feldman 1992, pp. 135–138 for a clear explanation of what is wrong with premise 2.


It could be argued that betrayals do prevent good things from happening—viz., they prevent the victim from having the intrinsic good of not being betrayed. (Thanks to two referees for suggesting this interpretation of the Nagel/Fischer view). This does not strike me as very much more plausible than the claim that pain is bad in virtue of preventing the victim from having the intrinsic good of not being in pain. The most straightforward interpretation of Fischer and Nagel has them endorsing the view that death causes intrinsically bad things to happen to its victim. In further support of this interpretation: if Nagel and Fischer merely intended to give examples of misfortunes that deprive their subjects of good things, much less controversial examples were available to them, such as my example of Derk and the tickets; such examples need not presuppose the falsity of hedonism, as Nagel’s and Fischer’s do.


For a similar argument in a different context see Sinnott-Armstrong 2009.


Posthumous events can make a difference to whether the desires the victim had while alive are fulfilled or not; so given a desire satisfactionist axiology, there need not be quite so big a difference between death and posthumous events. But the deprivation account is compatible with the falsity of desire satisfactionism.


Copyright information

© Springer Science+Business Media B.V. 2010