American Philosophical Quarterly (forthcoming)

1
Death`s Distinctive Harm
Stephan Blatti
ABSTRACT: Despite widespread support for the claim that death can harm the one who dies, debate contin-
ues over how to rescue this harm thesis (HT) Irom Epicurus`s challenge. Disagreements Iocus on two oI the
three issues that any defense of HT must resolve: the subject oI death`s harm and the timing oI its injury.
About the nature oI death`s harm, however, a consensus has emerged around the view that death harms a
subject (when it does) by depriving her oI the goods liIe would`ve aIIorded had she continued living. This
deprivation view oI death`s harm (DV) derives some oI its credibility Irom the general deprivation theory
of which it is an instance: mortal harm is subject to the same kind of analysis plausibly given of other non-
mortal harms. Furthermore, note that the weak formulation of HTasserting only that death can inflict
harm, not that it always or necessarily doesaccommodates the intuition that instances of rational suicide
and justifiable euthanasia present cases in which death fails to harm. DV is equipped to explain how in the-
se cases the harms involved in continued existence outweigh the goods of which death deprives the subject.
I agree that suicide can be rational and that euthanasia can be justifiable. Likewise I accept both HT and
DV as far as they go. But they do not go far enough. Specifically, I argue here that death harms even those
who die as a result oI rational suicide or justiIiable euthanasia; that death`s harm is neither undiIIerentiated
nor wholly contingent, but multifaceted and partly necessary; that the necessary part oI death`s harm is dis-
tinctive, inflicting a peculiar restriction on the autonomy of one who dies; and, regarding the timing and
subject issues, that this restriction harm is inflicted on the antemortem subject prior to her death.
The Harm Thesis (HT) states that death can harm the one who dies. The Epicurean
rejection oI this claim relies on the Iollowing Iamiliar argument: beIore death there is
no harm (i.e., prior to its occurrence, death inIlicts no injury on its eventual victim);
aIter death there is no victim (since, it is assumed, death brings about the subject`s
nonexistence, and thus there exists no postmortem subject to whom a harm could ac-
crue); thereIore, HT is Ialse: death cannot harm the one who dies.
1
'|T|he most awIul
oI evils, as Epicurus notoriously put it in his Letter to Menoeceus, 'is nothing to us,
seeing that, when we are, death is not come, and, when death is come, we are not.
Indeed, insoIar as one`s Iear oI death is predicated upon a Iear oI harm to oneselI, Ep-

1
While I think my straightIorward gloss captures the spirit oI Epicurus` own argument, diIIerent commentators re-
construct this argument diIIerently. Bradley (2004: 1), Ior instance, suggests that Epicurus derives the Ialsity oI HT
Irom '(1) Anything that is bad Ior someone must be bad Ior that person at a particular time and '(2) There is no
time at which death is bad Ior the one who dies. Luper (2009: 67), by contrast, construes the Epicurean argument
as an extended modus tollens: 'iI death harms the individual who dies, there must be a subfect who is harmed by
death, a clear harm that is received, and a time when that harm is received. . Seeing that there is no coherent solu-
tion to all three issues, Epicurus rejects |HT|. See also Grey (1999: 359), Silverstein (1993: 96; 372 n.2), and War-
ren (2004) Ior pertinent commentary. The discussion here, however, will not depend on precisely how one renders
this argument. In something like the way we oIten use Cartesian` (rather than Descartes`s``), I use Epicurean` (in-
stead oI Epicurus``) to reIer disjunctively either to the view oI Epicurus himselI and/or to the view oIten attributed
to Epicurus himselI and/or to the view in the spirit oI the view held by Epicurus himselI.




DEATH`S DISTINCTIVE HARM



2
icurus held, the Ialsity oI HT exposes such Iear as irrational. Again Irom the Letter:
'He speaks idly who says that he Iears death, not because it will be painIul when pre-
sent but because it is painIul in anticipation. For iI something causes no distress when
present, it is Iruitless to be pained by the expectation oI it.
Like most participants in contemporary discussions oI the metaphysics and ethics
oI death, I resist the Epicurean argument and endorse HT. But anyone who endorses
HT conIronts several challenges implicit in that argument. My discussion here will
address two oI these challenges. The Iirst concerns timing: iI HT is true, when is
death`s harm incurred? The second concerns harm: iI HT is true, what is the nature oI
the harm brought about by death?
About the timing issue there has been considerable controversy, with apparently
all conceptual space now occupied. To wit, each oI the Iollowing answers to the tim-
ing question has been deIended recently: at all times (eternalism) (Feldman 1992);
beIore death occurs (priorism) (Luper 2007); at the moment death occurs (concur-
rentism) (Lamont 1998); aIter death occurs (subsequentism) (Bradley 2009; Feit
2002); at some indeterminate time (indefinitism) (Nagel 1970; Silverstein 2000).
2

About the harm issue, by contrast, there is remarkable consensus: when bad,
death`s badness consists in depriving us liIe`s goods. According to this view, death
harms the one who dies solely by preventing her Irom attaining those oI liIe`s goods
she would have enjoyed had she continued living: positive experiences, healthy rela-
tionships, signiIicant projects, etc. In the 1970 article that has become the locus clas-
sicus Ior this view, Nagel argues that one can be harmed by something unexperienced,
that deprivation may constitute a harm oI this sort (i.e., unexperienced but harmIul
nonetheless), and that death`s harm is an instance oI this species oI harm.
3
Since then,
and notwithstanding various intramural disputes about how best to Iormulate it,
4
the
deprivation view (DV) has come to be regarded as the standard view, enjoying broad
support amongst supporters oI HT.
5
Consequently, even advocates oI opposing an-
swers to the timing question nevertheless share a common commitment to the depriva-

2
This taxonomy oI positions is adapted Irom Luper (2007: 241). The lists oI expositors are illustrative, not compre-
hensive.
3
Nagel (1979: 45, 710).
4
For example, one point oI contention concerning the deprivation view has centered on whether (contra Nagel, et al.)
a person`s ability to experience something as bad is a necessary condition oI an event`s badness. On this point, see
Fischer (1993: 1823), Rosenbaum (1986) and Silverstein (1980). Other discussions may be Iound in Bradley
(2004), Feit (2002), Feldman (1991, 1992), McMahan (2002), and Silverstein (2000). I shall make no attempt in
what Iollows to represent all oI the proposed reIinements, but will instead concentrate on the core claim whose
rough Iormulation here should be suIIiciently encompassing as to be amenable to all parties.
5
In addition to Nagel, notable death theorists who accept (oIten explicitly, sometimes implicitly or tacitly) this view
include Bradley (2004, 2009), Brueckner and Fischer (1986), Feit (2002), Feinberg (1984), Feldman (1991, 1992),
Grey (1999), Lamont (1998), Luper (2007, 2009), McMahan (1988, 2002), Pitcher (1984), Williams (1973), and
Yourgrau (1987), among others.




DEATH`S DISTINCTIVE HARM



3
tionist answer to the harm question: eternalists are deprivationist eternalists, indeIi-
nitists are deprivationist indeIinitists, subsequentists are deprivationist subsequentists,
etc.
My primary Iocus in what Iollows will be the deprivationist answer to the harm
question. Although I accept DV as Iar as it goes, I will argue that its analysis oI
death`s harm is incomplete. The objective, thereIore, will be to isolate this lacuna and
to characterize the aspect oI death`s harm that cannot be accounted Ior by DV. But I
shall not ignore the timing question. For what we shall discover is that, whereas DV
can be paired with any oI the available answers to the timing question, the hitherto
overlooked dimension oI death`s harm jibes with one answer in particular.
The discussion unIolds in three steps. In section 1 I describe the priorist answer to
the timing question, show how it works with DV, and raise a couple oI misgivings. In
section 2 I present DV in greater detail, highlighting its advantages and drawing out
the respects in which I believe it to be incomplete. In section 3 I oIIer a proposal Ior
how to understand the aspect oI mortal harm that DV overlooks. This account accords
with priorism and avoids the objections raised in section 1.
1. Priorism
According to the priorist answer to the timing question, the victims oI death`s harm
are not the dead, but the living: the antemortem subjects whose interests death will
undermine.
6
II this is correct, then your Iuture death may constitute an injury to your
present selI. But how exactly does death reach backwards in time to bring about this
harm?
The priorist claims that death`s harm is brought about not directly, via backward
causation, but indirectly, by aIIecting the truth conditions oI propositions concerning
the living subject`s interests and projects. Feinberg, Ior instance, writes that
the Iact oI a person`s death 'makes it true that his antemortem interests were going
to be deIeated, and to that extent the antemortem person was harmed too, though his
impending death was still unknown to him.
7

Assuming the priorist adopts the deprivationist account oI death`s harm, it will be the
deIeat oI the antemortem subject`s intereststhe welIare whose enjoyment is pre-
cluded, the projects whose completion is thwarted, the relationships whose Ilowering
is prevented, etc.that constitutes the harmIulness oI death, on this view. But let us

6
Those who have deIended this view include Feinberg (1984), Luper (2004, 2007), Pitcher (1984), and possibly
Aristotle (cI. Scott 2006).
7
Feinberg (1993: 18788).




DEATH`S DISTINCTIVE HARM



4
concentrate Ior the moment on how, according to the priorist, death is supposed to
achieve this Ieat.
Consider a baseball analogy. In our imaginary game, the Yankees and the Red Sox
are tied in the sixth inning. A Red Sox batter then breaks the tie by hitting a homerun.
Failing to score in the Iinal innings, the Yankees lose the game. In some sense, it is the
Yankees` Iailure to score again that turns the sixth-inning homerun into the winning
run Ior the Red Sox. But oI course, no change in the homerun`s intrinsic properties
accounts Ior its status as the winning run; the scoreless seventh, eighth, and ninth in-
nings do not somehow reach backward in time to aIIect causally the sixth-inning
dinger. Better to say that the Yankees` Iailure to score in subsequent innings 'makes it
true that the sixth-inning homerun was the winning run. This suggests that the Yan-
kees` Iailure to score is the truth-making condition oI a proposition like The Yankees
will lose the game`. But notice that this proposition is true even prior to the game`s
end. Indeed, albeit unknown to anyone at the time, that proposition was made true the
instant the Red Sox batter crossed the plate in the sixth inning. In this way, the home-
run`s acquired status as the winning run reIlects a 'mere Cambridge change (on
which more shortly).
Similarly, says the priorist, death aIIects the living subject not by altering her in-
trinsic properties, but by being the truth-making condition Ior propositions which hold
oI the subject prior to her death. In particular, on this view, S`s death at t
2
impacts S at
t
1
by determining the truth values oI propositions which concern S`s interests at t
1
.
With the mechanics oI this view in place, it is not diIIicult to imagine how a pri-
orist who is also a deprivationist will deIend HT. Luper, Ior instance, writes:
Suppose I have made some project a central Iocus oI my liIe. Surely it is in my inter-
est now Ior my project to be brought to a successIul conclusion in the Iuture. Suppose
that only I can complete it, but I die prematurely. My dying ensures that it is true oI
me that my project will not be brought to Iruition, and I am harmed all the while I
have an interest in Iinishing my project, Ior during this time my welIare is lower than
it would have been had I succeeded at my project.
8

So, according to this combination oI answers to the timing and harm questions
deprivationist priorismS`s death at t
2
constitutes a harm to S at t
1
by depriving S at
t
1
oI goods that S would have enjoyed had she continued living. Put diIIerently, iI S`s
death at t
2
will render her eIIorts at t
1
ineIIectual (by precluding completion oI a sig-
niIicant project, Ior instance), then S`s death at t
2
harms S at t
1
by making S at t
1
com-
paratively worse oII than she would have been had she not attempted to complete her
project in the Iirst place. Note also that, because death is the truth-making condition
Ior propositions which hold oI the subject prior to her death, and because those propo-

8
Luper (2007: 24849). Similar statements may be Iound in Feinberg (1993: 18388) and Pitcher (1993: 16268).




DEATH`S DISTINCTIVE HARM



5
sitions become true or Ialse at whatever point the living subject acquires the interests
that her subsequent death will deIeat, the subject is harmed (albeit unwittingly) imme-
diately upon acquiring the interests that will be deIeated (just as, albeit unknown to
anyone at the time, the sixth-inning homerun immediately became the winning run).
9

In this connection, note that (contra Grey) the priorist is not claiming that death`s
harm is 'preIigured in the destiny oI the Iuture decedent.
10
The harm that S`s death at
t
2
inIlicts on S at t
1
is no more destined than is the Yankees` Iailure to score in the Iinal
innings. But given its occurrence at t
2
, the priorist claims, S`s death constitutes a harm
to S at t
1
.
To some, this view has seemed inadequate, even incredible. But on its behalI it
should be acknowledged that scenarios like the one described by Luper above are nei-
ther unrealistic nor unIamiliar. Consider, Ior instance, Ferdinand Magellan whose
liIe`s ambition it was to become the Iirst human being ever to circumnavigate the
world. In 1519, an armada under Magellan`s command set sail Irom Spain to accom-
plish this mission. Although the expedition was eventually completed in 1522, Magel-
lan himselI was killed en route in 1521. According to the deprivationist priorist view,
by precluding the completion oI his journey, Magellan`s death in 1521 constituted a
harm to Magellan prior to 1521, Ior Magellan shall never complete his project to be-
come the Iirst human being ever to circumnavigate the world` was true oI Magellan
even beIore he set sail. Since he sought to complete the expedition, and since an event
that prevents the completion oI a treasured project constitutes a misIortune Ior the in-
dividual whose project it is, death`s harm is thus attributed to Magellan prior to his
death.
Because my aim here is not to adjudicate the general debate between competing
answers to the timing question, I will reIrain Irom weighing the balance oI considera-
tions counting Ior and against the priorist part oI the deprivationist priorism combo.
Here I want simply to mention two misgivings about this view. The relevance oI these
concerns to my more limited objectives will emerge in section 3.
First, I noted previously that the harm that S`s death at t
2
inIlicts on S at t
1
is sup-
posed to be brought about by a mere Cambridge change. Something undergoes a
'Cambridge change, Geach writes, when 'there are two propositions about |that
thing|, diIIering only in that one mentions an earlier and other a later time, and one is
true, and the other Ialse.
11
This account appears to accommodate all cases oI change,
including not only the Iamiliar cases in which a thing undergoes an alteration in its

9
'|T|he harmed condition, Feinberg (1993: 186) writes, 'began at the moment |the antemortem subject| Iirst ac-
quired the interests which death deIeats.
10
Grey (1999: 364). Grey is objecting to Feinberg and Pitcher`s priorist account oI posthumous harm, but the objec-
tion reIlects a common conIusion about the priorist account oI death`s harm.
11
Geach (1979: 9091); see also Geach (1969: 7172; 1972: 32122).




DEATH`S DISTINCTIVE HARM



6
intrinsic or monadic properties (i.e., 'real or 'genuine changes), but also cases in
which there is a shiIt in the truth-conditions oI propositions that once held oI some-
thing and cases in which there is a diIIerence in a thing`s relational predicates. But
while all real changes are Cambridge changes, not all Cambridge changes are real
changes.
12
This is because something can undergo a Cambridge change without un-
dergoing any change in its intrinsic propertiesas in the baseball game example de-
scribed above, or as when widower` goes Irom being Ialse to being true oI Xanthippe
upon Socrates`s death. These last cases illustrate mere Cambridge changes; and ac-
cording to a deprivationist priorist, death`s harm is a mere Cambridge change.
But is it enough, in deIending HT, to concede that death`s harm eIIects no real
change in its victim, but only a mere Cambridge changeno alteration in the victim`s
intrinsic properties, but only a change in her Cambridge properties? It seems at this
point that the Epicurean opponent oI HT might well accuse the deprivationist priorist
oI moving the goalpostsoI construing harm in a way that departs Irom the more Ia-
miliar notion originally in play. While hardly a knockdown objection, certainly it must
be conceded that the source oI HT`s intuitive appeal is not obviously captured Iully by
the view that death harms its victim in this way.
A second worry concerns the Iact that, according to deprivationist priorism,
death`s harm is subjectively inaccessible to its victim. To be sure, DV represents a
welcome expansion oI the Epicurean view that harm must involve a condition experi-
enced as painIul or otherwise hurtIul. Nevertheless, it is surprising that so proIound a
harm could aIIect the living subject and yet be utterly inaccessible epistemically or
phenomenologically: neither known to, nor Ielt by, its antemortem victim. To be clear,
the objection here does not indict the broad claim that some harms may be unwittingly
incurred; oI course they can. The worry is narrower in scope; it concerns whether the
claim that deaths harm is unwittingly incurred could possibly Iigure in an adequate
deIense oI HTin a view that not only answers the Epicurean challenge, but also re-
spects the Iull Iorce and spirit oI that challenge. Put more suggestively, might there be
a way oI answering the Epicurean challenge that both retains the priorist answer to the
timing question and respects the intuition behind HT that the subject oI death`s harm
really is a victim?
Having simply registered these concerns, let us temporarily bracket the priorist
part oI the deprivationist priorism combo and consider what might be said on behalI
oI DV itselI.

12
E.g., Geach (1969: 72).




DEATH`S DISTINCTIVE HARM



7
2. Deprivation Harm
In Iact, there is much to recommend DV. First oI all, DV is a species oI a widely held
theory oI individual interests, according to which an event or state oI aIIairs is overall
bad Ior someone when, and to the extent that, it makes her liIe worse than it otherwise
would be, as measured in terms oI the intrinsic goods she attains and the intrinsic evils
she suIIers. DV adapts this comparativist approach to the particular case oI mortal
harm, claiming that, death is overall bad Ior us only when, and to the extent that, it
prevents us Irom having certain intrinsic goodsgoods we would have enjoyed had
death not occurred. Because it seems to many (including myselI) that a comparativist
approach to individual interests is broadly correct, DV`s application oI this approach
to the case oI mortal harm inherits some oI this credibility.
Additional credibility stems Irom the Iact that DV`s diagnosis is plausibly given
not only oI mortal harms, but oI non-mortal harms as well. Suppose, Ior instance, that
I planned to meet a Iriend Ior lunch yesterday but was unable to reach the restaurant in
time due to a traIIic jam in which I became ensnared along the way. In this case, the
harm inIlicted on me by the traIIic jam is plausibly accounted Ior in terms oI the en-
joyable experience I was prevented Irom havingoI the experience, in other words, I
would have had were it not Ior the traIIic jam. In this wayby assimilating the expla-
nation oI mortal harm to an explanation plausibly given oI harms whose standing is
beyond disputethe deprivationist not only oIIers a reasonable answer to the harm
question but also strengthens the plausibility oI the case Ior HT.
There is a third advantage oI DV that I want to register. But beIore I do so, it will
help Iirst to draw out a Ieature oI HT that may have passed unnoticed. Whereas the
conclusion oI the Epicurean argument is quite strongdeath cannot harm the one who
diesthe claim expressing its rejection is quite weak: death can harm the one who
dies. This contingency oI HT is reIlected in typical statements oI DV:
death is bad Ior a person . provided that the liIe that is thereby lost would on bal-
ance have been worth living;
13

when death is bad for a person, it is bad Ior him because he will be worse oII dead
than he would have been iI he had lived;
14

death, when bad for someone, is bad in virtue oI the goodness it keeps that person
Irom having;
15


13
McMahan (1993: 262); emphasis added.
14
Feldman (1993: 308); emphasis added.
15
Bradley (2004: 2); emphasis added.




DEATH`S DISTINCTIVE HARM



8
|death`s| deprivation harm can be made clear by comparing the actual world, in
which I have just died, to the closest possible world in which I live longer. If the lat-
ter is more valuable for me than the former, dying has harmed me.
16

The rationale behind these conditionalized, deprivationist statements oI HT can be
illustrated, I believe, by considering an important objection to Nagel Iirst raised by
McMahan.
17
McMahan notes that Nagel`s initial Iormulation oI DV has it simply that
death is bad because oI the possible Iuture goods oI which its victim is deprived.
18

Whilst he is an advocate oI DV, McMahan points out the Iollowing important wrinkle:
the badness oI death is determined not by the deprivation oI goods the victim other-
wise might have enjoyed, but by whatever goods the victim in fact would have en-
joyed. Since merely possible goods are inIinite in number, McMahan recognizes that,
iI death`s harm consisted in the deprivation oI possible goods, no death would count
as worse than any other, since all victims would be deprived oI an inIinite number oI
possible goods. But that conIlicts with some commonly held belieIse.g., that, ceter-
is paribus, earlier deaths are worse than later deaths. According to McMahan`s re-
Iinement oI Nagel`s initial statement oI DV, then, death is bad (iI it is) in proportion to
the quality and quantity oI goods its victim actually would have enjoyed had she not
died when and how she did. Subsequent developments oI DV preserve this reIine-
ment, and most deprivationists seem to agree with McMahan when he writes that a
signiIicant virtue oI his revision to Nagel`s view is that it preserves
the view that most oI us take in cases in which it seems that death is not a bad thing
Ior the person who dies, or is even a good thing. These are cases in which we Ieel that
suicide would be rational, or euthanasia justiIied. The reason why death is not bad in
these cases is that, iI the person were not to die, the liIe that he would subsequently
have would not be worth living. II, however, we assume with Nagel that death should
be evaluated relative to the possibilities Ior good that would be imaginable in its ab-
sence, then it seems that we should regard death as an evil even in these cases. This
makes it diIIicult to see how those who take Nagel`s view can Iind cases in which su-
icide would be rational or in which euthanasia would even be conceptually possible.
19

I agree. And this seems to me to be a third and important strength oI the deprivationist
approach, vi:. that, once revised in the way McMahan recommends, it can accommo-
date the robust convictions that some deaths are worse than others and that death
might not be overall bad Ior some oI those who die. For my part, I would suggest that
we should not accept any answer to the harm questionnor, more generally, any de-
Iense oI HTthat does not preserve this commitment.

16
Luper (2007: 240); emphasis added.
17
McMahan (1993: 24244).
18
Nagel (1993: 6264).
19
McMahan (1993: 244).




DEATH`S DISTINCTIVE HARM



9
Nevertheless, I think it is in DV`s handling oI just these and similar cases that we
can catch a glimpse oI its weaknessor, more precisely, its incompleteness. To see
this, let us begin by considering a diIIerent example altogether.
One explanation Ior why involuntary servitude harms the enslaved might be given
purely in terms oI deprivation harm.
20
On this view, slavery harms only by depriving
the enslaved oI the many and various goods they otherwise would have enjoyed. To be
sure, many actual slaves were (and are) deprived oI such goods, and these depriva-
tions certainly constitute harms. Nevertheless, deprivation harm is only a contingent
Ieature oI the concept oI slavery: a Iorm oI slavery in which individuals are enslaved
but capable oI partaking in all oI the goods enjoyed by Iree individuals is not incon-
ceivable. About this atypical Iorm oI slavery the deprivation theorist would have to
say that the non-deprived slave suIIers no harm. And yet the very idea oI an unharmed
slave will strike many oI us as a misnomer: slaves are harmed even iI they are not de-
prived. The harm that slavery inIlicts, we might say, is a necessary Ieature oI en-
slavement: not merely a possible eIIect, but an essential concomitant oI the institution
oI slavery
21

What is the nature oI this necessary aspect oI slavery`s harm? I am inclined to
think that it has something to do with the degradation oI dignity owing to slaves` sta-
tus as inIeriors. Through its conIerral and enIorcement oI an inIerior status, the institu-
tion oI slavery cannot Iail to violate the dignity oI those who are enslaved, and that
violation cannot be explained entirely in terms oI deprivation. Enslavement would
harm these individuals in this way even iI they were not deprived oI the goods en-
joyed by Iree individuals.
But a proper deIense oI this suggestion about slavery`s harm is well beyond the
scope oI the present discussion, and nothing in what Iollows hinges on the substance
oI the Ioregoing line oI thought.
22
Whatever the appropriate corrective to a purely dep-
rivationist analysis oI slavery`s harm may be, the purpose oI this example is simply to
illustrate by analogy the kind oI criticism to which that analysis will be subject when
we turn our attention to mortal harm.
23
For the cases analogous to the case oI non-
deprived slaves include those who lack meaningIul projects, relationships, and inter-

20
Although, to my knowledge, this view has not been advanced at quite this level oI abstraction, some oI the best
known theories oI slavery`s harm are clearly deprivation theories. According to Patterson`s 'natal alienation view,
Ior instance, slavery`s harm is to be understood as 'the loss oI ties oI birth in both ascending and descending gener-
ations and the 'loss oI native status, oI deracination (1982: 8). Nevertheless, oI primary interest here is not
whether such a view has in Iact been endorsed, but why such a view (whether or not endorsed) would Iail to ac-
count Ior the totality oI slavery`s harm.
21
Lawson (1992): 912.
22
The interested reader is commended to Debes (2009), Lawson (1992) and Roberts-Thomson (2008), all oI which
have sparked my thinking in this context. Other relevant work includes Cohen (1998), Elkins (1959), Hare (1979),
Kershnar (2003), Patterson (1982), Stampp (1956), and the readings Iound in Lott (1998).
23
Nor is this analogy in this context original to me. See Rosenbaum (1989) and Luper (2009: 9697).




DEATH`S DISTINCTIVE HARM



10
ests Ior death to undermine; those who invest in only those projects, relationships, and
interests that death cannot thwart; and those who have an interest in dying. These are
the subjects whom the deprivationist rules out as unharmed, or even beneIitted, by
death. On the contrary, I want now to argue that even these subjects do not escape
death`s harmthat, as slavery does to the enslaved, death harms its victims necessari-
ly.
To this end, let us return to the group oI individuals with perhaps the strongest
claim to being unharmed by death, vi:. those who have an interest in dying. As we saw
above, McMahan expresses the deprivationist position when he writes that 'the reason
why death is not bad in cases oI justiIiable euthanasia and rational suicide 'is that, iI
the person were not to die, the liIe that he would subsequently have would not be
worth living.
24
But this line oI thinking is mistaken: it conIlates the possible rationali-
ty oI suicide or justiIiability oI euthanasia, on the one hand, with the alleged non-
badness oI death-by-suicide or death-by-euthanasia, on the other hand. We can and
should agree that there are circumstances in which suicide may be rational and eutha-
nasia justiIied. But Irom the Iact that some lives are not worth continuing it does not
Iollow that the subjects oI these lives go unharmed by their deathseven when those
deaths are elective.
II the overall goodness oI these deaths were suIIicient to establish that the victims
were not harmed, why do we nevertheless mourn them? We do not mourn the contin-
ued suIIering oI the subject, since no such suIIering continues. Nor is our mourning in
these cases exhausted by our recognition oI the setbacks and sadness that we or others
will endure as a result.
25
The deprivationist, I think, would likely suggest that what we
mourn are the goods that death prevented the individual Irom acquiring, even though
on balance those goods are outweighed by the evils that would have been suIIered had
she continued living. But I doubt this is a coherent global attitude Ior a deprivationist
to adopt, because one who acknowledges that a death was overall good cannot Iail to
see the evils the individual would have suIIered as outweighing the goods oI which
she was deprived. For a deprivationist to mourn an overall good death would seem to
be tantamount to denying that the death was overall good.
But be that as it may. The Iact is that we do mourn even these deaths. And the rea-
son, I think, does not principally have to do with the ledger oI goods that would have
been enjoyed and evils that would have been suIIered. Rather we mourn because death

24
McMahan (op cit.).
25
Is it impossible Ior one`s mourning oI an overall good death to be exhausted by this recognition? Might not A
mourn the overall good death oI B not Ior selIish reasons among others, but Ior selIish reasons onlvsolely be-
cause B`s death results in a setback Ior A? About such a case I am inclined to say that A is not, in Iact, mourning B`s
passing at all. A`s emotion would not be grieI for B, but rather something more like regret Ior a consequence oI B`s
passing. For Iurther discussion, see Solomon (2004).




DEATH`S DISTINCTIVE HARM



11
harms even those Ior whom continued liIe would be overall bad.
26
The harm that
death inIlicts, I suggest, is maniIest in the very Iact that hastened nonexistence could
possibly represent a superior alternative to continued existence.
27
In other words, the
harm is enIolded into the Iact that euthanasia could be justiIiable, into the very Iact
that suicide might represent the most rational course oI action.
28

To say only this much does little to clariIy the aspect oI death`s harm that, I claim,
cannot be explained in deprivationist terms. But beIore going any Iurther, note that iI
death harms even those Ior whom death is overall good, this strongly suggests that, in
general, death harms the one who dies whether or not the nearest possible world in
which the subject`s counterpart lives longer is comparatively superior to the actual
world in which the subject dies. And in that case, even Ior subjects whose death is
overall bad, the nature oI the harm inIlicted is not exhausted by the welIare whose
enjoyment is precluded, the projects whose completion is thwarted, the relationships
whose Ilowering is prevented, etc. Even when death`s victims are thus deprived, in
other words, these deprivations do not Iully account Ior the harm that death inIlicts.
Whatever this non-deprivationist Iacet oI death`s harm is, it would seem to be indiI-
Ierent to the valence oI death`s overall impact.
3. Restriction Harm
In the space remaining, I want at least to begin to characterize the nature oI this unno-
ticed aspect oI mortal harm. In so doing, I will indicate both why it accords with the
priorist answer to the timing question and how it avoids the challenges to deprivation-
ist priorism raised above.
I believe that the overlooked Iacet oI death`s harm can be understood as a kind oI
restriction oI, or constraint on, the autonomy oI the one who dies and that this re-
striction harm is distinct Irom (although complementary to) the deprivation harm that

26
OI course, our reasons Ior mourning even good deaths might be conIused or inapt in some way, and so we can
hardly assume them to be dispositive. My aim here is only, Iirst, to occasion in you an intuition that I myselI have
about these cases; and second, to note that, in so Iar as this intuition is widely shared, it would be better iI we were
able to explain it.
27
It might be suggested that the remorse Ielt in these cases is not (primarily) Ior death`s harm, but Ior the harms in-
Ilicted by whatever condition makes the suicide rational or the euthanasia justiIiable (e.g., a painIul and terminal
illness). But while it is entirely plausible that this grieI might extend to include sorrow Ior the harms inIlicted by
whatever state oI aIIairs grounds the individual`s interest in dying, my claim is only that no complete account oI our
grieI in these cases would overlook our remorse Ior the harm inIlicted by death itselIby the Iact that continued
existence could Iail to represent the best all-things-considered alternative.
28
This claim that death inIlicts a harm even on those Ior whom euthanasia is justiIied does not settle the ethical, legal,
and political questions regarding euthanasia. What matters in answering these questions is whether death constitutes
an overall harm. For my part, I believe that euthanasia can be morally justiIiable (vi:. Ior subjects Ior whom death
would be overall good), and I believe that this commitment is widely shared. But beyond registering these belieIs
here, I have no interest in engaging the larger, practical questions concerning euthanasia.




DEATH`S DISTINCTIVE HARM



12
death can also inIlict. Death`s restriction harm, I suggest, consists in the Iact that every
exercise oI a subject`s autonomy is possibly thwarted by her death. This restriction
can be (and oIten is) maniIest in our awareness oI our Iinite nature: oI the Iact that
death will bring about one`s nonexistenceoI the apparent truth, in other words, oI
what is sometimes labeled the 'termination thesis.
29
For lurking at the edge oI every
exercise oI one`s autonomy is the knowledge that 'I could cease to exist at any mo-
ment, and this might be that momentthe knowledge that we are limited in our
agency by an end that is certain to come, oIten unIoreseen in arriving, and irrevocable
once past. OI course, this knowledge may inIorm determinations about when, how,
and why to exercise one`s autonomy. And iI particular exercises oI autonomy (the
adoption oI certain projects, the pursuit oI speciIic interests, etc.) turn out to be actual-
ly thwarted by death, then death will be rightly said to deprivation-harm the subject in
question. But what this deprivationist analysis oI death`s harm does not take into ac-
count is the Iact that, at every moment oI our lives, we are endowed with only a lim-
ited autonomy, every exercise oI which, whether ultimately thwarted or not, is possi-
blv thwarted, and known to be so. It is in the boundedness oI our capacity Ior selI-
determination that the distinctive aspect oI death`s harm consists, and this restriction
harm is inIlicted (wittingly or not) on every subject endowed with that capacity at eve-
ry moment in which they are so endowed. Notably, then, these victims oI death`s re-
striction harm are antemortem, autonomous subjects, and thereIore the restriction-
harm part oI death`s overall harm accords with the priorist answer to the timing ques-
tion.
30

About the 'restriction view that I am advocating here, note that although death`s
restriction-harm oI all living, autonomous subjects goes some way toward vindicating
the priorist answer to the timing question, this vindication may be only partial. Noth-
ing I have said rules out the possibility that the two aspects oI death`s harm are inIlict-
ed at diIIerent times; hybrid viewssay, restrictionist priorism plus deprivationist
subsequentismare possible at least in principle.
Note Iurther that death`s restriction harm is suIIered only by those mortal beings
that are also autonomous. Although death might still deprivation-harm them, human

29
The termination thesis asserts simply that death brings about the nonexistence oI the one who dies. See Feldman
(2000: 98101) Ior a useIul introduction. Advocates oI the termination thesis'terminators, as Feldman memora-
bly describes theminclude Feinberg (1984), Hershenov (2005), Luper (1987), Nagel (1979), Olson (1997), Ros-
enberg (1983), Rosenbaum (1986), Silverstein (1980), and Yourgrau (1987), not to mention, oI course, Epicurus and
Lucretius. Admittedly, the termination thesis is not universally endorsed; notable 'continuators, as we might call
them, include Ayers (1991), Carter (1999), Feldman (1992, 2000), Mackie (1999), Persson (1999), and Thomson
(1997).
30
Hauntingly evocative both oI my criticism oI DV and oI my proposal concerning restriction harm is Philip Larkin`s
poem 'Aubade, the Iirst lines oI which read: 'I work all day, and get halI drunk at night./Waking at Iour to sound-
less dark, I stare./In time the curtain edges will grow light./Till then I see what`s really always there:/Unresting
death, a whole day nearer now,/Making all thought impossible but how/And where and when I shall myselI die./




DEATH`S DISTINCTIVE HARM



13
and nonhuman animals (not to mention plants and other living organisms) lacking this
capacity Ior selI-determination are not restriction-harmed by death. An autonomous
human being is necessarily restriction-harmed by her Iuture death so long as she re-
mains autonomous. Should she lose her autonomy due to some non-Iatal condition or
injuryor, Ior that matter, should she somehow become immortalshe would cease
to be the sort oI being who could be restriction-harmed.
It is important to emphasize that restriction harm is not just a disguised Iorm oI
deprivation harm. The claim is not that death harms its victim by depriving her oI the
ability to exercise her autonomy; iI it were, then restriction harms could simply be
subsumed under DV. The claim is rather that every living, autonomous subject who
dies is harmed by death in the Iollowing sense: no exercise oI her autonomy Iails to be
possibly limited, and this limitation is intrinsically harmIul. It is not that the living
subject lacks autonomy, oI course. It is rather that the Iuture death oI every living, au-
tonomous subject ensures that every exercise oI that autonomy is possibly limited.
Indeed, unlike the deprivation harms which may or may not accompany death, the
restriction harm that death inIlicts is not contingent, but necessary. It is necessary not
because it is a necessary truth that all autonomous, living subjects will one day cease
to exist (that is hardly a necessary truth), but because the harm visited upon such sub-
jects by their Iuture deaths is inIlicted simply as a Iunction oI death`s occurrence: giv-
en its Iuture occurrence, the subject cannot Iail to be harmed.
31

Furthermore, death`s restriction harm is distinctive. Whereas a subject`s interests
may be deIeated by a variety oI non-mortal circumstances, only death aIIects every
exercise oI a subject`s autonomy in quite this way. For while there might be other cir-
cumstances in which every excercise oI a mortal being`s autonomy is possibly lim-
itede.g., circmustances imposed Irom without (e.g., slavery) and Irom within (e.g.,
dementia)only death imposes this restriction necessarily, solely as a Iunction oI the
subject`s mortality and autonomy.
Returning to the objections to deprivationist priorism raised earlier, note Iirst that
living subjects may be, and oIten are, Iully aware oI death`s restriction harm. On the
view sketched here, any selI-conscious, autonomous subject may be cognizant oI her
constrained autonomy at any moment oI her liIe. Recall also the concern that death`s
deprivation harm eIIects no real change in the living subject who subsequently dies,
but only a mere Cambridge change. InsoIar as this view Iails to satisIy the Epicurean,
the restriction view may oIIer a welcome supplement Ior advocates oI HT seeing a

31
The Iirst stanza oI Larkin`s poem continues: 'Arid interrogation: yet the dread/OI dying, and being dead,/Flashes
aIresh to hold and horriIy./The mind blanks at the glare. Not in remorse/The good not used, the love not given,
time/Torn oII unusednor wretchedly because/An only liIe can take so long to climb/Clear oI its wrong begin-
nings, and may never:/But at the total emptiness Iorever,/The sure extinction that we travel to/And shall be lost in
always. Not to be here,/Not to be anywhere,/And soon; nothing more terrible, nothing more true.//




DEATH`S DISTINCTIVE HARM



14
view according to which the harm incurred is metaphysically more robust. On this
view, part oI what it is to be a mortal, autonomous subject is to be encumbered with
the harm inIlicted by Iuture death. Death`s restriction harm is a Iunction oI a living
being`s status as autonomous and so is no mere Cambridge change, but a real change
toor, better, a real state ofthis subject.
It might be asked whether the view advocated herespeciIically, the claim that
death`s limitation oI one`s autonomy is intrinsically harmIulassumes that autonomy
is intrinsically valuable. The short answer is that, in the interest oI simplicity, I am
indeed making this assumption. But it is not, I think, an altogether implausible or un-
popular assumption. Quite apart Irom Kant, Ior instance, no less than Mill himselI
held that autonomy is 'one oI the elements oI well-being.
32
And nothing I have said
thus Iar would seem to preclude proponents oI competing theories oI autonomy Irom
endorsing my account oI death`s restriction harm.
Note Iurther that the notion oI restriction harm helps to illuminate a distinction
drawn earlier between the possible rationality oI suicide or justiIiability oI euthanasia,
on the one hand, and the alleged non-badness oI death-by-suicide or death-by-
euthanasia, on the other hand. For it might be asked: iI death necessarily harms the
one who dies, how then could suicide be rational or euthanasia justiIiable? Here, while
continuing to insist that death necessarily restriction-harms all autonomous subjects
who die, we can allow that death`s restriction harm is only a pro tanto harm. Doing so
enables us to hive oII the question oI whether the death oI an individual who has an
interest in dying does or does not constitute an overall, all-things-considered harm to
that individualand thus whether or not the death in question is justiIiable or not,
rational or notwithout thereby denying that it constitutes a pro tanto, restriction
harm in any case. The suggestion, in other words, is that the justiIiability oI an in-
stance oI euthanasia (Ior example) depends in part on whether death`s pro tanto re-
striction harm is outweighed by death`s beneIit to the individual in questionwhere
the measurement oI that beneIit is established in the way Iamiliar Irom the depriva-
tionist analysis, namely by determining whether the nearest possible world in which
the subject`s counterpart lives longer is comparatively superior to the actual world in
which the subject dies.
33


32
Mill (1859): ch. 3.
33
The concluding stanzas read: 'And so it stays just on the edge oI vision,/A small unIocused blur, a standing
chill/That slows each impulse down to indecision./Most things may never happen: this one will,/And realisation oI
it rages out/In Iurnace Iear when we are caught without/People or drink. Courage is no good:/It means not scaring
others. Being brave/Lets no-one oII the grave./Death is no diIIerent whined at than withstood.//Slowly light
strengthens, and the room takes shape./It stands plain as a wardrobe, what we know,/Have always known, know that
we can`t escape/Yet can`t accept. One side will have to go./Meanwhile telephones crouch, getting ready to ring/In
locked-up oIIices, and all the uncaring/Intricate rented world begins to rouse./The sky is white as clay, with no
sun./Work has to be done./Postmen like doctors go Irom house to house.




DEATH`S DISTINCTIVE HARM



15
The distinction between pro tanto and all-things-considered harms also helps in
addressing another question that might be asked: iI S is harmed because every exer-
cise oI her autonomy is possibly thwarted by her Iuture death, does it Iollow that S
would be unharmed iI her death never occurrediI, in other words, S were immortal?
As with euthanasia and suicide, the question oI whether immortality is an all-things-
considered harm can be answered only by consulting the complete tally oI goods and
evils aIIorded to the deathless subject. The debate surrounding this calculation is not
one that I can engage here.
34
But even skeptics concerning the overall desirability oI
immortality should agree that our limited liIespan constitutes a pro tanto harm to us. II
this were not the case, it is diIIicult to see why the debate over immortality would be
oI any interest whatsoever.
4. Conclusion
I have argued that death`s harm is not homogenous, but multiIaceted; not entirely con-
tingent, but partly necessary. While death may deprivation-harm some, not all are dep-
rivation-harmed by death. On this basis, however, we should not conclude that death
does not harm all who die. Rather, we should inIer that death`s harm is not Iully ac-
counted Ior in terms oI the loss oI those oI liIe`s goods one would have enjoyed had
one continued living. On the view sketched here, each and every autonomous, living
subject who dies is harmed because her every exercise oI autonomy is possibly un-
dermined by death. II this is correct, then HT is unnecessarily weak: not only can
death harm the one who dies, it always does sowhether or not the death is overall
good Ior that individual.
35

ReIerences
Ayers, M. (1991) Locke, 2 vols. (Routledge).
Bortolotti, L, and Y. Nagasawa (2009) 'Immortality Without Boredom, Ratio 22: 26177.
Bradley, B. (2004) 'When Is Death Bad Ior the One Who Dies? Nous 38: 128.
(2009) Well-Being and Death (OxIord University Press).
Brueckner, A. and J. M. Fischer (1986) 'Why is Death Bad? Philosophical Studies 50: 21323; reprint-
ed in Fischer (1993), 22129.

34
Beyond Williams (1973), there has been some interesting work on this topic oI late; e.g., Bortolotti and Nagasawa
(2009), Fisher (2006), and Moore (2006).
35
Previous versions oI this paper were presented at the American Philosophical Association (Central Division),
Southern Society Ior Philosophy and Psychology, Animals & Society Institute (Wesleyan University), and at New-
castle University. An expanded version will appear in Death. Its Meaning, Metaphvsics, and Moralitv, ed. Thom
Brooks (Princeton University Press, Iorthcoming). For helpIul Ieedback and conversations, I am grateIul to JenniIer
Baker, Ben Bradley, Remy Debes, Neil Feit, Lori Gruen, Bill Lawson, JeII McMahan, Alastair Norcross, Tim
Roche, Paul Snowdon, James Stacey Taylor, students in my seminar on death at the University oI Memphis, and es-
pecially Steve Luper.




DEATH`S DISTINCTIVE HARM



16
Carter, W. (1999) 'Will I Be a Dead Person? Philosophv and Phenomenological Research 59: 16772.
Cohen, J. (1998) 'The Arc oI the Moral Universe, in Lott (1998): 281327.
Debes, R. (2009) 'Dignity`s Gauntlet, Philosophical Perspectives 23: 4578.
DeGrazia, D. (2005) Human Identitv and Bioethics (Cambridge University Press).
Donnelly, J., ed. (1994) Language, Metaphvsics, and Death, 2
nd
ed. (Fordham University Press).
Elkins, S. (1959) Slaverv. A Problem in American Institutional and Intellectual LiIe (University oI Chi-
cago Press; 3
rd
rev. ed., 1976).
Epicurus, 'Letter to Menoeceus, in Epicurus. The Extant Remains, trans. C. Bailey (Clarendon Press,
1926).
Feinberg, J. (1984) 'Harm to Others, in Harm to Others (OxIord University Press, 1984), 7995; re-
printed in Fischer (1993), 17190.
Feit, N. (2002) 'The Time oI Death`s MisIortune, Nous 36: 35983.
Feldman, F. (1991) 'Some Puzzles about the Evil oI Death, Philosophical Review 100: 20527; reprint-
ed in Fischer (1993): 30726.
(1992) Confrontations with the Reaper (OxIord University Press).
(2000) 'The Termination Thesis, Midwest Studies in Philosophv 24: 98115.
Fischer, J. M., ed. (1993) The Metaphvsics of Death (StanIord University Press).
(2006) 'Epicureanism About Death and Immortality, Journal of Ethics 10: 35581.
Geach, P. T. (1969) God and the Soul (Routledge & Kegan Paul).
(1972) Logic Matters (Basil Blackwell).
(1979) Truth, Love, and Immortalitv. An Introduction to McTaggarts Philosophv (Hutchison).
Grey, W. (1999) 'Epicurus and the Harm oI Death, Australasian Journal of Philosophv 77: 35864.
Hare, R. M. (1979) 'What is Wrong with Slavery, Philosophv and Public Affairs 8: 10321.
Hershenov, D. (2005) 'Do Dead Bodies Pose a Problem Ior Biological Approaches to Personal Identity?
Mind 114: 3159.
Kershner, S. (2003) 'A Liberal Argument Ior Slavery, Journal of Social Philosophv 34: 51036.
Lamont, J. (1998) 'A Solution to the Puzzle oI When Death Harms its Victims, Australasian Journal of
Philosophv 76: 198212.
Lawson, B. E. (1992) 'Oppression and Slavery, in Between Slaverv and Freedom. Philosophv and
American Slaverv (Indiana University Press, 1992), ed. H. McGary and B. E. Lawson, 115.
Lott, T. L., ed.(1998) Subfugation and Bondage. Critical Essavs on Slaverv and Social Philosophv
(Rowman and LittleIield, 1998).
Luper, S. (1987) 'Annihilation, The Philosophical Quarterlv 37: 23352; reprinted in Fischer (1993),
26990.
(2004) 'Posthumous Harm, American Philosophical Quarterlv 41: 6372.
(2006) 'Death, Stanford Encvclopedia of Philosophv, Fall 2008 edition, ed. E. Zalta,
http://plato.stanIord.edu/entries/death.
(2007) 'Mortal Harm, The Philosophical Quarterlv 57: 23951.
(2009) The Philosophv of Death (Cambridge University Press).
Mackie, D. (1999) 'Personal Identity and Dead People, Philosophical Studies 95: 21942.
McMahan, J. (1988) 'Death and the Value oI LiIe, Ethics 99: 3261; reprinted in Fischer (1993): 233
66.
(2002) The Ethics of Killing. Problems at the Margins of Life (OxIord University Press).
Mill, J. S. (1859) On Libertv (various editions).
Moore, A. W. (2006) 'Williams, Nietzsche, and the Meaninglessness oI Immortality, Mind 115: 31130.
Nagel, T. (1970) 'Death, Nous 4: 7380; reprinted in Mortal Questions (Cambridge University Press,
1979), 110; and in Fischer (1993), 5970.
Olson, E. (1997) The Human Animal. Personal Identitv Without Psvchologv (OxIord University Press).
Patterson, O. (1982) Slaverv and Social Death (Harvard University Press).
Persson, I. (1999) 'Our Identity and Separability oI Persons and Organisms, Dialogue 38: 51933.
Pitcher, G. (1984) 'The MisIortunes oI the Dead, American Philosophical Quarterlv 21: 18388; re-
printed in Fischer (1993), 15968.
Roberts-Thomson, S. (2008) 'An Explanation oI the Injustice oI Slavery, Res Publica 14: 6982.




DEATH`S DISTINCTIVE HARM



17
Rorty, R. (1989) 'The Contingency oI SelIhood, in Contingencv, Ironv, and Solidaritv (Cambridge Uni-
versity Press), 2343.
Rosenbaum, S. (1986) 'How to Be Dead and Not Care, American Philosophical Quarterlv 23: 21725;
reprinted in Fischer (1993), 11934.
(1989) 'Epicurus and Annihilation, The Philosophical Quarterlv 39: 8190; reprinted in
Fischer (1993): 293304.
Rosenberg, J. (1983) Thinking Clearlv about Death (Prentice-Hall).
Ruben, D-H. (1988) 'A Puzzle about Posthumous Prediction, Philosophical Review 97: 21136.
Scott, D. (2006) 'Aristotle on Posthumous Fortune, Oxford Studies in Ancient Philosophv 18: 21129.
Silverstein, H. (1980) 'The Evil oI Death, Journal of Philosophv 77: 401 24; reprinted in Fischer
(1993), 95116.
(2000) 'The Evil oI Death Revisited, Midwest Studies in Philosophv 24: 11635.
Solomon, R. C. (2004) 'On GrieI and Gratitude, in In Defense of Sentimentalitv (OxIord University
Press), 75107.
Stampp, K. (1956) The Peculiar Institution. Slaverv in the Antebellum South (KnopI).
Sumner, L. (1976) 'A Matter oI LiIe and Death, Nous 10: 14571.
Thomson, J. J. (1997) 'People and their Bodies, in Reading Parfit, ed. J. Dancy (Blackwell), 20229.
Warren, J. (2004) Facing Death. Epicurus and His Critics (OxIord University Press).
Williams, B. (1973) 'The Makropulos Case: ReIlections on the Tedium oI Immortality, in Problems of
the Self (Cambridge University Press), 82100; reprinted in Fischer (1993), 7392.
Yourgrau, P. (1987) 'The Dead, Journal of Philosophv 86: 84101; reprinted in Fischer (1993), 13756.