THE INSIGNIFICANCE OF PERSONAL IDENTITY FOR BIOETHICS

There has long been consensus that personal identity and bioethics are importantly

intertwined, in particular that certain key bioethical positions depend heavily on the truth of

certain metaphysical accounts of identity. In 2003, David DeGrazia forcefully concluded an

essay on the topic in Philosophy & Public Affairs by saying, “[W]e cannot ignore personal

identity theory in examining the marginal cases [in bioethics]….”1 I think, to the contrary, that

we can for the most part, that personal identity is far less significant to bioethics than is usually

thought. To show this, I’m going to examine arguments on three main bioethical issues where

personal identity has been thought to be non-derivatively important—abortion, the definition of

death, and advance directives—and argue that in each case it is something other than identity

that’s doing the relevant work. I leave open whether or not there might be other examples of a

bioethical argument plausibly depending non-derivatively on personal identity, but one might

think of this paper as both a challenge to present such a case and an expression of skepticism

about its prospects.

The Nature of the Personal Identity at Stake

Before beginning the main line of argumentation, I need to take a moment to discuss the

precise nature of personal identity at stake in the following claim, the one I will deny:

considerations of personal identity are non-derivatively significant for the bioethical issues of

abortion, definition of death, and advance directives. There are in general two closely related

questions currently pursued in personal identity theory that are allegedly relevant here. First,

there is the question of what preserves our identity across time. Second, there is the question of

what our essence actually is. Let me briefly discuss each.

Regarding the first question, I will be exploring the significance of criteria of diachronic

numerical personal identity to bioethics, criteria which identify the conditions under which a

person at one time is one and the same being as some individual at a different time. Now up

until recently, criteria of personal identity were typically about what makes a person at one time

identical to a person at another time. But primarily because of the careful work of advocates of

biological criteria of identity2, the nature of the debate has been transformed. One important

consideration here is that I was a newborn infant, it seems—perhaps even a fetus—but neither

infants nor fetuses are persons, i.e., self-conscious entities with highly complex psychologies.

But if I was indeed an infant, then personhood is a concept befitting me for only a phase of my

life, akin perhaps to adolescence, adulthood, fatherhood, and the like. If so, then the boundaries

of my identity may be broader than the boundaries of my identity as a person, in which case we

should be more neutral in articulating the general formula: instead of looking for conditions of

person-to-person identity, we ought to be looking for conditions of identity between a person at

one time and an individual being at another, where of course such individual beings might well

be persons (or not).

“Personal identity” may be something of a misnomer, then, if it is taken to be about the

conditions for the preservation of identity across time for all and only persons. Indeed, the

identity conditions for “individuals like us” is now the most favored target of identity theorists,

and such identity is typically taken to be a function of our essence, which is the second relevant

element of personal identity theory for our purposes. That is, a determination of the essence of

an individual like you and me will determine what the persistence conditions actually are for

such entities. If, for example, you and I are essentially biological creatures, then our identity

across time will consist in continuity of that biological essence. Similarly, if we are essentially

2

psychological creatures (persons?), then our diachronic identity will consist in psychological

continuity. As a result, questions of essence have, for most personal identity theorists, become a

crucial component of investigations into the nature of diachronic numerical identity. As

DeGrazia notes, a prominent question in personal identity theory has become “‘What are we

human persons, most fundamentally: persons, human animals, or something else?’”3

There is one type of personal identity I will not be discussing, namely, the recently

developed and deployed conception of narrative identity.4 On this view, various experiences and

actions are gathered together into the life of one person via their being part of a coherent self-told

story about one’s life. This sort of identity has been taken to be relevant to bioethics as well, but

because discussion of it would require a great deal of exposition and would raise a host of other

issues, I will set it aside here. Instead, I will focus simply on whether or not questions of our

essence or our diachronic numerical identity have the significance for bioethics they have been

claimed to have.

Abortion

Generally, there are two ways personal identity has been thought to be significant to the

abortion debate. One is that it allegedly can provide support to a theory of moral status. The

other is that it allegedly can be used to distinguish abortion from contraception. Let me begin

with the first sort of move.

1 David DeGrazia, “Identity, Killing, and the Boundaries of Our Existence,” Philosophy & Public Affairs 31 (2003): 413-442, p. 442.

2 For a very good example, see Eric Olson, The Human Animal (Oxford: Oxford University Press, 1997).

3 DeGrazia, p. 414.

4 For sustained development and discussion of the notion of narrative identity, see Marya Schechtman, The Constitution of Selves (Ithaca, NY: Cornell University Press, 1996). For an application of narrative identity to bioethical concerns, see David DeGrazia, Human Identity and Bioethics (Cambridge: Cambridge University Press, 2005).

3

Those engaged in this project tend to favor a moderate to liberal pro-choice conclusion,

one that actually likens early abortion to contraception. Jeff McMahan, for example, ostensibly

rests part of his view of moral status on his Embodied Mind account of personal identity,

according to which you and I, who are essentially embodied minds, don’t begin to exist until the

organisms we inherit develop the capacity for consciousness, and from that point on what

preserves our identity “is the continued existence and functioning, in nonbranching form, of

enough of the same brain to be capable of generating consciousness or mental activity.”5 Now

you and I, of course, have significant moral status. An early fetus could never be someone like

you or I because an entity that’s essentially minded like we are is always minded and the early

fetus lacks the physical substrate supporting the capacity for being minded. This, McMahan

suggests, implies that an early fetus lacks the “special moral status” you and I have “sufficient to

make it seriously wrong to kill it.”6 It is, in his terminology, a something rather than a someone.

As a result:

An early abortion does not kill anyone; it merely prevents someone from coming

into existence. In this respect, it is relevantly like contraception and wholly

unlike the killing of a person. For there is, again, no one there to be killed.7

Nevertheless, the fact that some entity has a different essence than one of us—and so

could never be numerically identical with one of us—means neither that it has a different moral

status from us nor that, if it does, its different moral status is a function of that different essence.

Now McMahan openly admits the first point, noting that the conclusion about the early-stage

fetus having a different moral status also depends both on its not having “a special sanctity that

5 Jeff McMahan, The Ethics of Killing (Oxford: Oxford University Press, 2002), p. 68.

6 Ibid., p. 269.

7 Ibid, p. 267.

4

otherwise comparable nonhuman organisms lack” and on its not having “the relevant sort of

potential to become a person.”8 Suppose, though, that his later independent arguments succeed

in eliminating both exceptions. This result then might suggest that an appeal to our essence

could at least do some non-negligible work in supporting a theory of moral status, for such an

appeal reveals a clear-cut foundational difference between us and early fetuses—what seems an

analogous difference between us and a sperm or ovum—and so with buttressing by the non-

sanctity and non-potential-persons arguments such a point looks to be morally significant.

Now of course it should be obvious that metaphysical status implies nothing on its own

about moral status—this is a very old point and one that I’m not interested in rehashing here.

Indeed, all that McMahan needs here to bridge the “is/ought gap” is the principle he is likely

assuming, namely, if an entity lacks our essence (an embodied mind), it lacks our moral status

(assuming also that it has no independent moral status in virtue of its sanctity or potential

personhood). But if this is the relevant bridge principle, it raises my second question above,

namely, why should we think that moral status is a function of essence? As it turns out,

McMahan can’t really believe that it is. His focus is actually on the source of interests in a

creature, for they, absent actual personhood, are what determine the moral wrongness of killing

or harming it.9 But then what fundamentally matters is whether or not a creature has interests,

not whether or not it shares our essence. As it turns out, of course, anything with an embodied

mind has interests, but this is only a contingent matter: it’s possible for there to be disembodied

minds with interests, or embodied minds without interests. But presumably were either to be the

case, McMahan’s moral radar would surely continue to track those creatures with interests, the

fact that they did or didn’t have an embodied mind rendered irrelevant. As a result, even though 8 Ibid., p. 269.

9 See, e.g., ibid., pp. 245-265.

5

it’s contingently true that all and only those creatures with embodied minds have interests, what

matters for morality is the interests part, not the embodied minds part. What our essential nature

consists in is thus only derivatively significant here, significant only in virtue of the contingent

fact that it delivers the interests that are of genuine moral significance.

This point is brought out more clearly in McMahan’s treatment of later-term abortions,

the killing of more developed fetuses that are one of us, having passed the point at which their

organism’s capacity for consciousness has been activated. One might think that once one of us

has been brought into existence it will have the same moral status as the rest of us, but this isn’t

yet the case for McMahan. Rather, you and I have the high moral status we enjoy because we

are persons—entities with the capacity for self-consciousness—and so deserve respect.10 But

there are entities that, while individuals like us in virtue of a common essence, are not persons,

and so lack our high moral status; they are in fact governed solely by a different account of the

morality of killing.11 Any moral status they have—determining the seriousness of the wrongness

of killing them—depends entirely on their time-relative interest in continuing to live, itself a

function of the value of their future and their expected psychological unity with the embodied

mind that will undergo that future good. But because they lack the ability to anticipate,

contemplate, and form intentions about their future good, their psychological unity with that

future self is extremely weak, and so their time-relative interest in continuing to live is itself

weak, rendering the wrongness of killing them far less serious than the wrongness of killing

persons like you and me.12

10 Ibid., p. 275.

11 Thus McMahan calls his view a Two-Tiered Account of the morality of killing. See ibid., pp. 245-265.

12 See, e.g., ibid., pp. 275-276.

6

So what role do the conditions of our essence and/or our numerical identity play here?

As it turns out, having an embodied mind—being a someone who meets the conditions for

personal identity across time—isn’t what does any of the work to generate moral status in the

arena of abortion. For one thing, as we have just seen, being an embodied mind isn’t what

generates full moral status; for that, one needs to be a person, an entity deserving of respect. For

another, as we saw earlier, being an embodied mind isn’t even what generates partial moral

status, which is generated instead merely by the having of interests. Furthermore, the degree to

which one’s interests determine one’s moral status depends on one’s psychological unity with

some future beneficiary of value, but psychological unity just isn’t numerical identity.

Now McMahan explicitly assumes that identity should “coincide as closely as possible

with our sense of what matters,”13 but he also claims that the degree of warranted egoistic

concern for one’s future (part of what matters) may rationally vary in accordance with the degree

to which one will be psychologically unified with that future self. So insofar as the degree of

one’s prudential concern (partially) determines one’s time-relative interests, and insofar as the

degree of said concern may diverge widely from one’s numerical identity (which admits of no

degrees), what determines one’s moral status with respect to later-term abortions—namely, one’s

time-relative interests—does so independently of one’s numerical identity.

David DeGrazia explicitly rejects the idea that what matters—presumably, what grounds

egoistic concern—is numerical identity.14 This is because his essentialist-grounded criterion of

numerical identity is biological: the essence of individuals like you and me is our animal nature,

our biological life, such that X (a person) at one time is one and the same as any Y at another

time just in case X’s biological life is Y’s biological life.15 But one can easily see that a criterion

13 Ibid., p. 54.

7

like this will have a poor fit with our practical concerns, which more or less track psychological

relations (as he essentially admits16). For instance, special self-concern (a present-future

relation) and moral responsibility (a present-past relation) are surely grounded in psychological

relations of some sort, not biological ones, so while biological continuity is perhaps necessary to

sustain them, it isn’t the sort of thing that can make sense of them.17 As a result, DeGrazia

appeals to the notion of narrative identity to ground some bioethical matters, an account of the

different sense of “identity” I am setting aside here.

Nevertheless, he does claim to make use of the biological criterion of numerical identity

in the abortion case. On his view, unlike on McMahan’s, the early fetus is in fact an individual-

like-us, for its essence—its biological organism—is in existence and individuated roughly two

weeks after conception (once the possibility of twinning is gone). In this respect, he agrees with

one of the constituent parts of Don Marquis’ famous “future like ours” account of the wrongness

of killing, or FLOA.18 Nevertheless, DeGrazia tries to deny Marquis’ conclusion—that if a fetus

has a valuable future like ours then it has an equal interest to ours in not being deprived of it—by

adopting a version of McMahan’s time-relative interests account. He argues that what matters

for determining the moral permissibility of depriving someone of his or her future is that entity’s

time-relative interest in staying alive, itself determined by that entity’s psychological unity with

its future, beneficiary self. But “the complete lack of psychological unity between the early fetus

and later minded being requires a very heavy discounting of the value of its future in considering

14 DeGrazia, “Identity, Killing, and the Boundaries of Our Existence,” e.g., p. 425.

15 Ibid., p. 421.

16 See ibid., p. 425. For a more hedged view, see DeGrazia, Human Identity and Bioethics, pp. 60-65. See also Olson, pp. 70-72.

17 See my [SELF-IDENTIFYING REFERENCE DELETED].

8

the fetus’s stake in continuing life,”19 and so the fetus’s interest in staying alive could be

outweighed by virtually any conflicting interest of the mother (or anyone else, I suppose).

Assuming no other relevant impersonal considerations, then, early abortion looks to be justified

with ease.

The question under consideration is what identity has to do with the argument or verdict

here, and the answer is obviously none. The only real disagreement between DeGrazia and

McMahan is over whether or not the early fetus is an individual like us: DeGrazia says it is;

McMahan says it isn’t. But in neither case does this turn out to be relevant for their arguments

justifying abortion. Instead, what is relevant (and is the only relevant thing for DeGrazia) is the

relation that matters for prudential concern, namely, psychological unity, which is neither a

numerical identity relation itself nor a tracker of the numerical identity relation for either party.

Nevertheless, DeGrazia insists that “personal identity theory can illuminate the marginal

cases and the connections between them,”20 but it turns out that what he means by this is that “[a]

plausible theory of what matters in survival—a part of personal identity theory, broadly

construed—proves very important.”21 So while questions of essence or numerical identity

themselves may not turn out to be (non-derivatively) important for bioethical concerns, what

matters in identity may, and if that’s the case, then we can still say that personal identity theory

is important for bioethics.22

This is far too broad a construal of personal identity theory, though. Suppose one were to

follow Parfit (and his reasoning) in abandoning identity as what matters in survival.23 When

investigating certain questions of prudential rationality and morality, then, one might focus

18 Don Marquis, “Why Abortion is Immoral,” Journal of Philosophy 86 (1989): 183-202.

19 DeGrazia, “Identity, Killing, and the Boundaries of Our Existence,” p. 433; emphasis in the original.

9

solely on the psychological relations of connectedness and continuity that hold (or don’t)

intrapersonally as grounding the relevant practical reasons. A Parfitian might take this to be the

correct strategy, regardless of the truth of any particular theory of personal identity. Indeed,

Parfit himself is agnostic about whether or not a psychological criterion or a version of the

physical criterion of personal identity is true.24 But if the true theory of identity is just irrelevant

to our practical concerns, one might think there to be no real point to figuring out which one is

true. Yet if one takes that attitude into a study of what matters in egoistic concern with respect to

bioethical questions, say, how can we say that what one is doing has anything at all to do with

the study of personal identity theory anymore? Nevertheless, this is essentially what McMahan

and DeGrazia are doing: the relation that matters for both—psychological unity—neither is nor

tracks their favored numerical identity relations, in which case it becomes very difficult to see

how putting all the ethical weight on that (non-identity) relation actually fits into an account of

personal identity theory at all. One could easily just come to place ethical weight on the relation

of psychological unity utterly independently of any investigation at all into the nature of personal

identity, in which case one would openly be doing what McMahan and DeGrazia are more

obliquely doing, namely straightforward ethical theory.25

Turn now to the second general way in which personal identity has been thought to be

important to the abortion debate, namely, as a way to distinguish abortion from contraception.

Recall that DeGrazia denies Marquis’s anti-abortion conclusion in part by taking ethical weight

20 Ibid., p. 416.

21 Ibid.; emphasis mine.

22 Indeed, DeGrazia claims that personal identity theory is generally the investigation of three things: (a) the conditions for our persistence across time, (b) the conditions of our essence, and (c) what matters in survival. (See “Identity, Killing, and the Boundaries of Our Existence,” p. 414.) In what follows, I question the inclusion of (c) as a proper part of personal identity theory.

10

off of the biological criterion of personal identity. If we want to see whether or not a moral

conclusion about abortion can rest on a theory of personal identity, therefore, we should see what

happens if we try to put the weight back on something like a biological criterion by returning to

Marquis.

In his reply to Earl Conee’s argument that “there is no metaphysical support for a moral

conclusion about abortion,”26 Marquis tries to show precisely where metaphysics, and in

particular personal identity theory, supports his famous verdict that abortion is seriously prima

facie immoral.27 He does so by essentially reiterating his originally stated view distinguishing

contraception from abortion. In abortion, what’s deprived is the fetus’s valuable future-like-

ours, and that’s what makes it wrong. One might then worry that the valuable future of the

sperm and/or unfertilized ovum would be deprived in contraception too, making it also prima

facie wrong, a result that would, Marquis insists, constitute a reductio of his view. But he claims

instead that the two cases are quite different, insofar as what makes killing someone (an adult

human or a fetus) wrong is the loss to the victim of her future life.28 But a “necessary condition

of this being so is that the future life that is lost would have been the actual life of the same

23 Derek Parfit, Reasons and Persons (Oxford: Oxford University Press, 1984), esp. pp. 253-266.

24 Ibid., p. x (note added in 1985), p. 208, and elsewhere.

25 Another way to think about this point: what matters in survival, in identity across time, may be very different from what matters in egoistic concern. So what matters in preserving what we ordinarily think of as survival might be some biological relation, whereas what matters for purposes of anticipation and self-concern might well be some psychological relation. One might, then, easily adopt the latter view independently of any investigation whatsoever into the nature of identity, and if so it would be clear that one was engaged squarely in ethical theorizing. My point here is that this is essentially what DeGrazia and McMahan are doing.

11

individual who dies prematurely….”29 Killing the sperm or unfertilized ovum that were my

precursors, then, could have constituted a loss to them only if they would have been numerically

identical with me. But neither could have been me insofar as that would have made them (by

transitivity) numerically identical with each other, which they obviously were not. As a result,

neither could have been deprived of the valuable future that is my life had my parents engaged in

contraception at the time I was conceived.30

Here is an argument that seems on its face to get real ethical mileage out of a

metaphysical view of personal identity, specifically, any view of identity that renders early

fetuses one and the same individuals with the adult human beings into which they grow. Such a

view of identity is readily available, either by drawing from the sort of biological criterion

DeGrazia defends of course, or by drawing from a soul criterion, according to which what makes

the early fetus and later adult one and the same individual is their possession of one and the same

soul.

Nevertheless, despite appearances, Marquis’ ethical view just isn’t non-derivatively

dependent on conclusions about numerical identity. To see why, note that what makes killing

the fetus wrong is that doing so deprives it of its own valuable future. Marquis then takes a

fetus’s ownership of a valuable future to entail the numerical identity of the fetus with the

individual who would otherwise have lived through that future. But there is no such entailment

between ownership and numerical identity.31 For instance, ownership—proper attributability—26 “Metaphysics and the Morality of Abortion,” Mind 108 (1999): 619-646, p. 644.

27 Don Marquis, “Does Metaphysics Have Implications for the Morality of Abortion?” Southwest Philosophy Review 18 (2002): 73-78.

28 Ibid., pp. 77-78.

29 Ibid., p. 78; emphasis in original.

30 Ibid.

12

doesn’t necessarily obtain uniquely, as must identity. To say that some X is mine, in other

words, doesn’t mean that X is mine exclusively. Just as one may jointly own property with

another, so too one may jointly own a valuable future with another. This may be so in cases of

marriage, business partnerships, team sporting ventures, and so on, where one enters into a

relationship with other individuals, together creating and constituting a joint entity to which

various valuable things accrue, e.g., tax deductions, profits, victories, and so on.32

The reason the ownership relation may obtain independently of the numerical identity

relation is the fact that their relata are just different. What Marquis wants is an account of what

makes some valuable future mine, but that simply consists in a relation between me-now and

some set of future experiences, say, not a relation between me-now and some future experiencer.

This difference leaves room for the possibility of some valuable future being mine, where my

relation to the future experiencer is non-unique. To take a Parfitian science-fiction case, suppose

I were to be fused with you tomorrow.33 Depending on the details of the case (including the

psychological make-up of the resultant fused person), the future of the two-days-from-now

person might truly be said to be mine, or at least partially mine, pre-fusion, despite the fact that

either I am not numerically identical with the fused person or the identity of that person is

indeterminate.34

Consequently, if ownership is the relation that matters morally, and ownership doesn’t

entail numerical identity, then there’s no reason in principle why a sperm and an unfertilized 31 I’ve argued for this point in [SELF-IDENTIFYING REFERENCE DELETED].

32 See, e.g., Bennett Helm’s excellent “Plural Agents,” Nous 42 (2008): 17-49.

33 Parfit discusses various fusion cases on pp. 298-299.

34 This would be metaphysical indeterminacy, not epistemological indeterminacy: it’s not that we just wouldn’t know (or have no way of knowing) the identity of the fused person; rather, it’s that there would just be no facts of the matter regarding whether or not this person would be me, you, or someone else.

13

ovum couldn’t jointly own a valuable future, regardless of their individual lack of numerical

identity with that future experiencer, in which case the alleged disanalogy between contraception

and abortion is lost.35 Marquis’ moral conclusion directly rests, not on a view of numerical

identity, but on a theory of ownership-of-future-experiences, a theory that remains to be worked

out. And even if it turns out that ownership of this sort does (contingently) depend on numerical

identity, identity would still have only derivative importance to Marquis’ argument—carrying

weight only in virtue of its delivering the ownership relation—not the nonderivative importance

he assumes it to have.

Death

Turn now briefly to the other end of life. What might seem to be a purely conceptual

matter—determining the definition of death—is actually motivated by some major bioethical

concerns. Probably the most pressing is the question of when it is morally permissible to remove

organs from someone for transplantation. The answer often given to this question is “only when

the patient is dead.” What does it mean, though, for a patient to be dead? The fresher organs

are, the more viable they are for transplant, so we may have pragmatic reasons to understand

death as ending the life of persons, that is, of psychological creatures. But if that’s the case, then

what are we to say about the human organisms surviving the persons who had inhabited them?

Isn’t what ends their lives the true “death”? Or are there perhaps multiple concepts of “death” in

play?

I am going to focus on three different definitions of “death,” each one alleged to depend

squarely on a different theory of personal identity. As we will see, none of them clearly do.

35 My own thought is that there’s no difference between the early fetus and the sperm/ovum case with respect to ownership of a valuable future, but instead of both having such a future, neither do. This is because, for one thing, I suspect a proper account of ownership would require owners to possess some basic sort of psychological capacities, which these entities altogether lack. But I’m not prepared to defend such a view here.

14

The first is Green and Wikler’s famous ontological defense of brain death—irreversible

cessation of brain function—as constituting the proper understanding of death.36 They base this

view on a psychological criterion of the identity of persons, a criterion they think is favored by

our intuitions in body-switching thought experiments. Consequently, in order for Jones, a

patient, to be alive, then the patient must be alive and the patient must be Jones, and given that

Jones is essentially a being with psychological properties whose identity over time is preserved

by psychological continuity and connectedness, the irreversible loss of this psychological

capacity via irreversible loss of brain function signifies the cessation of that person’s existence,

which “of course” means that Jones is dead.37

The second and third accounts of death come from, respectively, DeGrazia and

McMahan. Both are alleged to be grounded in their essentialist views about identity. DeGrazia

insists that you and I are essentially living human animals, biological organisms, such that our

ceasing to exist just consists in the deaths of our organisms, and he thinks the most plausible

account of organismic death is the circulatory-respiratory standard, according to which “human

death is the permanent cessation of circulatory-respiratory function.”38

By contrast, while McMahan does agree that a human organism dies “when it irreversibly

loses the capacity for integrated functioning among its various major organs and subsystems,”39 36 Michael B. Green and Daniel Wikler, “Brain Death and Personal Identity,” Philosophy & Public

Affairs 9 (1980): 105-133.

37 Ibid. See p. 118 for the “of course” comment. Does this result mean, then, that Jones’ organs are fair game? Not necessarily, for it’s possible that Jones has “turned into” Smith (through some radical psychological discontinuity), yet it wouldn’t thus be permissible to harvest Smith’s organs (insofar as he’s still alive). But this is just to say that, while the motivation for investigation into the definition of death has been the concern over the timing of organ harvesting, it’s not at all the case that a determination of the definition of death will settle the ethical issues at stake (nor do the advocates of this methodology believe it will).

38 DeGrazia, Human Identity and Bioethics, p. 149; emphasis in original.

39 Jeff McMahan, “An Alternative to Brain Death,” Journal of Law, Medicine & Ethics (Spring 2006): 44-48, p. 47.

15

this won’t be what my death consists in, because I’m not an organism; rather, I’m essentially an

embodied mind, so I cease to exist—that is, I die40—when the capacity for consciousness is

irreversibly lost, and this happens as a result of loss of function in the higher brain, or cerebral

death.41 This leaves us with two concepts of death, one for the death of organisms, the other for

the death of persons. But given the practical concerns related to our interest in the nature of

death—regarding the morality of organ transplants, life-prolonging treatments, and so forth—the

concept that matters is cerebral death, the death of persons like you and me.42

My worry about each of these three definitions has to do with the relation each theorist

assumes holds between numerical identity, ceasing to exist, and death. The general reasoning

advanced by each view goes as follows (with each specific variation in brackets):

1. X exists only insofar as X’s numerical identity is preserved across time, i.e., X at

t1 ceases to exist at or by t2 just in case there is no Y at t2 with whom X is

numerically identical.

2. What preserves the identity of some individual across time is preservation of that

individual’s essence {psychological continuity, mind, biological organism}.

3. If X ceases to exist, X dies. ________________________________________

4. Thus, if X’s essence {psychological continuity, mind, biological organism} is not

preserved, X dies.

This is the sort of argument many have thought ensures the relevance of personal identity

to the concept of death, but this conclusion is unwarranted because premise 3 is far from

obvious. A powerful reason to doubt it comes from consideration of fission cases, of both the

40 Ibid. McMahan explicitly uses the phrases “die” and “cease to exist” interchangeably.

41 Ibid., pp. 47-48. See also McMahan, The Ethics of Killing, pp. 423-426.

42 McMahan, “An Alternative to Brain Death, p. 48.

16

non-fiction and science fiction varieties. When one amoeba splits into two, it seems the original

ceases to exist without dying.43 This is also true of the embryo that twins and the sci-fi person

who enters the fission machine. In such cases, there is no Y at the time of the split, twinning, or

fission with whom the original X is identical, precisely because uniqueness, an essential

constituent of numerical identity, has been lost. Nevertheless, it seems bizarre to say that X died

at that point, that fission killed him, given that everything else involved in ordinary survival

remains completely intact.44

This is an important point, for it makes clear that preservation of one’s essence can at

most be one necessary condition for the preservation of one’s numerical identity. The inclusion

of uniqueness as another necessary condition, however, reveals the conceptual gap between

ceasing to exist and dying: one may cease to exist where either uniqueness or one’s essence is

lost, whereas dying has nothing to do with the loss of uniqueness at all. If there’s such a gap,

then, it’s difficult to see what relevance appeals to either our essence or our numerical identity

could have in this arena.

There may be objections, though, to the use of the fission case to block premise 3. For

instance, one might insist that what’s really at stake in considerations of fission are the normative

evaluations typically attached to death, and not the conceptual issue of prizing apart “ceasing to

exist” from “death.” In other words, we might think that our concept of death has been shaped

by considering ordinary cases of death, which are typically bad for the dyer. As a result, we

have closely associated badness with death. Perhaps, though, the lesson of the fission case is that

we were wrong to do so: some deaths (e.g., deaths by fission) may not be so bad at all. Indeed,

43 Strangely, McMahan admits as much (see The Ethics of Killing, pp. 27 and 425), but he doesn’t appreciate the force of the admission against his view.

44 Obviously, this argument draws from Parfit, pp. 253-263.

17

the primary deployer of the fission case, Derek Parfit, himself calls fission “a way of dying.”45

So it might well be, then, that ceasing to exist, which is still dying, just isn’t as bad as we have

long thought, that what’s prized apart is the concept of cessation-of-existence/death from its

longstanding normative associations. Ceasing to exist/death, on this view, may be “about as

good as ordinary survival.”46

This is an interesting possibility. Our task, then, is to see whether or not there really are

such normative associations with the concept of death. True enough, we typically think the

death of a person is bad for that person. But this isn’t the case with non-persons generally,

where it seems the same concept of death is operative. Consider again the amoeba that splits.

We don’t typically think death is bad for amoebas, yet it still seems strange to say that the

amoeba dies when it fissions out of existence.47 Perhaps similarly, suppose a human being in a

permanent vegetative state (PVS) were fissioned. Once more, it seems bizarre to say that she has

died, even though she has ceased to exist. What this suggests, then, is that when we remove the

typical normative associations from death, we still get a substantive result from the fission case,

namely, there does seem to be genuine conceptual gap between ceasing to exist and dying.48

A different sort of objection would be to insist that, in the real world at least, fission (of

full-fledged human beings, anyway) just doesn’t occur, so we can safely set such considerations

45 Ibid., p. 264.

46 Ibid.

47 Thanks to an anonymous referee for raising both the “normative associations” worry as well as the amoeba reply.

48 One might think the response here, relying as it does on considerations of death to non-persons, leaves McMahan at least with a version of the “normative associations” objection intact, for he wants to ground the definition of death for persons on personal identity theory. Nevertheless, insofar as the default view of death is surely that it is a unified concept (applying equally to all entities capable of dying), and insofar as a conceptual gap between ceasing to exist and dying is revealed when considering non-persons in fission cases, the burden of proof is squarely on McMahan to show why the “normative associations” argument altogether prevents any conceptual gap in the person-based fission case as well.

18

aside and assume that cessation of existence in all ordinary cases equals death. But this reply

misses the point, for the bioethical debate about death is a conceptual debate, an exchange about

the proper definition of death, and there’s no reason at all to think the relevance or application of

our conceptual intuitions must be restricted to the everyday or the likely (“water” and “XYZ,”

anyone?).

One might, finally, still resist the conclusion by pointing to the obvious fact that, if a

living X ceases to exist, then X is clearly no longer alive. One might then think it naturally

follows that (a) if X is no longer alive, X must be dead, and (b) if X is dead, then X must have

died.49 But the first inference doesn’t necessarily follow. Suppose you magically popped out of

existence. It would no longer be true of you that you are alive, certainly enough, but it would

also not necessarily be true of you that you are dead: you would more likely be, it seems, neither.

In any event, it’s an open question whether or not you would be dead, and to admit as much

where it would also not be an open question that you had ceased to exist reveals the conceptual

gap at issue. Ceasing to exist doesn’t entail dying, and unless that’s the case it seems that what’s

relevant for the definition of death remains independent of considerations of personal identity.50

Advance Directives

The basic methodology should be clear by now, so my treatment of the final issue will be

brief. It is the case of advance directives preceding severe dementia. The relevant question is

usually thought to be, “Is the pre-demented signer of the directive numerically the same

49 An assumption David Hershenov deploys in “The Death of a Person,” Journal of Medicine and Philosophy 31 (2006): 107-120, p. 113.

50 There are other possible objections I don’t have the space to canvass here. One might be to draw from Lynn Rudder Baker’s work to insist that uniqueness and essence cannot come apart in fission, that the pre-fission person’s unique first-person perspective will survive into one of the fission products or into neither (but not into both). (See Persons and Bodies: A Constitution View (Cambridge: Cambridge University Press, 2000). There is much theoretical machinery involved in this view that would take a great deal of time to discuss and dissect, however, so I will set it aside here.

19

individual as the later demented patient (someone who is by definition a non-person, let’s say)?”

This is particularly a problem in the case in which the younger signer (YS) directs treatment (or

non-treatment) that the contented demented patient (DP) claims not to want. Our strong intuition

is that YS’s directive is authoritative over the wishes of DP. Call this the Intuition. Identity-

based arguments on the topic typically go as follows:

1. YS’s preferences are authoritative over DP’s if and only if YS is numerically

identical to DP.

2. YS [is/is not] numerically identical to DP. _________________________

3. Thus, YS’s preferences [are/are not] authoritative over DP’s.

Start with the negative version of premise 2, something a person-essentialist or

psychological continuity theorist would likely maintain.51 This view yields what DeGrazia calls

the “someone else problem.”52 Here, notice that one can still deny the conclusion that YS’s

preferences aren’t authoritative, and thus rescue the Intuition, by denying premise 1’s assertion

that numerical identity is necessary for authority, and indeed this is what many have done. One

way to do this is by appealing to “surviving interests,” interests people have regarding certain

states of affairs whose (dis)satisfaction depends on what happens after they cease to exist.53

Another way is by appeal to “substituted judgment,” which depends on close family members or

loved ones to determine what treatment the formerly competent patient would have wanted

51 See, e.g., Rebecca Dresser, “Life, Death, and Incompetent Patients: Conceptual Infirmities and Hidden Values in the Law,” Arizona Law Review 28 (1986): 379-381; and “Advance Directives, Self-Determination, and Personal Identity,” in Chris Hacker, Ray Moseley, and Dorothy Vawter, eds., Advance Directives in Medicine (New York: Praeger, 1989).

52 See David DeGrazia, “Advance Directives, Dementia, and ‘the Someone Else Problem’,” Bioethics 13 (1999): 373-91; “Identity, Killing, and the Boundaries of Our Existence,” pp. 440-441; and Human Identity and Bioethics, pp. 164-167.

53 See, e.g., Allen Buchanan, “Advance Directives and the Personal Identity Problem,” Philosophy & Public Affairs 17 (1988): 277-302, p. 287.

20

(where the advance directive counts as authoritative evidence for that). But in either case, the

loss of identity is irrelevant to the preservation of the Intuition.

On the other hand, one might claim to preserve the Intuition by embracing the positive

version of premise 2, something a mind-essentialist or biological continuity theorist would likely

do. This would allegedly render YS’s preferences authoritative over DP’s. But for those who

adopt this option (e.g., McMahan and DeGrazia), the problem now is to figure out why this one

and the same individual’s earlier preferences are to be respected over her current preferences,

when this is the opposite of ordinary practice in other arenas. There are various replies here,

having to do with how to place precedent autonomy into the hands of YS. McMahan appeals to

the time-relative interest account again.54 DeGrazia appeals to considerations of narrative

identity as a way to show how DP’s experiences may or may not be unified into YS’s life in

terms of what matters.55 But neither view depends on numerical identity to preserve the

Intuition, precisely because doing so depends on a particular account of what makes certain

preferences authoritative over others, and identity seems to be neither here nor there with respect

to that.56

This point is made quite clear once we realize that some authors actually counsel

abandoning the Intuition itself, arguing instead that YS’s preferences aren’t authoritative over

DP’s, given that DP may still have an important sort of autonomy.57 On this view, identity is

explicitly irrelevant to a determination of the bindingness of the advance directive.

The debate isn’t over identity, then, but rather over the nature of preferential authority

directly, over what it is that renders some preferences authoritative when in tension with others.

Depending on what that consists in, then, YS’s directive may be authoritative for DP regardless

of their (non-)identity. And alternatively, her directive may not be authoritative for DP,

21

regardless of their (non-)identity. But in either case, numerical identity simply isn’t doing the

relevant work. Whether numerical identity is in fact relevant for some other issues in bioethics,

or whether there’s some other sense of “identity” that is relevant to these issues instead, that will

have to be a discussion for another day.58

54 McMahan, The Ethics of Killing, pp. 496-503.

55 DeGrazia, “Identity, Killing, and the Boundaries of Our Existence,” pp. 441-442; Human Identity and Bioethics, pp. 173-186.

56 Two points are relevant here. First, Steven Wall has suggested a possible account of advance directives in which YS binds herself via the directive, such that no matter what she might think or feel when demented, her earlier wishes are to be authoritative. On such an account, it seems as if numerical identity might do some real work: what renders the later self subject to the directive is precisely her numerical identity with the earlier, binding self. I don’t want to rule out the possibility of such an account, but I haven’t seen one like it developed before, so I’ll remain agnostic until I can assess the details. (One initial worry is that we let ourselves off the hook sometimes in such self-binding arrangements, so it would be unclear whether or not (a) there would be legitimate instances of “letting off the hook” in cases of advanced directives, or (b) a hypothetical later competent self might have done so where the actual later self was pleasantly demented.)

Second, Marya Schechtman has pointed out to me that it looks as if considerations of numerical identity actually do at least some minimal work for DeGrazia and McMahan (and perhaps others). After all, they’re left with the problem of saying why YS’s preferences should be given precedence over DP’s only because they are the same person. So appeals to identity constitute at least some part of the argument. Fair enough. I don’t necessarily want to insist here that identity plays no role whatsoever in these bioethical arguments; rather, I merely want to show that appeals to identity are far less significant to bioethics than have often been thought, and in particular they do little to none of the heavy lifting in these arguments, which is actually done by other relations. In the case of advance directives, what’s doing the heavy lifting is the establishment of authoritativeness in preferences. Given that if YS and DP aren’t numerically identical, YS’s preferences for DP may still be authoritative, and given that if YS and DP are identical, YS’s preferences may not be authoritative over DP’s (as I point out in the next paragraph above), identity considerations play very little role in addressing the main issues.

57 See, e.g., Agnieszka Jaworska, “Respecting the Margins of Agency: Alzheimer’s Patients and the Capacity to Value,” Philosophy & Public Affairs 28 (1999): 105-138.

58 ACKNOWLEDGMENTS.

22