ArticlePDF Available


The ethics of treating future persons.
Persons in ProspectJ. DAVID VELLEMAN
i. the identity problem
Derek Parfit calls it the non-identity problem.
It’s the problem of how to
treat future persons given that any attempt to treat them better may
result instead in their never being born. For example, the people who will
have inadequate resources in the twenty-second century because of our
wastefulness today will owe their existence to human couplings that
never would have occurred if we had lowered our thermostats and show-
ered less often. As those future people commute by bicycle or read by
candlelight, they will have to acknowledge that we couldn’t have con-
served resources for them, since our conserving would have prevented
them from existing. Because the people affected by our wastefulness will
not be identical to those who would have been affected by our conser-
vation, there appear to be no future individuals for us to harm or benefit,
whatever we do.
This description of the problem depends on an empirical assumption
about the effects of our environmental policies on the makeup of the
These three essays were written in conjunction with an undergraduate course on the
topic “Future Persons,” which I taught at NYU in the fall of . My thanks to the students
enrolled in the course for helping me to think through the issues. The entire series has been
much improved by close reading and detailed commentary from Imogen Dickie, Jeff Sebo,
and the Editors of Philosophy & Public Affairs.
. Parfit first discussed the problem in “On Doing the Best for Our Children,” in Ethics
and Population, ed. Michael Bayles (Cambridge, Mass.: Schenkman, ), pp. . See
also Robert Merrihew Adams, “Existence, Self-Interest, and the Problem of Evil,” Nous 
(): ; Gregory Kavka, “The Paradox of Future Individuals,” Philosophy & Public
Affairs  (): ; Thomas Schwartz, “Obligations to Posterity,” in Obligations to
Future Generations, ed. Richard Sikora and Brain Barry (Philadelphia: Temple University
Press, ), pp. . I will deal primarily with Parfit’s discussion of the problem in Part IV
of Reasons and Persons (Oxford: Oxford University Press, ).
©  Wiley Periodicals, Inc. Philosophy & Public Affairs , no.
population. In this part I will argue that even if this empirical assumption
were false, the problem would remain. Even if we could ensure that
the people affected by our conservation were identical to the people
affected by our wastefulness, neither group could be harmed or ben-
efited by what we do. I call it the identity problem, to indicate that it is a
variant of Parfit’s.
In subsequent parts I will develop a solution to the problem in both of
its forms. In Part II (“The Gift of Life”), I will articulate a conception of
what we do to someone by bringing him into existence, and how we
thereby incur personal obligations to him.
The resources developed in
Part II will enable me to address the (non-)identity problem directly in
Part III (“Love and Nonexistence”).
A fault-line runs between Parfit’s discussion of personal identity in
Part III of Reasons and Persons and his subsequent discussion of the
non-identity problem in Part IV. The discussion in Part III is largely
devoted to the question what matters in survival: what it is about our
relation to past and future selves that grounds our self-interested
concern for them, including our concern for the latter’s very existence.
The discussion in Part IV is devoted to the obligations that every gen-
eration owes to its successors, the obligations that appear to be under-
mined by the non-identity problem. As Parfit points out, this problem
does not involve personal identity over time; it involves “personal
identity in different possible histories of the world.”
The discussions
in Parts III and IV therefore end up relying on different conceptions
of personal identity.
This discontinuity creates a puzzle, because transworld personal
identity bears on obligations between generations by way of the same
self-interested concern that was at issue in Parfit’s discussion of iden-
tity through time. The reason why future generations will be in no posi-
tion to blame us for consuming the earth’s resources, according to
Parfit, is that they will not be the same as the people who would have
benefited from those resources, had we conserved. The puzzle is
whether a lack of identity with those possible beneficiaries will restrict
future people from having any reactions to which they would have
. I use ‘he’ as the unmarked pronoun, which is semantically gender-neutral.
. Reasons and Persons, p. .
 Philosophy & Public Affairs
been entitled otherwise. Even if they could have received the benefits
of our conservation themselves, their complaint for not actually receiv-
ing them will have to rest on damage to their interests, and will thus
presuppose an identification of interests between their actual selves
and the merely possible counterparts who would have benefited. If
their relation to those more fortunate counterparts doesn’t include the
basis for self-concern, they will still have no grounds for feeling that
their own interests have suffered. And then it will turn out to make no
difference whether they could have received the benefits, given that
they couldn’t have identified with the interests of the potential benefi-
ciaries in any case.
(This puzzle is initially difficult to grasp, because it involves a double
counterfactual about what future generations could have felt if they
could have benefited, if we had conserved resources. I hope that the
puzzle will become clearer as I proceed.)
Parfit never questions the grounds of first-personal concern for other
possible selves. As it turns out, his account of such grounds in the case of
past and future selves cannot be extended to cover the counterfactual
case. And without knowing how self-interest might extend to other
possible selves, we cannot understand why it should matter to future
persons whether they could have been the beneficiaries of more respon-
sible behavior on our part.
I am going to analyze Parfit’s view of what matters in survival and
then propose a crucial revision of the view. After defending the revised
view, I will point out that the relation mediating our first-personal
concern for past and future selves is absent in the case of other
possible so-called selves—not just absent, in fact, but metaphysically
impossible. I will therefore argue that none of the people who would
have existed, if the past history of the world had been different,
merit first-personal concern from us. This result has various implica-
tions for the rationality of regret as well as for the problem of
obligations between generations.
According to Parfit, what matters in survival is a relation of
psychological connectedness and continuity between present
and future selves. Parfit’s definition of psychological connectedness
begins with Locke’s memory-theory of personal identity: “Let us
say that, between X today and Y twenty years ago, there are direct
 Persons in Prospect
memory connections if X can now remember having some of the expe-
riences that Y had twenty years ago.”
Parfit then expands on Locke’s
theory like this:
We should . . . revise the view so that it appeals to other facts. Besides
direct memories, there are several other kinds of direct psychological
connection. One such connection is that which holds between an
intention and the later act in which this intention is carried out. Other
such direct connections are those which hold when a belief or a
desire, or any other psychological feature, continues to be had.
Parfit defines psychological continuity as the ancestral of connected-
ness. That is, X’s being psychologically continuous with Y consists
in there being some (possibly empty) series of subjects S,S,...such
that X is directly connected to S, who is directly connected to
S, . . . who is directly connected to Y. Parfit describes this relation
between X and Y as consisting in “chains of psychological connected-
ness,” which may overlap.
Initially, Parfit says that what matters in survival is a relation labeled
R, which is a combination of psychological connectedness and continu-
ity. Parfit subsequently qualifies his view, by claiming that some psycho-
logical connections are more important than others. The more
important connections, he claims, are the ones that involve features
that are distinctive of the individual, or features that the individual
values in himself.
I suspect that Parfit introduces these qualifications partly because he
equivocates on the phrase “what matters in survival.”
Sometimes Parfit
. Reasons and Persons, p. . Parfit modifies this definition by adopting Shoemaker’s
concept of “Q memory” to cancel the possible implication that X’s remembering Y’s expe-
riences entails X’s being the same person as Y (pp. ). I will assume that “memory”
means “Q memory.”
. For the importance of a feature’s distinctiveness, see pp.  and n. on p. .
For the importance of a feature’s value to the subject, see p. . See also the discussion of
“The Nineteenth Century Russian” on pp. .
. This point was made by Paul Torek in an unpublished paper and in his Ph.D. dis-
sertation, Something to Look Forward To: Personal Identity, Prudence, and Ethics, Univer-
sity of Michigan, . For the idea that “what matters in survival” is ambiguous in Parfit’s
usage, see also Tamar Szabó Gendler, “Personal Identity and Thought Experiments,” Philo-
sophical Quarterly  (): .
I also suspect that Parfit equivocates on the term ‘continuity’. In some contexts, he
uses ‘continuity’ for the ancestral of connectedness. But because he emphasizes the
 Philosophy & Public Affairs
interprets the question “What matters in survival?” to mean “Why should
one have a first-personal interest in surviving?”
Sometimes he takes the
question to mean “Why should one have any first-personal concern for
the self who will survive?”
These two questions exhaust Parfit’s osten-
sible topic, but he obscures this topic with other readings of the question
“What matters in survival?” Sometimes he takes the question to mean
“What is it about one’s present self whose survival in future selves is
worth wanting?”
Sometimes he even takes it to mean “What kind of
survival is worth wanting?”
The latter readings of the question are not equivalent to the former.
One’s grounds for taking a self-regarding interest in future persons may
not depend on their having features of oneself that one has an interest in
preserving, or their living lives that one has an interest in living. Confla-
tion of these issues crucially affects Parfit’s discussion of problem cases,
in particular, the one that he calls the “Branch Line case.”
In the Branch Line case, Parfit imagines a “scanner” that, at the press
of a green button, destroys and analyzes his entire body, including his
brain. The scanner is linked to a “replicator” that assembles a molecule-
by-molecule copy of him on Mars. He then imagines that the scanner is
upgraded to a model that leaves his original body intact, so that there are
duplicate versions of him, one on each planet. Finally, he imagines that
the upgraded scanner has damaged his heart and that he will conse-
quently die within a few days. Having received this dire prognosis, he
speaks with his replica on Mars by interplanetary videophone:
. . . Since my Replica knows that I am about to die, he tries to console
me with the same thoughts with which I tried to console a dying
friend. It is sad to learn, on the receiving end, how unconsoling these
thoughts are. My Replica then assures me that he will take up my life
where I leave off. He loves my wife, and together they will care for my
connections that consist in the mere persistence of a trait or an attitude, he sometimes
understands ‘continuity’ to mean “qualitative continuity,” in the sense that denotes the
absence of abrupt qualitative changes. See, e.g., p.  of Reasons and Persons.
. See, e.g., p. .
. See, e.g., pp. .
. See, e.g., pp. , , .
. See p. .
. This case is first introduced on pp. . It is discussed again on pp. .
. P. .
 Persons in Prospect
children. And he will finish the book that I am writing. Besides having
all of my drafts, he has all of my intentions. I must admit that he can
finish my book as well as I could....
If we believe that my Replica is not me, it is natural to assume that
my prospect, on the Branch Line, is almost as bad as ordinary death.
I shall deny this assumption. As I shall argue later, being destroyed
and Replicated is about as good as ordinary survival.
Parfit later explains his view of the case as follows:
It may be slightly inconvenient that my Replica will be psychologically
continuous, not with me as I am now, but with me as I was this
morning when I pressed the green button. But these relations are
substantially the same. It makes little difference that my life briefly
overlaps with that of my Replica.
If the overlap was large, this would make a difference. Suppose that
I am an old man, who is about to die. I shall be outlived by someone
who was once a Replica of me. When this person started to exist forty
years ago, he was psychologically continuous with me as I was then.
He has since lived his own life for forty years. I agree that my relation
to this Replica, though better than ordinary death, is not nearly as
good as ordinary survival. But this relation would be about as good if
my Replica would be psychologically continuous with me as I was ten
days or ten minutes ago.
Parfit does not explain why the survival of a forty-year-old replica would
be less desirable than that of a replica produced within the past ten
minutes. He seems to imply that the survival of the forty-year-old replica
would be less desirable because he has “lived his own life for forty years”
and would be less likely to carry on the life that will be cut short at Parfit’s
death. At the replica’s creation forty years ago, he might have finished
the book that Parfit was writing then, but he now lacks the beliefs,
desires, and intentions that would enable him to finish the book that
Parfit is writing now and will not survive to finish. Parfit’s judgment in
this case thus illustrates his view that what matters in survival is the
continuation of that in oneself or one’s life which one finds important.
. P. .
 Philosophy & Public Affairs
Parfit concludes his discussion of the Branch Line case with the
admission that his judgment is counterintuitive:
...Iadmit that this is one of the cases where my view is hardest to
believe. Before I press the green button, I can more easily believe that
my relation to my Replica contains what fundamentally matters in
ordinary survival. I can look forward down the Main Line where there
are forty years of life ahead. After I have pressed the green button, and
have talked to my Replica, I cannot in the same way look forward
down the Main Line. My concern for the future needs to be redirected.
I must try to direct this concern backwards up the Branch Line beyond
the point of division, and then forward down the Main Line. This
psychological manoeuvre would be difficult. But this is not surprising.
And, since it is not surprising, this difficulty does not provide a suffi-
cient argument against what I have claimed about this case.
Reading this passage, one wonders why a difficulty should have to be
surprising in order to be philosophically significant. One rather suspects
that Parfit’s talk of “redirecting” his concern up one “line” and down the
other—talk of a kind that appears nowhere else in Parfit’s discussion—
reveals an important feature of self-concern, a feature that will help to
explain the rationality both of wanting to have future selves and of caring
about their fate.
In what sense does Parfit find himself directing his concern “up” one line
and “down” the other in the Branch Line case? The answer is that “up”
and “down” in this case represent the direction of time. Parfit later says
that psychological continuity is a transitive relation in either temporal
direction but not “if we allow it to take both directions in a single
That is, the reason why Parfit is not psychologically
continuous with any of his replicas is that the psychological connections
between them run first backwards in time, up to the point of division,
and then forwards, down the “main line.” But why should this change of
temporal direction make any difference? Parfit doesn’t say. He simply
admits that when directing his self-concern through time, he has
difficulty switching directions.
. Ibid.
. P. .
 Persons in Prospect
I suggest that concern for his “main line” replica is difficult for Parfit
because the direction of time is also the direction of causation, and the
change of direction severs internal communication between Parfit and
his replica, in the sense that their psychological connections cannot
carry information between them. Parfit’s conception of psychological
connections has all along implied that they are channels of information,
but he has chosen instead to emphasize the relations of resemblance
between their input and output—between experiences and the
corresponding memories, intentions and the corresponding actions,
psychological features and their subsequent instantiation—rather than
the fact that these inputs and outputs are connected in ways that convey
Yet Parfit’s difficulty in feeling concern for his replica
seems to indicate that internal communication with earlier or later
selves is significant.
I now turn to an explanation of why such communication is
significant. After that, I will return to the puzzle of obligations
between generations.
Last night I dreamed that I was Wittgenstein brandishing a poker at Karl
Popper. (I am prone to nightmares.)
. See pp. , where Parfit discusses the case in which psychological continuity
and connectedness have a cause that isn’t reliable. Parfit says that a replica to whom one is
unreliably connected is just as good as one to whom one’s connection is reliable. He
compares this case to that of a medication that effects a cure sometimes but not reliably:
“This effect is just as good, even though its cause was unreliable.” This analogy suggests
that what matters in survival are the effects of one’s causal connections to future selves, not
the connections themselves.
. This and the next several sections draw on material from my “Self to Self,” Philo-
sophical Review  (): ; reprinted in Self to Self : Selected Essays (Cambridge:
Cambridge University Press, ), pp. . That paper draws in turn on Bernard
Williams’s “The Imagination and the Self,” in Problems of the Self (Cambridge: Cambridge
University Press, ), pp. . Parfit makes a similar point on p. :
Since Jane seems to remember seeing the lightning, she seems to remember herself
seeing the lightning. Her apparent memory may tell her accurately what Paul’s experi-
ence was like, but it tells her, falsely, that it was she who had this experience.
There may be a sense in which this claim is true. Jane’s apparent memories may come
to her in what Peacocke calls the first-person mode of presentation. Thus, when she
seems to remember walking across the Piazza, she might seem to remember seeing a
child running towards her. If this is what she seems to remember, she must be seeming
to remember herself seeing this child running towards her.
We might deny these claims. In a dream, I can seem to see myself from a point of view
outside my own body. I might seem to see myself running towards that point of view.
 Philosophy & Public Affairs
Now, when I say, “I dreamed that I was Wittgenstein,” my first use of
‘I’, in “I dreamed,” refers to me, David Velleman, who groaned in his
sleep and then woke up remembering a nightmare. What about my
second use of ‘I’, the one in “I was Wittgenstein”? That use of ‘I’ occurs
within a that-clause reporting the content of my dream, and so it refers,
in the first instance, to a component of the content being reported—
specifically, the first-personal component. The dream is thus reported to
have represented someone first-personally. But who?
I did not dream that David Velleman was Wittgenstein. While dream-
ing, I was temporarily oblivious to the existence of David Velleman—
oblivious, in fact, to my own actual existence under any name or
description whatsoever. But my dream contained a first-personal mode
of presentation, since it represented an experience from the perspective
of its subject—an experience of brandishing a poker, from the perspec-
tive of the poker-brandishing experient. Who was he?
The dreamed-of experience, though similar to one that Wittgenstein
might have had, was merely dreamed of, not real. There was no poker, no
Popper, and no experience of brandishing the one at the other—only a
dream of such an experience. Hence there was no subject of the
dreamed-of experience, either: there was only a dreamed-of subject. In
my dream-report, then, the second ‘I’ refers to a first-personal mode of
presentation that, in its original occurrence, failed to pick out a referent
at all. My dream had the content “I am Wittgenstein . . . ,” but there was
no one of whom I was dreaming that he was Wittgenstein.
Here is another nightmarish scene. I am six years old and my grandfather
is holding me up to look over the parapet of the observation deck on the
Empire State Building. I am terrified of falling. (I wet my pants.)
Since it is myself that I seem to see running in this direction, this direction cannot be
towards myself. I might say that I seem to see myself running towards the seer’s point of
view. And this could be said to be the direction in which Jane seems to remember seeing
this child run. So described, Jane’s apparent memory would include no references to
Though we could deny that Jane’s apparent memories must seem, in part, to be about
herself, there is no need to do so. Even if her apparent memories are presented in the
first-person mode, Jane need not assume that, if they are not delusions, they must be
memories of her own experiences....
. For this analysis of first-person reference in dreams, I am indebted to Imogen
 Persons in Prospect
This report and the last are similar in giving the content of a second-
order experience: a dream of an experience in the first report, a
memory of an experience in the second. I can bring out this similarity
by saying, “I had a dream about an experience in which I was Wittgen-
stein brandishing a poker,” and “I have the memory of an experience in
which I am a six-year-old visiting the Empire State Building.” And now
we can ask about my second report the same question that we asked
about the first. I have the memory of an experience in which someone
is a six-year-old visiting the Empire State Building. But who? Who
is that six-year-old, according to the content of the remembered
Conveniently for present purposes, that six-year-old was called Jamie,
and so I can distinguish him from the fifty-five-year-old who remembers
his visit to the Empire State Building. I can say, “I, David, have the
memory of an experience in which I am Jamie visiting the Empire State
Building.” Obviously, I cannot be reporting the memory of an experience
in which it is I, David, who am Jamie. The experience occurred in ,
and in the content of the experience, its subject was the six-year-
old Jamie. Fifty-five-year-old David was still forty-nine years in the
future—an impossibly old man who Jamie could not imagine becoming.
And the experience that occurred in  at the top of the Empire State
Building was not an experience as of a fifty-five-year-old stranger
inhabiting a six-year-old’s body. (I would have done something even
worse in my pants.)
Although David wasn’t represented as being six-year-old Jamie in the
remembered experience, perhaps he is so represented in my memory of
the experience. Maybe the memory has a memorial mode of presenta-
tion, by which it presents a visit to the Empire State Building as formerly
experienced by me.
I think that there is a distinctively memorial mode of presentation, but
I do not think that it is first-personal with respect to the remembering
subject, in addition to the subject of the remembered experience. Rather,
a memory presents earlier events as experienced and recorded herein.
When I remember visiting the Empire State Building, I have an image of
that visit as it appeared at the initial receipt of this very image.
Questions about the memorial mode of presentation must be distin-
guished from questions about the concept of memory. Some philoso-
phers think that classifying a mental image as a memory implies not only
 Philosophy & Public Affairs
that it represents an experience that actually occurred, but also that it
represents an experience that occurred to the present subject of the
image. If these conditions are not met, according to these philosophers,
then the image must be reclassified as a merely apparent memory. But
what appears in a memory, real or apparent, is that the experience
occurred and was herein recorded, not that it occurred to the remem-
bering subject. A memory appears to be a record of the experience; it
does not necessarily appear to be a memory—not, at least, if the concept
of memory includes an identity between the remembering and remem-
bered subjects. If that identity is thought of at all, the thought of it is
a further inference.
I can express this inference by saying, “I must have visited the Empire
State Building as a child, because I remember visiting it.” The second
clause of my statement can be expanded to ...Iremember that I visited
it,” but the second ‘I’ in the clause still refers to the first-personal mode
of presentation by which the memory-image presents the experience as
recorded from the perspective of the experient. In other words,
“I remember that I visited it” says no more than “I remember visiting it”:
it says that I have an experiential record of a visit from the visitor’s point
of view.
In this respect, my memory report resembles my dream report.
crucial difference is that, whereas the first-personal mode of presenta-
tion in my dream failed to pick out a subject, the corresponding mode of
presentation in my memory succeeds in picking out the subject of the
remembered experience, the six-year-old Jamie. The memory-image
inherits the references of that experience, thus providing me with a
context in which I can, by directing my attention, make particular
objects perceptually salient, so that I can refer in thought to “this [pic-
tured] building,” “that [pictured] parapet,” and so on. It also inherits the
reference of the first person, providing a context in which I can think of
Jamie as “this person [not pictured],” or “me.”
When I say that the memory-image provides a context for referring to
Jamie as “me,” I do not mean that it establishes my numerical identity
with him. My ability to pick him out with the first-person pronoun has
. I am overlooking the fact that the dream report is mediated by a memory of the
dream, whereas the memory report is not mediated by a memory of the memory. That
difference is irrelevant for present purposes.
 Persons in Prospect
nothing to do with whether he was the same person as my present,
remembering self. I can pick him out in the first person because he was
the subject of the experience from which my memory-image inherits its
references. The image has an egocentric point of view, which was the
point of view of the original experience, in which Jamie was the ego at its
center. In entertaining the image, then, I am thinking of Jamie as “me.”
When I infer from the memory that I must have visited the Empire State
Building as a child, I take a step beyond the content of a first-personal
record of the experience. What do I infer?
In moments of philosophical alertness, I infer that I, David, am a
segment of a temporally extended entity of which Jamie was a much
earlier segment—or something of that sort. For when David and Jamie
are considered as a fifty-five-year-old and a six-year-old, respectively, as
I have been considering them, they are by no means identical; they are
merely unified across time, by intervening person-stages (one of whom
changed his name).
Ordinarily, however, no thought of their being unified in this manner
ever crosses my mind. What does cross my mind, I think, is that I am
literally identical to Jamie, because both of us are “me”—not just
dreamed or imagined to be “me” but accurately experienced (David) and
remembered (Jamie) as “me,” and consequently one and the same.
Before I consider any sameness of person between Jamie and me, I find
myself able to pick him out in memory as “me,” and for that very reason
I infer us to be the same person, and not just different stages of the same
person but numerically identical.
This inference involves an error that is easy to explain.
I have occur-
rent experiences in which there is a first-personal mode of presentation
that picks out the subject, who is David; I also have a memory-image in
which the first-personal mode of presentation picks out the subject of
the experience from which it was recorded, who was Jamie. When I have
. Richard Wollheim would call Jamie the “internal subject” of the image. See Lecture
III of Painting as an Art (Princeton, N.J.: Princeton University Press, ); see also Woll-
heim’s “Imagination and Identification,” in On Art and the Mind (Cambridge, Mass.:
Harvard University Press, ), pp. .
. I discuss this error in “So It Goes,” The Amherst Lecture in Philosophy (): , It is also discussed in Thomas Hofweber and J. David
Velleman, “How to Endure” (unpublished manuscript).
 Philosophy & Public Affairs
multiple images such as these, I have to correlate their frames of refer-
ence so as to figure out what (in one image) is what (in the other), and
who is who. That’s what I do when turning a map around so as to line it
up with the coordinates of my perceived environment. “That’s Fifth
Avenue,” I say (pointing to a line on the map), “and we’re at Washington
Square Arch” (pointing above my head), “so Broadway must be this way”
(pointing to my right). Similarly, I think, That’s me (Jamie experiencing
the remembered scene) and this is me (David remembering it), so they
must be one and the same.
The relata of this identity relation are momentary subjects, each exist-
ing only for the duration of an experience. In thinking of them as strictly
identical, rather than as distinct segments of a temporally extended
entity, I think of myself as persisting in the manner that metaphysicians
call endurance. (The segment-wise manner of persisting is called perdu-
rance.) That is, I think of myself as an entity of a kind that is wholly present
in each moment of its existence rather than spread out over time in the
form of distinct stages or temporal segments. I think of myself as here in
my entirety now, here all over again now, and so on, a successively
re-existent whole rather than interconnected parts. Thinking of myself as
enduring in this sense is a mistake, but the mistake is understandable.
Consider now experiential anticipation, which is the mirror image of
experiential memory. Suppose that I plan to revisit the Empire State
Building next weekend, and I picture how the city will look from the
observation deck. I compose a mental image of a future experience in
which I am looking out over the parapet.
This kind of experiential anticipation has a complex structure. Unlike
the past self who couldn’t imagine becoming a fifty-five-year-old
remembering his experience, I know that the subject of next weekend’s
experience will remember having imagined it here and now. In fact, I
now imagine him not only viewing the cityscape but doing so in the
awareness of its being the view that he remembers having hereby imag-
ined. I compose a prospective image of an experience within which this
very image will be accessible in retrospect. It’s as if I were composing a
picture with the intention of climbing through it into the depicted scene
and walking off with an awareness of the picture at my back.
This structure is essential to the rhythm of our mental lives. The air of
expectancy in experiential anticipation comes from its representing a
 Persons in Prospect
future experience as its felt completion—as destined to provide a sense
of closure, that is, to this very expectation, like the notes that are
expected to be heard as the cadence of a currently developing musical
phrase, the exhalation that is expected to resolve an expectant holding of
the breath. The future experience will provide this sense of closure, of
course, only if it includes a retrospective awareness of the current phrase
that it completes, and indeed of that phrase as having been hereby
experienced expectantly, in anticipation of that very closure. So in order
to represent the future experience as its felt completion, the expectation
must represent that experience as including a memory of having
been hereby expected.
Does my expectation represent the future subject of the expected
experience as identical to its own, present subject? Not explicitly. But the
two subjects naturally become conflated, as they did in experiential
memory. As in that case, the ‘I’ of the image in which I am viewing the
city next weekend and the ‘I’ of my present context seem to be strictly
identical—a single subject expecting to climb through his own mental
image into the future—rather than distinct segments along the temporal
extent of a perduring entity.
To whom does ‘I’ refer in the context of this image? Is this image like
a memory, in which the first-personal mode of presentation picks out
the remote subject of the originating experience? Or is it like a dream,
in which there is a first-personal mode of presentation that fails to
determine a reference?
Surely, the subject of this image is my future self visiting the Empire
State Building. But the anticipatory image cannot refer to him in the
same way as the memory-image refers to my past self at the same loca-
tion forty-nine years ago. The memory-image inherits its reference from
the remembered experience, via a channel of information connecting it
to that experience and, through it, to the referent. The anticipatory
image cannot inherit its reference from the future.
But as we have seen, the latter image has an anticipatory mode of
presentation that implies a causal connection to the experience antici-
pated, just as the former has a memorial mode of presentation implying
. The phenomena described here are the ones to which Husserl applied the terms
‘retention’ and ‘protention’. Of course, these prospective and retrospective awarenesses
need not be available to articulate thought.
 Philosophy & Public Affairs
a connection to the experience remembered. Whereas the memory-
image represents an event as experienced and recorded in this image, the
anticipatory image represents an event as to be experienced in the light of
this image. Both modes of presentation are implicitly token-reflexive, in
that they represent an event in its relation to the images themselves. In
that sense, both images present themselves in their relation to experi-
ences of the events represented: the one presents itself as conveying an
impression from a corresponding experience; the other presents itself as
due to make an impression on a corresponding experience.
Although the latter image cannot inherit its references from the expe-
rience on which it promises to make an impression, it can rely on that
experience to determine its references. The subject of that experience is
expected to remember having hereby anticipated it and to draw the
connections between things represented in the remembered anticipa-
tion and things that he experiences. He will associate the view that he
remembers envisioning with the view spread before him, and his doing
so is already part of what is envisioned. The image therefore represents
things as they will be associated with this very representation by a future
subject who will thereby secure its references to those things.
In some cases, my anticipatory image needn’t rely on a future subject
to determine its references. It already represents the Empire State Build-
ing, whether or not anyone who remembers the image ever draws a
connection between its representation of a building and that building in
particular. But the image also represents an event, for example, and the
question is initially open whether there is a particular, concrete event
that I am imagining; I may simply be imagining that there will be some
event or other of the imagined kind. When my future self connects this
remembered anticipation with a particular event, however, he can say
that I anticipated, not just some such event, but that very one. If I expect
to run across you at the Empire State Building and I do run across you,
then I can say, “This is just as I envisioned it” or “This is not as I envi-
sioned it,” “it” being the particular meeting with which I compared my
vision, as I had already envisioned doing. If I don’t run across you, then
there will have been no particular meeting that I envisioned.
What if I am going to run across you but also forget having envi-
sioned doing so? Well, I won’t then wonder whether our meeting is the
. Here again I am indebted to Imogen Dickie.
 Persons in Prospect
one that I previously envisioned, since I will have forgotten having envi-
sioned it. But from my present perspective of envisioning our meeting,
will a meeting not then associated with its having been hereby envi-
sioned count as the meeting that I now have in mind? From my present
perspective, there seems to be no fact of the matter. If I am imagining an
encounter on the observation deck at noon but I end up running across
you earlier in the subway, will that be the meeting that I am imagining?
The answer would be yes if I am going to remember this imagining and
think, This is not as I imagined it, thereby completing the referential
chain. The answer could be no if I was going to remember the imagining
and think, This isn’t even the meeting that I had in mind. But if I won’t be
in a position either to complete or to break the referential chain, I would
seem to be imagining a meeting but no meeting in particular.
Thus, anticipating experiences that include the memory of having
been anticipated enables me to secure singular reference to future epi-
sodes in my life. I can also attempt singular reference to future epi-
sodes by means of definite descriptions, for example, by framing an
image of an encounter under the description ‘the next time I run across
you’. But then I may run across you prematurely on Thursday, where-
upon my definite description will have sent my reference astray. I could
jury-rig my definite description to insure against such mishaps, but
that is more trouble than I need to take in thinking about my future.
For I can simply anticipate future episodes and send those anticipa-
tions forward to be assigned future referents by my future selves.
Unlike definite descriptions, these references aren’t hostage to future
vicissitudes, because they are entrusted to a future subject who will
manage them on my behalf.
I can do the same with beliefs that do not involve mental images. I can
open a mental file and fill it with anticipatory beliefs about a future
episode, without including any images of what the episode will be like.
Provided that I can rely on a future subject to retrieve the file and asso-
ciate it with a particular episode in the future, I need not frame a definite
description sufficient to pick out which episode I have in mind. The file
will include the belief or intention that its references will be determined
by a future retriever of the file, but it need not include an image of what
it will be like for that subject to remember what it was like to open the
file. My expectations about the episode may therefore lack the phenom-
enology of climbing through an image into the future, but they will still
 Philosophy & Public Affairs
include the expectation that the episode will be experienced as having
been hereby expected.
This telegraphic form of downstream reference is what makes a future
subject into my future self. As the visitor to the Empire State Building
associates my anticipatory image with his ongoing experience, attach-
ing features of the image to their referents, he will implicitly attach its
first-personal mode of presentation to himself, thereby assigning
himself the role of “me” in the context of the image. I can therefore
frame the image egocentrically and rely on him to provide the ego in
that context, without my having specified further who I have in mind.
I can of course envision an experience as undergone by someone
explicitly identified as me, just as I envision it as occurring at a location
explicitly identified as the Empire State Building. But such an envision-
ing will not be fully first-personal if it does not represent the experience
as colored by the memory of having been hereby envisioned. And if it
does represent the experience in that way, then it need not identify the
subject of the experience any further, since it will already have identified
him as the subject who will associate it with the experience and thereby
determine its references, including its reference to himself.
In this respect, my experiential anticipation is like a message
in a bottle.
Suppose that I am stranded on a desert island and I launch a bottle
containing a note that says, “If you find this message and bring it to my
wife in New York, she will reward you with $,.” To whom does ‘you’
refer in the context of my note? It refers to whoever finds the note. (If the
note is never found, my use of ‘you’ fails to refer.) Alone on my desert
island, I have no one to whom I can refer in the second person—no one
with whom I am, so to speak, on second-personal terms. In casting my
message on the waters, I am hoping to get onto second-personal terms
with someone, by succeeding in my attempt to refer with the pronoun
‘you’. That referential hope is part and parcel of my hope to communi-
cate with someone by way of the message.
So it is with my anticipatory image of a future experience. I compose
an image of a scene as experienced by a subject remembering this image
and acknowledging various features of his experience as the ones that
were hereby anticipated. I hope that there will be a subject to take
himself as “me” in the context of the image, just as I hope that there will
 Persons in Prospect
be a reader to intercept the second-personal reference in my message.
If my hope for the image is fulfilled, I will have succeeded in thinking
with it about a future subject as “me.” What I am hoping, then, is to get
onto first-personal terms with him, or to communicate with him in
the first person.
What matters in survival, I think, is being able to anticipate experiences
as providing closure to these anticipations. Of course, I care whether the
future will answer my expectations in the sense of bearing them out, by
making them true. But I care even more, and more fundamentally,
whether my expectations will be answered in the sense that future
experiences will be felt as resolving them, as if concluding a musical
I also care about communicating on first-personal terms with the
subjects of those experiences. I imagine their experiences “from the
inside” not only in the sense of imagining them from the future subjects’
point of view; I imagine them from a future point of view at which this
very imagining will be accessible in memory, so that my present, antici-
patory point of view will be “inside” the anticipated experiences, and my
present self will be internally accessible to their subjects as “me.” This
“inside” view of their experiences has an intimacy that is heightened by
my reliance on them to determine its references. One important aspect
of intimacy is the ability to dispense with referential cues. We recognize
long-married couples, for example, by their telegraphic style of conver-
sation, in which they use pronouns without antecedents—without even
following one another’s gaze—because of already knowing what is
salient to one another. Similarly, I can think of a future self as “me” and
rely on him to know that I meant him, that is, the self to whom he will
naturally attach the reference.
Because referential cues are the means
of coordinating different points of view, doing without them gives the
impression of occupying a single point of view. Like a long-married
couple, then, I and my future self seem to share a single point of view
because of being referentially in sync.
If I am right about what matters in survival, then the relevant aspect of
psychological connectedness is not the one that interests Parfit. What
. This intimacy would be lacking if I were going to undergo fission, as in the Branch
Line case. I discuss this issue in “Self to Self.”
 Philosophy & Public Affairs
interests Parfit, I have argued, is the relation of resemblance between the
termini of psychological connections: the experiences and their corre-
sponding memories, the intentions and the corresponding actions, the
acquired attitudes or traits and their persisting instantiation. These psy-
chological causes and effects often perpetuate various features of mine,
and Parfit believes that those features that are distinctive of me, or valu-
able to me, count more than others in constituting what matters in
survival. As I have just argued, however, the aspect of psychological
connectedness that really counts is the causal relation that establishes
an informational channel to carry anticipations forward to their antici-
pated cadences, and to carry future-directed references forward to find
their referents, including the future “me.” Whether the same connec-
tions preserve any of my features is relatively unimportant.
My account of what matters in survival thus explains why Parfit has
difficulty caring first-personally about his replica in the Branch Line
case. He can neither store thoughts to be retrieved by his replica nor
retrieve thoughts that are stored by him, and so he can neither expe-
rience his future as responding to the replica’s expectations nor expect
the replica’s future to be experienced as responding to his. The causal
tides can carry no internal messages between them. Even if the replica
finishes Parfit’s current book-project, he will not experience its
completion as fulfilling the remembered hopes of Parfit’s present self,
and so Parfit can no longer aim his hopes at such an experienced
fulfillment. As far as he is concerned, his book will be finished by
someone else—someone who is like him, perhaps, but who is not
himself, because of being in no position to complete the phrases of his
current mental life.
I thus arrive at the conclusion that Parfit’s difficulty in caring first-
personally about his replica is unsurprising for reasons that do not
deprive it of philosophical significance. On the contrary, his difficulty in
caring first-personally about his replica is unsurprising for the very good
reason that he and his replica are not on first-personal terms.
My conclusion has implications that reach beyond the realm of science
fiction. A person’s replicas are not the only candidate selves from whom
he is causally isolated in a way that blocks internal communication. Also
causally isolated from the person are his other possible selves—that is,
himself as he is in other possible worlds.
 Persons in Prospect
Before embarking on this topic, I will need to regiment my language
carefully. From now on, I will use the term ‘selves’ for those subjects who
are connected to me by the relation that conveys first-personal concern.
Since the present question is whether inhabitants of other possible
worlds can bear that relation to me, it amounts to the question whether
I have other possible selves at all. The candidates for selfhood in this case
are inhabitants of other possible worlds who are numerically identical
to the person I am—that is, to David Velleman. I am not questioning
whether David Velleman exists in other possible worlds: clearly he does.
What I am questioning is whether David Velleman as he might have been
is any self of mine.
But what shall we call him in relation to me? We can’t refer to him as
another possible self, since his selfhood with me is the relation that is
currently in question. I propose that we call him my identical.
I often wonder what would have become of me if I hadn’t decided to go
back to graduate school in . Maybe I would have become a
freelance writer. There are possible worlds in which I did become a
freelance writer: in some of them I am living just a few blocks from
where I live today. I wonder whether I have children in those worlds.
And so on.
The James David Velleman living in those worlds diverged from my
actual path in , and since then he has followed a very different path
from mine. My relation to this identical is therefore similar to Parfit’s
relation to his replica in the Branch Line case. In order for my concern to
reach the other possible David Velleman, it would have to travel “back-
wards up the Branch Line,” rewinding my years as a philosopher, back to
the moment of my decision to go to graduate school; “and then forwards
down the Main Line,” playing out the life I would have lived if I had
decided to become a writer instead. In my view, this maneuver cannot
convey genuine self-concern, because it does not follow a possible
. In considering this question, I needn’t worry about the metaphysical dispute
between counterpart theorists and theorists who posit strict transworld identities. Whether
another possible David Velleman is related to me in the way that justifies self-concern
doesn’t depend on whether he is strictly identical to me or merely more similar to me than
anyone else in his world. (I do wonder, however, whether the inaccessibility of transworld
identicals to self-concern played a role in David Lewis’s intuitions when he developed his
counterpart theory.)
 Philosophy & Public Affairs
channel of information between me and its object, and so it cannot
direct my concern to someone meaningfully conceived of as “me.”
In short, my relation to the person I would have been in another
possible world does not include what matters in survival. Although I am
the same person as the David Velleman who became a freelance writer,
he is not a self of mine in the sense that calls for me to identify with him,
or to identify my interests with his. He and I may be numerically identi-
cal, but—as Parfit himself puts it—identity is not what matters.
Back in , of course, I was in a position to look down many alternative
paths and to compose first-personal thoughts that would have suc-
ceeded in picking out the traveler on any one of them, had he been the
one to end up carrying the records of those thoughts along with him,
available for later retrieval from memory. Looking forward, then, I could
have entertained hypothetical self-concern for many different possible
future selves, concern that might in each case have turned out to be
about a future person.
After the point of decision, however, alternative paths were closed to
me not only in practice but also in first-personal thought. Whatever
befalls the travelers on those paths is what would have befallen David
Velleman, if I had decided differently, but his being David Velleman is, so
to speak, nothing to me: it doesn’t matter in the same way as my being
the one who might undergo different fates in the future depending on
what I now decide.
Here is a reason for not regretting what might have been—a reason
other than the ordinary, pragmatic reason that nothing can be done
about it. Regretting what I actually did is perfectly rational, since
memory puts me on first-personal terms with the agent who did it: he is
my past self in every respect relevant to self-concern. But the person who
. Thanks to Elena Weinstein for making this connection. As Matthew Hanser has
pointed out to me, this conception of what matters in survival helps to explain why, as
Lucretius observed, we want to die later but don’t regret not having been born earlier. We
cannot complete psychological cadences for the past selves who would have been born
earlier, but we can start psychological cadences for our longer-lived future selves to com-
plete. Some explain the difference by arguing that we would not have been identical with a
person born earlier. (See Thomas Nagel, “Death,” in Mortal Questions [New York: Cam-
bridge University Press, ].) I think that the explanation does not depend on the neces-
sity of our origins. Even if we could have been identical with an earlier-born person, our
relation to that person lacks what matters in survival.
 Persons in Prospect
might be better off today if I had done differently in the past—that person
is inaccessible to my self-concern. Of course he is who I might have been,
but his fate is no more pertinent to me than anyone else’s, since I can
only imagine undergoing that fate.
Yet if regretting what might have been is truly irrational, then being
relieved about what is may be irrational as well, insofar as it involves a
comparison between the two. The merely possible people whose misfor-
tunes I have skirted by making wise decisions are identical with me in
some metaphysical sense, but our identity does not give me first-
personal access to them. I can rationally celebrate what I have done
and who I have become, but not in a way that involves comparison
with the alternatives—not, at least, if my celebration is to be
rationally self-interested.
My view bears implications for our concepts of harm and benefit. These
concepts are implicitly comparative: a person is benefited when he is
made better off, harmed when he is made worse off. The question is,
Better or worse than what? According to my view, the answer should not
be, Better or worse than he might have been.
A harm should be something that makes sense for the person to
regret, in proportion to the degree of harm; a benefit should be some-
thing about which it makes sense for him to feel gratified, in proportion
to the degree of benefit.
And these emotions make sense only as the
person’s responses to comparisons of his actual welfare with that of his
other possible selves. Does it make sense for him to feel gratified that his
interests are better fulfilled than someone’s might have been? Does it
make sense for him to regret that his interests are less fulfilled than
someone’s might have been? Not unless the latter person is an object of
self-concern. Faring better or worse than a possible person for whom he
lacks self-concern might provide grounds for feeling smug or envious,
perhaps, but not for feeling regret or gratification. And differences that
do not provide grounds for regret or gratification should not qualify as
harms or benefits.
Harm and benefit should rather be conceived as making a person
better or worse off than he formerly was, and hence as involving a tem-
poral rather than counterfactual comparison. Now, this temporal
. These statements are supported by a theory of value that I will expound in Part III.
 Philosophy & Public Affairs
conception of harm and benefit might seem to imply that people cannot
be harmed if they are very badly off to begin with—in other words, if they
have nothing to lose. But even as my view plays down the importance of
counterfactual possibilities, it plays up the importance of future possi-
bilities, where multiple alternatives are still accessible to self-concern.
And we can assess how something affects a person’s interests by consid-
ering not only his current quality of life but also his future prospects.
Something can harm a person by damaging his prospects, or benefit
him by improving them. Thus, the comparisons relevant to a person’s
interests are comparisons of possible futures, and events impinge on
the person’s interests by altering those possibilities. He can be benefited
or harmed by being caused to have better or worse possible futures than
he previously had. Someone who has nothing to lose may yet have plenty
to gain, and we can easily harm him by putting those possible gains
out of reach.
This conception of harm and benefit enhances our understanding of
the non-identity problem. Surely, the right way to assess how our
despoiling the earth and depleting its resources will affect future genera-
tions is to consider how we will thereby constrain their prospects, the
possible futures accessible to them when they come into existence. Just
as surely, however, we cannot make their prospects any worse than they
will formerly have been, since people have no prospects at all until they
exist. How, then, can our irresponsible behavior affect the interests of
future generations at birth?
. This conception of harm and benefit raises questions that I will not be able to
answer here. For example: Can I harm someone by creating a dire prospect for him, or
benefit him by creating a favorable possibility, even if it is never realized? In one sense, the
possibility’s not being realized means that no harm (or benefit) has been done. In another
sense, however, I have already done him a bad or a good turn, no matter what happens.
In general, I doubt whether we can be satisfied with a single, unambiguous conception of
harm or benefit. For discussion of these concepts, see Seana Valentine Shiffrin, “Wrongful
Life, Procreative Responsibility, and the Significance of Harm,” Legal Theory (): 
; Matthew Hanser, “Harming Future People,” Philosophy & Public Affairs  (): ,
and “The Metaphysics of Harm” (forthcoming in Philosophy and Phenomenological
Research); Elizabeth Harman, “Can We Harm and Benefit in Creating?” Philosophical Per-
spectives  (): .
. Of course, our irresponsible behavior can cause future generations to become pro-
gressively worse off during their lives, and so it can cause them harm in the diachronic
sense that I have defined. What it cannot do is cause them to be harmed at birth
via their inheritance.
 Persons in Prospect
The non-identity problem now looks more problematic than before.
The problem no longer depends on the improbability of each person’s
existence, in light of which alternative policies cannot harm or benefit
future generations because of being likely to change their membership
instead. As it now appears, even if our depleting or conserving resources
would not affect the membership of future generations—even if the
same people would exist in either alternative—those people would have
no self-interested grounds on which to compare their actual lot to what
it would otherwise have been, because self-concern does not extend
to other possible worlds; and they would of course be in no position to
compare their lot to what it was before they ever existed. Our power
to make a difference to future generations via their inheritance therefore
seems even more limited than it did before; indeed, it seems to be nil.
Precisely because we cannot harm or benefit future persons via their
inheritance, however, our moral relation to them should not be con-
ceived in terms of harm and benefit in the first place. When affecting
their interests seemed merely improbable, because of the likelihood that
we would affect their existence instead, the concepts of harm and benefit
still seemed applicable. But if harming or benefiting them via their inher-
itance is metaphysically impossible, then our role as malefactors or
benefactors is morally irrelevant, and our moral relation to them must be
conceived in different terms.
In some sense, then, the non-identity problem looks less problematic.
We cannot rationalize our irresponsible behavior on the grounds that its
ill consequences for future generations will be outweighed by the
accompanying benefits of existence. Nor can we rationalize it on the
grounds that future generations will at least be no worse off than they
might have been. The rationalization “No harm done” will be not so
much ineffective as irrelevant. Where harm is strictly impossible, not
having done any is nothing to brag about.
In short, we have now eliminated both the inculpating and the excul-
pating considerations out of which the non-identity problem was con-
structed. The question remains, of course, how to understand our moral
relations to future persons. I will turn to that question in Parts II and III.
 Philosophy & Public Affairs
ii. the gift of life
They will arrange for the suckling of the children by bringing their
mothers to the nursery when their breasts are still full, taking every
precaution to see that no mother recognizes her child.
—Plato, Republic V.ii.e
Nor is there any way of preventing brothers and children and fathers
and mothers from sometimes recognizing one another; for children
are born like their parents, and they will necessarily be finding indi-
cations of their relationship to one another.
—Aristotle, Politics II.iii.a
Many people are grateful to their parents for giving them a gift consisting
in life itself. Life itself is an odd sort of gift, since there is no one around
antecedently to serve as its intended recipient.
Life is at best a benefit
that prospective parents toss into the void in the hope that someone will
turn out to have snagged it, to his own surprise as much as anyone’s. But
once parents have performed this random act of kindness, they may be
thought to have no further obligation to the future beneficiary, for whom
they have already done more than anyone will ever again be able to do.
Of course, babies are needy creatures, and their biological parents
generally bear the burden of seeing to it that their needs are met. This
allocation of childcare duties may be no more than a social convenience,
however, taking advantage of the biological fact that at least one of the
parents is bound to be on the scene when the needy creature makes
its appearance. Maybe alternative childcare arrangements would be just
as good, if only they could be institutionalized, as Plato famously imag-
ined. If proximity to the birth is all that biological parents have going
for them as caregivers, Plato’s scheme for community nurseries may
be worth considering.
This essay was presented to the Legal Theory Working Group at the Baldy Center,
University of Buffalo, and to the Legal Theory Workshop at the University of Toronto Law
School. I also had helpful discussions or correspondence on the topic with Jules Coleman,
Daniela Dover, Robin Jeshion, Arthur Ripstein, Brian Slattery, and Paul F. Velleman.
. As Matthew Hanser pointed out, no one can act with the intention of bringing a
particular person into existence (“Harming Future People,” Philosophy & Public Affairs 
[]: , at p. ).
 Persons in Prospect
Aristotle criticized this scheme as unrealistic. Children who are not
seen as the sons and daughters of anyone in particular will not be prop-
erly cared for, he thought; and in any case, people will seek out their own
parents, children, and siblings, despite all efforts to keep them apart. As
Aristotle realized, human beings have a natural tendency to find and
associate with their biological relatives.
Today we can explain this tendency in evolutionary terms, since it
enables each human organism to promote the propagation of his geno-
type and to benefit from the like tendency of his relatives. But the aims of
natural selection need not be ours. If the human tendency to congregate
in biological families is a vestige of natural selection, then it may be like
the capacity for murderous jealousy, for example—a natural tendency
that human society has no reason to accommodate. Certainly, the
human affinity for consanguines is implicated in such regrettable human
phenomena as racism and xenophobia. Maybe it should be killed in the
cradle, as Plato suggested.
Still, that’s not what modern-day readers of The Republic think; they
think that Plato’s scheme for child rearing is inhumane. Why do they
think so? What would be wrong with permanently separating parents
and children at birth?
I think that associating with relatives is more than a biological impera-
tive; it’s a personal need, imposed on persons like us by our predicament
as human beings. Because I believe that biological ties have value, I also
believe that there are good reasons for assigning the duties of childrear-
ing to biological parents in the first instance. Indeed, I believe that the act
of procreation generates parental obligations that cannot be contracted
out to others, except when doing so is in the best interests of the child.
These obligations arise because being begotten is not, as many
believe, the original birthday present. As Seana Shiffrin has argued in a
brilliant paper on claims of “wrongful life,” being brought into existence
is at best a mixed blessing, and those who confer it are not entitled to
walk away congratulating themselves on a job well done.
. For a different defense of the same position, see Rivka Weinberg, “The Moral Com-
plexity of Sperm Donation” (forthcoming in Bioethics).
. Seana Valentine Shiffrin, “Wrongful Life, Procreative Responsibility, and the Signifi-
cance of Harm,” Legal Theory (): . Brad Inwood has directed me to Seneca’s De
Beneficiis, Book , Sections . For example: “[I]t is a pretty trivial benefit for a father and
 Philosophy & Public Affairs
Shiffrin argues that bringing someone into existence is a morally
equivocal act, because it entails imposing harms on the person as well as
bestowing benefits. Shiffrin argues further that a fundamental asymme-
try between harms and benefits prevents the harm imposed by procre-
ation from being justified by the benefit bestowed. And Shiffrin attempts
to explain the asymmetry by proposing a philosophical account of harm,
although she does not develop it fully.
Now, although I agree with Shiffrin that bringing someone into exist-
ence is a morally equivocal act, I do not think that it can be equivocal
because of conferring a mixture of harms and benefits. For as I explained
in Part I, I believe that a person can be neither harmed nor benefited by
being brought into existence. I will therefore devote the first half of this
part to paraphrasing Shiffrin’s arguments in slightly different terms,
by drawing out elements, already implicit in them, of an Aristotelian
conception of human well-being. I will then draw some conclusions
that are congruent with Shiffrin’s and a few more that I doubt whether
she would endorse.
The best way to explain Shiffrin’s conception of harm, I think, is to apply
it, not to cases of harm per se, but to the philosophical problem of
distinguishing between pain and suffering. That pain and suffering are
distinct is obvious from the many cases of pain that do not occasion
suffering (stubbed toes, skinned knees), as well as cases of suffering that
do not necessarily involve pain (loneliness, boredom).
What makes the difference between pain and suffering is coping.
Suffering occurs when someone cannot or does not cope with adversity
of some kind. To cope with pain or other adversity is to exercise, or to
give oneself the sense of exercising, some degree of control over the
mother to sleep together unless there are additional benefits to follow up on this initial
gift and to consolidate it with additional services to the child. It is not living which is
the good, but living well. And I do live well. But I could have lived badly” [Section ,
Inwood’s translation].
. That the goods and ills of existence are in some sense asymmetric is an intuition
discussed by several philosophers. See, e.g., Trudy Govier, “What Should We Do About
Future People?” American Philosophical Quarterly  (): ; David Benatar, “Why It
Is Better Never to Come Into Existence,” American Philosophical Quarterly  (): 
; Michael Tooley, “Value, Obligation and the Asymmetry Question,” Bioethics  ():
. The issue is discussed by Parfit, Reasons and Persons, p. .
 Persons in Prospect
adversity itself or, at least, over one’s reactions to it. Coping is therefore
a way of exercising one’s will in the face of adverse circumstances, by
managing one’s response to them and maybe also by managing the
circumstances themselves.
When someone fails to cope, we describe him as going to pieces,
falling apart, breaking down—all expressions that reflect damage not
just to the body or to personal projects but to the self.
Failure to cope
entails damage to the self because it entails a defeat or disabling of the
will. The person is thrown into a condition of helplessness in the face of
some obstacle or assault. Stripped of his agency, he is damaged in his
very personhood. The fact and the experience of this damage to the self
are constitutive of suffering.
This brief account of suffering echoes Shiffrin’s account of harm. She
suggests that harm consists in a condition toward which a person finds
himself in a position of passive subjection—the position, as Shiffrin puts
it, of an “endurer.” She thus reverses the order of explanation between
the badness of harm and our unwillingness to undergo it. It’s not that
we’re unwilling to undergo something harmful because it’s bad; rather,
something is bad enough to qualify as harmful if and because we find
ourselves undergoing it unwillingly.
Shiffrin also briefly suggests a corresponding account of benefit. What
she says is that unsought benefits are not as good as benefits that
the recipient has chosen to pursue and has succeeded in obtaining.
She thereby suggests that, while being passively withstood is
constitutive of harm, being actively sought and attained is at least
characteristic of benefit.
These remarks about harm and benefit ground Shiffrin’s explanation
of the asymmetry between the harms and benefits entailed in the gift of
life. In Shiffrin’s view, the asymmetry arises from the fact that the gift of
life is never sought or even accepted by its recipient. He simply becomes
. For this account of suffering, see Eric J. Cassell, “Recognizing Suffering,” Hastings
Center Report  (): . See also Kathy Charmaz, “Loss of Self: A Fundamental Form
of Suffering in the Chronically Ill,” Sociology of Health and Illness (): .
. Because coping is an exercise of the will, it requires choice on the part of the subject.
That’s why we can sometimes think that people have chosen to suffer, although we’re never
quite sure. There is no clear line between inability and unwillingness to cope, but there
certainly are cases in which someone could cope but chooses not to; or maybe he cannot
choose to cope.
 Philosophy & Public Affairs
aware, long after the fact, of having been stuck with it. Even if the recipi-
ent welcomes this gift retrospectively, his will was nevertheless pre-
empted when it was given to him, since he had no chance to refuse or
accept. The harms that accompany this gift are consequently aggravated
by having been imposed on him willy-nilly, with the result that he is
already in a relation of passive subjection to them from the start. And the
associated benefits are somewhat undermined by lacking the features of
choice and successful effort that belong to the most significant benefits.
Thus Shiffrin. Much as I admire her attempt to explain the asymmetry
between the goods and ills of existence, I do not believe that a balance of
goods and ills can account for what is morally equivocal about procre-
ation. Still, I think that her explanation points us in the right direction, by
pointing us toward an Aristotelian conception of human well-being.
According to Aristotle, human well-being consists in the exercise of
capacities that are in excellent condition, and pleasure is that complete
absorption in the exercise of one’s capacities which their being in excel-
lent condition tends to facilitate.
The excellent condition of one’s
capacities is what Aristotle called aretê. His claim that pleasure consists
in being absorbed or engrossed in exercising one’s capacities has been
confirmed by research into what psychologists call “flow.”
Aristotle’s conceptions of well-being and pleasure are hospitable to
Shiffrin’s account of harm and its asymmetrical relation to benefit. Any-
thing that casts a person into a state of passive subjection will prevent
him from exercising his capacities, and it will also deprive him of the
enjoyment of becoming absorbed in their exercise. Conversely, any good
that is acquired through the exercise of the relevant capacities will bring
with it a bonus of flourishing and “flow,” like a destination that lies at the
end of an engrossing journey.
I think that Aristotle’s conceptions of human well-being and pleasure
also carry implications for the value of the so-called gift of life, because
they imply that human happiness takes work. It takes work in the form
of exercising one’s proper capacities; more importantly, it takes work
. ‘Well-being’ and ‘flourishing’ are not precise equivalents for Aristotle’s eudaimonia,
since they can be achieved at a particular time, whereas eudaimonia can be achieved only
over the course of an entire life.
. Mihaly Csikszentmihaly, Flow: The Psychology of Optimal Experience (New York:
Harper and Row, ).
 Persons in Prospect
because the relevant capacities must be acquired by practice and
habituation. In this respect, humans are unlike other animals, whose
well-being consists mostly in the exercise of capacities that are innate. A
cat is born already equipped for the activities that will constitute its
flourishing; a human being must be educated and trained for his
most rewarding activities.
According to the Aristotelian view, then, a human child is born with
the general, second-order capacity to acquire the further, specific
capacities whose exercise will eventually constitute its flourishing as
an adult. The gift of life is therefore the gift of an opportunity—the
opportunity to do the work and thereby gain the reward of human
This opportunity is accompanied by a corresponding threat and a
corresponding risk. The threat is that if the child doesn’t undertake the
work prerequisite to flourishing, it will suffer harm. And we can now see
that it will be harmed quite literally, because without the capacities
needed for human flourishing, the child will find itself in a position of
passive subjection to its circumstances, lacking the resources to cope
with them. The corresponding risk is that even if the child accepts
the challenge of flourishing, it may nevertheless fail. (The streets of
every large U.S. city are littered with individuals who are not coping
with their circumstances, or are coping only poorly, and who are
consequently faring poorly.)
The opportunity wrapped up in the gift of life is thus an offer of the
sort that the child cannot refuse. To be born as a human being is to be
handed a job of work, with a promise of great rewards for success, a
threat of great harm for refusal, and a risk of similar harm for failure. The
scene on which a human child appears willy-nilly is the scene of a pre-
dicament, a challenge with very high stakes. Hence the so-called gift of
life is indeed a mixed blessing, as Shiffrin claims.
Shiffrin and other philosophers tend to view parental obligations
as arising from the harms and benefits that parents confer on children
by bringing them into existence. As I argued in the previous part,
however, parents are metaphysically incapable of conferring either
harms or benefits in that way. The Aristotelian spin that I have now put
on Shiffrin’s arguments enables me to conceive of parental obligations
in different terms.
 Philosophy & Public Affairs
What is equivocal about procreation is not that it confers both ben-
efits and harms on the resulting child; what’s equivocal is that it throws
that child into a predicament, confronts it with a challenge in which the
stakes are high, both for good and for ill. Moreover, it is a challenge that
no child can meet without the daily assistance of others over the course
of many years, since the human infant is not at all equipped to acquire
the necessary capacities on its own.
Consider the hackneyed example of a child who is drowning at the
deep end of a swimming pool. People lounging around the pool obvi-
ously have an obligation to rescue the child. But the obligation to fish the
child out doesn’t fall on the bystanders equally if one of them pushed the
child in. The one responsible for the child’s predicament is not just a
bystander like the others, and he bears the principal obligation.
Obviously, if the responsible party cannot or will not help the child,
then others are obligated to act. The child has a right to be saved by
somebody if not by the person who caused its predicament. But just as
obviously, the person who pushed the child into the pool should have
considered beforehand, not just whether someone or other would come
to its assistance, but whether he himself was willing and able to fulfill the
obligation of assistance that he was about to incur. You shouldn’t go
pushing children into the deep end if you aren’t willing to get wet.
Likewise with procreation and parenting. In my view, parents who
throw a child into the predicament of human life have an obligation to
lend the assistance it needs to cope with that predicament, by helping it to
acquire the capacities whose exercise will enable it to flourish and whose
lack would cause it to suffer. By choosing to create a child, perhaps even by
choosing to have sex, adults take the chance of incurring this obligation.
To risk incurring the obligation without intending to fulfill it is irrespon-
sible; actually to incur it and then not to fulfill it is immoral.
I will shortly consider whether it is morally permissible for biological
parents to delegate this obligation to others. Is the obligation incurred
through the act of procreation an obligation to see that the child receives
the needed assistance in coping with the human predicament? Or is it an
obligation to render that assistance oneself, in person?
. Jeff Sebo has directed me to Sidgwick’s remarks on the subject: “This . . . we might
partly classify under . . . duties arising out of special needs: for no doubt children are
 Persons in Prospect
Of course, parental obligations must sometimes be transferable in
practice. A child has a right to grow up in the care of parents who are
willing and able to care for it. If its biological parents do not rise to the
task, then the child has a right to adoptive parents who are willing and
able to take their place. Thus, the mere unwillingness of biological
parents to discharge their obligations may be sufficient to ensure that
those obligations may be transferred to others, in deference to the rights
of the child.
But this practical accommodation does not mean that the biological
parents are morally permitted to abdicate their responsibilities at will.
We do not think that parents are permitted to relinquish a newborn for
adoption because of a last-minute social engagement, for example, or
dismay at the size of its ears.
More importantly, we don’t think that adults are permitted to con-
ceive a child with the prior intention to put it up for adoption. A woman
may not decide to conceive simply in order to have the experience or
health benefits of pregnancy, we think, no matter how confident she may
be of finding suitable adoptive parents to take over her subsequent
responsibilities. Thus, we regard parental obligations as transferable,
morally speaking, only under exigencies that make their transfer benefi-
cial for the child rather than convenient for the parents.
In one case, however, we tolerate a practice equivalent to creating a
child for adoption. Those who “donate” their sperm and eggs play their
role in conceiving children whom they have no intention of parenting.
Indeed, they play their role in conception precisely on the condition that
they will never be called upon to deal with the resulting children, a
naturally objects of compassion, on account of their helplessness, to others besides their
parents. On the latter they have a claim of a different kind, springing from the universally
recognized duty of not causing pain or any harm to other human beings, directly or
indirectly, except in the way of deserved punishment: for the parent, being the cause of the
child’s existing in a helpless condition, would be indirectly the cause of the suffering and
death that would result to it if neglected. Still this does not seem an adequate explanation
of parental duty, as recognised by Common Sense. For we commonly blame a parent who
leaves his children entirely to the care of others, even if he makes ample provision for their
being nourished and trained up to the time at which they can support themselves by their
own labour. We think that he owes them affection (as far as this can be said to be a duty)
and the tender and watchful care that naturally springs from affection: and, if he can afford
it, somewhat more than the necessary minimum of food, clothing, and education” (The
Methods of Ethics [Indianapolis: Hackett Publishing, ], p. ).
 Philosophy & Public Affairs
condition readily accepted by those who purchase their gametes, which
would be unacceptable if they came with parental strings attached.
Why do we condone the antecedent intention to transfer parental obli-
gations in this case?
Before I discuss the transferability of parental obligations, I want to
discuss a different question raised by donor conception, about the pro-
vision that one must be able to make for future children in order to be
justified in creating them. People should not create children for whom
they cannot provide adequately; but what is an adequate provision? In
particular, does an adequate provision require an opportunity for the
child to know and be reared by its biological parents?
Here I am using the word ‘adequate’ in a sense that is relativized to a
particular decision, namely, the decision whether to create a child. Most
adoptive parents make more than adequate provision for their adopted
children, but the relevant standard of adequacy is premised on the chil-
dren’s already existing and needing a home. My question is what provi-
sion for a child is adequate to justify the decision to create it, in the first
place. And my view is that the standard of adequacy applicable to the
procreative decision is different from the standard applicable to deci-
sions made after the child already exists.
My arguments in Part I imply that the adequacy of a child’s initial
provision is not relative to what could have been provided to the self-
same child. The child will not be in a position to identify his interests
with those of the better- or worse-provisioned children he might have
been. From the child’s perspective, the better or worse starts he could
have had in life will not be a matter of self-interest, because his self-
concern will extend only to his actual present and possible future selves,
not to children inhabiting possible histories that will already have
diverged from reality. When the child compares the hand he has been
dealt at birth with those he might have been dealt, he will not be able to
see himself as ahead or behind in the game of life; he will only see himself
as starting a life that amounts to a whole new game. Hence what could
. My discussion of donor conception will be confined to the typical case of anony-
mous donation between strangers. Cases of donation within families, or of “open” dona-
tion, are significantly different in respects that would call for different treatment.
 Persons in Prospect
have been provided to him in particular is not especially relevant to the
standard of an adequate provision.
A standard that philosophers sometimes apply to procreative deci-
sions is whether the resulting child would have “a life worth living.” In
Part III of this series, I will argue that this phrase has no meaning that can
apply to procreative decisions. ‘A life worth living’ can mean “a life worth
continuing,” but procreative decisions concern whether a life should be
started, not whether to continue it. Alternatively, ‘a life worth living’ can
mean “a life not to be regretted,” but I will argue that people are biased
against regretting their existence by considerations that depend on their
already existing, considerations that are irrelevant to the decision
whether to bring them into existence. In any case, what’s barely prefer-
able to nonexistence is not enough for a child by the standard of
adequacy that I consider appropriate.
The standard that I consider appropriate does not peg a child’s initial
provision at any particular level of happiness or well-being. Hence it is
not what philosophers call a person-affecting standard; it is rather a
personhood-respecting standard.
An adequate initial provision for a
child, in my view, is one that expresses due consideration for the impor-
tance of human life.
When creating human life, we are obligated to show due consideration
for it, not just for its individual possessors. The importance of human life
itself forbids us to treat it lightly in creating it.
Human life is important because it is a predicament faced by a creature
that matters—that is, by a person, whose success at facing it will entail the
flowering of personhood, and whose failure will entail a disfigurement of
that value, in the form of damage to the self. Just as we are obligated to
realize the value of personhood in ourselves, so we are obligated, in
creating human lives, to create ones in which that value is most likely to
flower and least likely to be disfigured. In this respect, the importance of
human life is like the importance of art—the kind of importance that
makes something worth creating well if worth creating at all.
Due consideration for the importance of human life requires us to
ensure that the human race does not go extinct, but it does not require us
. For a similar view, see Rahul Kumar, “Who Can Be Wronged?” Philosophy & Public
Affairs  (): .
 Philosophy & Public Affairs
to create any particular human lives, or any particular number of them.
With respect to individual lives, it mainly requires that we avoid creating
lives that will already be truncated or damaged in ways that seriously
affect the prospects for personhood to be fully realized within them.
I claim that a life estranged from its ancestry is already truncated in this
way. This claim is no less than universal common sense—though it is
also no more, I readily admit. I cannot derive it from moral principles; I
can at best offer some reflections on why we should trust rather than
override common sense in this instance.
When I say that my claim is universal common sense, I mean that
people everywhere and always have based their social relationships, in
the first instance, on relations of kinship, of which the basic building
block is the relation between parent and child. Not every society has
favored the nuclear family, of course, but virtually every society has
reared children among their kin and in the knowledge of who their bio-
logical parents are. The universal consensus on this matter is enshrined
in the United Nations Convention on the Rights of the Child. Article ,
paragraph , states: “The child shall be registered immediately after birth
and shall have the right from birth to a name, the right to acquire a
nationality and, as far as possible, the right to know and be cared for by
his or her parents.”
. The Convention is posted at
See Eric Blyth and Abigail Farrand, “Anonymity in Donor-Assisted Conception and the UN
Convention on the Rights of the Child,” International Journal of Children’s Rights  ():
. The Implementation Handbook for the Convention on the Rights of the Child makes
clear that the term ‘parents’ in this clause includes biological parents in the first instance,
and that the Convention therefore militates against the practice of anonymous gamete
donation (Rachel Hodgkin and Peter Newell, Implementation Handbook for the Conven-
tion on the Rights of the Child [UNICEF, revised edition ], pp. ).
For some social-scientific and legal perspectives, with further references, see Michael
Freeman, “The New Birth Right? Identity and the Child of the Reproductive Revolution,”
The International Journal of Children’s Rights (): ; Amanda J. Turner and Adrian
Coyle, “What Does It Mean to Be a Donor Offspring? The Identity Experiences of Adults
Conceived by Donor Insemination and the Implications for Counselling and Therapy,”
Human Reproduction  (): ; Lucy Frith, “Gamete Donation and Anonymity,”
Human Reproduction  (): ; Truth and the Child: A Contribution to the Debate on
the Warnock Report, ed. N. Bruce, A. Mitchell, and K. Priestley (Edinburgh: Family Care,
); Truth and the Child  Years On: Information Exchange in Donor Assisted Conception,
ed. Eric Blyth, Marilyn Crawshaw, and Jennifer Speirs (Birmingham: British Association of
Social Workers, ).
 Persons in Prospect
When people deny the importance of biological ties, I wonder how
they can read world literature with any comprehension. How do they
make any sense of Telemachus, who goes in search of a father he cannot
remember? What do they think is the dramatic engine of the Oedipus
story? When the adoptive grandson of Pharaoh says, “I have been a
stranger in a strange land,” what do they think he means? How can they
even understand the colloquy between Darth Vader and Luke Skywalker?
Surely, the revelation “I am your father” should strike them as a piece of
dramatic stupidity—a remark to be answered, “So what?”
As the stories of Telemachus, Oedipus, Moses, and even Luke Sky-
walker illustrate, people unacquainted with their origins have been seen
throughout history as dramatically, even tragically disadvantaged. There
must be some reason why people living at different places and times,
under very different conditions, have converged on the opinion that
a relationship with biological parents is essential to the minimally
adequate provision for a child. To be sure, other articles of age-old con-
sensus have been rejected fairly recently in history—the permissibility of
slavery, for example. But they have been rejected on the basis of soul-
searching reflection, whereas the rise of donor conception has been
driven by the procreative preferences of adults, with little thought for
the children involved.
Ironically, the preferences of these adults are often based on the same
common sense that ought to raise questions on behalf of the children.
The reason for resorting to donor conception, after all, is usually the
desire of an adult to have a biologically related child despite lacking a
partner with whom he or she can conceive. Yet whereas the parent will
be just as fully related to the child as any mother or father, the child will
know only half of its biological ancestry. These adults seek to enlarge
their own circle of consanguinity by creating children who will never
know half of theirs. Where is the common sense in that?
The material cited here argues that donor-conceived offspring should have access to
information about their biological parents. In this part I argue for a stronger conclusion—
that donor conception as usually practiced is wrong. In my view, the reasons for conclud-
ing that children should have access to information about their biological parents support
the stronger conclusion that, other things being equal, children should be reared by their
biological parents. For many children already born, other things are not at all equal, and
adoption is therefore desirable; but as I argue below, other things are indeed equal for
children who have not yet been conceived.
 Philosophy & Public Affairs
As I have said, I cannot prove that knowing and being reared by biologi-
cal parents is part of the minimally adequate provision for a child;
the best I can do is to make plausible the venerable and worldwide
conviction to that effect. People have tried living in vastly diverse ways,
but they have almost always settled on lifeways that accord central
importance to biological family ties. Let me offer some considerations
that may explain why.
Part of the task facing a human being is to find goals and activities in
pursuit of which to develop and exercise the capacities relevant to
human flourishing. A human being needs to find work, employment: he
needs, as we say, to get a life. A cat does not need to get a life: it instinc-
tively does what it needs to do in order to do well. Getting a life is a task
peculiar to the human being, who is not born to do anything in particu-
lar, and must therefore figure out what to do with himself.
A human being accomplishes this task by becoming a self worth doing
one thing rather than another with. That is, he forms an identity—a
complement of traits and attitudes, reflected in a self-image by which to
guide their expression in practice. The task of identity formation is not
optional for a human being. As soon as he acquires the cognitive where-
withal to ask, “Who am I?” and “What am I like?,” he is obliged to start
coming up with answers, in order to form a specific identity for which
there will be specific ways of flourishing.
The task of forming an identity is carried out on raw materials that
are not infinitely plastic. A human being begins life with a somewhat
determinate temperament and set of aptitudes, which can be kneaded
into many different shapes but not into just any shape whatsoever.
These individual raw materials are present at birth, as determined by
the child’s genetic endowment (and perhaps by the intrauterine
environment as well).
Research on twins and adoptees has shown that many psychological
characteristics are heritable to a considerable degree. Genetic differ-
ences are responsible for a proportion of the variance between people
. As Sophia Moreau has pointed out to me, there are cultures in which one’s identity
is largely dictated by social convention. Even within these cultures, however, the individual
remains responsible for a significant degree of self-definition. From our cultural distance,
the nineteenth-century British housemaid seems to have been stamped with a prefabri-
cated identity; below stairs, however, that housemaid may have been no less self-defined
than we are today.
 Persons in Prospect
not only in IQ (somewhere above  percent) but also for the variance in
their traits of personality such as extraversion, conscientiousness, agree-
ableness, and openness to experience (around  percent); in whether
their interests are artistic, social, enterprising, or conventional (around
 percent); in their inclination toward authoritarian or conservative
attitudes (around  percent); and even in their degree of religiousness
(around  percent).
These measures of heritability are manifested, for
example, in greater similarity between identical twins than between fra-
ternal twins, or between biological siblings reared apart than between
unrelated children reared together. In many cases, the effects of genetic
endowment tend to increase with age, possibly because the influence of
guardians wanes. As people approach adulthood, in other words, they
come into their genetic inheritance.
Thus, the predicament into which you were born, though generically
human in many respects, was also highly individual, because it required
you to fashion an identity out of a genetically inherited supply of
raw materials. The possibilities and constraints inherent in those
materials gradually came to the fore as you grew up and formed
your adult identity.
A few people in the world had already coped or were already coping
with predicaments similar to yours in its distinctive features—namely,
your biological parents and siblings. Not only did each of your parents
form an identity out of a genetic endowment half of which was
to become half of yours, but also they jointly forged an identity as a
couple, by reconciling between themselves the manifestations of what
were to become the two halves of your genetic endowment. Or that’s
. My argument does not rest on any particular quantitative measures of heritability. I
cite these statistics only for the sake of suggesting a rough order of magnitude to which
psychological traits are probably heritable. In considering the statistics, keep in mind that
what accounts for variance among individuals does not necessarily account for variance
among groups. For example, individual variance in skin color is largely heritable, but the
variance between lifeguards and coal miners is almost entirely due to environment.
The statistics cited here are drawn from Thomas J. Bouchard Jr., “Genetic Influence on
Human Psychological Traits: A Survey,” Current Directions in Psychological Science 
(): . On the heritability of values and religious attitudes, see Laura B. Koenig and
Thomas J. Bouchard Jr., “Genetic and Environmental Influences on the Traditional Moral
Values Triad—Authoritarianism, Conservatism, and Religiousness—as Assessed by Quan-
titative Behavior Genetic Methods,” in Where God and Science Meet; How Brain and Evo-
lutionary Studies Alter Our Understanding of Religion, Volume I: Evolution, Genes, and the
Religious Brain, ed. Patrick McNamara (Westport, Conn.: Praeger, ), pp. .
 Philosophy & Public Affairs
what they did if they were a couple. For that very reason, however, you
stood to benefit from their being a couple; and for similar reasons, you
stood to benefit from their rearing you together with your biological
siblings, if any.
This claim depends on an assumption about heritability that is politi-
cally incorrect, I know. We are supposed to believe that every child is
born with the capacity to fulfill any arbitrary human aspiration. In
private, however, most parents realize that part of their job is to help
their child form realistic aspirations, folded into an identity in which it
can truly flourish; and they realize that their ability to do so is greatly
enhanced by their ability to recognize in the child various traits, inclina-
tions, and aptitudes that they have seen before, either in themselves or in
other members of the family.
In the first instance, of course, family resemblance is physical, and family
members usually value the physical resemblances among them. There is
a temptation to dismiss this attitude as shallow, but I think that it
expresses a deep human need. For as human beings, we need to recon-
cile our identities as persons with our identities as animals.
The structure of human memory is such as to elicit an identification
between the self who remembers and the self of the experience retrieved
from memory.
Locke thought of that identification as constituting per-
sonal identity. Even if his metaphysics was shaky, his phenomenology
was impeccable: we certainly seem to have existed at whatever times and
places we remember experiencing, so that our sense of persisting
through time does not depend on reidentifying our bodies on different
occasions. Our relation to our bodies can therefore seem to be contin-
gent. We feel embodied in but not identical to our bodies, and so we can
imagine, for example, swapping bodies with other people.
. My arguments in Part I imply that the benefit in question consisted, not in a coun-
terfactual life-history that would have been preferable, but rather in an improvement that
could have been brought about in your actual future prospects. Of course, if your parents
conceived you with the intention of transferring their parental obligations to others, then
this benefit may have been ruled out before you existed, hence before you had any future
prospects to be improved. As I explain at the end of this part, however, it would have been
wrong of your parents to conceive a child with the intention of refusing to provide the
relevant benefit when it became possible to provide it.
. I discussed this phenomenon in Part I, and I have discussed it before in “Self to Self”
and “So It Goes.”
 Persons in Prospect
To be born in a human body is thus to be susceptible to alienation
from it. We are probably the only animals capable of feeling uncomfort-
able in our own bodies, even hating them—and loving them, too, for that
matter. Coming to terms with our bodily selves is thus a part of
the human predicament.
A connection to biological parents helps us to cope with this aspect
of our predicament. In infancy we learn to love human faces whose
features will eventually be blended in the face that emerges in the mirror
as we reach adulthood. We grow into a body akin to the bodies from
which we came, while growing into a personality akin to the ones that
animate those other bodies. We thus repeatedly have the sense of
becoming our own parents, a common form of intergenerational déjà vu.
Those who do not know their parents can only wonder who they are
becoming. Hence they can only wonder, How did someone like me come
to be living in a body like this?
Some people say that they have nothing in common with their parents
and siblings. I think that they are speaking figuratively; or maybe they are
just in denial. Almost all of us look and sound and feel and move and
think like the people from whom we came: a genuine sport of nature is
very rare. What is more likely is that a person’s similarities to his relatives
lie in aspects of himself that don’t matter to him, or that he dislikes and
rejects. Not valuing commonalities is indeed a way of not having any-
thing in common, figuratively speaking; it just isn’t a way of literally
having nothing in common.
Someone who doesn’t value what he has in common with his relatives
may think that he need never have known them in order to forge his
independent identity. I doubt it. This person is likely to have defined
himself as different from his relatives precisely because they exemplified
aspects of himself that he would otherwise have been unable to discern
clearly enough to disdain. Learning not to be like his relatives has still
involved learning from them: if he had never known them, he might well
have ended up more like them.
The point is that biological origins needn’t be worth embracing in
order to be worth knowing. Someone who doesn’t know his relatives
cannot even turn up his nose at them. The question for him is not “Shall
I follow my progenitors?” but “Am I following them?” and to this latter
question he can never know the answer. He can have neither the
 Philosophy & Public Affairs
satisfaction of continuing in their footsteps nor that of striking out on his
own, because their footsteps have been effaced.
Even if a child never knows its biological parents, they usually remain
significant figures in its life, figures to whom it is likely to develop an
attachment. That’s why roughly half of adopted children search for their
biological families at some point,
and it is why the children
of donor conception are now starting to search for their families as well.
In my view, the tendency to become attached to unknown parents bears
on whether parental obligations are transferable, a question to which
I now turn.
Why do these children search for absent parents who can no longer
rear them and are unlikely to form a significant relationship with them?
Having reached adulthood, haven’t they finally made these parents
redundant? Apparently not, although we can only speculate why. Here
are my speculations.
. A recent literature review concludes: “Following conservative estimates of more
recent studies in countries with open records policies, about % of all adopted persons
will, at some point in their life, search for their birth parents” (Ulrich Müller and Barbara
Perry, “Adopted Persons’ Search for and Contact With Their Birth Parents I: Who Searches
and Why?” Adoption Quarterly (): , at p. ). These numbers have recently
been increasing (p. ), perhaps in response to greater awareness and acceptance of
such searches.
. See, for example, the Donor Sibling Registry (http://www.donorsiblingregistry
.com/); the Donor Offspring/Parents Registry and Search Page (
DonorOffspring/); the “Donor Offspring” page of the Donor Conception Support Group
of Australia (; the UK Voluntary Information Exchange and
Contact Register (; and New Zealand’s Human Assisted
Reproductive Technology (HART) Register (
HARTbrochure.pdf/$file/HARTbrochure.pdf). A series by David Plotz in the online maga-
zine Slate resulted in many inquiries from donor offspring seeking their biological families
(/); Plotz discusses these inquiries, and many other aspects
of donor conception, in The Genius Factory: The Curious History of the Nobel Prize Sperm
Bank (New York: Random House, ). See also an op-ed entitled “Give Me My Own
History” by David Gollancz (The Guardian, May , ,
comment/story/,,,.html); and a series of slides from the Oprah Winfrey Show
On the similarities between donor conception and adoption, see Eric Blyth, Marilyn
Crawshaw, Jean Haase, and Jennifer Speirs, “The Implications of Adoption for Donor
Offpsring Following Donor-Assisted Conception,” Child and Family Social Work
(): .
 Persons in Prospect
Humans are unlike other creatures in being at risk for feeling un-
moored. We have both an egocentric conception of the world and an
objective conception of a creature whose conception it is, a creature who
is identical with the “I” at the center of that egocentric conception. Seeing
the world from within our own point of view and also from without makes
us susceptible to a sense of alienation. Unless we can reconcile these two
conceptions of ourselves, we may suffer what might be called existential
insecurity—an insecure sense of our own concrete reality.
The creature who I am is securely rooted in the objective order. It is
rooted in the objective order not only by being located in time and space
but also by its location in the web of causality. It didn’t just appear out of
nowhere: it is the result of causal antecedents that tie it to the rest of
spatiotemporal existence. Of course, I am that creature, and so I didn’t
just appear out of nowhere, either: I came from the same origins. Yet in
order to feel that its connections to the rest of reality are mine, I must
find a way of translating them into my egocentric perspective—a way of
seeing them from my point of view.
The challenge, in other words, is to identify subjectively with the
objective reality of the creature who I am, by seeing how that creature’s
place in reality can possibly be mine. In order to make that identification,
I must see how the connections anchoring that creature in the objective
order can have, from my personal point of view, the subjective signifi-
cance of connections.
But of course, the “I” of my egocentric perspective is a person, for
whom connections are most real when they are personal connections,
consisting in felt attachments. The way to identify subjectively with the
creature who I am objectively is to see its place in the objective world as
my place in a personal world. Personal attachments to my causal origins,
in the form of my biological parents and ancestors, enable me to expe-
rience firsthand the objective reality of the creature who I am. If I lack
such subjective correlates for the connections anchoring that creature in
objective reality, I am existentially insecure, because I am unable to see
from my perspective how its place in reality is mine. That’s why people
who don’t know their origins speak of feeling adrift in the cosmos, out of
place in the world.
This sense of rootlessness is especially acute in light of elementary
knowledge about the realm of living things. That realm is structured by
the life-function of self-replication, which locates every living thing in a
 Philosophy & Public Affairs
chain of progenitors and progeny. To be a living thing is to be a link in
that chain. Not to experience oneself as located in that chain is to lack a
sense of one’s membership in the realm of life, which is the locus of one’s
membership in reality.
Most people feel a need for a connection to that realm. It can be
expressed as a need for roots, for a home—for a family. It is manifested in
religious creation stories and cosmologies, in the perpetuation of
traditions, and in the ceremonies surrounding ancestors and memorials.
The same need naturally leads children to seek an attachment to their
biological parents. And it is another peculiarity of human beings to
be capable of becoming attached even to figures with whom they
are not acquainted.
Many animals become attached to members of their family or group,
and they appear to experience grief when these attachments are severed.
But they become attached only to others with whom they are acquainted
and whom they can recognize by sight or sound or smell. Humans, too,
become attached to one another by acquaintance, of course; but they
have the unique capacity for attachment to others whom they have never
met and wouldn’t recognize.
Those who study and counsel adoptees believe that they feel the loss
of the birth parents they never knew, and that their sense of loss is
comparable to that of children who experience parental death or
How can a child experience the loss of parents with whom it
has had no relationship to begin with? The answer is that a child is
capable of forming attachments to absent figures, provided that they are
present to its thoughts as real objects.
Typically, an object is first presented in thought when it is perceived,
whereupon a mental file may be opened to store information received
from it via perception.
Such a file is used for thinking about the thing
. See David M. Brodzinsky, “A Stress and Coping Model of Adoption Adjustment,” in
The Psychology of Adoption, ed. David M. Brodzinsky and Marshall D. Schechter (New York:
Oxford University Press, ); Steven L. Nickman, “Retroactive Loss in Adopted Persons,”
in Continuing Bonds: New Understandings of Grief, ed. Dennis Klass, Phyllis R. Silverman,
and Steven L. Nickman (Washington, D.C.: Taylor and Francis, ), pp. .
. See Robin Jeshion, “Acquaintanceless De Re Belief,” in Meaning and Truth: Inves-
tigations in Philosophical Semantics, ed. Joseph Keim Campbell, Michael O’Rourke, and
David Shier (New York: Seven Bridges Press, ), pp. . I am grateful to Jeshion for
suggesting this way of expressing what was a vague intuition on my part.
 Persons in Prospect
directly, in a way that is not mediated by a description or a concept.
One does not merely have an existentially quantified belief to the effect
that something satisfies various predicates; one does not merely have
various beliefs whose subject-terms pick out the same thing under
various descriptions; one has a mental file that stands for the thing and
collects predicates descriptive of it, much as the thing itself unifies a
bundle of properties.
Though a mental file is typically connected to its object by a channel
of perceptual information, it can also stand for an object without such a
If a creature can have intentions with respect to its own
mental representations, then it can open and maintain a file intended to
stand for a single thing. It must somehow pick out what the file is to stand
for, but thereafter it can use the file to treat the thing as an immediate
object of thought.
Of course, there is no point in opening a mental file for something that
probably doesn’t exist or cannot be picked out as its intended referent.
But no such risks can deter a child from opening mental files for a bio-
logical mother or father with whom it is unacquainted. Every child can
be certain of having one and only one such mother and father, to whom
it can refer as “my mother” and “my father,” and for whom it can there-
fore open files in the assurance of their standing for unique individuals.
The child can fill these files with speculations about its parents, and it
can become attached to those parents by thinking about them in this
distinctively objectual way.
These considerations about the need and the capacity for attachment to
biological parents are what lead me to think that parental obligations are
nontransferable. The obligations are nontransferable, I think, because
they arise in the context of a personal relationship.
Let us consider the daughter of a sperm donor, so that we can rely on
pronomial gender to keep the parties straight. If the mother is like other
recipients of donated sperm, she may insist that the girl has no use for
her biological father, because he is “nobody to her.” This statement is
demonstrably false. The daughter may be nobody to him, because he can
think of her only under the description “my possible children,” never
. Again, see Jeshion, “Acquaintanceless De Re Belief.”
 Philosophy & Public Affairs
knowing whether he is referring to anyone at all. But to her he is a real
person, locatable in thought, no matter how elusive he may be in time
and space. Like every human child, she knows that with the word
‘father’, she can reach down a causal chain to address a single other
human who is partly responsible for her existence.
In trying to cope with the predicament entailed by her existence, the
daughter can want to be helped, not just by some paternal figure or
other, but by the particular father who introduced her into that predica-
ment; who links her to humanity, the realm of life, the causal order; who
is her prototype and precursor in personal development; and who could
give her a hint of how psyche and soma might be reconciled in her case.
Out of those needs, the child can establish a mental representation
capable of sustaining an emotional attachment to her father, and she can
then frame a demand addressed directly to him, whether or not she
knows his earthly address. So personal a demand, so obviously justified,
deserves to be answered in person.
I know that my view seems grossly unenlightened. What passes for
enlightenment today, however, strikes me as the mirror image of the
purported enlightenment of the eugenics movement a century ago. Back
then, the people who claimed to know better than common sense
believed that a person’s biological heritage was all-important; today they
believe that it is utterly insignificant. Neither belief is true; either belief
can lead to a wholesale violation of rights. The rights violated in the
present case are the rights of children.
One objection to arguments like mine is that they seem to cast asper-
sions on donor-conceived children, by implying that they should never
have been born. I do not think that my arguments yield that implication
in a form that should give offense; in Part III of this series, I explain why.
Another objection is that the children of donor conception are likely to
waive any claims they may have on their biological parents. I deal with
this objection in Part III as well.
A final objection to my arguments is that donor-conceived offspring
have received the gift of life, which they wouldn’t have received without
the help of a sperm or egg donor. But I have argued that life is not
a gratuitous benefit but a predicament with which the recipients require
a kind of assistance that they will justifiably call on their biological
parents to provide.
 Persons in Prospect
Note, moreover, that an obligation undertaken in bad faith cannot be
excused by the fact that the party to whom it is owed was better off for its
having been undertaken. If my promise to assist you with a risky project
gives you the necessary confidence to begin it, then I am still obligated to
assist you even if, in retrospect, my defaulting on the promise would not
cause you to regret having begun. And if I know in advance that I am
going to default on my promise, then I cannot justify issuing it on the
grounds that it will induce you to begin a project that you will subse-
quently be glad to have begun, despite my expected default.
In this last example, my behavior is somewhat analogous to that of a
sperm donor, only not quite as bad. The sperm donor doesn’t induce his
offspring voluntarily to enter the predicament of human life, on the
grounds that they will be glad to have entered it; and he doesn’t just
expect to have all-things-considered reason to default on those obliga-
tions. The sperm donor throws his offspring into the human predica-
ment willy-nilly, on the basis of a positive intention to default on the
obligations that he thereby undertakes, since he wouldn’t have under-
taken them, in the first place, if he hadn’t planned to default on them. I
don’t think that he is morally entitled to bank on his children’s forgive-
ness in this way, even if they do eventually forgive him.
iii. love and nonexistence
The birth of a child can move us to value judgments that seem inconsis-
tent. Consider, for example, a fourteen-year-old girl who decides to have
This part was presented to the graduate student colloquium at NYU (February ); at
The Fourth Steven Humphrey Excellence in Philosophy Conference at the University of
California, Santa Barbara (February ), where the commentator was Mark Schroeder; to
an ethics conference at Northwestern University (May ), where the commentator was
Richard Kraut; to a seminar on the ethical significance of the emotions at the Centre for the
Study of Mind in Nature (Oslo, June ); and to the philosophy department at the
University of Illinois, Urbana-Champaign. For comments and suggestions, I am grateful to
the participants in these events and to Paul Boghossian, Caspar Hare, Robin Jeshion, Nishi
Shah, and Sharon Street.
This part bears some similarity to Larry Temkin’s “Intransitivity and the Mere Addition
Paradox,” Philosophy & Public Affairs  (): . Both seek to show that a combina-
tion of views about future persons is not as paradoxical as it seems. The difference between
the papers is this. Temkin focuses on failures of transitivity among comparative judgments;
I address a different problem, in which the value of a general state of affairs appears
inconsistent with the values of all possible instances. I am unsure whether the metaethical
solution that I propose for the latter problem is called for by the former.
 Philosophy & Public Affairs
a baby.
We think that the birth of a child to a fourteen-year-old mother
will be unfortunate, even tragic, and hence that she should not decide to
have one. But after the birth, we are loath to say that the child should not
have been born. Indeed, we now think that the birth is something to
celebrate—once a year, on the child’s birthday.
We may be tempted to say that we have simply changed our minds
in light of better information. Before the birth, we didn’t know how
things would turn out, and now we know more. But the birth did not
bring to light any previously unknown information relevant to our
Or, at least, I mean to restrict my attention to cases in
which it didn’t. There may be cases in which we feared specific calami-
ties, such as a birth defect or a descent into juvenile delinquency; and
then if such possibilities don’t materialize, we change our minds. I am
not speaking of such cases; I am speaking of cases in which we
disapproved of the girl’s decision for reasons that are not falsified by
subsequent developments, and yet we are subsequently glad about the
birth. The child is raised under serious disadvantages of the very sort
that we anticipated, but the severely disadvantaged child is still a child
to be cherished.
We knew in advance how we would feel. Even as we deplored the girl’s
decision, we knew that we would welcome the child. We may even have
cited this prospect to ourselves as a reason for softening our opinion:
“Don’t condemn her for deciding to have a child.” We might have said:
The paper also overlaps in important respects with Caspar Hare’s “Voices From Another
World: Must We Respect the Interests of People Who Do Not, and Will Never, Exist?” Ethics
 (): . In the last section of that paper, Hare discusses the difference between
de re and de dicto concern for persons, which is more or less the same difference that I
discuss here.
Finally, Jeff McMahan discusses many of the same issues in “Preventing the Existence of
People With Disabilities,” in Quality of Life and Human Difference: Genetic Testing, Health
Care, and Disability, ed. David Wasserman, Jerome Bickenbach, and Robert Wachbroit
(New York: Cambridge University Press, ), pp. . My approach to these issues is
similar to McMahan’s in some respects and different in others. The closest similarity is to
remarks that he makes about “attachments” on pp. ff. The greatest difference is that
McMahan analyzes cases of this kind as involving changes of evaluative judgment, whereas
I analyze them as involving pairs of judgments that seem inconsistent only if understood in
mistakenly realist terms.
. This case is discussed by Derek Parfit in Reasons and Persons (Oxford: Oxford Uni-
versity Press, ), chap. .
. McMahan makes the same point, on p. .
 Persons in Prospect
“Once it is born, you’ll be delighted.” But such arguments could not
dispel our sense that something was wrong.
One might think that these judgments can be reconciled, after all,
because their objects are not the same. When we think that the girl
should not have a baby, the object of our judgment is a quantified propo-
sition, about her having some baby or other, whereas it is the birth of a
particular baby that we will celebrate. And of course we can consistently
think that her having a baby is unfortunate in general but not in the case
of her having some particular one, since the general rule affirmed by our
first judgment may allow for an exception noted by the second.
Yet the attempted reconciliation appears to be blocked by the fact,
which was known to us in advance, that any baby she had would be
welcomed. How can we judge that a fourteen-year-old’s having a baby
would be unfortunate as a rule, if we also judge that any particular
instance of the rule would be an exception?
I do not think that we actually change our minds after the birth of
this child, if a change of mind would entail giving up our antecedent
judgment. We still think that the girl should not have had a baby,
delighted though we are with the baby she has had. That one judgment
predominated beforehand and the other afterwards should not be
allowed to obscure the fact that we are of two minds about the case.
One might hope to dispel the appearance of inconsistency by claim-
ing that the former is a prima facie judgment, deploring any birth only
insofar as the mother was underage and thus leaving open the possibility
of mitigating circumstances. But we don’t just think that the girl should
not have had a baby insofar as she is underage; we think that she should
not have had a baby all things considered; and yet we are glad about the
birth of this baby all things considered as well.
The mother herself may regret her decision. She may wish that she
hadn’t had a baby, may believe that she shouldn’t have had one. But of
course she still loves the baby and is thankful that it was born. As in our
case, her judgments persist in light of one another. That is, she regrets
having had a baby when she did even though it was this baby; and she is
thankful for this baby even though she had it when she did.
This conundrum is one of several that Derek Parfit considers in Part IV
of Reasons and Persons, the Part devoted to “Future Generations.” I want
to suggest a solution that Parfit doesn’t consider. Parfit’s entire
 Philosophy & Public Affairs
discussion presupposes that our value judgments must be consistent as
descriptions of the things they evaluate: they must be satisfiable by some
distribution of positive and negative value across the possibilities. I think
that the present case gives us reason to reject this assumption.
How could it be rational to have such different attitudes toward one and
the same event? The answer lies in the different modes of presentation
under which the event is viewed.
Our unfavorable judgment is about the baby under a description.
What makes this judgment tenable despite our countervailing judgment
is not, as we initially suspected, that it is general rather than singular.
We think not only that the girl should not have had a baby at fourteen but
that she should not have had the baby she had at fourteen, thus con-
sidered under a definite description that picks it out uniquely. The
reason why these judgments withstand our favorable judgment about
the baby is that, whereas they rely on descriptions, the favorable judg-
ment is about the baby considered demonstratively, as “this baby,”
“him,” or “her.”
Why does it matter whether we can make judgments about the baby
considered demonstratively? The reason is that such judgments are
guided by emotions that depend on acquaintance-based thought. One
such emotion is love.
In the context of its mother’s love, the child is
presented to her mind as it is known to her directly via sight and touch.
She does not love it under descriptions of the form “such-and-such a
child” or “the so-and-so” or even as “Fred” or “Sue.” The latter modes of
presentation would have been available to her even if she had merely
heard the child described or referred to by name, in which case she
would be in no position to love it. Unlike those modes of presentation,
acquaintance-based thought is a way of being mentally in touch or en
rapport with an object; and the rapport it entails is prerequisite to the
emotion of love.
. On the role of perception in love, see my “Love as a Moral Emotion,” Ethics  ():
, reprinted in Self to Self (New York: Cambridge University Press, ), pp. ;
and “Beyond Price,” Ethics  (): .
. Thus, an expectant mother who says that she already loves her future child may not
be speaking the truth, in philosophical strictness. She may be imagining how she will love
the child, mentally simulating what it will be like to love the child, or having fantasies of
loving it. But until she becomes acquainted with it, her emotion cannot be love.
 Persons in Prospect
Our mental relation to something can determine which attitudes
toward it are rational. Before we are acquainted with a baby, we can
approve or disapprove of it, but loving it is quite impossible, in my view,
and hence not rational, either; whereas loving a baby after being
acquainted with it is the easiest thing in the world; rational, too.
The different responses that are rational to have toward the baby, as
we think of it under different modes of presentation, account for our
different value judgments about its birth. We should feel free to experi-
ence these responses and hold the corresponding judgments, because
value is the shadow of such attitudes, not an independent standard of
their correctness. If the attitudes make sense, then the fact that they cast
conflicting shadows cannot undermine their authority. And they make
sense, despite the conflict between their shadows, because their inten-
tional objects are different in ways that rationally affect the emotions
informing our judgments.
When does a prospective mother become acquainted with her child? I would say that she
becomes acquainted with it when she first perceives it. And when does she first perceive
her child? I would say that she perceives the child at the point traditionally called quick-
ening, when the fetus begins to make movements that she can feel. Thus, the tradition that
interpreted quickening to be a morally relevant threshold was not just a superstition, in my
view; it drew what may