1. Prenatal injury

Consider the following case:

Consider the following claim:

There are conditions whose obtaining would arguably render (1) implausible. For example, if Amy’s using the drugs significantly reduces the extent to which pregnancy is burdensome for Amy, then (1) is arguably implausible. But suppose no such conditions obtain in Already Pregnant. Then, I assume, it is a natural thought that (1) is true. More generally, it is a natural thought that people who perform acts of prenatal injury, such as the one that Amy performs in Already Pregnant, thereby wrong the patients of these acts, absent adequate justification.

Note the difference between (1) and the following claim:

It is arguable that neither (1) nor (1*) entails the other. It is arguable that (1) does not entail (1*) because it is arguable that there are wrongings that are not wrong (e.g., certain acts of harming the innocent that prevent catastrophes), and it is arguable that (1*) does not entail (1) because it is arguable that there are wrong acts that are not wrongings (e.g., certain acts that raise the nonidentity problem, which I shall describe in the next section). Of course, many would defend both (1) and (1*), but my argument will bear much more directly on (1) than on (1*).

Note also that (1) does not (on its face) imply that Ben was wronged by Amy while he was a fetus (assuming Ben ever was a fetus). One could endorse (1) while insisting that Ben is wronged by Amy only once he has acquired certain properties that make him a potential wrongee and that no fetus has such properties. So insisting would presumably allow the proponent of (1) to hold, as many would, that Amy would not have wronged Ben if Amy had, in using the drugs, killed Ben while he was a fetus. Thus, the view that acts of prenatal injury are wrongings, absent adequate justification, does not (on its face) entail that prenatal killings are (ever) wrongings.Footnote ¹

Suppose (1) is true. What makes (1) true? Consider the following claim:

There is something clearly appealing about the thought that the truth of (2), at least in combination with some other facts that we presumably may assume to obtain (e.g., that Amy knew the risks to Ben of using the drugs when she used them, that Amy’s using the drugs does not significantly reduce the extent to which pregnancy is burdensome for her), makes (1) true. Thus, the following claim seems very attractive on the assumption that (1) is true:

Because it seems that if (1) is true, then (1←2) is true, I assume that views in tension with (1←2) are also in tension with (1). More generally, I assume that views in tension with (1←2) are in tension with the view that people who perform acts of prenatal injury thereby wrong the patients of these acts, absent adequate justification.

In what follows, I shall argue that two prominent philosophical views are each in tension with (1←2). The first, the psychological continuity view of personal identity across time, hereafter the “psychological continuity view,” implies that (2) is false. The second, the Parfitian view of prudence, hereafter “Parfitianism,” recommends the conclusion that (2), even if true, does not support (1). Thus, these views are each in tension with (1←2), with (1), and with the view that people who perform acts of prenatal injury thereby wrong the patients of these acts, absent adequate justification.

One consequence of my argument is that we appear to be stuck with a choice: either deny some widely endorsed theses concerning personal identity and prudence or accept a revisionary view of the morality of prenatal injury. But the main intended takeaway of my argument, to be clarified in due course, is that these theses recommend the conclusion that the nonidentity problem, or a very similar problem, can arise postconception.

2. The nonidentity problem

In this section, I shall describe the well-known nonidentity problem. Footnote ² Doing so will set the stage for the argument of the rest of the paper.

Consider the following case:

Ex hypothesi, the following claim is true:

But the following claim is also very plausible:

Why accept (4)? Call the particular pair of gametes from which Yorick actually originates “GG”; and suppose (extremely plausibly) that if Zelda had waited a month to have a child, then this child would not have originated from GG. Then it seems extremely plausible that Zelda’s child would in that case not have been Yorick. So, (4) seems very plausible.

Many will be tempted to endorse the following claim:

Furthermore, many will be tempted to say that (5) is true at least partly because the following claim, which might seem prima facie to be recommended by (3), is true:

However, given (4), it is not the case that: if Zelda had waited a month to have a child, then Yorick would have been better off than he actually is. For, given (4), if Zelda had waited a month to have a child, then Yorick would not have existed.Footnote ³ So, (4) implies that (6) is false. So, it is not the case that: (5) is true at least partly because (6) is true.

Of course, it does not follow that (5) is false. It might be that (5) is true, just not even partly because (6) is true. (After all, it is widely thought that we can wrong people without making them worse off than they otherwise would have been, e.g., by paternalizing them.) However, once we have (6)’s falsity in view, it seems that we ought to find (5) seriously doubtful. Given that Yorick lives a life worth living and would not have existed if Zelda had not procreated as she did, how could Zelda plausibly be thought to wrong Yorick in creating him? Presumably, if Zelda wrongs Yorick in creating him, then it would be appropriate for Yorick to say, concerning Zelda’s act of creating Yorick, the following to Zelda: “How could you have done this to me?” But even if it would be appropriate for Yorick to say to Zelda “How could you have done this?”, it seems that it would be inappropriate for Yorick to say to Zelda “How could you have done this to me?” Adapting some terms from Michael Thompson (Reference Thompson, Jay Wallace, Pettit, Scheffler and Smith2004), we can say that even if a monadic complaint against Zelda on Yorick’s part would be appropriate, a bipolar complaint against Zelda on Yorick’s part would be inappropriate.

Considerations such as the ones just raised have brought many to the conclusion that claims such as (5) are false and that agents who act as Zelda does in Not Yet Pregnant do not thereby wrong the people they create.Footnote ⁴ But few have drawn from this conclusion the further lesson that agents who act as Zelda does do not thereby act wrongly. Footnote ⁵ On the contrary, many would defend the following claim, despite denying (5):

Now I can state the nonidentity problem. The nonidentity problem, or at least one prominent version of it, is the following question: Do agents in circumstances like Zelda’s who act as Zelda does thereby act wrongly; and if so, why; and if not, why not? The nonidentity problem would presumably be a fairly easy problem if claims such as (5) were plausible, for in that case, claims such as (5*) could presumably be explained at least partly by appeal to claims such as (5). But because claims such as (5) are implausible or, at any rate, are widely taken to be hard to defend, the nonidentity problem is widely taken to be a difficult problem.

The nonidentity problem will strike some readers as a significant change of subject from that of the previous section. For Already Pregnant, unlike Not Yet Pregnant, seems not to raise (anything like) the nonidentity problem. Recall the following claims, introduced in section 1:

It seems that Amy’s use of the drugs affects her child’s wellbeing for the worse, and that Amy’s child would have been Ben even if Amy had not used the drugs. Thus, (2), unlike (6), seems plausible, and consequently (1←2) also seems plausible (given the assumption that [1] is true). Thus, Already Pregnant seems not to raise (anything like) the nonidentity problem.

But in what follows, I shall argue that the psychological continuity view and Parfitianism each recommend the conclusion that things are not as they appear. The psychological continuity view, I shall argue, implies that Already Pregnant raises the nonidentity problem and that (2) is false. And Parfitianism, I shall argue, recommends the conclusions that Already Pregnant raises a problem very similar to the nonidentity problem and that (2), even if true, does not support (1). It follows that these views each recommend the conclusions that (1←2) is false, that Already Pregnant raises (something very similar to) the nonidentity problem, and, more generally, that (something very similar to) the nonidentity problem can arise postconception.

3. The psychological continuity view

In this section, I shall identify and clarify a claim that I take to be an implication of the psychological continuity view.

The psychological continuity view is very popular, and there are many versions of it.Footnote ⁶ But all of them, as I understand them in this paper, hold that psychological continuity is a necessary condition on personal identity across time. More precisely, we can formulate this implication of the psychological continuity view as follows:

Psychological continuity is ordinarily defined in terms of psychological connectedness. Footnote ⁸ What psychological connectedness is, is a contestable matter. On one possible view, for P, as P is at t, and x, as x is at t*, to be psychologically connected is for P to have at t and x to have at t* some of the same mental items. On another possible view, for P, as P is at t, and x, as x is at t*, to be psychologically connected is for some of the mental items that P has at t and some of the mental items that x has at t* to be causally related to one another in the right way. And there are other possible views.

For P, as P is at t, to be psychologically continuous with x, as x is at t*, is for there to be a sequence of entities as they are at particular times, beginning with P as P is at t and ending with x as x is at t*, any two adjacent members of which are suitably psychologically connected.Footnote ⁹ What counts as suitable psychological connectedness is also a contestable matter. One could hold that suitable psychological connectedness is a matter of a sufficient number of psychological connections (say, enough of the same mental items), or of sufficiently strong psychological connections (say, enough mental items that have very strong causal relations to one another), or of the right kinds of psychological connections (say, memory connections), or … We could thus reformulate (7) as the claim that some person, P, who exists at some time, t, and some entity, x, that exists at some other time, t*, are identical only if there exists some sequence of entities as they are at particular times beginning with P, as P is at t, and ending with x, as x is at t*, any two adjacent members of which are suitably psychologically connected.

To illustrate: (7) seems clearly to imply that a person cannot survive an accident resulting in sudden total memory loss along with a very dramatic personality change. Whatever psychological continuity involves exactly, it seems clear that, in such a case, the person as she is “postaccident” and the person as she is “preaccident” are not psychologically continuous and, thus, that, by (7)’s lights, the person who exists after the accident and the person who existed before the accident are nonidentical. (7) does not, however, imply that an old man who now exists and the child whom we would ordinarily say the old man “once was” are nonidentical, even if the old man, as he is now, is not psychologically connected to the child, as the child was, provided the old man, as he is now, is psychologically continuous with the child, as he was. That people cannot survive the sort of accident involved in the first sort of case strikes many philosophers as intuitively plausible. That people can survive the sort of gradual psychological change that occurs in the second sort of case strikes many of the same philosophers as intuitively plausible. That these two verdicts strike many philosophers as intuitively plausible accounts for much of (7)’s popularity and, thus, for much of the psychological continuity view’s popularity.

More could be said about these matters, but I believe I have clarified (7) enough for my purposes. In the next section, I shall argue that (7) implies that Already Pregnant raises the nonidentity problem, that the nonidentity problem can arise postconception, and that (2) is false. It follows that the psychological continuity view has these implications, too.

4. The psychological continuity view and the nonidentity problem

First, some setup.

In what follows, I shall use the expression “Ben*” to refer to the child Amy would have had if she had not used the drugs in Already Pregnant. Introducing this term might seem odd, for it might seem obvious that the child Amy would have had if she had not used the drugs would have been Ben, for whom we already have an excellent term (namely, “Ben”). But I shall argue that (7) (the claim that I suggested in section 3 is an implication of the psychological continuity view), combined with some additional plausible premises, implies that Ben* would not have been Ben.

I shall assume that there is at least one time at which Ben is a person, and I shall assume that there would have been at least one time at which Ben* would have been a person. These assumptions are neutral about whether an entity can be a person at one time and a nonperson at another time and about what a person is.

I assume that Ben is associated in some very deep way with a human organism, and similarly for Ben*. I assume that Ben is not, for example, an immaterial soul with no relation whatsoever to any human organism, and similarly for Ben*. In making these assumptions, I make no further assumptions about the nature of the “association” that obtains between Ben and the organism with which he is associated in the sense at issue, and similarly for Ben*. For all my present assumptions imply, these associations might be ones of identity, parthood, constitution, grounding, causation … I shall use “Ben’s organism” to refer to the human organism with which Ben is actually associated, and “Ben*’s organism” to refer to the human organism with which Ben* would have been associated.

I shall use the term “psychological life” to refer to the mental items (if any) that an entity (e.g., a human organism) sustains over the course of its existence that are relevant to personal identity across time according to the most plausible version of the psychological continuity view (whatever that version happens to be). To illustrate: Some versions of the psychological continuity view imply that memories (or quasimemories) are the only mental items relevant to personal identity across time. That is, some versions of the psychological continuity view imply that memory (or quasimemory) connections are the only psychological connections that bear upon whether a given person, as she is at some time, and a given entity, as it is at some other time, are psychologically continuous. By contrast, some less restrictive versions of the psychological continuity view imply that (at least some members of) each of the following categories of mental items, and perhaps others, are relevant to personal identity across time: memories (or quasimemories), conscious experiences, desires, beliefs, intentions. The components of a psychological life (as I use the term “psychological life”) are whatever mental items the most plausible version of the psychological continuity view says are relevant to personal identity across time. I shall assume that the most plausible version of the psychological continuity view is quite unrestrictive, but nothing I say in what follows turns on this assumption. For ease of exposition, I shall sometimes speak of the psychological life sustained by a person’s organism, and sometimes of a person’s psychological life. These expressions, as I mean them, pick out the same thing.

Finally, I shall assume that, for any human organism, HO, there is a period of time beginning when HO comes into existence during which HO does not sustain a psychological life.Footnote ¹⁰

Now for my argument. My argument contains four premises, which jointly imply that if (7) is true, then Already Pregnant raises the nonidentity problem, the nonidentity problem can arise postconception, and (2) is false.

The first premise of my argument is the following assumption about Already Pregnant:

To see what I mean by (8), consider the following two figures:

Figure 1 represents Ben’s actual psychological life. Figure 2 represents the psychological life that Ben* would have had. The carets in figure 2 represent the fact, posited by (8), that Ben*’s psychological life would have been massively different from the start from Ben’s. Ben* would have had, from the beginning of his psychological life, massively different conscious experiences, memories, beliefs, desires, and so forth from the ones that Ben actually has over the course of his life.

Figure 1. Psychological life actually sustained by Ben’s organism

Figure 2. Psychological life that would have been sustained by Ben*’s organism

Three clarificatory points about (8). First, (8) does not imply that Ben* would not have been Ben (or that Ben* would have been Ben).

Second, (8) does not imply that Ben*’s psychological life would have had nothing in common with Ben’s actual psychological life. To illustrate: (8) is compatible with its being the case both that Ben has a belief that his mother’s name is “Amy” and that Ben* would have had a belief that his mother’s name is “Amy.” (8) posits a massive difference, not a complete difference, between the two psychological lives at issue.

Third, according to (8), Ben*’s psychological life would have been massively different from the start from Ben’s actual psychological life. To see what I mean by this, consider the following case:

There is (we can suppose) a massive difference, but not a massive difference from the start, between the actual psychological life of Cate’s child and the psychological life that Cate’s child would have had if Cate had not used the drugs. For (we can suppose) these two psychological lives are just like one another at least until Cate starts using drugs while breastfeeding, and “deviate” from one another only after that. Thus, these two psychological lives can be represented by the following figures:

(8) implies that the beginning of Ben’s actual psychological life and the beginning of Ben*’s psychological life do not “overlap” as do the beginnings of the psychological lives represented in figures 3 and 4.

Figure 3. Psychological life actually sustained by Cate’s child’s organism

Figure 4. Psychological life that Cate’s child’s organism would have sustained if Cate had not used the drugs

The lack of an “overlap” between the beginning of Ben’s actual psychological life and the beginning of Ben*’s psychological life requires that we assume that Amy used the drugs sufficiently early in pregnancy. For if Amy had used the drugs sufficiently late in pregnancy—in particular, after her child’s psychological life had begun (assuming, contestably, that Amy’s child’s psychological life began during Amy’s pregnancy)—then presumably there would have been an overlap between the beginning of Ben’s actual psychological life and the beginning of Ben*’s psychological life.

Some might wonder how typical it is for a real-world variant of Already Pregnant to be such that the analogue of (8) applied to it is true. More generally, some might wonder how often prenatal injury causes the child who ends up being born to have a psychological life massively different from the start from the psychological life that would have been had by the child who otherwise would have been born. I shall not speculate about this issue beyond registering my suspicion that in many real-world variants of Already Pregnant, prenatal injury results in a psychological life massively different from the start from the one that otherwise would have occurred. At any rate, if the argument of this section is successful, then (7) implies that how often prenatal injury results in a psychological life massively different from the start from the psychological life that otherwise would have occurred is potentially a matter of considerable moral importance. For if the argument of this section is successful, then (7) implies that cases involving prenatal injuries that result in massively-different-from-the-start psychological lives are cases in which the nonidentity problem arises postconception.

Here is the second premise of my argument:

Why accept (9)? Suppose, for argument’s sake, that (7)—the claim I have suggested is an implication of the psychological continuity view—is true. In addition, suppose for illustration’s sake that Ben is currently a twenty-year-old man. Was Ben ever, for example, a three-week-old fetus with no, or an extremely rudimentary, mental life? (7) seems clearly to imply that he was not, for Ben, as he is now, is not psychologically continuous with any such fetus, as it was. In general, (7) implies that Ben did not start existing at least until his organism’s psychological life began, and similarly for Ben*. Thus, (9) seems very plausible.Footnote ¹¹

Before I proceed, I shall introduce a new term and make a new assumption. Let the original intrinsic properties of some entity, x, be the intrinsic properties (if there are any) that x has at the first moment of x’s existence. I shall assume that Ben has original intrinsic properties and that Ben* would have had original intrinsic properties.Footnote ¹²

Here is the third premise of my argument:

Why think (10) is true? Let an organism’s psychological interval be the period (if any) during which this organism sustains a psychological life. Suppose Ben*’s organism would have sustained a psychological life massively different from the start from the psychological life actually sustained by Ben’s organism. Then presumably Ben*’s organism would have had throughout Ben*’s organism’s psychological interval intrinsic properties very different from the ones that Ben’s organism actually has during Ben’s organism’s psychological interval. But if this is so, then it seems very plausible that Ben* himself would have had original intrinsic properties very different from the ones that Ben actually has, given the assumptions that Ben came into existence when or after the psychological life sustained by Ben’s organism began and that Ben* would have come into existence when or after the psychological life that would have been sustained by Ben*’s organism would have begun. Thus, (10) seems very plausible.

It is worth acknowledging two assumptions on which my defense of (10) rests. The first is that a very great difference between the psychological lives sustained by two organisms implies, or makes very probable, a very great intrinsic property difference between these organisms. The second is that a very great intrinsic property difference between two person-associated organisms implies, or makes very probable, a very great intrinsic property difference between the persons associated with these organisms. Perhaps these assumptions could be challenged, but they strike me as very plausible. At any rate, if it should turn out that the best way to resist my argument is to resist one of these assumptions, then this would be an interesting result.

Now for the fourth and final premise of my argument:

Why accept (11)? Suppose God creates ex nihilo a bronze perfect sphere. Call this object “Brawny.” I think it plausible that Brawny could eventually become conic, given sufficiently gradual transformations over time. To deny this would presumably require denying the possibility of many diachronic intrinsic property changes that are intuitively possible. (I assume that shapes are intrinsic properties.Footnote ¹³) However, I think it extremely implausible that Brawny could have begun its existence conic. This is not to say that God could not have created a bronze cone instead of creating a bronze perfect sphere. Obviously, God could have done this. But even God could not have created a bronze cone that would have been identical to Brawny instead of creating a bronze perfect sphere. Nothing could have begun its existence with conicality that would have been identical to Brawny. Or so I am strongly inclined to believe.

Suppose Brawny could not have had conicality at its first moment of existence. Why could Brawny not have had conicality at its first moment of existence? Some might be tempted to explain this fact by appeal to the thought that Brawny (for some reason) de re necessarily lacks conicality. But recall my suggestion that Brawny could become conic, given sufficiently gradual transformations over time. If this is so, then Brawny does not de re necessarily lack conicality. Thus, the explanation just mentioned strikes me as unpromising.

Here, I think, is a much more plausible explanation of the fact that Brawny could not have had conicality at the first moment of its existence: in order to have had conicality at its first moment of existence, Brawny would have had to have original intrinsic properties very different from its actual ones; but nothing could have had original intrinsic properties very different from its actual ones; so, Brawny could not have had conicality at its first moment of existence. If this explanation is correct, then it appears that (11) is true. Thus, I believe that reflection on the case of Brawny lends considerable support to (11).Footnote ¹⁴

Still, some might think that (11) has some implausible implications. For now, I shall take (11)’s truth for granted, but, in a moment, I shall argue that the core lessons of this section can be defended without appeal to (11).

I have defended the four premises of my argument. As a reminder:

Now I shall show that these claims jointly entail my desired conclusion.

Suppose (8)–(10) are true. It follows that if (7) is true, then Ben* would have had very different original intrinsic properties from Ben. So, given (11), it follows that if (7) is true, then Ben* would not have been Ben. But if this is so, then (7) implies that Already Pregnant raises the nonidentity problem, more generally that the nonidentity problem can arise postconception (e.g., in Already Pregnant), and that (2) is false. For if Ben* would not have been Ben, then Amy does not, in using the drugs, make Ben worse off than he otherwise would have been. And because the psychological continuity view implies that (7) is true, the psychological continuity view has the implications of (7) just described.

I said above that some might think that (11) has implausible implications. Consider No Legs, a variant of Already Pregnant in which (i) Amy’s drug use causes Ben’s organism never to have legs, (ii) Ben*’s organism would have had legs (if Amy had not used the drugs), but (iii) Ben*’s psychological life would have been just like Ben’s (and, thus, (8) is false). (No Legs is hard to imagine but is arguably possible.) It might be thought that, in No Legs, Ben* would have had original intrinsic properties very different from the ones Ben actually has but would have been identical to Ben. So, No Legs might be thought to undermine (11).Footnote ¹⁵

I do not think that No Legs undermines (11), for I can think of no combination of (i) a way of filling in the details of No Legs and (ii) a respectable view of personal ontology that supports both of the following conclusions: (a) Ben* would have been identical to Ben, and (b) Ben* would have had original intrinsic properties very different from the ones that Ben actually has. Any combination of (i) and (ii) that makes (a) plausible (and there are many) seems to make (b) implausible; and any combination of (i) and (ii) that makes (b) plausible (and there are some) seems to make (a) implausible.

However, even if you find (11) implausible (whether because you think a case like No Legs undermines it or for another reason), I think you ought to accept the essentials of the argument of this section. It just seems to be a very plausible thought that, given the psychological continuity view, a person could not have had a psychological life massively different from the start from her actual one. (11) lends support to this thought, and indeed might partly explain its truth, but this thought’s plausibility seems not to depend on the plausibility of (11). But this thought, when combined with (8) (i.e., the assumption that in Already Pregnant Ben*’s organism would have sustained a psychological life massively different from the start from the psychological life actually sustained by Ben’s organism), is enough to yield the conclusions that Already Pregnant raises the nonidentity problem, more generally that the nonidentity problem can arise postconception, and that (2) is false. So, even if you reject (11), I think you ought to accept that the psychological continuity view yields these conclusions.

I shall close this section by zooming out and saying informally how I have arrived at the conclusion of this section. I shall also identify some further lessons revealed by my argument.

The nonidentity problem arises from the fact that actions that affect the wellbeing of future people can also affect which people will exist in the future and have this well-being. Commonsense metaphysics holds that the only way to affect which human people will exist in the future is to affect which human organisms will exist in the future. This is presumably at least partly why standard nonidentity cases, such as Not Yet Pregnant, are cases in which people face a choice between creating one human organism and creating a different human organism.

But not all metaphysics is commonsense metaphysics. Not every view of the metaphysics of personal identity implies that the only way to affect which human people will exist in the future is to affect which human organisms will exist in the future. Some noncommonsensical views of the metaphysics of personal identity, such as the psychological continuity view, at least strongly support the conclusion that one way to affect which human people will exist in the future is to affect the character of future psychological lives sustained by human organisms in certain ways, even without affecting which human organisms will exist in the future. Thus, there are cases, such as Already Pregnant, that do not raise the nonidentity problem given commonsensical metaphysical assumptions but that do raise the nonidentity problem given certain noncommonsensical metaphysical assumptions. In particular, given certain noncommonsensical metaphysical assumptions, e.g., the assumption that the psychological continuity view is true, there are cases, such as Already Pregnant, in which the nonidentity problem arises postconception, i.e., after the relevant human organism has been created.

I have argued that, given the psychological continuity view, cases in which prenatal injury occurs shortly after conception, such as Already Pregnant, are excellent candidates for cases in which the nonidentity problem arises postconception. But some versions of the psychological continuity view will imply that the nonidentity problem can also arise quite a while after conception. For example, a version of the psychological continuity view according to which the components of a psychological life are sufficiently sophisticated that a human organism does not begin sustaining them until some time into infancy will imply that the nonidentity problem can arise after birth. Given such a view, some cases of postnatal injury will raise the nonidentity problem. These points bring out a general lesson: given the psychological continuity view, until the time at which some human organism’s psychological life begins, it is possible for the nonidentity problem to arise for this organism.

Furthermore, I have suggested that the psychological continuity view implies that the nonidentity problem arises postconception in cases, like Already Pregnant, in which prenatal injury brings about a psychological life massively different from the start from the one that otherwise would have occurred. But a “reverse” nonidentity problem arises in some cases of prenatal benefit, given the psychological continuity view. Suppose a pregnant woman uses fetus-enhancing drugs before her fetus begins sustaining a psychological life, which make the child she ends up delivering better off than the child she otherwise would have delivered, but which also make the child she ends up delivering have a psychological life massively different from the start from the one that would have been had by the child whom the woman otherwise would have delivered. Given the psychological continuity view, this woman does not, in using the fetus-enhancing drugs, make the child she ends up delivering better off than this very child otherwise would have been, for, given the psychological continuity view, this very child would not have existed had the woman not used the drugs (or so the argument of this section suggests). This fact would seem to have some revisionary implications. For example, it might seem plausible, on the face of things, that it would be appropriate for the child whom the woman ends up delivering to be grateful to the woman for making this very child better off than this very child otherwise would have been. But, given the psychological continuity view, such gratitude would seem to rest on a mistake.

Furthermore, given the psychological continuity view, it might turn out that some severe prenatal injuries raise the nonidentity problem (because they result in psychological lives sufficiently different from the ones that otherwise would have occurred) but some not-so-severe prenatal injuries do not raise the nonidentity problem (because they do not so result). If so, it might turn out that some prenatal injuries are wrongings despite being much less severe than some otherwise similar prenatal injuries that are nonwrongings. Footnote ¹⁶

Finally, I have focused in this section on cases in which some way of affecting a human organism brings about a psychological life massively different from the start from the one that otherwise would have been sustained by this organism. But the psychological continuity view seems to imply that the nonidentity problem can arise postconception in other cases, too. Consider the following (farfetched, but presumably metaphysically possible) case: Dara, who is pregnant, uses drugs which make the child she actually ends up delivering have a psychological life massively different at the start from the psychological life that would have been had by the child Dara otherwise would have had. However, although the two psychological lives at issue are massively different from one another at their beginnings, these psychological lives “converge” not long after their beginnings and become just like one another. Thus, the psychological lives at issue in the case could be represented by figures 5 and 6:

Figure 5. Psychological life actually sustained by Dara’s child’s organism

Figure 6. Psychological life that Dara’s child’s organism would have sustained if Dara had not used the drugs

This case is a “reverse” version of Breastfeeding, discussed above.

Presumably, given the psychological continuity view, Dara’s actual child has original intrinsic properties very different from the original intrinsic properties that Dara’s “counterfactual” child would have had. If so, then the argument of this section recommends the conclusion that, given the psychological continuity view, Dara’s “counterfactual” child would not have been identical to Dara’s actual child. Thus, given the psychological continuity view, it seems sufficient for the nonidentity problem to arise postconception that the relevant psychological lives be massively different from one another at the start. Massive difference from the start, as in Already Pregnant, is sufficient but not necessary. (Compare: suppose Brawny, the bronze object discussed earlier in this section, actually has sphericality for all of its existence. I find it very hard to believe that God could have created an entity that would have lacked sphericality for even the first millisecond of its existence and that would have been identical to Brawny, even if this entity would have had sphericality for the rest of its existence.)

5. Parfitianism

One response that some might give to the argument of section 4 would be to reject the psychological continuity view and all views that imply that (7) is true, and to endorse (say) a bodily continuity view of personal identity across time instead. Doing this would presumably allow one to hold that Ben* would have been Ben, that Already Pregnant does not raise the nonidentity problem, and that (2) is true. Doing this might also allow one to insist that the nonidentity problem cannot arise postconception. I shall take no stand on whether this is a good response. I shall merely insist that if this is the best response to the argument of section 4, then this is an interesting result, given the psychological continuity view’s popularity.

However, in this section and the next, I shall argue that the Parfitian view of prudence, hereafter “Parfitianism,” supports the conclusion that (2), even if true, does not support (1). Thus, Parfitianism is in serious tension with (1←2) and with (1). I shall also argue that Parfitianism supports the conclusion that Already Pregnant raises, if not the nonidentity problem, a very similar problem, and more generally supports the conclusion that a problem very similar to the nonidentity problem can arise postconception.

In part III of Reasons and Persons (Reference Parfit1984), Derek Parfit famously distinguishes between identity and what matters prudentially, and argues that identity is not what matters prudentially. Parfit takes the following case (among others) to support his view:

Do you survive your operation in Down the Middle? Parfit argues that this is an “empty question.”Footnote ¹⁷ What Parfit means when he calls this question “empty” and why he thinks this are not important for my purposes. What matters for my purposes is that Parfit holds that even though this question is empty, the twenty good years that Lefty and Righty get are good for you from the point of view of prudence. Furthermore, Parfit takes the fact (assuming it is a fact) that the twenty good years that Lefty and Righty get are prudentially good for you, even though the question whether you survive your operation is empty, to support the following conclusion: what matters prudentially is not (even partly) identity but is instead Relation R, i.e., “psychological connectedness and/or continuity, with the right kind of cause,”Footnote ¹⁸ which Parfit takes each of Lefty and Righty, throughout their twenty years of postoperation existence, to bear to you as you were just prior to the operation.

Parfitianism is the claim that Relation R is what matters prudentially. How best to understand this claim is a difficult matter. But the following claim seems clearly to be implied by Parfitianism, however Parfitianism is best understood:

To illustrate: (12) implies, for example, that you now have prudential reason to care about how you will be in twenty years only if you, as you will be in twenty years, and you, as you are now, are psychologically continuous.

(12) is what one gets from “transforming” (7)—the thesis I claimed in section 3 is an implication of the psychological continuity view—into a thesis about what a given person has prudential reason to care about at a given time. Furthermore, Parfitianism is at least very similar to what one gets from “transforming” the psychological continuity view into a thesis about what matters prudentially. Thus, we ought to expect that Parfitianism will have implications for Already Pregnant similar to the implications for Already Pregnant of the psychological continuity view that I identified in section 4. And in the next section, I shall argue that Parfitianism does indeed have some such implications. I shall argue that (12) supports the following conclusions: that (2), even if true, does not support (1); that Already Pregnant raises a problem very similar to the nonidentity problem (if not the nonidentity problem); and more generally that a problem very similar to the nonidentity problem can arise postconception. If my argument succeeds, then Parfitianism supports these conclusions, too.

6. Parfitianism and the quasinonidentity problem

I shall first argue that (12) supports the conclusion that (2), even if true, does not support (1). My argument contains four premises. The first premise was also the first premise of my argument in section 4:

Before I introduce the second premise, I shall introduce a term. For any given person, A, say that a given nonactual psychological life, PL, would have been “prudentially-not” A’s just in case A ought, from the point of view of prudence, to regard PL as though PL would, if PL had occurred, not have been A’s.Footnote ¹⁹ This terminology is neutral about what makes it the case that a given nonactual psychological life would have been prudentially-not a given person’s. One natural view is that a given nonactual psychological life, PL, would have been prudentially-not A’s if and only if, and because: if PL had occurred, then PL would have (in fact) not been A’s. But this is not the only possible view of the matter, as will emerge in what follows.

Here is the second premise of my argument:

Less formally, and a bit imprecisely, but not so imprecisely as to cause confusion, we might say that, according to (13), (12) supports the conclusion that a person ought, from the point of view of prudence, to regard an outcome in which his psychological life would have been massively different from the start from his actual one as though he would never have existed in such an outcome.

(13) is not a claim about what (12) entails. For all (12) entails, it might be (for example) that I ought, from the point of view of prudence, to regard any nonactual psychological life that would in fact have been mine as though it would have been mine, regardless of the extent to which it is dissimilar from my actual psychological life. Nevertheless, I take the clear spirit of (12) to recommend the thought that if I had had a psychological life massively different from the start from my actual one, then things would have been, from a prudential perspective, as though I had never existed. Hence (13).

In defense of (13), suppose (12) is true. Now imagine some way that your life could have gone that would have been psychologically massively different from the start from your actual one but in which you would have been better off than you actually are. What reason, of a broadly self-interested character, could you have to wish that you had had this counterfactual life instead of your actual one? The most obvious candidate for such a reason would seem to be the fact that you would have been better off in this counterfactual life than you are in your actual life. But, if you accept (12), then it would seem strange for you to take this fact to be such a reason. After all, if you accept (12), then you will not take the bare fact that you would be better off in some epistemically possible future or past to be a reason, of an even broadly self-interested character, to want to have or to have already had this future or past, or to prefer it to another one, or to do any other such thing. Taking this fact to be such a reason seems to involve presupposing that identity matters, contrary to (12). Rather, if you accept (12), then you will take the fact that there would be or would have been some individual appropriately psychologically related to you as you are now to be a reason to want some epistemically possible future or past, to prefer it to some other one, and so on. Given that you accept (12), then, why would you take the bare fact that you would have been better off in some counterfactual life than in your actual life to be a reason of a broadly self-interested character to wish that you had had this counterfactual life? It seems that to do this while accepting (12) would be to combine an “identity doesn’t matter” attitude toward how it is epistemically possible one’s life will go in the future or went in the past with an “identity matters” attitude toward how one’s life could have gone but did not. Although there is no contradiction involved in this combination of attitudes, I do not think that it is an attractive combination of attitudes, for it seems to apply different standards of prudential rationality to different things without justification. Thus, I think that (12) supports the thought that one ought to have a broadly Parfitian attitude not just toward (say) one’s epistemically possible futures and pasts, but also toward how one’s life could otherwise have gone. Thus, I think (13) is true.Footnote ²⁰

Note that (8) and (13) both employ the expression “massively different from the start.” It is important for the validity of the argument to follow that this expression is employed univocally between (8) and (13). However, some might worry that, given some legitimate readings of “massively different from the start,” there are versions of Already Pregnant about which (8) is plausible but for which (13) has implausible implications. Consider a version of Already Pregnant in which Ben actually suffers severe pain for all of his psychological life, Ben* would not have suffered severe pain for all of his psychological life, but otherwise these two psychological lives are very similar to one another from the start.Footnote ²¹ (This case might be hard to imagine, but it is very arguably possible.) This is clearly a version of Already Pregnant about which (8) is true, given some legitimate readings of “massively different from the start.” However, it might be thought that even if (12) is true, this is not a version of Already Pregnant in which Ben*’s psychological life would have been prudentially-not Ben’s. For it might be thought that even if (12) is true, Ben has (say) good prudential reason to wish that Ben*’s psychological life had occurred instead of Ben’s actual one. Thus, it might be that given some legitimate readings of “massively different from the start,” there are versions of Already Pregnant about which (8) is plausible but for which (13) has implausible implications.

If you are moved by the worry just presented, then please assume that the first premise of my argument in this section is not (8) but the following, arguably more demanding claim:

In general, if you think that there are legitimate readings of “massively different from the start” such that, for some versions of Already Pregnant, (8) is true but (13) has implausible implications, then please assume in what follows that the first premise of my argument in this section is (8+), not (8). I shall ignore this complication in what follows.

The third premise of my argument involves the expression “bipolar complaint” introduced in section 2:

I take (14) to be very plausible. Suppose (for whatever reason) Ben*’s psychological life would have been prudentially-not Ben’s. And suppose (contrary, I have argued, to the psychological continuity view) that (2) is true. Then Ben ought, from the point of view of prudence, to regard Amy’s making him worse off than he otherwise would have been as she did as though it were the reason why he ever came into existence at all. But if this is so, then, given that Ben lives a life worth living, it is hard to believe that Ben has a justified bipolar complaint against Amy that is grounded in the fact that Amy did what she did. That is, it is hard to believe that it would be appropriate for Ben to say to Amy, concerning Amy’s making him worse off than he otherwise would have been as she did, anything like “How could you have done this to me?” On the contrary, perhaps Ben ought to be grateful to Amy for doing what she did. Thus, I take (14) to be very plausible.

My present point does not rely on the (highly controversial) principle that if A performs some action, X, the patient of which is B, then B has a justified bipolar complaint against A that is grounded in the fact that A performed X only if B has prudential reason to wish that A had not performed X. If Wanda paternalizes Xavier in a way that is unqualifiedly good for Xavier, then, for all I assume here, Xavier has a justified bipolar complaint against Wanda even though Xavier has no prudential reason to wish that Wanda had not done what she did. But in ordinary cases of paternalism, it is neither the metaphysical literal truth of things nor the prudential nonliteral truth of things that the paternalizee owes his very existence to the paternalizer’s act. By contrast, if Ben*’s psychological life would have been prudentially-not Ben’s, then it is the prudential nonliteral truth of things (even if not the metaphysical literal truth of things) that Ben owes his very existence to Amy’s act. Thus, I think we ought not say about Already Pregnant anything like what we ought to say about ordinary cases of paternalism if Ben*’s psychological life would have been prudentially-not Ben’s.

Now for the fourth and final premise of my argument:

I take (15) to be extremely plausible. It is extremely hard to see how (2) could plausibly be thought to support (1) if Ben has no justified bipolar complaint against Amy that is grounded in (2)’s truth (assuming [2] is true).

I shall now show that the four premises of my argument jointly yield my desired conclusion. (8) and (13) jointly yield the following subconclusion:

(13.5) and (14) jointly yield the following subconclusion:

And (14.5) and (15) jointly yield the following conclusion:

But if (16) is true, then Parfitianism supports the conclusion that (2), even if true, does not support (1).

Now let the term “quasinonidentity problem” refer to the following question: Does a person, A, who performs an act, X, that adversely affects the well-being of a future person, B, thereby do anything wrong if (i) B lives a life worth living and (ii) the psychological life that B would have had if A had not performed X would have been prudentially-not B’s? If the argument of this section is successful, then Parfitianism supports the conclusion that Already Pregnant raises the quasinonidentity problem (if not the nonidentity problem). Furthermore, because Parfitianism supports the conclusion that the quasinonidentity problem (if not the nonidentity problem) can arise postconception and the quasinonidentity problem is very similar to the nonidentity problem, Parfitianism supports the conclusion that a problem very similar to the nonidentity problem (if not the nonidentity problem) can arise postconception.

Some might think that the argument of this section ought to be of interest only to philosophers who endorse Parfitianism. But many philosophers have endorsed the view that a person who lives a life worth living lacks prudential reason to wish that she had had a life very different from her actual one even if she would, in that case, have had a better life.Footnote ²² And one can endorse this view without accepting Parfitianism.

Consider, for example, the following passage from a classic essay by Robert Adams:

Presumably a person sympathetic to the view that Adams defends here will have great sympathy for the view that people who live lives worth living cannot have good prudential reason to wish that they had had psychological lives massively different from the start from their actual ones. And it seems that a person can be sympathetic to Adams’s view without being a thoroughgoing Parfitian.

Views similar to Adams’s also appear in some philosophers’ writings on Lucretius’ Puzzle, i.e., the question whether we have good reason to prefer later death to earlier death but not to prefer earlier creation to later creation. Some philosophers have argued, by way of partial response to Lucretius’ Puzzle, that we lack prudential reason to wish that we had come into existence substantially earlier than we actually did, even if our having come into existence then would have resulted in our having lives better than our actual ones, because in that case we would (very probably) have had lives very different from our actual ones.Footnote ²⁴ Again, presumably a proponent of a view of this sort will be sympathetic to the view that people who live lives worth living lack prudential reason to wish that they had had psychological lives better than but massively different from the start from their actual ones. And it seems that a proponent of a view of this sort need not be a thoroughgoing Parfitian.

In general, I do not think that one must unqualifiedly hold that Relation R is what matters prudentially in order to be attracted to the view that a person can have decisive prudential reason to regard certain lives that he could have lived as though they would not have been his own. Furthermore, it is natural to think that if this view is correct, then, even if Ben* would have been Ben, the life that Ben would have lived if Amy had not used the drugs would have been a life of this sort. If so, then, even if Parfitianism is not exactly correct, it seems that Already Pregnant raises the quasinonidentity problem; more generally, that a problem very similar to the nonidentity problem can arise postconception; and that (2), even if true, does not support (1).

One final point. Parfit’s main defense of his view of prudence is in part III of Reasons and Persons, and the best-known presentation of the nonidentity problem occurs in part IV of that book. The argument of this section suggests that part III is in tension with part IV, for the very idea that there is a nonidentity problem of the sort displayed by part IV seems to presuppose that identity has an ethical significance that part III seems to imply it does not. To be sure, it is plausible that the vast majority of nonidentity cases are also cases in which the psychological lives of the creatable individuals are massively different from one another from their starts. Consequently, one who affirms the conclusions of part III can say that a special moral problem arises in the vast majority of nonidentity cases. However, Parfitianism recommends the conclusion that when this special moral problem arises, this is not because the creatable individuals are nonidentical but, rather, because the psychological lives of the creatable individuals are sufficiently different from one another. Furthermore, there are arguably metaphysically possible nonidentity cases that Parfitianism suggests do not raise any special moral problem, e.g., nonidentity cases in which the psychological lives of the creatable individuals are just like one another from the start. At any rate, part III seems to imply that Parfit ought not have spoken in part IV of the “nonidentity problem” except as a bit of loose talk.