Three Problems with the Impairment Argument

Abstract

In his recent article “Even if the fetus is not a person, abortion is immoral: The Impairment Argument,” Perry Hendricks sets out to sidestep thorny metaphysical questions regarding human fetuses and present a new argument against abortion – if impairing a fetus with fetal alcohol syndrome is immoral, then killing the fetus is immoral. Hendricks takes inspiration from Judith Jarvis Thomson’s defense of abortion – that even if fetuses are persons with a right to life, the right to life is not the right to use others, so it is acceptable to induce abortion. Together with Bruce Blackshaw, Hendricks set out to strengthen the impairment argument by appealing to Don Marquis’s future like ours (FLO) account of the wrongness of killing. Here I argue the impairment argument falls short in three ways. First, Hendricks and Blackshaw fail to assume fetuses aren’t persons, broadly construed. Second, they fail to show that impairing a fetus is immoral. Third, they overlook abortions that (merely) let the fetus die. Finally, I argue Thomson’s defense of abortion preempts the significance of the impairment argument; Thomson seems to show that even if killing a fetus is prima facie immoral, women still have the right to induce abortion.

Introduction

With his impairment argument, Perry Hendricks attempts to show that (most) induced abortion is immoral (Hendricks 2019a, b; Blackshaw and Hendricks 2020). Hendricks takes inspiration from Judith Jarvis Thomson (1972) in an attempt to bypass thorny metaphysical questions regarding the identity and moral status of human fetuses. Thomson assumes what anti-abortion theorists argue—that fetuses are persons, broadly construed, from conception—and sets out to show that women still have a right to induce abortion. Hendricks purports to assume fetuses are not persons and sets out to show that induced abortion would (usually) be immoral.

The Impairment Argument

• 0.Assumption: a human fetus is not a person.
• 1.It is immoral to impair a human fetus by giving it fetal alcohol syndrome (FAS).
• 2.If it is immoral to impair something to the n^th degree, it is immoral to impair it to the n + 1^th degree.
• 3.Being killed is a greater impairment than being given FAS.
• 4.Induced abortion (usually) kills the fetus.
• 5.Therefore, induced abortion is (usually) immoral.not

Here I discuss three substantive problems for the impairment argument. First, I argue Hendricks fails to assume that fetuses are not persons, broadly construed, and thus fails to bypass the difficult metaphysical questions at the heart of the abortion debate. Hendricks’ (2019a, b) first account of the wrongness of FAS seems to assume the fetus is numerically identical to the person it would become if it is allowed to develop, and the revised argument of Bruce Blackshaw and Hendricks (2020; Blackshaw 2021) appeals to Don Marquis’s (1989) “future like ours” (FLO) view, and Marquis’s view also assumes just this.1

Second, I argue Hendricks fails to show that impairing a fetus by giving it FAS is immoral. Blackshaw and Hendricks (2020; Hendricks 2019a, b; Blackshaw 2021) attempt to explain the wrongness of giving a fetus FAS purely in terms of the harm it causes to a future person, but either the fetus is numerically identical to that future person, or it is not. If it is, then Hendricks has failed to assume the fetus is not a person, broadly construed. If it is not, then the argument runs afoul of Derek Parfit's (1976, 1984) non-identity problem; as risking FAS does not harm any existing person and might change which person comes into existence. Furthermore, if the non-person fetus is aborted, or FAS cured before (or after) the person who would be disadvantaged comes into existence, then it is not clear risking FAS is immoral.

Third, the impairment argument fails to draw a distinction between killing abortions and those that (merely) let the fetus die by disconnecting it. Surgical disconnect abortions are substantively more invasive and medically risky than killing abortion, and thus killing abortions are the medical default. But even if the impairment argument were to succeed in showing that killing abortions are (usually) immoral, it would have failed to show that inducing abortion by disconnect is immoral.

Now suppose that disconnect abortions “count” as killing, rather than letting die; even so, the impairment argument falls short. Thomson (1972) argues the right to life is a (negative) right not to be unjustly killed, not a (positive) right to be kept alive. Similarly, Jane English (1975) argues that one has the right to kill in self-defense and David Boonin (2002) argues that one has the right to withdraw ongoing assistance. Thus, the impairment argument fails to show that killing abortions are instances of unjust killing.

Personhood, Broadly Construed

Hendricks (2019a) contends:

Much of the debate on the ethics of abortion has centered around the notion of personhood. This is because many philosophers hold that the morality of abortion is contingent on whether or not the fetus is a person.

However, different philosophers mean different things by the term “person.” In “Why Abortion is Immoral,” Marquis (1989) discusses three views regarding the moral status of the fetus—the biological view, the psychological view, and the FLO view. Many critics of abortion embrace the biological view, according to which a fetus is a person merely if it is biologically human. Many defenders of abortion embrace a psychological view, according to which a fetus is a person only if it possesses a robust set of psychological characteristics. Finally, Marquis’s FLO view proposes a happy medium; what gives us moral status is our capacity to (psychologically) value a possible future, so if a fetus is numerically identical to the person it may become (because it is the same human organism), it has that person’s future, and thus moral status. The FLO view is comparable to another view, the substance view (Beckwith 2007; George and Tollefsen 2008), which contends that fetuses, from conception, possess the capacity for rationality (psychological personhood) and are thus as much persons as normal sleeping adults.

In “A Defense of Abortion,” Thomson (1972) sets out to sidestep metaphysical questions regarding numerical identity and moral questions regarding the moral status. She begins by assuming what critics of abortion argue—that fetuses are persons from the moment of conception, where “person” is meant to be broadly construed as a rightsholder, specifically one with the right to life. (She then goes on to argue the right to life is not the right to use others to survive without their permission; thus, even if fetuses have a right to life, gestational mothers may induce abortion.)

Although Hendricks (2019a) purports to sidestep the same metaphysical and moral questions by assuming fetuses are not persons, upon closer inspection, he fails to assume that fetuses are not persons, broadly construed. In their defense of premise 1, Hendricks (2019a, b) and Blackshaw (2021; Blackshaw and Hendricks 2020) assume the fetus is numerically identical to the person it may develop into.

Hendricks (2019a) seeks to illustrate the wrongness of impairing a fetus with FAS by appealing to the case of Angela, explaining that because of her mother’s drinking while pregnant, Angela developed a variety of psychological and physical conditions that significantly diminished her quality of life, including an inability to count or tell time in the seventh grade. Here Hendricks assumes the fetus is numerically identical to whatever person it develops into. Blackshaw and Hendricks (2020; Blackshaw 2021) explicitly embrace this view by appealing to Marquis’s FLO view to explain the wrongness of impairment; for Marquis, what gives a fetus moral status is a possible future “like ours”—a possible future it can come to value; as such fetuses can only have a future like ours if it is numerically identical to a person like us.

Hendricks can be said to assume that fetuses are neither psychological persons nor rational human substances, but fails to assume that fetuses are not persons, broadly construed to be rightsholders (Marquis’s (1989, 2007, 2013) FLO view, of course, makes no such assumption; he argues that grounding moral status in biological or psychological personhood is inherently arbitrary; but he still concludes that fetuses are human organisms and rightholders).

Blackshaw and Hendricks (2020; Hendricks 2019a, b; Blackshaw 2021) treat premise 0, the assumption that fetuses are not persons, as independent to the argument, but critics (Crummett 2021; Räsänen 2020; Simkulet 2021a; Gillham 2021) note that the novelty of the impairment argument stands or falls on this assumption; if the impairment argument relies on the hidden premise that a fetus is a person, or rightsholder, from conception, then it begs the question; assuming what anti-abortion theorists seek to argue. But if a fetus is not a person, broadly construed, then Hendricks needs an alternative explanation for why impairing the fetus with FAS is immoral.

Is Giving a Fetus FAS Immoral?

Hendricks (2019a) contends that when a pregnant woman risks impairing her fetus with FAS, she acts immorally, but confesses “I know of no argument to convince someone of this position, and regard it as fairly self‐evident.”

Let us briefly consider two arguments against risking FAS—one that turns on the assumption that fetuses are numerically identical to the person they will become, and one that seeks to avoid the non-identity problem:

The Harm Argument

1. Assumption: fetuses are numerically identical to whatever person they might develop into later in life.

2. Giving something FAS harms it.

3. It is immoral to harm a person.

4. The risk principle: if it is immoral to do x, it is immoral to risk doing x.

5. Drinking alcohol while pregnant risks FAS.

6. Therefore, drinking alcohol while pregnant is immoral.

Hendricks (2019a, b) seems committed to the truth of premise 2; contending that giving a fetus FAS harms it because it causes numerous impairments in the future, as a born person. Alternatively, Blackshaw and Hendricks (2020; Blackshaw 2021) propose that giving a fetus FAS harms it because it deprives the fetus of a future it might otherwise value in which it is not impaired. In either case, this argument turns on an identity-dependent account of harm, where an action is harmful if and only if it harms a specific person.

However, it is far from obvious that a fetus is numerically identical to any person that might develop from it. Even if we assume a fetus is a person, broadly construed, at conception, if it undergoes twinning (when one mass of cell divides into two or more distinct masses of cells) or chimerism (when two masses of cells merge into a single mass of cells), it ceases to exist and is replaced by one or more numerically distinct fetuses. Suppose we give a fetus FAS before it twins or chimeras; because that fetus ceases to exist, we have not harmed any specific person.

Many philosophers believe that we are inherently thinking things, or embodied minds (Descartes 2008; McMahan 2002, 2007; Parfit 2012). One virtue of the embodied mind view is that it is prima facie more consistent with our intuitions in cases of conjoined twins, brain death, brain transplant, and many theories of the afterlife than the view that we are inherently human organisms. If impairing a fetus with FAS prevents it from developing a mind, then it harms no person. Some cases of FAS result in miscarriage, perhaps before the fetus forms a mind, but Hendricks restricts his discussion to non-lethal FAS, which he contends impairs the person that develops from the fetus. But this is far from obvious; suppose that giving a fetus FAS changes the identity of the mind that comes to inhabit the fetus’s body. If this were the case, then giving a fetus FAS would be no different than conceiving with a different pair of gametes, changing who exists, but not impairing anyone. For example, consider the case of Angela; suppose her mother had chosen not to drink while pregnant, the person that develops from that fetus might be a completely different, numerically distinct person, Alice, whom while in seventh grade can read and write but perhaps is far more selfish and cold than Angela. If FAS changes which person comes into existence, then it does not make sense to say that Angela’s mother causes identity-dependent harm to Angela, as had she refrained form drinking, Angela would never have existed! This is an example of Parfit’s non-identity problem (Parfit 1976, 1984).

The problem with the harm argument is that it cannot explain why those who reject premise 1’s assumption would oppose risking FAS. To do this, we need an identity-independent account of harm. Consequentialist theories like Utilitarianism can easily provide such an account, but consider the following (non-consequentialist) argument:

The Parental Environment Argument

1. Assumption:fetuses are not persons, broadly construed.

2. Prospective parents take on a special moral obligation to provide an environment in which their future children would thrive.

3. Giving a fetus FAS prevents any future child from thriving.

4. The risk principle: if it is immoral to do x, it is immoral to risk doing x.

5. Drinking alcohol while pregnant risks FAS.

6. Therefore, drinking alcohol while pregnant is immoral.

This turns on what I (Simkulet 2021b) call an identity-independent account of harm in an attempt to avoid the non-identity problem. Consider the practice of prescribing folic acid to women who may become pregnant; this practice, along with other preventative interventions, does not prevent harm to specific existing persons, but furthermore they might actually cause different people to come into existence than otherwise would. Still, it makes sense to say that FAS would cause identity-independent harm, even if risking FAS ends up creating a different person than would exist otherwise. The Parental Environment Argument uses an evaluation of the environment, rather than harm to a specific person, to explain the wrongness of risking FAS. Note, though, that this argument only applies to prospective parents, it is far from obvious that this argument would apply to pregnant women seeking an abortion, although I (Simkulet 2021a) have also argued that as long as a gestational mother seeking abortion might reasonably change her mind, she might be reasonably understood as a prospective parent.

However, I (Simkulet 2021a) argue that risking FAS is not always immoral; consider the following case:

Parasite: Darcy knows that she is pregnant, that consuming alcohol risks causing FAS, and that FAS is bad. However, her physician informs her that her fetus has a parasite that will impair her fetus worse than FAS if nothing is done. The parasite has one weakness—alcohol—and the physician recommends consuming alcohol to kill the parasite. Darcy does so, the parasite is killed, and her child suffers from FAS.

In this case, risking FAS is the lesser of two evils. But now suppose that we can have FAS without preventing a child from thriving. Consider the following case:

NFAS: After discovering a secluded civilization subsisting entirely on alcoholic beverages where no born persons show signs of FAS, scientists discover that their fetuses exhibit signs of FAS, but chemical N found in the local wine negates the harms normally associated with FAS

Carol knows that she is pregnant, knows consuming alcohol risks causing FAS, and knows that without chemical N, FAS can prevent a born person from thriving. However, she also knows that with chemical N, FAS has no ill effects drinks wine with chemical N, causing harmless FAS.

Blackshaw and Hendricks (2020; Hendricks 2019a, b; Blackshaw 2021) argue risking FAS is immoral because it impairs the born person later in life, not because it impairs the fetus. By stipulation, chemical N does not prevent FAS from impairing the fetus as a fetus, but it prevents FAS from harming impairing the born person later in life. Although FAS can be said to impair the fetus, it does not make sense to say that is immoral with chemical N in the environment.

Similarly, without chemical N, if we assume that fetuses are not persons, broadly construed, then the fetus is numerically distinct from the person it will become, so it would not make sense to say that FAS harmfully impairs the fetus; it merely causes identity-independent harm to any person who develops from that fetus. In light of this, the first premise of the impairment argument is false because giving a fetus FAS either does not harm it or does not impair it!

But things only get worse from here for the impairment argument; not only does FAS not impair a fetus, but inducing abortion works very much like chemical N in that it prevents the identity-independent harm of FAS by preventing a person with FAS from coming into existence, as by assumption the fetus is not numerically identical to whatever person would be born if not aborted. On this view, inducing abortion is not relevantly different than destroying a specific egg or sperm cell, or risking FAS, or a comparable condition, to a sperm or egg cell one does not intend to fertilize is prima facie unobjectionable.

Killing and Letting Die

Many philosophers draw a distinction between acting and refraining, with the former usually being understood as more morally impactful than the latter. Many who hold this view contend it is (usually) worse to kill a person than to merely let them die. Many critics of abortion contend that abortion is immoral because it involves killing, but that merely letting others die would less bad… perhaps not even bad at all!2

According to the impairment argument, killing a fetus is a greater impairment than giving it FAS. Some critics question whether killing impairs at all (Blackshaw 2019, 2020; Pickard 2020). Another criticism is that even if killing is prima facie immoral, other factors might make it morally acceptable, all things considered (Blackshaw 2019, 2020; Pickard 2020; Räsänen 2020; Crummett 2021). However, there is another problem—the impairment argument restricts its focus to killing, but while many induced abortions do kill the fetus, it is possible to induce abortion without killing the fetus. A fetus can be disconnected from the mother and left to die, or others can act to try to save its life with ectogenesis technology, surrogate motherhood, or the like.

At this point, it will be practical to distinguish between three ways in which a pregnancy can be aborted. In some cases, pregnancy can be ended even by preventing the fetus from forming an attachment to the prospective gestational mother, usually by means of taking a medication. Such abortions do not kill the fetus, they merely let it die.

However, once the fetus has formed an attachment to the gestational mother’s womb there are several surgical interventions to induce abortion. The medical default is to induce abortion by killing the fetus in the womb. Alternatively, one might disconnect the fetus from the gestational mother by surgically removing it, or by performing a hysterectomy and taking the fetus with the uterus. These disconnect abortions are (usually) far more invasive and medically risky than killing abortions. Because of this, empirically most surgical-induced abortion is a matter of killing, rather than disconnecting the fetus to it die.

In light of this distinction, it seems that the impairment argument is invalid, or at least misleading. Premise 4 of the argument makes an empirical claim about abortion—most induced abortion kills the fetus, because many contemporary induced abortions are surgical, killing abortions. But inducing abortion by disconnect, whether surgical or medical, does not kill the fetus; they merely let it die.

However, the conclusion of the argument appears to be a claim about the morality of abortion in general; but at best, all the argument shows is that killing abortions are immoral, not all induced abortions. There is no contradiction in accepting premises 0–4 of the impairment argument and choosing to abort in a way that “merely” lets the fetus die. In other words, the impairment argument cannot show that inducing abortion is immoral, at best it can show that how we often currently go about inducing abortion, opting for safer, less invasive killing abortions, is immoral.

Suppose a defender of the impairment argument rejected drawing a distinction between killing and letting die or believed that disconnect abortions are better understood as instances of killing. Even this will not help, as the impairment argument similarly fails to draw a distinction between just and unjust killings. Many defenders of abortion argue that inducing abortion may be an instance of moral acceptable, even just, killing.

Notably, Thomson (1972) argues that even if the fetus is a person from conception with a full right to life, this right to life is a (negative) right not to be unjustly killed, not a (positive) right to be kept alive. Hendricks directly compares the impairment argument to Thomson’s defense of abortion:

[S]ince my argument is predicated on a concrete ethical fact – [that giving a fetus FAS is immoral] – it has the virtue of not forcing the reader to entertain abstract thought experiments (such as Jarvis‐Thomson’s famous violinist) to arrive at its conclusion.

Thomson, like many philosophers, uses abstract thought experiments to illustrate relevant philosophical issues, but while some thought experiments, like that of the violinist, expanding baby, and people seeds can be abstract, others are fairly “concrete”; for example, Thomson (1972) says:

But it cannot seriously be thought to be murder if the mother performs an abortion on herself to save her life.

Self-defense cases like this are far less abstract than, say, Jane English’s (1975) case of innocent, but hypnotized attackers, but both demonstrate that the right to life is not the right to never be killed, but merely the right not to be killed unjustly.

Similarly, David Boonin (2002) asks us to imagine a case in which a bone marrow donor begins life-saving donations but chooses to withdraw ongoing assistance. This case is not abstract, but as few of us donate bone marrow, it is unfamiliar; though it is easy enough to imagine a case in which a parent “cuts off” their adult child, forcing them to move out and get a job. Even if a parent consented to help their adult child “get back on their feet,” certainly they have the right to withdraw aid.

Blackshaw and Hendricks (2020) recognize that the impairment argument is vulnerable to such objections, and seek to strengthen it by appealing to Marquis’s FLO view; the problem is that Marquis’s view is equally vulnerable to these objections; Marquis (1989) assumes that morality of abortion stands or falls on the moral status of the fetus, but this is exactly what Thomson, English, and Boonin argue against; even if a fetus is a person, broadly construed from conception with a right to life, that right is not the right to use others without their permission. We can illustrate this by way of abstract and/or unfamiliar thought experiments or real, familiar cases of self-defense and withdrawal of aid.

Although Thomson (1972) argues that women have a right to induce abortion, she does not argue that induced abortion is morally acceptable in all cases. She says:

It would be indecent in the woman to request an abortion, and indecent in a doctor to perform it, if she is in her seventh month, and wants the abortion just to avoid the nuisance of postponing a trip abroad.

This should not be surprising; many defenders of abortion believe some abortions are immoral—often those done in the later stages of pregnancy, once the fetus develops somewhat robust psychological capacities and/or has become viable. Few defenders of abortion defend late, frivolous abortions like the one Thomson discusses here.

Note, however, that writers on both sides of the abortion debate often agree that it is morally acceptable to induce abortion in cases of rape and medical risk. Indeed, many people who believe that abortion is usually immoral reject anti-abortion restrictivism, the view that we ought to legally restrict access to abortion, because restrictivist laws fail to allow these commonsense exceptions. Indeed, because both exceptions turn on deeply private facts—about how the fetus was conceived and nuanced estimates of medical risk that medical professionals might sensibly disagree about, it is unclear how a restrictivist social policy could make these commonsense, life-saving exceptions.

Conclusion

Hendricks’ impairment argument purports to show that most induced abortion is immoral. Here I have argued the argument fails for three reasons. First, it begs the question by assuming the fetus is a person, or rightsholder. Second, it fails to show giving a fetus FAS is immoral. Third, it fails to distinguish between either killing and letting die abortions or unjust and just killings; even if the former is immoral, it is far from obvious the latter are.

Author Contribution

I am the sole author of this work.

Data availability

Not applicable.

Declarations

Ethics Approval and Consent to Participate

Not Applicable.

Consent for Publication

Consent given.

Competing Interests

The author declares no competing interests.