Examined Life

In April 2019, Korea’s Constitutional Court ruled that the provisions of the Criminal Code and the Mother and Child Health Act — which criminalized abortion without regard to gestational age — unconstitutionally infringed upon pregnant women’s rights to self-determination. At the time, I expressed my support for the decision, framing the central issue as “a question of when the fetus acquires the status of a living human being.” As a socio-political matter, I still stand by this liberal commitment. But I have lately grown seriously troubled by whether the reasoning deployed in defense of this conclusion is conceptually coherent, and whether it is philosophically defensible.

Let us begin by examining part of the Court’s argument. In holding that a blanket criminalization of abortion regardless of gestational stage was wrong, the Court invoked the concept of “fetal viability” — the capacity of a fetus to survive independently outside the womb:

“After a certain point in development, a fetus becomes capable of surviving independently outside the mother’s body. While advances in medical technology may shift this threshold, the World Health Organization places it at 22 weeks of gestation, and the obstetric community likewise holds that, with optimal medical care, independent survival becomes possible at around 22 weeks. A fetus that has reached the point of independent viability may be considered to have attained a state far closer to that of a human being than one that has not.”¹

On this basis, the Court concluded that abortion, depending on whether it occurred before or after 22 weeks, could receive differential criminal treatment. But the argument has a conspicuous flaw.

In saying that a viable fetus has attained a state “far closer to that of a human being,” the Court quietly smuggles in the concept of being human — a concept that appears nowhere else in the reasoning. It cannot mean biological humanity, since a fetus is biologically human at any stage. What the Court must have had in mind is what philosophers call personhood: the qualification to be a “subject of moral consideration,” or, in the words philosopher Mary Anne Warren uses in her essay “On the Moral and Legal Status of Abortion,”² a “member of our moral community.” This is a normative concept, not a biological one. Just as a legal person is something capable of bearing legal rights and duties, a person in the ethical sense is something capable of being a bearer of moral rights.

The problem is that the Court defined the criterion for personhood as the capacity for independent survival. This is a plainly absurd criterion. Not only does the Court offer no argument for why viability should ground personhood, but the standard flatly contradicts our ordinary moral intuitions. In contemporary societies, those who have lost the capacity for independent survival — the severely disabled, the gravely ill, the dependent elderly — are not excluded from moral consideration. If anything, they are accorded greater care and protection.

Warren herself proposes five criteria for personhood, which seem much more promising: consciousness, reasoning, self-motivated activity, the capacity to communicate, and self-conception and self-awareness. She argues that personhood is constituted on all or some of these criteria, sort of in a family resemblance fashion. Warren is vague on exactly where personhood begins, but argues that that’s irrelevant to the question at hand, since fetuses satisfy none of the five criteria anyway. And if a fetus is not a person, it has no moral rights and cannot be a subject of moral consideration — from which follows the conclusion that abortion is morally permissible at any stage of pregnancy.

The Court, it seems to me, was drawn to this kind of personhood- or rights- based approach because it wanted to permit early abortions across the board without the need for justification. The most famous pro-choice argument in the philosophical literature is Judith Jarvis Thomson’s violinist thought experiment,³⁴⁵ but that argument derives its force specifically from cases where pregnancy was not consensual. A personhood-based argument, by contrast, denies that the fetus has a right to life in the first place. Therefore the reasons for the pregnancy, or the motivations for the abortion, become entirely irrelevant to our moral judgment.

Warren’s personhood-based argument, however, seems to run into serious difficulties of its own. The most troublesome is the problem of infanticide. A late-term fetus and a newborn infant both fail to satisfy Warren’s five criteria for personhood. It would then follow that infanticide, too, is morally permissible, which seems absurd. Warren addresses this in her paper, arguing that infanticide is nonetheless impermissible — unlike abortion — because an infant, unlike a fetus, poses no threat to the mother’s body or life, and because there are alternatives to killing, such as giving the child up for adoption. But this response seems at best shoddy. Under Warren’s framework, we can break down the question of whether abortion is morally permissible into two steps: (1) Is the fetus/infant a subject of moral consideration? And (2) (assuming we grant that it is) is abortion/infanticide morally permissible? If we do not grant that the fetus/infant is a subject of moral consideration, there is no need to discuss (2) at all. That is precisely Warren’s point about abortion. Yet when it comes to infants, Warren denies personhood just as she does for fetuses, but then proceeds to argue against infanticide on what seems to amount to a form of utilitarianism, addressing a question — (2) — she had already rendered moot. For the personhood-based approach to hold together, one must either revise the criteria for personhood so that fetuses or infants beyond a certain stage qualify as persons, or else accept the monstrous conclusion that infanticide is permissible. The latter is exactly the position Peter Singer, although he gets there through a different — utilitarian — reasoning. That’s a big fat bullet to bite, in my opinion.

The second problem is what prompted me to write this piece in the first place. The proposition that “the motivation or timing of an abortion does not affect its moral evaluation” seems to me to violate our ordinary moral intuitions. Recently, I came across a story online: a woman posted online about her decision to have an abortion because her husband, who desperately wanted a daughter, had pressured her upon learning the fetus was male. The comments section, predictably, erupted in condemnation of the couple.

Until I encountered this story, I had been going along, without much thought, with the Constitutional Court’s logic that a fetus below a certain gestational age lacks personhood. But the story gave me pause. If the fetus is indeed not a subject of moral consideration before a certain point (or at any point, according to Warren’s view), then there seems to be no moral basis for condemning the couple at all. And yet I found myself feeling a strong moral repulsion at their sex-selective abortion. I was left with two possibilities: either my moral intuition is simply mistaken, a kind of moral optical illusion; or the personhood- or rights-based approach has a serious defect that prevents it from capturing what our moral intuitions are actually tracking.

One possible resolution is to shift focus from the act itself to the agent. A virtue ethics approach, for instance, would hold that choosing to abort for a trivial reason reveals a certain vice in the agent, regardless of whether the act itself is right or wrong. to my knowledge, philosophers like Rosalind Hursthouse, in her essay “Virtue Theory and Abortion,” have taken this kind of approach (though I have not yet read it). I have two reservations. First, it is not clear that the virtue ethics framework can avoid the prior question of whether the fetus is a subject of moral consideration — since whether the act reveals a vice may depend on that answer. Second, granted that the act of abortion is itself morally unproblematic, the virtue theorist may still argue that deciding to abort for a trivial reason nonetheless reveals a kind of vice (rashness, perhaps?). But then it is hard to see how something like rashness constitutes a vice serious enough to induce the kind of deep moral repulsion most people feel in cases like this. And beyond that, I am not sure whether it is appropriate to import agent-centered ethics into a domain where it is practically impossible to avoid judgments about the moral and legal permissibility of the act itself. These are questions I intend to sit with as I read further.

One final thought. The question of AI personhood strikes me as one of the most pressing moral problems we now face, given the pace of AI development. Whether it is permissible to work a domestic AI robot without limit, or to address an AI with quasi-racial slurs (such as “Clanker”), no longer feel like questions for the distant future. And once we start taking these questions seriously, we find ourselves drawn back into the older debates — about abortion, and about the moral status of animals.

1. Constitutional Court Decision 2017heonba127, rendered on April 11, 2019. Emphasis added by yours truly ↩︎
2. Warren, M. A. (2017). On the moral and legal status of abortion. In Applied Ethics (pp. 360-367). Routledge. ↩︎
3. Thomson, J. J. (1971). A defense of abortion. Philosophy & Public Affairs, 1(1), 47–66. ↩︎
4. One morning you wake up in a hospital bed, connected to an unconscious man lying beside you. He is a famous violinist suffering from a fatal kidney disease. Yours is the only blood type that can save him. The Society of Music Lovers has kidnapped you and plugged your circulatory system into his. If you disconnect yourself, he will die. But if you remain connected for nine months, he will recover and be able to live on his own. Do you have a moral obligation to remain bedridden for nine months to save this violinist? ↩︎
5. As a side note: under the Court’s viability-based reasoning, the violinist would not qualify as a person, since he cannot survive independently, and so you could be morally justified in disconnecting at any time. This is the same conclusion that Thompson draws, but the reasoning seems, to me, absurd, as the violinist clearly is a person. ↩︎