Non-Standard Doing of Harm Without a Justificatory Bypass – Phil Reed and David Hershenov 

Kingma and Woollard insist that some action is a standard doing of harm when it has the normative profile similar to Push. However, in Push, the victim is independent of the agent imposing harm. This means that, trivially, neither in pregnancy cases nor in any so-called barrier contamination does the agent do harm in the exact way that the agent does in Push. In pregnancy cases and in Plate, the victim is dependent on the agent providing aid. But this is hardly enough to show that the fetal harm is not the doing of a wrongful harm nor that a so-called justificatory bypass is in effect. The harm that befalls the fetus might still be a wrongful harming such that the pregnant women still deserves blame and censor (and sometimes coercion), but less blame and censor than when harm is done to the independent. If this is right, those concerned with gestational harm are not guilty of treating such harm as a “standard doing of harm” nor engaging in a justificatory bypass. 

Kingma and Woollard claim that the justificatory bypass is in effect when people frame the pregnant woman’s behavior as doing harm. They intend to provide evidence of this framing, from public health advice and philosophers, in the first section of their article. We believe, however, that this evidence is ambiguous: merely pointing out that some action “threatens” another (which could mean merely exposing him to risk) or even “harms” another could mean either doing or allowing harm. Consider that it seems perfectly natural to say that removing your barrier-car “harms” the victim, even though this is a case of allowing. It’s likewise perfectly natural (including in one of the pieces of evidence they provide) to say that not taking prenatal vitamins or folic acid threatens or harms the fetus, even though this too, as an inaction case, is strictly allowing harm. None of the authors in the examples that Kingma and Woollard provide were trying to say anything about the relevant actions as a doing rather than an allowing, so it is not obvious that they were taking the so-called justificatory bypass. Consider that some parents’ choice to skip or delay vaccinations for their children is commonly framed in terms of harm; do Kingma and Woollard think that such framing invokes a widely recognized injunction against “doing harm” as opposed to allowing it and so this framing commits a justificatory bypass?

Kingma and Woollard cite Wilkinson et al. as evidence of a justificatory bypass but they neglect to mention that Wilkinson et al. are explicitly countering what they take to be a prevalent view, which is roughly that the law does not typically treat pregnant women’s behavior as a standard doing of harm. “In most jurisdictions, the fetus in-utero has no legal right to protection,” Wilkinson et al. write. “While the actions of pregnant women can cause harm to their future children, even if the possible harm is serious and likely to occur, the law will generally not intervene.” If Wilkinson et al. are right, the consensus legal view is that there is no justificatory bypass: bodily-mediated fetal harms are often ignored, excused, or lightly sanctioned.

Suppose, however, that Wilkinson et al. overstate the consensus legal view. After all, there are plenty of cases where pregnant women are on the hook for behavior that causes harm to their fetuses. Nevertheless, it still seems clear to us that bodily-mediated fetal harms are often ignored, excused, or lightly sanctioned compared to similar harms to the independent by third parties. Consider, for example, that historically, even in jurisdictions where abortion was illegal, only a “small number” of women who procured abortion were subsequently fined or imprisoned. Andrea Rowan, “Prosecuting Women for Self-Inducing Abortion: Counterproductive and Lacking Compassion,” Guttmacher Policy Review, September 25, 2015, https://www.guttmacher.oInfanticide, meanwhile, is a felony. To see this point in terms of public opinion, consider that more than half of people regardless of their views on abortion believe that women who procure illegal abortion should receive either no punishment, therapy/education, or a fine. Crawford et al., “An Exploratory Examination of Attitudes toward Illegal Abortion in the U.S. through Endorsement of Various Punishments.” We hypothesize that there is exactly zero chance that more than half of these same respondents would claim that infanticide should be met with no punishment, therapy/education, or a fine. In short, attitudes and sanctions toward pregnant women’s behavior that harms the fetus are not symmetrical to third party harming of the independent, so no “standard constraint” against doing harm is being invoked in pregnancy cases.

The typical advocate against abortion doesn’t use violence to prevent chemical abortions (assume the embryo can survive the pregnant woman’s death), despite supporting an abortion ban. Various, “Killing Abortionists: A Symposium,” First Things, December 1, 1994, https://firstthings.com/killing-abortionists-a-symposium/.They don’t take the same pacifist attitude to women committing infanticide. The best explanation is not a difference in moral status of the late-term fetus and a premature baby but that abortion is not as wrong as killing the non-dependent because the aborted were being kept alive by the woman who then aborted. Critics of abortion might implicitly be accepting Frances Kamm’s idea that abortion inherits some of the moral valence of letting die since it is the rare killing of someone who couldn’t continue to survive without their assistance. Kamm, Creation and Abortion, 30. The latter is a definitional feature of “letting die.” But anti-abortion advocates still believe abortion is bad enough for denunciation and a ban. This suggests a parallel for cases of harm to the future child (e.g., fetal alcohol syndrome) that doesn’t fall under the standard constraint of doing harm to the independent; in these cases, pregnant women still deserve condemnation, though perhaps not coercion. Such harms are not as bad as harming the independent, but still wrong, in part, because the pregnant woman was responsible for there being a choice between suffering the burdens of unwanted pregnancy or disabling her fetus.

That maternal harms to the fetus are not, all else being equal, as wrong as harms to independent children opens the possibility that Kingma and Woollard are mistaken to believe that an “incorrect” or “misleading” (310) judgment of doing harm and the justificatory bypass is typically in effect. Rather, there is the possibility that the difference in the facts between harming an individual to whom one was providing bodily support and harming an independent individual is recognized in the “widespread expectations.” That implicit understanding would explain the discrepancy in the responses to a woman drinking and causing fetal alcohol syndrome in her own child versus driving drunk and causing comparable cognitive defects in an unrelated child. There is much greater condemnation of the latter. Foster parents, daycare providers, and the like would surely lose custody, licenses to care for children, suffer stiffer fines, etc., if they repeatedly initiate toxins in their young charges, while pregnant women who do the same are treated less severely. We suspect that Kingma and Woollard might just be disagreeing with the severity of the widespread expectations placed upon pregnant woman rather than diagnosing a failure of others to implicitly distinguish between standard harms and barrier-mediated harms and engaging in a justificatory bypass.

Kingma and Woollard assume that when social judgments are made about fetal harm, there is no reference to special obligations or other substantive questions like the mother’s responsibilities. We propose that these factors are often implicit in such judgments, evidenced by the fact that fetal harm and childhood harm are not treated symmetrically. The overlooked maternal responsibility (as well, perhaps, as special obligations) is compatible with having more wrong-making features, making fetus-affecting maternal behavior have a moral valence that is closer to doing harm. We think the onus is on Kingma and Woollard to show not just that some pregnant women are condemned and even coerced for harming those they are gestating, but that they are treated as harshly as people who commit the same harms to the born whom they are not providing gestational support. Without providing this evidence, the central target of their argument falls apart.

Kurt B. September 18, 2025 at 9:04 AM

One main concern in the article was about the relationship between weakening a barrier and allowing harm. K&W argue that when someone who is pregnant drinks alcohol, smokes cigarettes, or eats sushi, for example, they do not do harm to the fetus because they are only modifying their body and so contaminating a barrier to harm (p. 304). The claim, if I understand it correctly, is that contaminating a barrier to harm is not wrong because "outside of pregnancy there are few no artificial cases where an agent is simultaneously providing a barrier to harm and facing a cost if they do not contaminant it" (p. 308). But consider blood donation. There is a widespread deferral category for possible blood donors who are on antibiotics because if a recipient who is allergic to antibiotics gets the blood with antibiotics in it, they will be worse off. Bracket the legal or policy reasons for the deferral, it seems like the donor on antibiotics is doing something morally wrong by harming the recipient. Now, K&W argue that consent doesn't happen in gestation and so maybe this lack of consent makes blood donation disanalogous because the recipient has to consent to the blood. Granting that some people do consent to a transfusion, many do not (e.g., accidents, emergencies) and so if the pregnant body is a barrier, it's seems like the blood donor's body is also a barrier. But if it's wrong for the donor to contaminate their blood (barrier), then it would also be wrong for the person who is pregnant to contaminate their body.

Another concern I had was about the claim that pregnancy is morally unique and singular case and so unlike other cases, hence our intuitions about other cases cannot map onto pregnancy (p. 314). But if this argument is true, then, strictly speaking, the entire position cannot get off the ground because the motivating cases in the beginning (Push, Block, Plate) cannot tell us anything relevant about pregnancy because, by hypothesis, pregnancy is unlike them as a singular case. So I wasn't sure how to unify the initial cultivation of moral intuitions with the later conclusions about pregnancy.

Last, K&W fruitfully shift the conversation away from a particular action and more to an emphasis on patterns of behavior (pp. 317-18). Focusing on patterns of behaviors helps clarify that an occasional alcoholic drink or higher consumption of caffeine might not be bad and so there is less reason to police or monitor what pregnant people consume. But the pattern of behavior points to an accumulative effect of each individual instance. While I think this framing makes sense of the general and common concerns about consumption, it does seem to have an unintuitive result for more concentrated concerns (especially given their earlier comments about contaminated barriers). Thalidomide, for example, was a drug prescribed to treat, inter alia, morning sickness. Thalidomide ended up causing significant birth defects and even a singular dose of 50 mg can cause the birth defects

(https://www.accessdata.fda.gov/drugsatfda_docs/label/2001/20785s12s14lbl.pdf)Download pdf.

If a bioethicist at the FDA suggested that prescribing Thalidomide was permissible during pregnancy because it merely contaminated a barrier to harm by offsetting morning sickness, and was not part of a more robust pattern of behavior, they would be laughed out of the room. So, while I'm on board with shifting the conversation about behavioral or lifestyle risks during pregnancy towards patterns of behavior, rather than single one-off instances of typical substances, I think the prior theoretical commitment to the contaminated barrier thesis interferes with a plausible verdict about dangerous single exposure behaviors (viz. Thalidomide). But, giving up the contaminated barrier thesis undermines much of what K&W say about doing and allowing.

 David H September 20, 2025 at 9:15 AM

Kurt.
Persuasive points. The thalidomide case was a great counter example. I don't think the K & W point about behavior is significant. It sometimes takes many individual actions in cooperation to causally produce an outcome. Say many people must lift poison and pour it into the county's water supply in order for anyone to be poisoned. We just say their many actions together did the harm. It doesn't even require us to defend there is a collective agent. I don't see why repeat behavior is metaphysically or morally different from many individual acts doing harm. Likewise, many individuals can each omit to save someone who is drowning in front of them. Or a group of individuals in the hospital may be needed to withdraw the high tech life support aid that they were collectively providing.

I would just add that I don't think pregnancy in the morally relevant ways is so unique that we can't use real and artificial cases to illuminate the morality of what pregnant woman do. Besides transfusions, there is breast feeding in which the mother's body is a barrier to starvation and can deliver a threat from say her drug use. Then parents who are providing support more indirectly through bodily actions like feeding, dressing, housing, etc. (rather than directly via gestating and nursing) to young children can be contagious.

Conjoined twins may also be useful for exploring what one can do to the other whose body is intertwined and such cases have the benefit that their personhood is not in doubt and their long-term entanglement makes them not appear to be trespassers. Incidentally, Kingma thinks fetuses are parts so they can't be trespassers anymore than one's limbs and organs can

There are also cases of rescue where's one body is providing burdensome support to another. I take it that the authors are not going to rule out Thomson style violinist thought experiments that are alleged to show the woman doesn't wrong the fetus by ending bodily aid. They are not contaminated by the culture's attitudes to woman and their bodies but designed to avoid a burden that only woman can provide as Thomson invites the readers, male and female, to imagine they were kidnapped and hooked up to the violinist. I would counter with similar examples of bodily rescue where it seems it is wrong to allow or do harm to avoid the burdens. After my students read about Thomson's violinist, I give them two scenarios, the first makes kidney failure as universal as gestation, everyone of us in the near future will need kidney support. This offsets any biases due to the probability being near zero that one will ever need another's support which might be making it too easy for people to bring about the violinist's death. Students are more hesitate to say they don't have to provide the support when they will need the same support in the future as that is just the human condition, not a rare disease.

The second example suggests a different distortional effect than low risk and identifying with just the kidnapped person providing support and not the needy violinist. This case involves an earthquake that throws every other student on top of the student sitting next to them. The quake has weakened the room's floor and it will soon give way. The student on top of the other student will be protected from the fall by the student's body beneath them. The student on the bottom will hit the floor first and suffer pains comparable to birth and then months of swelling, nausea and other pregnancy-like pains. I ask my students that if they were the student on the bottom, can they permissibly rotate their bodies so the student on top hits the floor first and is killed by the impact of the floor and the crushing force of their body on top. They say no. I then ask can they the rotate away before the floor gives away, letting the other student fall by themselves to their deaths. Many are reluctant to say they can permissibly allow that fatal harm.

 Phil Reed September 21, 2025 at 6:41 AM

Kurt, Good comments. I do think that they are not concerned here to say whether the behavior they discuss is "wrong" or "permissible." What they want to address is whether fetal-affecting maternal behavior is "doing harm." If it's merely allowing harm or closer to that, then the widespread judgment against these behaviors is misplaced.

They probably believe that all things considered, pregnant women shouldn't smoke and drink in excess. It's just that they don't like society imposing that obligation on pregnant women.

Your thalidomide example is particularly challenging though. Maybe they would say that this too is a kind of allowing harm (seeing that it works through the mother's body). But even if that's true, are we really supposed to believe that there should be no societal expectation that pregnant women ought to refrain from thalidomide? That we need an account of special obligations to say something like that?

Your points connect to my concern that their central thesis is irrelevant to the pregnant woman's real life situation (see below).

Kurt B. September 21, 2025 at 3:00 PM

David and Phil, thanks for the kind words about the Thalidomide example. Perhaps, as to Phil's point, K&W would say that it too is an example of allowing harm, but that seems ripe for a modus tollens rather than a modus ponens.

Phil, you're right that I should have been more careful with my phrasing and focusing on doing/allowing harm rather than the other issue of "wrong" or "permissible." Thanks for catching it.

Rephrasing the argument, then, it still seems like taking thalidomide during pregnancy is doing harm in a way that is unlike the drinking or smoking cases. My sense of the pattern of behavior issue was that medical practitioners and everyday folks will be hyper critical of someone who is pregnant having a glass of wine or can of beer during the pregnancy. But it's not immediately clear that a single or occasional consumption is as awful. Emily Oster (2013) suggests that it's not nearly as bad as it has been made out to be and further presses that these risk are often moralized by healthcare specialists. Revisiting Phil's comment, I don't think that K&W were objecting to others imposing the obligation per se, but more so they were pushing back on the hypervigilant and unsolicited advice and criticism about what pregnant people deicide (sometimes the group of intervening do-gooders are called the "pregnancy police," though they are more a self-deputizing posse).

Going back to your concern, David, I think the difference is that someone who drinks sporadically or occasionally during pregnancy is doing something morally different than someone who is drinking every day. The thalidomide case, by contrast, shows that there are other instances where a single act of consumption is sufficient to cause harm. So, I'm somewhat sympathetic about reframing the conversation about consumption during pregnancy to habits or repeated acts rather than single acts. My worry is that the "single acts are just as bad as habits" approach is that it does end up creating the pregnancy police mentality. Pregnant people in other countries have higher rates of alcohol consumption during pregnancy (https://www.sciencedirect.com/science/article/abs/pii/S1871519217300057). Oster interprets that kind of data to show that people who are pregnant can make the judgment call for themselves (which might drift back to Phil's objection about social imposition), but perhaps she is wrong (she has received quite a bit of criticism, both from healthcare professionals and from other health economists).

I hadn't thought about your conjoined twin example David, that's a great point too. I don't know how K&W would approach the conjoined twin case (I haven't read enough of Kingma to know), but I suspect they might have to take the same kind of stance. And, if that's true, then I think you're right that Twin A could say that Twin B is harming them when Twin B ingests something harmful even if, under the K&W framework it is only contaminating the barrier.

Phil Reed September 21, 2025 at 5:50 PM

Kurt, I wonder if they would say that ingesting thalidomide is not a "fetal affecting maternal behavior." See my comments on David's post as to why abortion might also not count as a fetal-affecting maternal behavior.
 

pure (314).

David H September 19, 2025 at 5:22 PM

I think three of the main claims of the paper are False.

First, Kingma and Woollard (henceforth K & W) write “Our main claim is that justificatory bypass is unwarranted: in many cases either the pregnant person does not count as doing harm…” (290)

Second, K & W claim "we argue that barrier -contamination cases are morally different from the standard case of doing harm …the reason is that the agent’s behavior only affect the victims because of the victims’ use of the barrier. If the victim was not using the barrier, then the agent would be able act on the barrier without anyone suffering harm.” (307)

Third, K & W falsely claim that “... and it is the key point for our argument, any such objections surely need to appeal to a substantial notion of the pregnant woman’s special obligations to the fetus.” (309)

David H September 19, 2025 at 5:22 PM

Regarding the first claim that woman is allowing harm rather than doing harm seems just false in the case of surgical abortions and I will suggest it is also false in the case of extraction abortions. Sure, the woman's body was keeping the fetus alive, but if she dismembers it that is killing and the doing of a harm, not allowing the harm of death, withdrawing barrier support . But many abortions today are chemical abortions and they could be construed as withdrawing aid and allowing harm rather than doing harm

Oddly, K & W say very little about abortions, noting McMahan views extractive abortions involve withdrawal of life support and let the fetus die rather than kill. They disagree with Thomson that abortion kills “…Thomson’s famous argument that the maternal relation should be seen as one of ongoing lifesaving rescue, rather than of potential threat there are also important differences. Thomson accepts the characterization of abortion as killing, i.e., as doing fatal harm. The argument in this article challenges the characterization of maternal behavior as doing harm."

Perhaps some barrier withdrawals like smoking involve providing the fetus with less oxygen and thus allow harms that the women's gestating bodies had been preventing "their suboptimal behavior, the pregnant person make the body less ideal; less effective at protecting, nourishing, and sustaining the fetus that it might have been. (304). But barrier contaminations don't involve the body doing less well in providing the support they had. Bodies aren't designed to provide a certain amount of toxins, not more or less. They are merely the medium for doing harm via poisoning. This can be seen clearly if one imagines a Modern day Medici poisoning her enemy through transfusion after developing tolerance to arsenic or some poison sent through her body when providing support for her anemic enemy in need of a blood transfusion. Surely she is doing harm as she is providing needed support. So support doesn’t mean one can’t do harm to the one that is being kept alive by the other’s body

Extraction abortions may be thought to be allowing death as the fetus is withdrawn from the mother's body, the latter no longer being a barrier to harm. But if the extraction is done via hysterotomy, the doctor is removing aid that she was not supporting. Allowing harm involves omission or withdrawing what one was providing. K & W recognize this writing "In Drive away and Plate it seemed to matter for our classification that the barrier was some important way the agent’s.” (308). If we change drive away so the car belongs to Vicky (the victim) it becomes more plausible that moving the car would be morally equivalent to a standard doing of harm" (303).
Mifepristone may not just detach the fetus like life support can be drawn from an ICU patient. it may be that when the uterine wall falls away it damages the placenta (a fetal organ). So this would make it more like damaging someone's throat preventing the ingestion of nutrients than removing an IV.

Ironically, if Kingma's thesis that the fetus is a part of the mother, that will make all extraction abortions killing and doing harm rather than allowing the harm of death. if something is a part of you and cut off, that is killing the part, not allowing it to die. You cut off your hand, your hand dies, or if hands can’t die, then the cells in them are killed not allowed to die. Only if you cut off a source of support that is external is it a case of allowing to die Consider cutting off your head and you survive briefly as the head or headless body. The cutting doesn’t allow you to die but you have been killed. So even extraction abortions will be killings, not death allowings.

David H September 19, 2025 at 5:48 PM

The second crucial but false claim quoted above involves the barrier cases being different from standard cases of doing harm because the pregnant woman is being prevented from doing things that won't be harms if the victim hadn't been using (nourished and protected) by the barrier. But alcoholics and drug addicts are harming themselves as well as the fetus. Imagine the reluctant addict who wants to stop but can't. It seems that if she harms the fetus while harming herself as well than there isn't a case of the pregnant woman losing out on some control of her body for the sake of others that Woollard uses to defend the doing/allowing distinction. So the harm to the fetus when the addict mother harmed herself is not "a normative imposition..." She is not losing out on some bodily benefit to serve as a bodily barrier to prevent harm to another.

But it may be only Woolard and not Kingma who can claim such harms are different from typical cases of doing harm to others.

Somewhat ironically, given Kingma's mereological claims about pregnancy, any injury to fetuses will also be injuries to pregnant woman as fetuses are parts of their mothers. So the mother is not being restricted by pregnancy rules from engaging in an activity that only harms someone else. Every injury to the fetus will be an injury to a part of her and thus herself, no different from a pathological organ or limb making unhealthy the individual composed of such parts.

David H September 19, 2025 at 6:13 PM

The third mistake at the heart of K & W's claims is that the justificatory bypass brushes aside the needed discussion of special obligations. It may be the mother is doing harm because she has special obligations. K & W think special obligations wont' require morally demanding sacrifices of pregnant women. Their mistake is to think it is special obligations that are doing the moral heavy lifting.

They are conflating or confusing responsibility objections with special obligations. If I cause you to bleed intentionally, or negligently or recklessly, then I owe you’re a blood transfusion. The mother is responsible for the fetus's predicament and neediness. This is the standard pro-life responsibility objection to abortion. The fetus is in a predicament because the mother willingly engaged in sexual activity knowing that a fetus could be conceived in a very needy state, needs that only the mother can fulfill. So it is not the special obligation that comes from kinship ties that grandparents and siblings have to some degree that is responsible for the mother to avoid harm to her fetus

The special obligation is the mother made the fetus dependent, like Frankenstein making someone with a broken neck who can use his lab.

That no special obligation is involved can be seen by looking at another case, that of a rape-induced pregnancy, where there is the kinship relationship constitutive of special obligations. Most pro-lifers believe women don't have to stay pregnant when the rape resulted from pregnancy. There is no special obligation despite the fetus being their own child. What is doing the moral heavy lifting is the woman is not responsible in the case of rape. So if women have a duty not to harm their fetus by drinking and smoking etc., it is because they are responsible for the fetus being in a needy state, located in their bodily barrier.

One can see special obligations is doing little work if one imagines a mix up where an IVF clinic puts someone else's fetus in a woman's body. She is not the genetic mother. Most pro-lifers will think she can't kill the fetus. Here it is not responsibility or special obligations (unless you think being the gestational mother introduces the special obligations), rather it is just standard prohibitions against killing persons. So by analogy, the harms of fetal alcoholism are not being condemned because of special obligation but are akin to secondary smoke harming people who aren't your kin. that suggests further that K & W have misdiagnosed what the so called justificatory bypass keeps out of play. I suspect considerations of responsibility and impartial morality (the IVF mix up case) have more force than special obligations in imposing duties on pregnant women

In fact, all the K and W cases of harms, even those that the pregnant allow (e.g. oxygen reduction) rather than do, are the result of the mother doing something like procreative sex or IVF that created the scenario where there was a choice between harms to the mother or the fetus. It is like my placing Phil precariously by a dangerous body of water and than allowing him to later drown.

David H September 20, 2025 at 9:48 AM

What I meant to write in the brief third passage consisting of just one sentence above is that if Dr. Frankenstein creates a monster with a broken neck, so moving the monster will be fatal , that monster can stay in his lab until he recovers and can be healed. While the monster doesn't have a nonderivative right to stay on the property of Dr. Frankenstein, the creature has a derivative right to stay in virtue of a right not to be killed. So I would say in response to K and W that the fetus can acquire rights to use the mother's body when the mother is responsible for them needing to do so. Yes, it is her body and she has the initial claim to control it but she is (partially) responsible for someone else being in her body in a needy state. They might respond that stopping support doesn't make the fetus worse off than if the support had never been provided in the first place But this isn't true in the case of an IVF pregnancy as the fetus could have been frozen or adopted and placed in another womb. Moreover, there is a procedure called lavage where the fetus is flushed from the womb before implantation in the uterine wall. Since every fetus could be so flushed, that makes it false to say that if the woman aborts and ends gestation than the fetus is no worse than it would have been if she hadn't been providing gestational support at that time. She could have been not providing support than if the fetus was flushed. Of course, lavage is impractical, not widely available, and burdensome now. But imagine that in the future this wasn't the case. Then such arguments that ceasing to provide support doesn't make the fetus worse off than if it had not been provided support would seem to have an expiration date.

I wonder if they think that such considerations of responsibility are just further instances of issues that the justificatory bypass keeps off the table. This just leads me to say that their thesis is a strange one as every philosopher knows or should know that one's duties will be determined by an all things considered judgment. So I don't think it is a major theoretical success if K and W defend the claim that people shouldn't jump from the occurrence of harm to the fetuses to demanding women prevent them and sanction women if they don't

 David H September 20, 2025 at 9:27 AM

I realize K & W will respond that my example doesn't involve comparing the activities of a woman without the fetus to the losses the woman experiences when she can't do something when pregnant with the fetus. But my point is that we should look at fetuses that are parts as we do other parts of the woman. If a woman grows breasts, which are undeniably parts, she then can suffer certain harms like breast cancer. It is perverse to then compare how her life would have gone without having parts such as breasts and infer the breasts were a burden imposed upon her. I am suggesting that one take the same attitude to fetuses that are parts. Any harms to the fetus are harms to the mother who is the whole containing such parts. It is a mistake to appeal to the justification of the distinction between doing and allowing in that the latter protects someone's

Phil Reed September 21, 2025 at 7:53 AM

I haven't read this section of the paper as closely. Your reply to your comment was helpful, as I see now the paragraph on 311-312 that seems to be an attempt to address your concern (which was also a referee's concern).

I guess their view is that these claims about causal responsibility and the duties they generate are controversial. And your attempt to argue for them here is a concession (in their minds) that the justificatory bypass is not warranted.

How controversial are your claims though? This is worth pressing. They mention that a pregnant woman's causal responsibility is only "purported" (311). Really?

In response to your last point, maybe they think that the standard constraint against doing harm is not a matter of an all things considered judgment. I have a pre-existing, pro tanto duty not to kill you. This could change when other considerations are involved.

In addition, they say to your objection "many of the cautionary notes we raise in the next section would still apply" (312). So, obviously, your view that mothers ought not to kill their unborn babies involves images of fetuses as intrinsically perfect and

 Phil Reed September 21, 2025 at 7:14 AM

One way to read their first claim is a reference to inaction cases (not taking prenatal vitamins).

In cases that are action cases (smoking, etc.), substantive questions are required. This is because barrier contaminations forces us to consider the way in which the mother is simultaneously harming and benefitting.

 Phil Reed September 21, 2025 at 7:27 AM

I think they would say: "we're not talking about abortion."

They don't define "fetus-affecting maternal behavior." But given their examples (smoking crack, abusing alcohol, etc.), it seems to be something like: when a woman does something she wants to do that indirectly has an effect on the fetus. If that is the definition, then maybe they can sidestep the abortion stuff?

Note this: "We do not dispute that pregnant people have special obligations toward their fetus. Indeed, we believe they do, at least as long as they intend to continue gestating" (310).

Phil Reed September 21, 2025 at 7:31 AM

They also say "There may be some very rare cases in which it is possible to act on the fetus directly, but in practically all cases the fetus is only affected through modifying the pregnant person's body" (305).

This suggests they're talking only about indirect actions. But, to your point: are surgical abortions supposed to be very rare cases?

David H September 22, 2025 at 10:17 AM

Part of problem is that K & W are just responding to a referee on 311-312. They had not thought, or at least had not included such approaches in their original submission. Like many of us, they tried to change as little as possible in response to referee. But the referee's three points are crucial, if not devastating. They involved a doing prior to pregnancy that puts the fetus in harms way. At least in the case of IVF, that might clearly be a harm if the fetus is killed or injured when it could have been frozen or remained frozen or adopted by someone else. Secondly, it could bestow a derivative right of the fetus to stay in the woman's body and stay there unharmed from drugs etc.

 David H September 22, 2025 at 10:20 AM

The third comment of the referee was that it might make the allowing into a doing. I doubt that. I think the wrongness of an action doesn't change it from a letting die into a killing or an allowing into a doing.

resources from being used by another. I suspect I have to think more about this.

David H September 22, 2025 at 10:09 AM

Why do K and W ignore abortion? Why don't they state explicitly up front that they are not considering abortion? In a footnote #12, they write "our strategy interprets harm to the fetus to mean harm to the future child." Very odd that this is relegated to a footnote and four pages into the paper and then in the passage Phil mentioned on page 310. Ignoring abortion makes doing harm to fetuses less frequent.

But it is arbitrary to limit the paper so. it is not as if surgical with the surgical tools to dislodge and dismember the fetus from the uterine wall don't modify the woman's body. Moreover, Keep in mind that the same substances that may affect adversely the born child can also cause miscarriages (spontaneous abortions.) Or the toxic substance could cause premature delivery or inducement of labor to treat the fetuses in distress. So some future children in the future are younger than fetuses. Some fetuses will die due to the prenatal harms soon after birth so they are cases of harming the future child. But they are not abortions as they were born alive. It would be odd to insist that if one dies shortly after birth due to a prenatal incident is an abortion. Such alleged abortions could be prevented by infanticide immediately after birth, preempting the later death from the prenatal harm. So it doesn't seem principled to limit the discussion to the future child. Why focus just on damage due to prenatal causes of the harmed newborn rather than prenatal causes of the individual's death?

Perhaps abortion is such a contentious issues, the author's wanted to avoid discussion of killing and keep to discussions of non-fatal harm. Maybe they don't believe the fetus has interests when a mindless fetus. Or maybe they want to keep the discussion to cases that can be more easily considered allowings than doings of harm. But they seem to think extraction abortions are cases of barrier removals as they don't take issue with McMahan's claims that certain abortions are barrier removals and they criticize Thomson writing "Thomson accepts the characterization of abortion as killing, i.e., doing fatal harm. The arguments in this article challenge the characterization of maternal behavior as doing harm." But this may have been a late addition, not given much thought, as they thank an anonymous referee

The difference from abortion though is that ingesting thalidomide wasn't an attempt to act directly on the fetus. It was an attempt to benefit the pregnant woman by alleviating morning sickness. This is why I think it's a good example.

Stephen Kershnar September 25, 2025 at 10:39 AM

K&W CLAIM #1: INCONSISTENT RESULTS FOR ABORTION-PILL AND SMOKING/DRINKING

Phil: Aw, Jesus, Bernie. Come on with the smoke. You know the doctor found nicotine in my urine again.
Bernie: Then keep your dick out of my ashtray.
[The Paper, 1994]

David and Phil:

I’m not sure I see the argument.

Let’s consider what might go on in an abortion-pill case.

Case #1: Abortion Pill
A fetus is kept alive by nutrients from the uterine lining. The woman owns the lining because it is a part of her. The woman sloughs off the uterine lining. As a result, the fetus dies.

Case #2: Machine – No Maintenance (Killing)
A fetus is kept alive by nutrients from a uterine-lining-resembling machine. The woman owns the machine. The machine is very expensive to keep running. The woman turns off the machine. As a result, the fetus dies.

Case #3: Machine – Maintenance (Letting Die)
A fetus is kept alive by nutrients from a uterine-lining-resembling machine. The woman owns the machine. The machine is very expensive to keep running. The machine only runs if the woman puts on 25 lbs. and donates blood to it once every five days (it’s a strange machine). After two weeks, the woman refrains from doing these things. As a result, the fetus dies.

One issue is whether Case #1 is similar to Case #2 or Case #3. Note that if a woman regularly takes a birth-control pill, it works in part by thinning the uterine lining (endrometrium). Let us consider three determinants of the deontic status of a pill-based abortion.

(1) Right-Maker #1: Moral Right. A woman owns her uterine lining. As a result, she may thin it so long as her doing so does not produce disproportionate injustice or harm. It doesn’t. See Jewish-woman-in-Auschwitz case.

(2) Right-Maker #2: Commission vs. Omission. I’m not sure what the baseline is here. The most plausible issue here is whether the woman did something that caused the fetus to die. If she was not already taking birth control pills, then she did. If she was already taking birth control pills, then she likely did not.

(3) Right-Maker #3: Intention. If double effect were true, then it would seem that it matters whether the woman intended to kill the fetus or merely remove her from her body.

Objection to Right-Makers #2& #3: Baseline Smoking and Drinking
If the woman smokes and drinks regardless of whether or not she’s pregnant, then it is unclear whether her refraining from stopping is a commission. It looks like an omission.

Machine Analogy
In Case #3, a woman who ordinarily runs and watches her weight refrains from putting on the 25 lbs. necessary for the machine to work. This looks like an omission, whether viewed in terms of commission/omission or intention/merely foreseeable.

Best.

Stephen Kershnar September 25, 2025 at 10:57 AM

K&W CLAIM #2: NORMATIVE IMPOSITION

David:

Here is what you say.

“[T]he barrier cases being different from standard cases of doing harm because the pregnant woman is being prevented from doing things that won't be harms if the victim hadn't been using (nourished and protected) by the barrier. … But alcoholics and drug addicts are harming themselves as well as the fetus. Imagine the reluctant addict who wants to stop but can't. It seems that if she harms the fetus while harming herself as well than there isn't a case of the pregnant woman losing out on some control of her body for the sake of others that Woollard uses to defend the doing/allowing distinction. So the harm to the fetus when the addict mother harmed herself is not "a normative imposition..." She is not losing out on some bodily benefit to serve as a bodily barrier to prevent harm to another.”

Here is what you seem to argue.

(1) If the barrier case differs from the normal doing-harm case, then the woman is prevented from doing something that wouldn’t be harmful if the victim hadn't been using by the barrier.
(2) In alcohol and drug use, the woman is not the woman is being prevented from doing something that wouldn’t be harmful if the victim hadn't been using by the barrier [because it also harms her and, so, there is no normative imposition.]
(3) Hence, the barrier case does not differ from the normal doing-harm case. [(1), (2)]

Premise (2) is odd. It requires that
(a) a woman have a claim against herself and
(b) that she does not waive it.

If one or both are false – and (b) is false – then (2) is false.

In addition, K&W can – and should – modify their normative-imposition claim so that it focuses on normative imposition on the basis of another person. Alternatively, they might focus on a new normative imposition.

On a separate note, I doubt the barrier case matters other than in challenging the what analogy best fits the pregnancy case. This, though, rests on a property-rights-based morality which they and you would reject.

Best.

 David H September 29, 2025 at 5:47 PM

Steve
The K & W claim is that woman is being limited by the harm her action imposes on the occupant of her bodily barrier. My claim is that there are occasions where she can't do what she wants to without harming herself. So there is no harm-based imposition brought just by the fetus, limiting what she could do without harm occurring. I didn't say anything about rights against herself and waiving those claims.

Stephen Kershnar September 26, 2025 at 2:33 PM

A CONCERN ABOUT CIRCULARITY

K&W say the following two things.

“Remember the main aim of our article: blocking the justificatory bypass. This bypass invokes the standard constraint against doing harm to others to move directly from negative fetal outcomes to condemning a pregnant
person’s behavior, thus avoiding engaging substantive questions such as what special duties, if any, the pregnant person owes their fetus.” [p. 309]

“We evaluate some of these arguments in the next section. What matters for now is that if they are successful, and if a pregnant person has the right kind of special duties to their fetus, then there is an argumentative
route leading from a pregnant person’s special obligations to the classification of their behavior as a standard doing of harm.” [p. 311]

Consider two possible dependencies.

(1) Duty to Harm. One person has a duty not to harm via doing a second person (except in the case of self-defense or punishment).
(2) Harm to Duty. Whether an act is harm via doing depends on whether it infringes a duty.

I wonder if K&W are committed to these claims and, if so, whether the argumentative route is circular.

Best.

Stephen Kershnar September 26, 2025 at 2:39 PM

AGAINST THE RESPONSIBILITY-CONDITION FOR RIGHT FORFEITURE

David says the following.

“[K&W] are conflating or confusing responsibility objections with special obligations. If I cause you to bleed intentionally, or negligently or recklessly, then I owe you’re a blood transfusion. The mother is responsible for the fetus's predicament and neediness. This is the standard pro-life responsibility objection to abortion.”

I have two concerns. First, I don’t see why the mental state matters for causal responsibility. If causal responsibility grounds a duty to save, it does not matter whether A intentionally, knowingly, or neither intentionally or knowingly endangers B.

Case: Epileptic
A has an epileptic fit – his first ever and entirely unpredictable – which results in his hitting B in the mouth. The necessary dental work costs $50 K. If A pays for B’s dental work, then he cannot afford to pay for his child’s going to college.

Perhaps this relates to David’s acceptance of McMahan’s responsibility thesis.

“McMahan has convinced me that if there is a choice between innocents facing a harm, justice demands that the person responsible for the harm suffer it. Consider driver who, non-culpably drives off the road and will injure someone in their house. The threat is due to his driving, he should veer his car away and take the hit rather than stop his car at no cost to himself by injuring the resident.”

Let’s consider non-promissory bases for a person incurring a duty to benefit or, perhaps, a duty save.

(1) Causation
(2) Blameworthiness
(3) Moral responsibility
(4) Endangerment (whether through one’s own or someone else’s responsibility)

I don’t understand how (3) is plausible.

I can see the intuition behind (1) and (4). Consider the intuitive right to use self-defense against a psychotic attacker (causation) or a small man thrown down a well at someone (the small man endangers but does not cause the endangerment).

I can see the intuition behind (2). Consider the notion that responsibility-valence determines one’s rights.

I can’t see the case for (3). Again, consider Epileptic above. How would it matter whether the epileptic had some small notice or not that he might have an epileptic fit and, thus, have some small amount of responsibility for being within arm’s range of another?

A responsibility basis for forfeiture (whether defensive or compensatory) is less plausible than a blameworthiness or causation basis for it.

Best.

David H September 29, 2025 at 6:08 PM

Steve
I believe that one doesn't have a right to defend oneself against an innocent threat. I would merely excuse one for doing so then claim they are within their rights.

Imagine that a tornado has sent someone hurtling towards you. The impact will kill you. Most people believe that you can shoot the innocent threat in self-defense. But imagine that he collides with you, leaving you not dead but dying. You can save yourself by getting to the hospital quickly. But the innocent man that caused your likely fatal injuries is lying in the road in front of you. You cannot run over and kill him to save your life. Surely, you can’t do that even though your killing him will have prevented him from having killed you. He is now an innocent bystander. I fail to see how it can matter that a few second earlier you could permissibly kill him and destroy his organs in the process.

I will try to bolster this response with a six-step argument.

1. Tornado blown mannequin: Tornado winds send a heavy metal mannequin hurling towards you. The impact will kill you. You can’t push an innocent bystander in front of the mannequin to save your life when doing so will kill that innocent bystander.

2. Detached Prosthetic Leg: Someone’s prosthetic leg - which is not a real part of their organic body in any biologically respectable account of composition - is detached by tornado winds and sent hurtling towards you. You aren’t entitled to grab the man whose prosthetic has been removed by the winds and throw him in the way of the prosthetic leg, killing him to save your own life as he is an innocent bystander.

3. Attached Prosthetic: Someone is still attached to their prosthetic leg when the tornado sends their prosthetic and their attached body hurtling towards you. The prosthetic is not a part of the man as parts of organisms must be caught up in life processes. The blow from prosthetic will kill you, the lethal threat is not the man attached to it as that man doesn’t even touch you or add any lethal force to the prosthetic’s impact. Surely, you have no right to fatally blow up the man attached to the prosthetic to stop the prosthetic from hitting and killing you as he is an innocent bystander.

4. Detached organic leg: Someone’s organic leg is detached from their organic body by a tornado and after numerous rotations within the tornado is expelled towards you. The detached organic leg will kill you unless you shove the now one-legged man in front of his detached leg, killing him to save yourself. You can’t morally do this as he is an innocent bystander.

5. Reattached dead organic leg: The tornado detaches a man’s leg. After ten minutes of rotations within the hurricane, the now dead leg is “reattached” to the man and will kill you while the rest of the body attached to the dead leg is not a threat. The dead leg is not part of the man, anymore than his clothes, glasses, or hearing aid. You can’t ethically bend the man’s body so the dead limb will impact his torso, killing him, but shielding and sparing you. He is an innocent bystander.

6. Reattached living leg: The tornado detaches and then quickly reattaches a living leg. The living leg will kill you, not the rest of the body attached to it. The leg is “alive”, so it is truly a part of the attached man. But his being attached to his leg, so to speak, can’t be justifying your killing him to stop his attached leg from fatally concussing you. It is morally arbitrary that he has been reattached to his leg and it is a part. He is, to paraphrase Alec Walen “aboard his own body” (1997, 1088) and thus is morally akin to an innocent bystander.

My conclusion – and I hope yours (dear reader other than Steve) - is that reattachment in #6 is not morally significant. Nor are any of the other differences in #1-5. Hence self-defense against non-responsible threat in #6 is no more justified than self-defense against an innocent bystander in #1.

Phil Reed September 21, 2025 at 6:39 AM

Philosophical Irrelevance
“whether you can do harm to your fetus” (the title of the paper) risks philosophical irrelevance. Your fetus obviously can be harmed by what you do. Here’s a justification for the expectation that mothers shouldn’t do things that harm their fetuses: fetal harm is bad. K&W say: well, if the badness arises from a mere allowing, it’s not obvious you have a duty to avoid actions that end up harming your fetus. But do we really need a justification for not wanting bad things to happen to our fetus? Maybe there is no obligation, strictly speaking, not to abuse alcohol. Or maybe society shouldn’t impose that obligation on pregnant women or shouldn’t judge them if they drink in excess. If true, that would still be completely irrelevant for most people who don’t want their fetuses to have fetal alcohol syndrome.

Giving an account of maternal obligations is hard. Deciding you don’t want your fetus to have preventable diseases is easy.

Phil Reed September 21, 2025 at 7:20 AM

All of the advice about how pregnant women's actions can harm their baby (292) is actually ambiguous on the technical point between doing and allowing. The point of that advice is that pregnant women can do things that will bring about harm to the fetus. K&W do not disagree (how could you?)...making the bulk of their article a philosophical concern that has no bearing on real life decision-making.

Kurt B. September 21, 2025 at 3:08 PM

These are really interesting points Phil. This might be more revisiting the issue in the thread above, but here I wonder what you think about cases where someone who is pregnant has, say, syphilis, which runs the risk of complications (fevers, jaundice, etc.) birth defects (skeletal issues, developmental delays, etc), and vertical transmission. Do you think that K&W would say that if someone is pregnant and doesn't get treatment for syphilis they are merely contaminating a barrier rather than harming the fetus? I take your point here seems to also fold in something like a character assessment--that even if there isn't an obligation to not harm the fetus, this obligation has nothing to do with why most people don't harm their fetus--why most people would get treated for syphilis while pregnant (rather than waiting until the second after delivery?). Or, do you see the motivation to not drink during pregnancy as drawing on some other kind of moral motivation that also won't draw on the special obligations K&W dispute?

Phil Reed September 21, 2025 at 5:54 PM

I think that not being treated for syphilis is an example of allowing harm to the fetus. The reason being that in this case, the mother doesn't initiate the harmful sequence. She doesn't contaminate her body in the way she does by ingesting noxious substances. See below where I discuss differences between barrier removals and barrier contaminations.

Phil Reed September 21, 2025 at 6:42 AM

Patterns of Behavior

Near the end of their paper, K&W raise what they take to be a puzzle that they believe has been inadequately addressed, which is how to analyze behavioral patterns (e.g. smoking during pregnancy). They claim that “this puzzle is interesting not just in pregnancy in particular but for the general literature on doing and allowing” (317). They believe that this puzzle constitutes a “contribution” of their paper because it identifies “a gap in the literature on action/inaction” (293).

K&W are right that patterns of behavior are different from discrete actions, but it is hard to see how such patterns are relevant to the doing/allowing distinction, much less a contribution to the discussion of that issue. This is simply because, however we analyze patterns of behavior, they have nothing to do with allowings or inactions. It might be an interesting question whether “a series of actions is an action” (318) but it has nothing to do with the general literature on doing and allowing (unless any kind of action analysis is supposed to be relevant to doing and allowing).

K&W might respond: patterns of behavior might be inactions, such as not feeding your child. However, there are two problems with this response. First, they seem to understand patterns of behavior as a series of actions. All of their examples (smoking, drinking, eating unhealthy food) are examples of doings or actions. Second, even if “behavior patterns” is supposed to include patterns of inaction, it’s not clear why this would be a distinct area of interest that advances the discussion of doing and allowing. Unless they think that series of actions and series of inactions are to be analyzed differently (and they give no indication of this), I don’t understand why patterns of behavior deserve special flagging as an overlooked puzzle for philosophers interested in doing and allowing.

Phil Reed September 21, 2025 at 6:47 AM

“Barrier Cases” as a Problematic Category

We can set to one side fetal-affecting maternal behavior that are cases of inaction. K&W are right that not taking prenatal vitamins or not getting vaccinated cannot count as doing harm to the fetus. I will only be concerned with fetal affecting maternal behavior that involve actions.

K&W’s central argument might be characterized in the following way.
1. In barrier removal cases, the standard constraint against doing harm does not apply.
2. Barrier removal cases are similar to barrier contamination cases.
3. Fetal affecting maternal behavior is a kind of barrier contamination case.
4. Therefore, in fetal affecting maternal behavior, the standard constraint against doing harm does not apply.

In what respect are barrier removal cases like barrier contamination cases? Well, they both involve barriers. But that can’t be the relevant respect. If I take a shield and beat someone to death with it, I’ve still done harm.

K&W aren’t so foolish to say that barrier cases are any cases that involve barriers. However, I do worry that their argument at times seems to approach this line of thinking. They point out that in barrier removal cases, how we classify the action (as a doing or allowing) depends on whether we own the barrier. They then claim in §4D that in fetal affecting maternal behavior, the agent owns the barrier because she owns her own body. Let’s grant this.

Still, the nature of the way the barrier is used is extremely important. There are actions where ownership of a barrier is irrelevant to the classification of doing v. allowing. If I use my shield or someone else’s shield to bludgeon an innocent person, I’ve done wrong. Here’s another way to put it: from that fact that ownership of the barrier matters for how to classify barrier removal cases it doesn’t follow that ownership of the barrier matters for barrier contamination cases for how to classify barrier contamination cases. All depends on the how the barrier is used. As K&W themselves say, “barriers can be acted upon in multiple ways” (304).

So, one thing the argument does is obscures the issue of fetal affecting maternal behavior by treating barrier contamination and barrier removal as one set of cases, which K&W call “barrier cases.” After explaining that barrier removal and barrier contamination both depend on whether the barrier is owned, K&W conclude “barrier cases in pregnancy are not standard doings of harm” (308). But this conclusion seems to rest on the idea that both barrier removals and barrier contaminations both involve barriers where ownership is relevant. That does not seem to me sufficient to support premise 2.

Everything depends on the nature of “barrier contamination” and how similar this is to barrier removal.

Phil Reed September 21, 2025 at 6:57 AM

Barrier Removals vs. Barrier Contaminations

Another difference between barrier removal cases and fetal affecting maternal behavior. In standard barrier removal cases, the agent stands to gain directly by removing the barrier. For example, Anna will have severe lifelong disabilities if she does not get to the hospital. However, in fetal affecting maternal behavior, the pregnant woman harms herself in addition to harming her fetus. All of the things that are unhealthy for the fetus are also unhealthy for the mother. Given this, these kinds of cases do not fit the normative profile of barrier removals (or of Push) because the agent does stand to gain by harming her fetus. I imagine K&W would say that the pregnant woman stands to gain in some sense, since she chooses the harmful behavior – she wants to drink alcohol or smoke crack. But this is a very different sense of “benefit” (in terms of well-being and not health, more strictly) than standard cases of barrier removal. In fetal affecting maternal behavior, the woman _contaminates_ not just a barrier, but her body.

Stephen Kershnar September 26, 2025 at 3:00 PM

BARRIER-CONTAMINATION TRADEOFF

Phil:

Given what you say above, I’m curious as to when you think barrier removal is permissible. Here is what you say.

“In standard barrier removal cases, the agent stands to gain directly by removing the barrier. For example, Anna will have severe lifelong disabilities if she does not get to the hospital. However, in fetal affecting maternal behavior, the pregnant woman harms herself in addition to harming her fetus. … In fetal affecting maternal behavior, the woman _contaminates_ not just a barrier, but her body.”

I can’t tell if you think what is relevant here is that the border contaminator causes harm to the fetus, is blameworthy for causing harm to the fetus, or intends to cause harm to the fetus.

Consider a pregnant woman who refuses to take vitamins that will prevent the fetus from becoming cognitively disabled because such vitamins will cause her to likely lose the ability to walk (because she has an odd genetic disease). Other than a consequentialist balance, what determines how much harm the woman must incur before she may refrain from taking the vitamin?

If it’s a consequentialist balance – or there’s no solution at all – why care about whether someone causes harm, is blameworthy for causing harm, or intends to cause harm? That is, without a tradeoff function, I’m not seeing why the way in which the harm is brought about matters much. Perhaps there is a sliding scale, but I’m skeptical.

Best,
Steve K

David H September 22, 2025 at 10:47 AM

I agree that there are many barriers where we don't say the removal or contamination make resulting harms into allowing harm than doing harm. You can do harm to yourself when wearing a vest and helmet. And likewise you can do harm to another contained within or who is a part of you.

The question is whether such doings don't warrant the standard constraint. Just as Phil and I argued in the paper "how not to defend the unborn" that abortion was not as bad as killing the independent whom one was not keeping alive, harms to the fetus are not as wrong. But we think they are still gravely wrong and so what one shouldn't do. What follows for non lethal harms to the future child is another question and there are too many other issues in play for general principles.

Phil Reed September 21, 2025 at 6:59 AM

Concerns About Plate

Besides fetal affecting maternal behavior, K&W offer Plate as another “barrier contamination” case. They admit that this example is forced. It is so far from a real world case that I find it very hard to have clear intuitions about Plate. There are a lot of distorting factors that could be present, even if we did have reliable intuitions.

How is the plate supposed to be a barrier to harm? Is it that without a plate, you suffer the harm of no lunch? But there are disanalogies here. A plate is necessary to avoid the harm of no lunch, but it is not sufficient. Whereas, in barrier removal cases, the barrier is sufficient to avoid the harm to the innocent party. At least with respect to certain functions, the woman’s pregnant body is sufficient to avoid certain harms to the fetus, such as the harm of oxygen deprivation. The fact that the plate is insufficient to avoid harm to the innocent party might explain why the case is not a standard case of doing harm (if we have that intuition).

K&W build into the plate case the way in which an agent benefits by contaminating the barrier: without the caffeine powder, the person produces fewer outputs and earns significantly less. However, as I said above, fetal affecting maternal behavior does not have a parallel benefit to the pregnant woman. What is the cost to the pregnant woman if she does not contaminate? This is another disanalogy.

With these problems, Plate does not seem to be able to do the work that K&W want it to do.

David H September 23, 2025 at 2:20 PM

Aren't the costs to the woman if she doesn't "contaminate the fetus" is that she loses out on the pleasures of those activities or foods or drink that are harmful to the fetus (assuming they are not harmful to her).

Phil Reed September 21, 2025 at 7:11 AM

A Charitable Interpretation

After developing the above points, I went back and studied the article more carefully to try to understand and reconstruct their central argument.

I made a big deal above about dissimilarities to barrier removals. I criticized Plate and the relevance of the relationship to the barrier. But maybe there is a charitable way of understanding their argument where these problems are avoided. Here's an attempt:

1. Fetal affecting maternal behavior is a barrier contamination case.
2. Barrier contamination cases are not standard cases of doing harm.
3. In cases that are not standard cases of doing harm, the standard constraint against doing harm does not automatically apply.
4. Therefore, the standard constraint against doing harm does not automatically apply.
5. If the standard constraint against doing harm doesn’t automatically apply, the justificatory bypass is unwarranted.
6. The justificatory bypass is unwarranted.

I wonder if this argument is in fact sound. The key premise is premise 2, and their idea in support of it is that in an action that involves harming, the mother's body is all the while also benefitting the fetus.

If they think that the "standard" case of doing harm involves no aspect where the victim is also benefitted in some way, then I'm not sure where they can be criticized.

Still, if this is right, their claims are very innocuous and limited: when women do things that harm their developing fetus, they are also still pregnant. If read this way, the article doesn't shed any light on whether fetal-affecting maternal behavior is (closer to) a doing or allowing.

David H September 22, 2025 at 12:49 PM

Maybe it depends upon the context. some barrier harms may have the same moral valence as non-barrier harms and thus elicit the constraint. consider a vindictive woman hurting her fetus to get back at her cad of a boyfriend. Or consider a pregnant woman who has Munchhausen degree by proxy and wants a sick child. Or a disabled mother who wants to raise a disabled child - assuming the disability is a bad difference and not a mere difference

In other cases, say when the barrier case involves the women not losing out on a good by avoiding the behavior that is harmful, the constraint against harm is plausible and there is no need to look at special obligations

As I suggested above, the justificatory bypass set up is a bit of a red herring as duties involving censure and prohibitions should be all things considered duties. So I don't think philosophers should ignore any relevant features.

David H September 23, 2025 at 9:26 AM

"Allowing Harm, Doing harm, Good barriers"

I don't think K & W analyze properly the moral relationship between barriers and harm. It is one thing when a woman's body is keeping the fetus alive or free from injuries due to any X or too little or too much of X and then she stops protecting the fetus against any X or too little or too much of X, or stops doing what she was doing to keep the fetus alive; but it is another thing when the women's body was not protecting the fetus against toxin Y, when it wasn't designed to, and she causes toxin Y to harm the fetus. The moral upshot is that this makes the toxin a doing of harm rather than allowing of harm

There is a more abstract or general description of the maternal body as protecting the fetus against pathology and death and so when she transmits a toxin to the fetus she can be described as not protecting the fetus as well as she was. That is different from a third party poisoning the fetus when that third party wasn't providing it with the support to avoid death and disease and injury. So maybe the mother's rendering the fetus less well off than she had been is a mitigating factor, making the mother's wrong less wrong than the third party as only she can say she wasn't providing as much health and safety as she was before.

Phil and I make a somewhat similar Kamm-inspired claim about why the mother's aborting isn't as bad as say her committing infanticide. (This has nothing to do with a difference in moral status of the unborn and the born. It just explains why violence against the woman aborting is not justified but violence against the woman committing infanticide could be.) The pregnant mother was keeping the fetus alive and then she stopped providing that needed support when she aborted and killed the fetus. Abortion is the rare type of killing which has a feature that is partly definitional of letting someone die - that is, the individual loses out on more life possible only with someone's else's aid. so the aborting woman kills the fetus whose life she was saving or preserving. But the induced surgical abortion was still a new threat, not the removal of support. So maybe the mother's poisoning the fetus that she was keeping healthy and alive is not as bad as the third party doing so. But it doesn't get morally annexed to barrier removal as the mother's body was not protecting the fetus against the toxin

The implications for the justificatory bypass? Perhaps the difference between maternal and third party harms of the fetus are morally different and so the constraint against harming doesn't automatically apply. But when we bring in the additional factors, unlike the case where the mother just doesn't provide oxygen as effectively and allows harm, the poisons the mother introduces and the body wasn't designed to protect against or allow within certain parameters, this makes the woman's actions worse than in barrier removal. So a case can be made for censure and criticism is certainly warranted just as if a lifeguard saving a drowning bather poisoned the drowning person in a way unrelated to the saving.

David H September 23, 2025 at 9:36 AM

"Baselines, harm, failure to benefit, and counterfactuals"

K & W raise the possibility that the mother isn't harming her fetus but failing to benefit. I think failures to benefit are harms, just not immoral harms in many cases. The authors overly moralize harm. Many harms to competitors in business, sports, professions, romance are not immoral harms. one person getting the financial or professional or amorous success at the expense of another is a harm but not a wrongful harm.

The authors wonder what would be the baseline for determining a harm has occurred since the alternatives depend upon what woman do. While it is hard to determine the nearest possible world - recall Steve Kershnar's discussions of what would have happened if the QB had not had his knee blown out by the tackler? Would he have been tackled by contact with a different part of the body, not tackled at all and so on - I would suggest that pregnancy doesn't bring in a new distinctive problem. My recommendation is health provides the base line. When the mother makes the fetus unhealthy, she has harmed it. when She doesn't enhance a healthy fetus, that failure to benefit is not even a morally legitimate harm as I don't think that is a harm at all since the fetus's interests when mindless are just in being healthy. The fetus has no other interest that is set back by not being enhanced.

Stephen Kershnar September 24, 2025 at 8:05 PM

OBJECTION #1: SEVEN OR MORE FACTORS

“Guys, that’s a lot of push-ups.”
Pete Mitchell, Top Gun

K&W’s list of factors include the following.

(1) Benefit vs. Harm. Refraining from benefitting vs. harming
(2) Doing vs. Allowing. Harming via doing vs. harming allowing
(3) Act vs. Collection of Acts. Harming via single action vs. harming via a collection of actions
(4) Cost and Harm-Function Curve. Cost to the harmer in the permissible-harm function to relevant parties.
(5) Cost and Harm-Function Curve. Cost to third parties, such as the woman’s other children.
(6) Barrier-Types. Barrier removal vs. barrier contamination

To this I would add the following.

(7) Right-Stringency. Stringency of fetus’ right against the woman (correlative of woman’s special obligation to the fetus). Consider, for example, whether the woman got pregnant intentionally, accidentally via voluntary sex, or via involuntary sex.

Consider the following competing claims.

(A) Consequentialism. What explains the permissibility of a woman doing something that purported harms the fetus is a consequentialist balance, perhaps one that includes desert.
(B) Seven-Factor Non-Consequentialism. What explains the permissibility of a woman doing something that purported harms the fetus is a seven-factor-plus non-consequentialist function.

My suspicion is that (A) explains the purported (B)-type explanation.

Best.

Stephen Kershnar September 24, 2025 at 8:06 PM

OBJECTION #2: PROPERTY RIGHTS UBER ALLES

“Get off my lawn.”
Walt Kowalski, Gran Torino

The barrier removal and contamination cases should be analyzed in terms of property rights. Once we figure out the answers to these three questions, we’ve discovered a lot of what matters regarding pregnancy-related harm.

(1) Right to be Inside the Woman. Does the fetus have a right to be inside the woman?
(2) Barrier-Right. If the fetus has a right to be inside the woman, then does the fetus own the barrier?
(3) Right-Infringement. If the fetus has a right to be inside the woman and the fetus owns the barrier, then does the woman’s engaging in various activities infringe the fetus’ rights?

If the answer to any of these three is “no,” then the woman does not wrong the fetus. Here the following is true.

(4) Right Infringement. One person wrongs a second person if and only if the first infringes the second’s right.
(5) Property-Right Infringement. One person infringes a second person’s right if and only if the first infringes the second’s right to her property.

Note that self-ownership is a property right over one’s self.

Best.

David H September 29, 2025 at 5:33 PM

Steve
I think it matters if the fetus can be safely removed and if the mother is responsible for the fetus's predicament.

Imagine if Dr. Frankenstein created the monster with a broken neck. He knew this might happen if he made the creature from some various spare parts. So the monster couldn't leave Dr. Frankenstein's lab until his neck mends. Given that the monster has a broken neck because of Dr. Frankenstein's actions, I think the monster has a right to stay in Frankenstein's lab. So too for the fetus

Stephen Kershnar September 24, 2025 at 8:08 PM

OBJECTION #3: A VIOLINIST AND A SQUATTER

"One good analogy is worth three hours discussion"
Dudley Field Malone

Consider Judith Jarvis Thomson’s violinist analogy.

The violinist complains that his benefactor doesn’t take the proper vitamins, drinks alcohol, and smokes cigarettes.

Offer: If you consent to waive your objection against (or claim to) my doing the above hedonic activities, then I hereby consent to remain hooked up to you.

Acceptance: If you consent to remain hooked up to me, then I hereby consent to waive your objection against (or claim to) my doing the above hedonic activities.

This seems to be a valid agreement.

Now consider if the beneficiary is unconscious and, so, cannot consent.

If there is (a) counterfactual, (b) rational, or (c) subsequent consent, then the benefactor can get the same valid agreement.

I doubt any of these are valid consent, but if so, then we need to know whether there is an additional right infringement if the woman chooses some harm less severe than killing the fetus, for example, blinding it.

I’m not sure what to say about this.

The problem is not the seven-factor-plus determiner of non-consequentialist permission. Rather, the problem is that I’m not sure what can be done to a trespasser short of coercing him to leave one’s property. Consider the following case.

Case #1: Squatters
A low-class squatter refuses to leave A’s house. Because of the state law, A cannot get the squatter out of her house. The squatter is damaging A’s house, selling some of A’s belongings, and driving down the value of the houses in the neighborhood by leaving her children’s toys and her car on the lawn. A plays music that is so obnoxious that the squatter’s children – involuntary squatters - cannot learn anything from school during the time during when they squat in A’s house. A knows that when the law finally evicts the squatters, they’ll neither get convicted nor pay any damages (they have no money). In fact, they’ll spitefully damage the house more before leaving. As a result, she gets enormous pleasure from playing this music. In fact, she says, each time she thinks about it, she gets more pleasure than she would from sex.

I do not know whether A’s playing the music is permissible. I suspect she may do so because it is her house. However, her music playing probably does not count as compensation (whether contract- or tort-like compensation), punishment, or relationship reparation.

I wonder if the violinist or squatter case – or both – is the best analogy.

David  H September 29, 2025 at 6:02 PM

The Moral Valence of Non-standard Doing and Non-standard Allowing Harm

K & W believe that doing harm to someone being preserved by your body makes that harm unlike those in which the standard constraint applies. So the moral valence is different as it is not as bad to harm someone depending upon your body than someone who is independent of your support.

But I would contend that doing something that puts someone in a situation where you then allow them to be harmed, gives such cases of allowing harm, a different moral valence, as it is worse to allow harm to those you put in a situation where they could have harm befall them. The woman who becomes pregnant has put her fetus in a precarious situation where she then allows the fetus to be harmed by partially withdrawing a barrier. So both the constraint against standard doing of harm and the lesser moral objection against standard allowing of harms don't apply in pregnancy. K & W just focus on how doing harm via barrier is different than doing harm to those one was not supporting and protecting. But they ignore how becoming pregnant makes it worse to allow harm to the person one put into a precarious position.

An analogy is that If I put Phil into a location where I am the only supplier of his basic nourishment needs, then my refusing him food and water is not the standard case of allowing harm. I have a responsibility not to allow such harm to befall him

David H September 29, 2025 at 6:42 PM

The irrelevance of special obligations, fair distributions of harms, and the duty not to allow harm

I mentioned earlier that K & W are misled by the role of special obligations in the non-pregnancy cases of removing barriers. Since the threat, as Phil points, out is not due to the mother, it is unlike pregnancy. In such non-pregnancy cases, then it may be the special obligations of mothers to fetuses is the only basis for her duty to save her child. It is, as they suspect, a weak claim. Mere biological ties are not as compelling source of special duties as the special obligations of having chosen to parent and also the non-special obligations of justice and fairness. Boonin backs up the former point where there are two parents refusing their kid a life saving bone marrow transplant, one is the parent who donated the sperm and the other is the parent who raised the child. It seems the stronger obligations to aid is found in the latter. This is even more obvious where the donation of sperm was for research not conception but the latter occurred due to a lob mistake

But K & W ignore the role of the mother in putting her fetus in a precarious position. She is responsible for the choice between harms to herself or the fetus.

The relative unimportance of special obligations is more evident in rape cases. Most pro-life lawmakers allow abortion in such cases. But the special obligations of mothers to their children would apply here in rape as just as in consensual sex. So one can see the much greater weight defenders of fetal rights are putting on responsibility as the raped woman is not responsible for the fetus's predicament and so considerations of fairness don't demand she suffer the harm of unwanted pregnancy.

Return to the case of the mix up in the IVF clinic. A woman ends up pregnant with a fetus that is not her genetic child. She has no special obligations if they are based upon her being a genetic mother and not merely a gestational mother. Still, most pro-lifers believe she should not abort the genetic child of another couple. So protectors of fetuses don't need special obligations to justify their opposition to harming fetuses. Of course, there is no responsibility for the choice between harms by going to the IVF clinic, the responsibility for the choice between harms is on the clinic staffer who made the mistake. So the objection to abortion will be based on the wrongness of intentionally killing the fetal person who has such great value (or if one is a new natural law advocated, the irrationality of trading off one incommensurable good such as life for another good such as pursuit of work, play, or family or whatever goods the unwanted pregnancy is preventing. ) This again suggests that in cases of biological kinship, it is not special obligations but the prohibition of killing and considerations of fairness of distributing harms to the responsibility that is doing the moral heavy lifting.

The upshot for allowing harm in cases of extraction abortions and non-fatal barrier cases is that the fairness considerations is doing the moral heavy lifting as there is no intentional doing of harm. So deontological and new natural law considerations against intentionally killing are not in play. One has to rely upon the wrongness of allowing harms in cases where one is responsible for the fetus being in a situation where you can allow it to be harmed. The wrongness of allowing someone to suffer harms that are a result of your earlier actions strikes me as stronger than special obligations. So the defender of fetuses against maternal harms is in better shape appealing to general duties than special obligations.