What Compliance Requires: Moral Obligation, Bodily Use, and Abortion Law
Even if a moral duty is granted, what gives the state authority over the bodily support that compliance would require?
Working draft v0.7.4 · 1 August 2026
An author manuscript released as a dated working paper. It asks what authority enforcement requires even when a moral duty is granted. No journal acceptance or current legal guidance is claimed.
Contents
- Abstract
- 1. Beginning After the Duty Is Granted
- 2. Why Authority Over an End Does Not Settle Every Means
- 3. Person-Use Authority
- 4. Granting the Duty to Gestate
- 5. Ending Support Does Not Automatically Authorise Death
- 6. Viability, Preservation, and Authorization
- 7. Stating the Facts: Courses, Times, and Refusal
- 8. Three Cases
- 9. “The State Is Only Prohibiting Killing”
- 10. The Baseline
- 11. Doing, Allowing, and Normative Imposition
- 12. What Survives If This Means of Enforcement Fails
- 13. What the Argument Establishes
- 14. What Would Defeat the Argument
- 15. Conclusion
- References
Abstract
Suppose a fetus has full moral standing; intentionally killing it is conduct a state may ordinarily prohibit; and the pregnant person has a genuine moral duty to continue gestation. What follows about enforcement?
Not the enforcement. This paper proposes a compliance test. It identifies the timely courses by which a bodily condition could be avoided, the public rules that close them, and whether those rules together make the condition unavoidable. To handle overlapping criminal, licensing, and facility restrictions, the test counts every rule that is a necessary element of at least one minimally sufficient closure set. Applied after viability, it distinguishes three cases. An authorized preserving course raises no further problem. A preserving course that is legally open but protectedly refused is not state-closed, yet the state has imposed a choice between two protected bodily interests and owes a justification of that conditional arrangement. Where the regulatory scheme closes every timely course, it owes an argument for authority over the continued bodily support thereby made unavoidable. Protected refusal, conditional choice, and legal closure are different institutional facts.
Keywords: abortion; state authority; bodily integrity; enforcement; person-use; viability
1. Beginning After the Duty Is Granted
Take a pregnancy in which the moral case against abortion has been made as strong as possible. Assume full fetal standing, voluntary responsibility for the pregnancy, a special parental duty, a genuine duty to gestate, and ordinary state authority to prohibit intentional killing. Set aside poverty, abandonment, deception, partner coercion, and inadequate care.
The question is what follows about what the state may make her body do.
Duties and enforcement are often linked: if I have a duty not to assault you, the state may prohibit assault. But the inference can fail at the level of means. A duty to repay a debt does not authorize removal and sale of the debtor’s kidney; a parent’s duty to keep a child alive does not by itself authorize compelled bone-marrow donation. The state may be justified in pursuing an end and not yet justified in using a particular means.
The compliance test. Identify (i) the timely courses the patient would authorize that could avoid the bodily condition at issue; (ii) the public rules that legally foreclose each course and whom they bind; (iii) the regulatory set, consisting of every rule that is a necessary element of at least one minimally sufficient set of public rules sufficient to close one of those courses; and (iv) whether the regulatory set contains a sufficient closure set for every timely authorized course. If it does, ask whether authority over what the rules formally prohibit also establishes authority over the bodily condition their closure makes unavoidable.
The set need not be unique. Borrowing only the formal structure of Wright’s NESS account of overdetermined causation, a rule qualifies when it is a necessary element of some minimally sufficient closure set, and the compliance test operates over the union of all qualifying rules (Wright 1985, 1775). Thus a criminal prohibition, a licensing rule, and a facility restriction may all belong where each independently forms, or helps form, a sufficient route of legal closure. Redundancy does not make closure disappear. The NESS analogy identifies membership; it does not import tort liability or decide the authority question.
Membership remains restricted to closures of courses the patient would authorize. Background consent law is not included when it merely gives effect to her refusal: it forecloses no course she seeks. This is what distinguishes a course the state has closed from one she has declined.
The test does not deny that a killing prohibition is a killing prohibition. It denies that a rule’s formal object exhausts the authority exercised by the regulatory scheme in which it operates.
1.1 Where this stands in the literature
The argument stands downstream of the body-use tradition. Thomson (1971) granted fetal personhood and denied that a right to life includes a right to another’s body; Boonin developed that tradition (2003) and argued that responsibility does not without more generate the bodily-use right the objection requires (1997, 288–90). Koppelman (1990) and Little (1999) developed legal registers.
Pallikkathayil (2025, 473–75) argues that pre-viability abortion prohibitions violate equality before the law because states do not otherwise require citizens to make their bodies available for others’ protection. Shiffrin (2025, 492–94), Gheaus (2025), Olsaretti (2025), and Viehoff (2025) respond from several directions; Hartley and Lindsley-Kim (2025) resist assimilating gestation to ordinary parental care.
The differentiator is not that this paper concerns law. It places the dispute farther downstream by granting a genuine directed duty to gestate and asking what means the state may use to enforce it. Nor is the argument comparative: it analyzes the authority exercised through legal closure without depending on a class of differently treated citizens.
Kingma and Woollard (2025, 296–98, 312–14) narrow the claim further. They treat the pregnant person’s bodily resources as within her protected sphere, apply Woollard’s account of normative imposition, and distinguish substantive obligations from sanctions and interference. The contribution here is institutional. A protected refusal and a legal closure may leave the same body gestating, but they are different state actions and generate different justificatory burdens. The compliance test distinguishes authorized preservation, protected refusal, and complete regulatory closure rather than resolving all three in one direction.
2. Why Authority Over an End Does Not Settle Every Means
Distinguish three things. Power is the capacity to produce compliance; authority is the standing to impose obligations within some domain; legitimacy asks whether that standing, and the means through which it is exercised, are justified. The question here concerns the second and third: not whether the state can produce continued gestation, but whether it has the standing to require it.
The premise needed is modest:
Means-specificity. Authority to require or enforce an outcome does not by itself include authority over every kind of means through which that outcome could be secured.
This is not a claim that outcomes do not matter, or that protective purposes are weak. It is a claim about institutional authorization: a state is not authorized to produce good outcomes by whatever route is available, but to act in certain ways toward persons who retain standing against it. Ripstein’s reconstruction of the Kantian account makes a stronger version of the point, organizing right around reciprocal independence and the form of relations between persons rather than the value of the results an institution pursues (2009, 30–34). The argument need not adopt that conclusion.
Nor does it require that constraints on means be absolute. A threshold view may hold that a sufficiently weighty end sometimes justifies a means otherwise excluded. Even then the end does not make the means question disappear; a further argument is needed to show why the threshold has been crossed and why this means is authorized (compare Quinn 1989, 298–99, on doing and allowing as a nonabsolute, rights-sensitive constraint at the level of individual agency).
Now add the observation that motivates the test. Suppose the state lacks authority to require condition Y by saying “you must maintain Y.” Suppose it then closes every timely route by which Y could be avoided and says: “we have required nothing; we have only prohibited some acts.” If the second raises no question the first raises, a means-limit can be evaded by regulatory form. The relevant inquiry is therefore not only what has been prohibited? but which routes around Y has this regulatory set legally closed?
The coercion literature is worth flagging without being made load-bearing: assessing an actor’s effect on another requires selecting a baseline (Nozick 1969), moralised baselines require an independent account of entitlements (Wertheimer 1987, 217–19), and Anderson (2010, 1–3) directs attention to the powers and actions of the coercing institution rather than to felt pressure. Section 10 returns to this; the argument does not depend on classifying abortion restriction as coercion in any technical sense.
Winston v. Lee (1985) deserves one sentence of precision, because it is often overstated. It applies a reasonableness balance, not a categorical limit on means; what it shows is only that the state’s valid law-enforcement objective did not settle whether compelled surgery could be used to pursue it, and that the bodily means received separate scrutiny. That is weaker than the argument needs and sufficient for it: distinguishing authority over an end from authority over a particular means is not an exotic move.
3. Person-Use Authority
The compliance test does not apply to every burden produced by law. A separate argument must show that the bodily condition is relevant to public authority. Continued gestation belongs to a family of person-use authority: public power assigning a person’s body, labor, or physiological capacities to another’s need or a public end. Conscription, quarantine, compelled procedures, compelled labor, and compelled gestation belong to that family.
Conscription is the hardest comparison. It is prolonged, dangerous service of an identified person for others’ protection; once substitution was abolished, the burden was institutionally nontransferable. States have claimed this authority.
The Selective Draft Law Cases (1918) make the comparison cut both ways. The Court relied on the enumerated army power and a reciprocal obligation of citizenship. That was not a bespoke doctrine of bodily use; taxation and jury service can also be defended through public power and reciprocal civic obligation. The analogy therefore sets a modest but real burden. Person-use needs a public-enforcement principle connecting a legally constituted role or relation to compulsory service.
The paper has granted a moral parental duty, and existing law shows that parenthood already can be a publicly enforceable role. Criminal neglect and nonsupport statutes require parents to furnish necessities or support under threat of punishment—for example, California Penal Code § 270 and New York Penal Law § 260.06. The argument therefore cannot rest on the claim that parental obligation is merely private.
Those laws also reveal the remaining step. They regulate externally dischargeable outputs—food, shelter, medical care, supervision, or financial support—which can ordinarily be supplied through purchase, delegation, shared care, or third parties. California § 270 includes a fetus-inclusive clause for fixing when the statutory support duty attaches, but the duty it enforces remains the furnishing of external necessities rather than continuation of gestation. Likewise, laws regulating harmful prenatal conduct prohibit acts or omissions; they do not by that fact establish authority to require the body’s continuing physiological support. These rules do not by themselves authorize compelled surgery, organ transfer, or the continued internal physiology of the parent. Conscription shows that a public role can sometimes support compulsory personal service; neglect law shows that parenthood can support coercive external-care duties. Together they define the opponent’s route: establish that the public parental role extends from externally dischargeable support to this nontransferable internal physiological service. If that principle is established, the paper loses under Section 14. The comparison does not show the bar is high; it identifies the bridge that must be defended.
Gestation remains a demanding instance of person-use. It requires not a discrete task but the continued cardiovascular, metabolic, endocrine, and immunological functioning of a particular body. Its nontransferability is biological, and one dependent life is sustained through the interior of one person—the structure Woollard’s account of normative imposition identifies and Kingma and Woollard apply to pregnancy (2025, 296–98). Imprisonment restricts liberty without placing the prisoner’s physiology at another’s use; a blood draw is invasive but discrete.
The closure analysis matters because conscription is imposed by direct command, while gestational use may be produced indirectly by closing alternatives. Indirect form does not eliminate the need for a role- or relation-based enforcement principle.
Capacity. The argument assumes a person who understands the options and communicates a stable choice free of deception or threat. The state may not decide whether her refusal counts by judging her reasons. Incapacity and proxy authority are outside this paper.
4. Granting the Duty to Gestate
The strongest objection to body-use arguments is not personhood. It is responsibility.
The objection observes that the fetus is not a stranger. Pregnancy standardly results from voluntary conduct, and the dependent being is one she has had a hand in bringing into existence and to which she stands as a parent. Thomson’s violinist is attached by strangers; that is what makes the case clean, and what makes critics doubt it transfers (Beckwith 2007; Lee and George 2007; Kaczor 2022).
Replies are available: Boonin (2003) argues that responsibility for another’s need does not generate a right to the responsible party’s body, and Hartley and Lindsley-Kim (2025) that gestation is not simply another instance of parental caregiving; Olsaretti (2025) presses the intentional-pregnancy cases, Gheaus (2025) presses the other way.
This paper does not adjudicate the dispute, because it grants the conclusion its opponents want.
Suppose responsibility generates a duty to continue the pregnancy. Three questions remain distinct: why the duty arose, what it requires, and what the state may compel so that it be fulfilled. The first two answers do not deliver the third. A directed claim identifies what one person may be owed by another; it does not, without an additional premise, specify every mechanism by which the state may secure performance. One can consistently hold that the fetus has moral standing, that she owes continued gestation, that refusal is wrongful, and that the state may ordinarily prohibit wrongful killing, while still asking whether public authority extends to securing the duty through continued internal physiological use.
Kingma and Woollard (2025, 312–14) make this separation harder to dismiss as a device invented for this paper, since they too distinguish the justification of obligations from the further justification of sanctions and interference. A rival can still defeat this step, but must supply the enforcement bridge: a general principle of enforceable directed parental duties, a pregnancy-specific principle of public jurisdiction over gestational physiology, or another principle reaching the same result.
For the hard case the paper also grants, for purposes of argument, that the relevant method intentionally kills the fetus.1 That concession strengthens the state’s ordinary reason for prohibition; it does not answer the means question.
One consequence is worth stating plainly. The argument does not require that a refusal to continue gestating be morally innocent. She may be acting wrongly; the fetus may have a genuine complaint; a serious duty may go unfulfilled. Nothing that follows depends on denying that. The argument is about what the state may do, not what she should do.
5. Ending Support Does Not Automatically Authorise Death
The argument becomes too easy if ending a pregnancy is treated as nothing but the withdrawal of aid.
Sometimes it is not. Some methods involve an intervention directed at the fetus, and a critic may reasonably insist that ending bodily support and causing death are different things that coincide in most cases. Thomson recognised the distinction: her argument licenses unplugging from the violinist, not killing him if he survives separation. That marks the boundary of what a body-use argument establishes — a right to stop one’s body being used is not a right to secure every outcome one might prefer once the use has ended.
This paper adopts the same boundary. If the fetus has moral standing, that supplies a reason against intentionally killing it which is independent of any claim to bodily support; Kamm (1992) develops the surrounding distinctions, which this paper does not settle. Suppose the critic is right that a given method intentionally kills, and that the state may ordinarily prohibit intentional killing. The question then becomes harder, not easier — and it becomes the question of Section 8.3.
6. Viability, Preservation, and Authorization
Viability matters here, but not because fetal life acquires value at a threshold, and not because the pregnant person’s authority over her own medical treatment lapses at one. It matters because it can change what is jointly achievable — but only through a further bodily intervention she must authorize. In one situation, every course that can begin ending the pregnancy now will result in fetal death. In another, a fetal-preserving course can begin now, but ending the pregnancy through it still requires induction, surgery, or another intervention on her body.
Those are structurally different problems, and the second is not one in which bodily authorization has disappeared. A claim against intentional fetal killing can sometimes be honoured without continued gestation, yet the preserving route remains a medical intervention that cannot be performed on a competent person who refuses it.
Kim (2026) supplies the strongest recent statement of the resulting complaint. Her Invasiveness Claim treats unwanted childbirth and fetal-preserving removal procedures, including induction and caesarean section, as capable of violating a stringent bodily right; combined with a Thomsonian account of self-defence against an innocent perpetrator, she conditionally defends feticidal abortion after viability (Kim 2026, 4–6). This paper accepts the first point and lets it structure the Protected-Refusal Case below: leaving a preserving course legally open may leave open only a course the patient has a stringent right to refuse. The preserving course is not treated here as costless or compulsory. The paper need not adopt Kim’s self-defence route, which depends on further premises about innocent perpetrators that the concession strategy is designed to avoid.
This is why the target range differs from Pallikkathayil’s. Her equality-based argument is expressly directed to pre-viability prohibition (Pallikkathayil 2025, 473–75); the compliance test does not treat viability as an authority threshold. It asks which timely courses exist, what bodily intervention each requires, what the patient authorizes, and what closes each. The post-viability range is therefore not one in which bodily conflict vanishes; it is the range in which preservation, authorization, and state closure come apart.
7. Stating the Facts: Courses, Times, and Refusal
Terms such as “reasonable,” “available,” and “medically appropriate” conceal the facts on which the argument turns.
Let T₀ be the earliest time at which the patient would seek and authorize a course and could clinically receive it absent the regulatory set. T₀ may be evidentially difficult because anticipated restrictions shape referral and care-seeking behavior; the observed presentation date cannot simply be assumed to be the counterfactual T₀. Applications require evidence from the patient’s plans, referral history, clinician testimony, and the clinical timeline.
Where the state says another course remains, that course must:
- be recognized in relevant clinical practice;
- apply to this patient’s medical circumstances;
- be capable of beginning at T₀ rather than a later T₁;
- end the pregnancy;
- not be prohibited by the regulatory set;
- be performable if authorized, apart from constraints created by that set; and
- be subject, under the governing law, to a legally controlling refusal.
Condition 3 prevents eventual delivery from masquerading as a present alternative. A course available only at T₁ does not answer closure at T₀. If the regulatory set closes the only course during [T₀, T₁], it produces interval-specific, delay-generated closure even if preservation later becomes possible.
Condition 7 is an application premise, not a universal background guarantee. In re A.C. (1990) held that a competent patient’s wishes govern in virtually all cases while leaving open extraordinary exceptions; Pemberton (1999) upheld a compelled cesarean in what it regarded as an overwhelming case. If refusal is overridden and B is performed over objection, the case leaves the closure taxonomy: it is direct compelled intervention, analyzed under Sections 2 and 3 because the intervention itself is the bodily means. If override is legally available but not exercised, the application must explain why B remains unused.
Viability is likewise a clinical range shaped by gestational duration and other circumstances (ACOG and SMFM 2017), not a single switch.
A limitation. Clinical pathways may not divide cleanly into rival courses. A disputed component can serve both procedural safety and fetal survival. Where the component is not severable, the cases below do not apply until the restricted object is identified.
8. Three Cases
Let A be a course that would end the pregnancy and intentionally cause fetal death, and B a fetal-preserving course satisfying conditions 1–7 at T₀. Retain the maximal assumptions throughout: full fetal standing, ordinary authority to prohibit intentional killing, a duty to gestate.
8.1 The Authorized-Preservation Case
She authorizes B. The regulatory set prohibits A; B proceeds; the pregnancy ends; the fetus survives.
Here the test returns nothing further for the state to justify. The set is not the legal closure reason for any timely course she would authorize, and no bodily condition is made necessary for compliance: the fetal claim has made a legal difference without generating a person-use problem. The case is worth stating separately because it shows what the test is — an inquiry that can clear a restriction, not a presumption against regulation wearing procedural clothes.
8.2 The Protected-Refusal Case
Now she refuses B, and by condition 7 her refusal is legally controlling. Then B cannot be performed: nothing about fetal standing or parental duty authorises operating on a competent person who refuses. That much holds without qualification.
But her refusal does not show that A must therefore be lawful. Her refusal answers: may B be performed on me? The answer is no. It does not answer: what has the state closed?
Attribution. If the regulatory set would permit B upon her authorization, that set has not closed B. Her refusal withholds the authorization B requires, and background law — excluded from the set for exactly this reason — gives that refusal effect. The obvious objection is causal: absent the prohibition on A the pregnancy would have ended, so the prohibition is a but-for cause of its continuing. True, and so is her refusal. Causal contribution does not determine the scope of authority exercised. The relevant distinction is between authorization and legal closure: her refusal determines whether B may be performed on her; the set determines whether B has been legally foreclosed. A refusal does not, without another legal act, add B to what the state has closed. This is set-specific, not a claim about the legal system as a whole: A’s prohibition, her refusal, and background consent law may together leave no executable route, while the set has still not closed every timely one.
The independent-ground condition. The state cannot rest on that attribution alone. Let H be the feature of A not shared by B that supplies the asserted ground for prohibition. Once there is a prima facie showing that restricting H predictably causes rejection of the remaining course and thereby leaves pregnancy continuing, the state bears the argumentative burden of showing that it would impose materially the same restriction on H in a relevantly matched case where restricting H does not produce continued pregnancy.2
That condition has been criticised as empty, and the criticism deserves a worked case rather than a reply.
Consider induction of fetal asystole, a discrete step performed before some abortions, for which current guidance identifies potassium chloride, lidocaine, and digoxin as acceptable agents (SFP and SMFM 2024). Construct the matched comparison: a patient has already authorized an immediate induction that will end the pregnancy at T₀ regardless of what else is done, and the state prohibits induced fetal asystole while permitting that induction to proceed. Restricting H there does not prolong anyone’s pregnancy by a day. Would the state restrict H in that configuration too?
If it would, fetal protection is independently operative, and the restriction is not merely the instrument through which continued gestation is produced. If the restriction appears only where removing H predictably leaves the patient pregnant, that pattern is evidence the other way. Regulatory coverage, exemptions, and the treatment of analogous instances of H are the evidence.
Three qualifications. The comparison is a clinically grounded hypothetical; I have not identified an enacted jurisdiction whose law has this structure. The severability limitation of Section 7 bites directly, since induction of fetal asystole can bear on the conduct of the procedure as well as on fetal survival, and where it is not severable H cannot be isolated. And if no matched case can be constructed at all, the condition cannot be satisfied and this case collapses toward Section 8.3 rather than standing as a separate result. An empty comparison class is not a gap in the test; it is a finding.
What remains after the attribution succeeds. Suppose the state satisfies the condition. It does not follow that no authority question arises.
Several things are true at once. Her continued pregnancy is not consented to; her options have been narrowed by law; and, by hypothesis, the one exit left open is an intervention she has a stringent right to refuse, which is what Kim’s Invasiveness Claim supplies. The state has therefore done more than leave an ordinary alternative standing. It has structured her option set so that every exit from continued gestation runs through a second protected bodily interest; she may avoid the first only by surrendering the second. That is not complete closure and should not be described as such. But it is not nothing:
May the state so arrange matters that avoiding continued gestation requires accepting a bodily intervention the person has a legally controlling right to refuse?
This is a substantive partial result. By removing A while leaving B available only if she waives a legally controlling refusal right, the state has imposed a conditional choice between two protected bodily interests. Means-specificity therefore requires a justification of that arrangement. The paper does not decide whether the justification succeeds; it establishes the burden and distinguishes it from complete closure.
Where the regulatory set leaves B legally open conditional on authorization, satisfies the independent-ground condition, and the courses are genuinely distinct, refusal does not establish complete closure. It establishes a different authority problem: whether the state may condition escape from continued gestation on acceptance of an intervention the person is entitled to refuse.
8.3 The Single-Course Closure Case
Now remove B.
Let Z be the only course that can, on the medical facts of this case and at T₀, begin to end the pregnancy, and suppose Z intentionally causes fetal death. The regulatory set prevents Z through rules that may bind the patient, the clinician, a facility, or another gatekeeper. The identity of the addressee matters and should be stated; for this inference the stipulated legal effect is the same.
Map the timely courses. There is no B and no other course that can presently end the pregnancy. Z is the only exit, and the set closes Z. Unlike Section 8.2, no second timely course has a nonuse attributable to her refusal or an independent constraint: the set is the legal closure reason for every timely route at T₀.
Her body consequently continues to perform the work that sustains the fetus. The claim is not that she personally violates the law by ceasing to gestate, but that the arrangement has removed the only legally available route through which she could presently stop.
Two descriptions are now both accurate. First: the state prohibits an act that intentionally kills a person. Second: if she obeys, her body must continue sustaining that person, because the state has closed the only course by which it could presently stop.
Authority for the first is granted. The question is whether it extends to the second. The answer cannot be that it does so automatically. If direct imposition of continued gestation would require an enforcement premise the state has not supplied, then reaching the same condition by closing the alternatives does not supply it either; the form of the arrangement is not a source of authorization. Whether direct imposition would require such a premise is not assumed here — Section 4 leaves the parental enforcement bridge open, and Section 14 lists its construction as a defeater. The argument for the conditional is in Section 9.
So the state owes a different argument here from Section 8.2’s independent-ground showing. Fetal standing does not supply it, because standing identifies the protected claim rather than the permissible means; the granted parental duty does not, because Section 4 leaves the enforcement premise open. The wrongness of killing supplies a reason for prohibiting Z; it does not, without that premise, settle authority to make continued gestation legally unavoidable.
9. “The State Is Only Prohibiting Killing”
The strongest objection should be stated at full strength.
The state does not require anyone to be pregnant. It prohibits an act. The pregnancy continues because the pregnant person refrains from killing another person, which is what everyone is required to do. Prohibitions routinely leave people in circumstances they would rather escape — the prohibition on theft leaves the hungry person hungry — and no one concludes that the state has appropriated their circumstances. To treat this prohibition differently is to smuggle in the conclusion that the fetus’s life does not count.
This is the deepest rival and deserves a direct answer rather than a redescription. Begin by granting what is true in it: the state really is prohibiting an act that, on our assumptions, intentionally kills a person, and Section 5 declined the move that would have made this easy. But one arrangement can have more than one feature bearing on the authority it exercises. On the facts of Section 8.3, it both protects the fetus against an act that would kill it and makes the continuation of a particular person’s bodily functioning necessary for compliance. The objection selects the first feature and treats it as exhausting what must be justified. That is a premise, not a datum, and it is the premise the compliance test denies.
Not every consequence of a rule enters the test. Three things must be established. First, complete closure: among the timely courses that would avoid the condition on the stipulated facts, the minimally sufficient regulatory set is the legal closure reason for each; a course closed instead by legally controlling refusal, an independent clinical fact, or genuinely independent provider unwillingness does not count. Provider unwillingness is independent only if it would persist absent the regulatory set. Where vague standards, overlapping sanctions, credible enforcement threats, licensing exposure, or facility penalties foreseeably cause providers to withdraw, that withdrawal is part of the set’s operation rather than an exogenous fact. Colautti v. Franklin (1979) illustrates the attribution problem: uncertain abortion standards backed by criminal liability created notice, enforcement, and chilling concerns for physicians. Second, set dependence: removing a member removes a legal closure necessary to the condition’s persistence. Third, the condition must have independent relevance to the scope of public authority — supplied here not by unpleasantness or causal consequence but by Section 3’s account of person-use authority.
The theft comparison should therefore be stated narrowly. Ordinarily a prohibition on theft is not the legal closure reason for every route to food, so the first condition fails. If a hypothetical regime truly closed every route by which a person could avert imminent starvation, the test would flag a further authority question only if an independent account also established the relevant bodily condition as authority-sensitive — and it would not follow that theft must be permitted. The test identifies an additional justificatory question; it does not answer it.
Stated generally: if the state has authority to prohibit X, the regulatory set is the legal closure reason for every timely course by which Y could be avoided, and an independent argument establishes Y as authority-sensitive, authority over X does not by itself answer the question about Y.
A critic can reject this at the foundation by defending a formal-object theory of public authority: only the act expressly prohibited matters, and necessary compliance conditions have no separate relevance. If that is right, the test fails. It also follows that a means-sensitive limit can be evaded whenever the state can secure the same practical condition by prohibiting alternatives rather than commanding it directly. The paper denies that regulatory form alone can do that justificatory work.3
10. The Baseline
The state did not cause the pregnancy. Why say its restriction makes pregnancy compulsory rather than merely refusing an escape from a condition it did not create?
Any claim that a rule changes a situation requires a baseline (Nozick 1969). The comparison here is narrow: holding the stipulated clinical, provider, authorization, and temporal facts fixed, identify which timely courses the regulatory set closes. In Section 8.3 a course would be available absent the set and is unavailable because of it. That is the change attributed to public power.
This baseline also limits the argument. If no clinician would provide the course, it is medically inapplicable, or another authorized course remains, the set has not produced complete closure. It is moralized in Wertheimer’s sense because it depends on what the actors are entitled to do, not prediction alone (1987, 217–19). Section 3 supplies the claimed entitlement against ungrounded person-use. If that argument fails, so do the baseline and the paper.
11. Doing, Allowing, and Normative Imposition
Nothing above requires the claim that abortion is merely the withdrawal of aid. Suppose Z is an intentional killing, the doing/allowing distinction is morally significant, and the state may prohibit killing far more readily than it may compel rescue. The argument is unaffected, because what remains to be shown is whether authority to prohibit this killing includes authority over the bodily condition compliance makes unavoidable.
The doing/allowing literature does more than leave that question standing; it sharpens it. Woollard (2015, 101) grounds the doctrine in imposition: constraints on doing harm, and permissions to allow it, are what make it true that a person’s body and resources are hers rather than a standing resource for others’ needs. She distinguishes causal imposition from normative imposition, which occurs when a moral requirement itself makes a person’s body available for others’ needs. Kingma and Woollard (2025, 296–98) bring this into pregnancy, treating the pregnant person’s bodily resources as within her protected sphere and asking when restrictions on that barrier amount to normative imposition.
The extension proposed here is from that moral-level analysis to public enforcement through legal compliance conditions. When a regulatory set makes continued bodily functioning necessary for compliance, it creates the institutional form of the same question — and the two levels come apart in a way that shows the extension does work rather than restating its source. At the moral level, the person in the Protected-Refusal Case and the person in the Single-Course Closure Case may face the same demand on the same body. Institutionally they do not face the same state: in the first it has imposed a choice between two protected interests, in the second it has removed the choice. The moral right to refuse B does not tell us which public-authority problem the arrangement has created.
The same point prevents the doing/allowing distinction from settling the paper by stipulation. The critic who moves from the intentional-killing characterization to the conclusion that the state may prohibit Z regardless of what compliance requires has supplied the protected end but not the enforcement bridge.
12. What Survives If This Means of Enforcement Fails
Suppose the argument succeeds on the facts of Section 8.3. What follows about the fetus’s claim?
It does not disappear. The conclusion is not that the fetus’s death does not matter, or that she has done nothing wrong, or that she had no duty, or that the loss is not a loss. The conclusion is that this means of enforcing the claim has not been justified.
A claim can remain valid when a particular route to enforcing it is unavailable: obligations of gratitude and fidelity are real and largely unenforced, and much serious wrongdoing lies outside what third parties may coercively prevent. Nor does a limit on what one person’s body may be made to do discharge everyone else.
At exactly this point a tempting inference appears. The worse the remaining harm, the stronger the pull of the thought: something terrible is unresolved; therefore someone must be required to resolve it; therefore the person whose body could resolve it must continue.
That inference is invalid, and identifying why is part of the argument. An unmet need shows that something of moral importance has not been secured; it does not show who owes what. Responsibility has to come from somewhere — a relationship, an undertaking, a role, a contribution to the danger, a legal assignment — and once established, the seriousness of the need can bear on how demanding it is. But seriousness does not generate the responsibility, and it does not regenerate an authority just found wanting.
The tradition on moral conflict supplies what is needed and no more (Williams 1965, 1981; Marcus 1980; Stocker 1990; Gowans 1994; Tessman 2015): a conclusion about what the state may enforce need not cancel the moral importance of what it cannot secure, and the persistence of that importance does not supply the missing authority. Otherwise the argument confronts a false choice — compel gestation, or deny anything of importance has been lost. Neither horn is required.
13. What the Argument Establishes
If the premises are granted, five propositions follow.
A moral duty to continue a pregnancy does not, without an additional premise, establish authority to compel the bodily performance through which the duty would be fulfilled.
Where a fetal-preserving course is authorized and performed, the compliance test returns no further authority question. The test can clear a restriction.
Where a preserving course remains legally open only on condition that the patient waive a controlling refusal right, refusal does not establish complete closure. The state has nonetheless imposed a conditional choice between two protected bodily interests and must justify that arrangement.
Where the set closes the only course by which the pregnancy could presently begin to end, continued pregnancy is what compliance requires on those facts.
Authority to prohibit the fetal death does not, by itself, establish authority over the continued bodily support that compliance then requires.
Propositions 2 and 3 are not concessions extracted from the argument; they are results of it, and they are what distinguish the compliance test from a general presumption against abortion regulation.
14. What Would Defeat the Argument
The general means premise fails. A critic might defend a principle under which authority to prohibit X automatically includes authority over every bodily condition compliance makes necessary.
Person-use authority is not a distinct category. A critic might deny that Section 3’s structure matters to the scope of public authority — that continued gestation is, for this purpose, on a footing with any burdensome consequence of obeying a law. The extension from normative imposition to means-indexed institutional authority would fail, and Section 10’s baseline with it.
A parental or procreative enforcement principle is established. Existing neglect and nonsupport law already establishes that parenthood can be a public-enforcement role, while conscription shows that some public roles can require personal service. A critic might extend those premises by showing that parental office authorizes not only externally dischargeable care and support but compulsory, nontransferable internal physiological service. If that extension is established, the argument fails; the paper does not assume the bar is high.
The formal object exhausts the authority inquiry. A critic might defend the view in Section 9. That rejects the test at its foundation, and also defeats the independent-ground condition, since why the state selected H would drop out once authority over H were established.
The regulatory set cannot be individuated. The NESS-style union resolves multiple overlapping sufficient sets, but the test fails if membership depends on arbitrary framing of courses, if public closure cannot be separated from background law, if regulation-induced provider withdrawal cannot be distinguished from independent unwillingness, or if the union becomes unbounded.
The facts fail in the particular case. An application fails if Section 7’s conditions are unmet: a lawful course remains at T₀, the supposed alternative can begin only at T₁, no willing qualified provider exists independently of the set, refusal is not legally controlling, or the clinical course cannot be decomposed enough to identify the restricted feature.
15. Conclusion
This paper has granted full fetal standing, parental responsibility, a duty to gestate, and ordinary authority against intentional killing. It has asked what remains to be shown about enforcement.
The answer depends on institutional structure. The compliance test identifies the timely authorized courses, the overlapping public rules that close them, and whether those rules contain sufficient closure sets for every route. Where preservation is authorized, no further problem arises. Where preservation is protectedly refused, the state has not completely closed the exit but has imposed a conditional choice between two protected bodily interests, which it must justify. Where the regulatory set closes every timely course, continued gestation is the bodily condition it makes unavoidable.
Fetal standing establishes the protected claim, not the permissible means. The granted parental duty does not by itself establish the public-enforcement relation; conscription shows what a role-based argument would have to supply. The intentional-killing classification supplies a reason to prohibit, not automatic authority over every bodily condition the regulatory scheme makes unavoidable.
Person-use authority requires a principle addressed to the use. Layering criminal, licensing, and facility rules cannot evade that requirement, and reaching the result indirectly does not answer the question direct compulsion would raise. The state is answerable both for the object of its prohibitions and for the bodily conditions its rules together make unavoidable.
References
American College of Obstetricians and Gynecologists and Society for Maternal–Fetal Medicine. 2017. “Obstetric Care Consensus No. 6: Periviable Birth.” Obstetrics & Gynecology 130(4): e187–e199. Reaffirmed 2025.
Anderson, Scott A. 2010. “The Enforcement Approach to Coercion.” Journal of Ethics and Social Philosophy 5(1): 1–31.
Beckwith, Francis J. 2007. Defending Life: A Moral and Legal Case Against Abortion Choice. Cambridge: Cambridge University Press.
Boonin, David. 1997. “A Defense of ‘A Defense of Abortion’: On the Responsibility Objection to Thomson’s Argument.” Ethics 107(2): 286–313. [Published as Boonin-Vail.]
Boonin, David. 2003. A Defense of Abortion. Cambridge: Cambridge University Press.
Gheaus, Anca. 2025. “Procreative Liability and Equality before the Law.” Journal of Applied Philosophy 42(2): 499–504.
Gowans, Christopher W. 1994. Innocence Lost: An Examination of Inescapable Moral Wrongdoing. New York: Oxford University Press.
Hartley, Christie, and Ashley Lindsley-Kim. 2025. “Pregnancy, Caregiving, and a Supposed Obligation to Gestate.” Journal of Applied Philosophy 42(4): 1301–1316.
Kaczor, Christopher. 2022. The Ethics of Abortion: Women’s Rights, Human Life, and the Question of Justice. 3rd ed. New York: Routledge.
Kamm, F. M. 1992. Creation and Abortion: A Study in Moral and Legal Philosophy. New York: Oxford University Press.
Kim, Jina. 2026. “A Defense of Post-Viability Abortion.” Bioethics, advance online publication, PDF pp. 1–7. https://doi.org/10.1111/bioe.70140.
Kingma, Elselijn, and Fiona Woollard. 2025. “Can You Do Harm to Your Fetus? Pregnancy, Barriers, and the Doing/Allowing Distinction.” Ethics 135(2): 290–319.
Koppelman, Andrew. 1990. “Forced Labor: A Thirteenth Amendment Defense of Abortion.” Northwestern University Law Review 84(2): 480–535.
Lee, Patrick, and Robert P. George. 2007. Body-Self Dualism in Contemporary Ethics and Politics. Cambridge: Cambridge University Press.
Little, Margaret Olivia. 1999. “Abortion, Intimacy, and the Duty to Gestate.” Ethical Theory and Moral Practice 2(3): 295–312.
Marcus, Ruth Barcan. 1980. “Moral Dilemmas and Consistency.” Journal of Philosophy 77(3): 121–136.
Nozick, Robert. 1969. “Coercion.” In Philosophy, Science, and Method: Essays in Honor of Ernest Nagel, edited by Sidney Morgenbesser, Patrick Suppes, and Morton White, 440–472. New York: St. Martin’s Press.
Olsaretti, Serena. 2025. “Bodily Rights in Intentional Pregnancies.” Journal of Applied Philosophy 42(2): 505–513.
Pallikkathayil, Japa. 2025. “Abortion and Democratic Equality.” Journal of Applied Philosophy 42(2): 473–491.
Quinn, Warren S. 1989. “Actions, Intentions, and Consequences: The Doctrine of Doing and Allowing.” Philosophical Review 98(3): 287–312.
Ripstein, Arthur. 2009. Force and Freedom: Kant’s Legal and Political Philosophy. Cambridge, MA: Harvard University Press.
Shiffrin, Seana Valentine. 2025. “Old Enough to Carry, Old Enough to Vote.” Journal of Applied Philosophy 42(2): 492–498.
Society of Family Planning and Society for Maternal–Fetal Medicine. 2024. “Clinical Recommendation: Induction of Fetal Asystole before Abortion.” Contraception 140: 110551.
Stocker, Michael. 1990. Plural and Conflicting Values. Oxford: Clarendon Press.
Tessman, Lisa. 2015. Moral Failure: On the Impossible Demands of Morality. New York: Oxford University Press.
Thomson, Judith Jarvis. 1971. “A Defense of Abortion.” Philosophy & Public Affairs 1(1): 47–66.
Viehoff, Daniel. 2025. “The Force of Equal Treatment.” Journal of Applied Philosophy 42(2): 514–524.
Wertheimer, Alan. 1987. Coercion. Princeton: Princeton University Press.
Williams, Bernard. 1965. “Ethical Consistency.” Proceedings of the Aristotelian Society, Supplementary Volume 39: 103–124.
Williams, Bernard. 1981. Moral Luck: Philosophical Papers 1973–1980. Cambridge: Cambridge University Press.
Wright, Richard W. 1985. “Causation in Tort Law.” California Law Review 73(6): 1735–1828.
Woollard, Fiona. 2015. Doing and Allowing Harm. Oxford: Oxford University Press.
California Penal Code § 270.
Colautti v. Franklin, 439 U.S. 379 (1979).
Church of the Lukumi Babalu Aye v. City of Hialeah, 508 U.S. 520 (1993).
Department of Commerce v. New York, 588 U.S. 752 (2019).
In re A.C., 573 A.2d 1235 (D.C. 1990) (en banc).
New York Penal Law § 260.06.
Mt. Healthy City School District Board of Education v. Doyle, 429 U.S. 274 (1977).
Pemberton v. Tallahassee Memorial Regional Medical Center, 66 F. Supp. 2d 1247 (N.D. Fla. 1999).
Selective Draft Law Cases, 245 U.S. 366 (1918).
Winston v. Lee, 470 U.S. 753 (1985).