Abortion, Abandonment, and Positive Rights:

The Limits of Compulsory Altruism*

by Roderick T. Long

[Originally published in Social Philosophy & Policy 10, no. 1 (Winter 1993), pp. 166-191.
© 1993 Social Philosophy & Policy Foundation; posted by permission of the Foundation.]

I. Introduction
II. Rights: Some Preliminaries
III. The Problem
IV. Basic Versus Derivative Positive Rights
V. Welfare Rights: A Digression
VI. The Right Not To Be Abandoned: A Derivative Positive Right
VII. The Right To Abort: A Derivative Positive Right
VIII. Summary
Notes]

I. Introduction

Let us imagine that we accept the following three propositions:

(1) Every person has a right not to be treated as a mere means to the ends of others.¹

(2) A woman who voluntarily becomes pregnant but later changes her mind has the right to abort her unborn child.

(3) A woman who voluntarily bears a child but later changes her mind does not have the right simply to abandon her child, but must care for it until she can arrange for a substitute caretaker.

This is particularly easy for me to imagine, since I do in fact accept them. Not everyone will be so inclined, for the propositions are hardly uncontroversial. Nevertheless, since many people do accept them, it is worthwhile finding out whether they are compatible; for there are good prima facie reasons for believing that there is, if not a logical inconsistency, then at least a serious tension among these three propositions – a tension that might well make us doubt the possibility of a plausible moral perspective that could accommodate all three. Defenders of the three propositions thus have an interest in showing that they are compatible; and opponents of (some or all of) the propositions might well have an interest in showing that they are incompatible.

My first aim in this essay, then, is to explain why the three propositions might seem to be incompatible, and to argue that there is nevertheless a plausible moral view which reconciles them.² My second and broader [167] aim is to take some preliminary steps toward answering the question: Under what circumstances is it legitimate to force someone to act on behalf of the welfare of others? In other words, my second aim is to sketch a theory of positive rights, and so to determine the limits of compulsory altruism. These two aims are connected: the first aim depends on the second because my solution to the apparent incompatibility of the three propositions depends on my views about positive rights; and conversely, my second aim depends on the first because I regard it as a desirable constraint on any theory of positive rights that it give (what I regard as) the “right answer” to the questions raised by the three propositions. Moreover, in treating the issues of abortion and abandonment I hope to give some content to my theory of positive rights by showing how it handles important moral dilemmas. The question “How much may we ask a woman to do for her unborn or newborn child?” is a special case of the broader question “How much, in general, may we ask one person to do for another?” An attempt to answer the general question should help to illuminate the specific one, and vice versa; and our responses to these two questions must be brought into harmony with one another if we are to have any confidence in either. Hence the dual focus of this essay.

II. Rights: Some Preliminaries

The three propositions under consideration employ the language of rights. We ought to get clear, then, about what we are saying when we ascribe (or deny) rights to people. I take it that having a right typically involves one or more of the following components:

(A) The moral value, or at least the moral permissibility, of one’s doing/having that to which one has a right.

(B) The obligation of others to let/help one do/have that to which one has a right.

(C) The legitimacy of forcing others to let/help one do/have that to which one has a right.

From these components it is possible to construct seven distinct varieties of rights: A-rights, B-rights, C-rights, AB-rights, AC-rights, BC-rights, and ABC-rights.³ [168]

Of which sort are the rights at issue in our three propositions? I think they are best understood as political rights, in the broad sense of being those rights that a just political system would respect and enforce.⁴ And if such rights are to be supported by legal sanctions, they will have to include component C. Furthermore, it would be difficult to characterize as just a political regime whose prescriptions and proscriptions one generally had no moral reason (apart from fear of punishment) to accept; so we may also assume that political rights (usually or always) include component B. I shall therefore take the rights at issue to be BC-rights,⁵ and by “right” without prefix I shall henceforth mean “BC-right.”

My having a right (i.e., a political or BC-right), then, involves two components: the obligation of others to treat me in a certain manner, and the legitimacy (moral permissibility) of forcing them to treat me in that manner. Let us call an obligation enforceable just in case it would be legitimate to force the bearer of the obligation to act in accordance with it.⁶ We may thus define the (relevant) notion of right as follows:

(4) S has a right against O to be treated in manner M if and only if O has an enforceable obligation to treat S in manner M.

(5) S has a right (simpliciter) to be treated in manner M if and only if, for every O, S has a right against O to be treated in manner M.

So defined, a right is a right to be treated in a certain way; the object of the right is a certain sort of behavior on the part of other people. [169]

Now we surely do sometimes think of rights as having quite different objects; sometimes the object of the right is a potential action of its bearer (I have the right to vote, or to speak freely), and sometimes the object is a thing (I have the right to food, or to the money I earn). But these rights may all be understood in terms of the right to certain sorts of treatment; the right to take certain actions is the right to be allowed, or helped, to perform those actions; and the right to a thing is the right to be given, or not to be prevented from obtaining, that thing. Hence every right must specify what (enforceable) obligations it entails for others.⁷

The rights claimed in propositions (1) and (2), and denied in proposition (3), are thus not grants of moral permission. Rather, they are rights that assign other people enforceable obligations to treat the bearers of such rights in certain specified ways.

Moreover, the rights if one person entail the absence of certain rights in other people. Suppose I have the right to be treated in manner M. That means that other people have an enforceable obligation to treat me in manner M. So it would be legitimate (morally permissible) to force everybody to treat me in manner M. So there is no obligation (and a fortiori no enforceable obligation) not to force everybody to treat me in manner M. And so nobody has a right not to treat me in manner M. If other people did have such a right, then I would have no right against them to be treated in manner M. Claims ascribing rights to one person automatically yield claims denying other rights to everybody else.⁸

Rights (i.e., BC-rights) are commonly divided into positive and negative rights. One natural way of drawing the distinction might look like this: A positive right is one whose correlative obligation is the obligation to take some positive action on behalf of the right’s bearer; a negative right, on the other hand, is none whose correlative obligation is merely the obligation to refrain from interfering in some way with the right’s bearer. Rights are positive or negative insofar as they involve positive or negative obligations on the part of others, where the objects of positive and negative obligations are doings and allowings, respectively.⁹

(6) S’s right to be treated in manner M is a positive right if and only if manner M involves taking positive action on S’s behalf. [170]

(7) S’s right to treated in manner M is a negative right if and only if manner M does not involve taking positive action on S’s behalf.

Nevertheless, I suspect that (6) and (7) do not succeed in capturing our intuitions about positive and negative rights.¹⁰ Suppose I claim that I have a right to cut off you arm – by which I mean that you have an enforceable obligation to let me cut off your arm. By definitions (6) and (7), the right I am claiming is a negative, not a positive right. But this seems strange: surely the point of the distinction is that positive rights go beyond the mere right to be left alone in one’s personal space, as it were. Negative rights build fences around people; positive rights break down fences. When I claim the right to cut off your arm, the right I am claiming is more plausibly thought of as a positive, fence-breaking right, rather than a negative, fence-building one. I suggest, then, that definitions (6) and (7) be replaced by (8) and (9), which do a better job of capturing the intuitions behind the positive/negative distinction:

(8) S’s right to be treated in manner M is a positive right if and only if (a) manner M involves taking positive action on S’s behalf, or (b) manner M involves letting oneself be used by S.

(9) S’s right to treated in manner M is a negative right if and only if (a) manner M does not involve taking positive action on S’s behalf, and (b) manner M does not involve letting oneself be used by S.

Note that (8) is a disjunction, while (9) is a conjunction. With these preliminaries out of the way, let us turn to the heart of our problem.

III. The Problem

Proposition (1) says that we never have a right to treat people as mere means to the ends of others. And this in turn seems to entail that nobody ever has an enforceable obligation to let himself or herself be treated as a mere means to the ends of others. But if that is true, it is hard to see how anyone could ever have any positive rights. For if I have a positive right against you, then you have an enforceable obligation (not merely to refrain from interfering with my activities but) either to let yourself be used by me or to take some positive action on my behalf; and it follows that it would be legitimate for me (or a third party) to forcibly use you, or to force you to take such action on my behalf. But in so reducing you to an object, or to involuntary servitude, it seems that I am treating you as a mere means – and that, proposition (1) tells us, can never be legitimate. Hence preposition (1), we may well infer, is inconsistent [171] with the existence of positive rights.¹¹ And if there are no positive rights, then (1) yields the following negative right: the right not to be forcibly used, or to be forced to take positive action to promote the welfare of others.

Now proposition (1), understood as denying the possibility of positive rights, sits rather oddly with propositions (2) and (3). For proposition (2) endorses the right to abortion, while proposition (3) denies the right of abandonment. The problem is that abortion – causing the death of the unborn child – seems like a case of doing or killing, whereas merely abandoning the child after it has been born seems like a case of allowing or letting die. If it is a violation of a person’s rights to treat him or her as a mere means to one’s own ends, then surely it is a violation of an unborn child’s rights to abort it; for in aborting her child, is not the mother treating it as a mere means to her own ends? And on the other hand, if it is a violation of a person’s rights to force him or her to take positive action to promote the welfare of others, is it not a violation of the mother’s rights – indeed, a form of involuntary servitude – to require her to care for her child once it is born?

Proposition (1), in short, appears to require the rejection of both (2) and (3). Proposition (1) yields a negative right not to be killed (and so not to be aborted), thus contradicting (2); and (1) also denies any positive right to be cared for (and so not to be abandoned), thus contradicting (3). That is, proposition (1) seems to give exactly the opposite answers on the abortion and abandonment questions from those expressed in (2) and (3). Our three propositions have apparently turned out to be incompatible, and with a vengeance!

IV. Basic Versus Derivative Positive Rights

I think this conclusion is too quick. To see why it is too quick, we need to take a closer look at the argument from (1) to the denial of positive rights. That argument claimed that since X has an enforceable obligation not to treat Y as a mere means¹² – not to violate O’s negative right to freedom from servitude – X can have no positive rights against Y. But [172] what if it should be the case that, since X has an enforceable obligation not to treat Y as a mere means, Y will turn out to have positive rights against X?

The right not to be run over is a negative right if anything is. To run over somebody is to treat that person as pavement for one’s wheels, to subordinate him or her to one’s ends.¹³ Well and good. But suppose Frieda is careening around the corner in her Lamborghini and suddenly sees Roscoe ambling across the road a short distance ahead. At her current speed, she has only two choices: (a) she can run over Roscoe, or (b) she can swerve around him. Running over Roscoe would be a violation of his negative rights; so Frieda has an enforceable obligation not to run over him. In the circumstances, what that amounts to is an enforceable obligation to swerve around him. But that obligation is not an obligation to sit back and do nothing; it is an obligation to take positive action, and a very specific positive action at that: she must turn the steering wheel in order to avoid running over Roscoe. Here we have a case, then, in which Frieda has an enforceable obligation to take positive action on Roscoe’s behalf – but only because that is the only way she can avoid violating her negative obligation not to treat Roscoe as a mere means. Roscoe’s right against Frieda not to be run over has generated, in the circumstances, a further right against her that she turn the wheel in a certain way. The first right is a negative right, one licensed by proposition (1); the second right is a positive right, but since it follows from a negative right endorsed by (1), it seems that it too is endorsed by (1). Let us call such a right a derivative positive right, as opposed to a basic positive right.

(10) A derivative positive right is a positive right that is generated by the requirement of some negative right(s).

(11) A basic positive right is a right that is not derivative.

Does (1) rule out derivative positive rights? If so, then (1) is simply incoherent. For if everyone has a right against everyone else not to be treated as a mere means, then by the same token everyone has an enforceable obligation to take whatever positive steps are necessary to avoid treating others as mere means. If the end is mandatory, the means to that end are mandatory as well. If forcing people to take [173] such positive steps counts as treating them as mere means, then (1) cannot license any rights at all, positive or negative.

Fortunately, we need not draw this conclusion. The enforcement of positive rights involves treating others as means, but it need not involve treating others as mere means. Some such distinction must be available if (1) is to be plausible at all; after all, when I pay a barber to cut my hair, the barber and I are suing each other as means, but (Marxist worries aside) there is presumably nothing wrong with this. In order to make sense of (1), then, we need to be able to draw a distinction between the use of people as mere means and the use of people as nonmere means.

The barber example might suggest one natural way of drawing the distinction:

(12) S uses O as a mere means if and only if S uses O as a means without O’s consent.

(13) S uses O as a nonmere means if and only if S uses O as a means with O’s consent.

According to this way of drawing the distinction, forcing the barber at gunpoint to cut my hair would be a violation of (1), while persuading him (e.g. with offers of money) to cut my hair voluntarily would not.

The problem with this way of drawing the distinction is that it does not do what we wanted it to do: it does not make room for derivative positive rights. If Frieda has an enforceable obligation to turn the steering wheel and so avoid killing Roscoe, then it would be legitimate to force her to turn the steering wheel; but in forcing her to turn the wheel we would be using her as a means without her consent, and so – according to (12) – we would be using her as a mere means, thus violating (1). If there are derivative positive rights, then it is sometime legitimate to use people as means without their consent; so if we want to maintain the truth of (1), we need to find a way of drawing the distinction between mere and nonmere means other than that expressed in (12) and (13).

I think we will have a better shot at the idea we are trying to hit if we think of (1) as specifying certain boundaries around people. All actions that involve using S as a means (whether mere or nonmere) fall within the boundary around S; all actions that do not involve using S as a means lie outside S’s boundary.

(14) S’s action lies within O’s boundary if and only if S’s action involves using O as a means (whether mere or nonmere) to the ends of others.¹⁴ [174]

Now we cannot interpret (1) as prohibiting all actions that lie within another person’s boundary, for that would exclude derivative positive rights. Nor can we interpret (1) as prohibiting all actions that lie within another person’s boundary and are undertaken without that person’s consent; for that too would exclude derivative positive rights. But among actions that lie within another perosn’s boundary, we can draw a distinction between those that violate that boundary and those that merely repel the person from one’s own boundary.

Let us draw some distinctions:

(15) S crosses O’s boundary if and only if S performs an action within O’s boundary.

(16) S invades O’s boundary if and only if S crosses O’s boundary without O’s consent.

(17) S violates O’s boundary if and only if (a) S invades O’s boundary and (b) invading O’s boundary (in that way) is not necessary to end any boundary-invasion on O’s part.¹⁵

(I shall be using “action” broadly here, to include involuntary as well as voluntary behavior.) Now boundary-violations seem to be a clear case of treating others as mere means, and so run afoul of (1) if anything does. But what about boundary-invasions that are not boundary-violations? Is it legitimate to end S’s boundary-violation by invading S’s boundary in turn?

The dilemma here is analogous to that in Shakespeare’s The Merchant of Venice. Antonio owes Shylock a pound of his flesh, but he does not owe Shylock any blood; yet of course Shylock cannot claim his pound of Antonio’s flesh without taking some of Antonio’s blood as well. Portia reasons that since Antonio does not owe Shylock any blood, Shylock has no right to take Antonio’s blood, and so has no right to take the pound of flesh either (since, although he ahs a right to the pound of flesh, he cannot exercise that right without doing something he has no right to do – namely, take Antonio’s blood). Shylock, on the other hand, reasons that sine Antonio does owe Shylock the pound of flesh, Shylock has the right [175] to take it, and so has the right to take the blood too (since that follows from something Shylock has a right to). By analogy, we might distinguish two possible principles to govern boundary-invasions:

Shylock’s Principle: If S violates O’s boundary, O (or O’s agent) has the right to invade S’s boundary in whatever way is necessary to end S’s violation of O’s boundary.

Portia’s Principle: O (or O’s agent) has no right to invade S’s boundary, even if invading S’s boundary is necessary to end S’s violation of O’s boundary.

Is Portia’s Principle correct? If so, then there are no derivative positive rights. But we have seen that in accepting (1) we are committed to accepting the existence of derivative positive rights. Since we are looking for a view that accommodates (1), we must therefore reject Portia’s Principle.

Should we then accept Shylock’s Principle? Shylock’s Principle may be consistent with the letter of (1), but it certainly conflicts with its spirit. Am I justified in killing you if that is the only way to stop you from violating my boundary in some trivial way (e.g., stepping on my toe)? My response seems disproportionate¹⁶ to the seriousness of the violation; and insofar as it does, I might justifiably be taken to treat you as a mere means to my convenience, thus violating (1). I therefore suggest that the following principle does a better job of capturing the intuitions behind (1) than either Portia’s or Shylock’s Principle does:

Principle of Proportion: If S violates O’s boundary, O (or O’s agent) has the right to invade S’s boundary in whatever way is necessary to end S’s violation of O’s boundary, so long as O’s (or O’ agent’s) invasion of S’s boundary is not disproportionate to the seriousness of S’s violation of O’s boundary.

And if we accept the Principle of Proportion, the most plausible way of drawing the mere/nonmere distinction is as follows:

(18) S uses O as a mere means if and only if either (a) S violates O’s boundary or (b) S invades O’s boundary and either (b1) S’s invasion of O’s boundary is not necessary to end any wrongful boundary-invasion on O’s part, or (b2) S’s invasion of O’s boundary is disproportionate to the seriousness of any wrongful boundary-invasion on O’s part. [176]

(19) S uses O as a nonmere means if and only if (a) S crosses O’s boundary and (b) S does not violate O’s boundary and either (c1) S does not invade O’s boundary or (c2) S invades O’s boundary and S’s invasion of O’s boundary is necessary to end some wrongful boundary-invasion on O’s part and S’s invasion of O’s boundary is not disproportionate to the seriousness of O’s boundary-violation.¹⁷

A “wrongful boundary-invasion” is one that either falls under (18a) – is an actual boundary-violation – or else falls under (18b).

In summary, then, Shylock’s Principle, Portia’s Principle, and the Principle of Proportion represent the three most plausible positions one could take in attempting to apply (1) to the issue of derivative positive rights, and the Principle of Proportion captures the spirit of (1) more faithfully than either of the other principles does. The Principle of Proportion furthermore yields (18) and (19) as the most plausible interpretations of the notion of being used as a means, as that notion appears in (1). So (1) may now be rendered as, or taken to be equivalent to:

(20) Every person has the right not to have her boundaries violated, and also not to have her boundaries invaded unless such invasion is necessary to end some wrongful boundary-invasion of hers, and such invasion is also not disproportionate to the seriousness of her boundary-invasion.¹⁸

So the problem of reconciling (2) and (3) with (1) is now the problem of reconciling (2) and (3) with (20).

V. Welfare Rights: A Digression

Before we proceed to that task, however, it is important to get clear about what the implications of (20) actually are. Proposition (20) licenses derivative positive rights, while denying the existence of basic positive [177] rights. Some may be reluctant to accept (20) because, in denying the existence of basic positive rights, it appears to deny the existence of welfare rights. Others may be eager to accept (20) for precisely the same reason. But I think either reaction would be mistaken, or at least premature.

Property rights are negative rights.¹⁹ That is, my right to my property is not in the first instance a right to be given positive aid; it is simply a right not to have my property taken from me, and not to have my use of t interfered with in certain ways. But these negative rights generate derivative positive rights. If you (intentionally or accidentally) take my umbrella, you acquire an enforceable obligation to take positive steps to return it to me. for example, you must walk back to my house, or you must mail the umbrella to me, or you must call me to let me know where I can pick it up. You cannot refuse to take these steps without violating my right not to be deprived of my property without my consent.

I suggest that welfare rights should be treated as a variety of property rights. We may regard theories of property rights as individuated, at least in large part, by principles of just acquisition and transfer. One way in which property rights are traditionally taken to be transferred is through voluntary gift or voluntary exchange. If Frieda gives Roscoe her Lamborghini, either gratis or for a consideration, we take Frieda to have lost, and Roscoe to have gained, a property right in the Lamborghini. Furthermore, such transfers of property rights need not involve an actual transfer of physical possession. If Frieda signs her title over to Roscoe, then Roscoe suddenly becomes the new rightful owner of Frieda’s Lamborghini, even if the Lamborghini is still sitting at home in Frieda’s garage. The car in Frieda’s garage is now Roscoe’s rightful property, and Frieda has an enforceable obligation to surrender it to him (which may involve a derivative positive obligation to deliver the car to him, or to give him the key to the garage).

Such transfers need not always be voluntary. Another way in which property rights can be transferred is through rectification or compensation. If Frieda uses her Lamborghini to run over Roscoe, she may acquire an enforceable obligation to compensate him (or his beneficiaries), and so Roscoe (or his beneficiaries) may acquire a property right to some share of Frieda’s resources. Once again, the money becomes Roscoe’s rightful property before she hands it over to him, so if she holds on to it she is guilty of theft; hence, she has a derivative positive obligation to hand the money over, stemming from her negative obligation not to hold on to someone else’ rightful property. [178]

In both these cases, the transfer is triggered by some action of Frieda’s; but other cases can be imagined. I suggest that proponents of welfare rights should be understood as claiming that property rights can be transferred in yet a third way. Suppose that the facts that (a) Roscoe is in need of financial assistance, and (b) Frieda is in the best position to provide such assistance, jointly generate in Roscoe a property right to some share of Frieda’s resources.²⁰ In that case, circumstances outside either Roscoe’s or Frieda’s control can trigger a transfer of property rights from Frieda to Roscoe. Some of the resources that previously were rightfully Frieda’s are now rightfully Roscoe’s, and Frieda – as in the previous cases – now has an enforceable obligation to take positive action to ensure that Roscoe receives his rightful property. If there are welfare rights, they are best understood as derivative positive rights arising out of negative property rights whose transfer conditions are sensitive to financial need. There is thus no need to invoke basic positive rights in order to ground the possibility of welfare rights.²¹

I have not argued that welfare rights exist, or indeed that property rights of any sort exist. My point is merely that the acceptance of (20) does not commit us one way or the other on the question of welfare rights, and thus that our positions – whether favorable or unfavorable – with regard to welfare rights should not affect our willingness to accept (20).

It originally seemed that (1) committed us to rejecting all positive rights per se. Since a woman’s right to abort her unborn child would have to be a positive right, and since a newborn infant’s right not to be abandoned would also have to be a positive right, we would then have been unable to countenance such rights, and so – in accepting (1) – would have to reject (2) and (3). But if the intuition behind (1) is best expressed as (20), then what seemed a general prohibition on positive rights turns out to be a prohibition on basic positive rights only; derivative positive rights are allowed. Propositions (2) and (3) may be salvaged, then, if it is possible to construe them as claiming derivative, rather than basic, positive rights. To that task I now turn. [179]

VI. The Right Not To Be Abandoned: A Derivative Positive Right

Let us first consider the case of abandonment. The proposition at issue is the following:

(3) A woman who voluntarily bears a child but later changes her mind does not have the right simply to abandon her child, but must care for it until she can arrange for a substitute caretaker.²²

The claim that a willing mother has no right of abandonment means that it is not the case that others have an enforceable obligation to let her abandon her child. This by itself is insufficient to generate in the child a right against its mother not to be abandoned; the child has such a right only if the mother has an enforceable obligation not to abandon the child. Now the absence of the right of abandonment means that others can legitimately use force to prevent her from abandoning her child, which in turn means that if she has an obligation not to abandon her child, that obligation is indeed enforceable. But this does not suffice to show that she in fact has such an obligation. However, the further claim that the mother “must” care for the child until a substitute caretaker is found does indicate an obligation (whether enforceable or otherwise); and her liability to legitimate coercion by others in this regard ensures this obligation’s enforceability. Therefore, (3) claims that the woman has an enforceable obligation not to abandon her child; and that does invest the child with a positive right against its mother not to be abandoned.

Now this right cannot be a basic positive right, for (20) rules out the existence of such rights. It must therefore derive from the requirements of some negative right; but what could that right be? (It cannot be a welfare or property right, because nobody can have property rights in another’s labor without running afoul of [1] or [20].)

The answer lies in the fact that the mother in (3) gave birth voluntarily.²³ In doing so, I suggest, she undertook an obligation not to abandon the child. Why so? And why is this obligation enforceable? [180]

Suppose that Stan is a pilot for Clouds-R-Us, a charter airline company. Now ordinarily Stan is under no obligation, enforceable or otherwise, to work as a pilot; he has a right to go on strike at any time. But now suppose that Stan decides to go on strike in mid-flight: he abandons the controls, dons his parachute, leaps out the door, and leaves his planeload of passengers (none of whom can pilot a plane) to fall to their doom. (Clouds-R-Us does not waste money on frills like copilots.) When Stan is accused of murder he is indignant: “What do you mean, murder? I didn’t kill my passengers; I merely let them die. Since there are no basic positive rights, I was not under any enforceable obligation to take positive action on behalf of my passengers’s welfare; I was merely obligated to leave them alone. And that’s precisely what I did: I left them alone. What am I, my passengers’ keeper?”

This response is clearly inadequate; but why? The answer, I think, is that Stan’s relation to his passengers is importantly different from, say, an innocent bystander’s relation to an accident victim. When one is merely a bystander, one’s failure to take positive action counts as letting die, not as killing. But it is a different story when one is not a bystander but the pilot.²⁴ The fact that all these passengers are traveling at a high speed, thousands of miles above the ground, is not simply an interesting situation to which Stan is a latecomer. The passengers are way up in the air because Stan brought them there. And the passengers consented to being brought there on the understanding that Stan would return them safely to the ground; they would not have consented to be carried upward if they had known that Stan was going to bail out. Thus, if Stan bails out, he has violated the conditions under which the passengers’ ascent was voluntary; and so Stan’s total behavior toward the passengers (carry them upward and then leaving them there) counts as a violation of their negative right not to be killed without their consent. Therefore, once Stan and the passengers are aloft, it would be legitimate to force Stan to return the passengers safely to the ground before resigning his post at Clouds-R-Us. (And this is because Stan carried the passengers aloft voluntarily; if a gun had been held to his head from the beginning, it is not clear that he would have had an enforceable obligation not to bail out.)

The moral we may extract from Stan’s story is this:

(21) If S voluntarily places O in a situation where S’s failure to take positive action on O’s behalf will result in O’s death, then such a failure on S’s part is a killing, not merely a letting-die.

Now when a woman voluntarily gives birth – i.e., voluntarily brings a child into the world – she is voluntarily moving it from a situation in which it has an automatic life-support system to a situation in which it [181] does not.²⁵ In other words, she is cutting it off from its life-support system and placing it in a situation where it depends on her care for survival – just as Stan, by carrying his passengers aloft, was cutting them off from earthbound safety and placing them in a situation where they depended on his piloting for survival. And since she is voluntarily placing her child in a situation where eh will perish without her positive assistance, if she were to refuse that assistance, her action – according to (21) – would be a killing rather than a letting-die; that is, it would be a boundary-crossing. But the child has not consented to be killed; so if she abandons her child, she will be not merely crossing but invading her child’s boundary. Since that invasion is not necessary to counteract any invasion on the child’s part (the child is merely lying there, newborn, not invading anything), her invasion of her child’s boundary, according to (17), counts as a violation of that boundary – and so is prohibited by (20). Hence, if a woman gives birth voluntarily (where the availability of safe, inexpensive abortion may be among the criteria of voluntariness), she has an enforceable obligation²⁶ not to abandon her child until she can find a substitute caretaker to take over for her – just as Stan has an enforceable obligation to ensure his passengers’s safety before he bails out.

VII. The Right to Abort: A Derivative Positive Right

I have argued that a newborn infant’s right not to be abandoned by its mother is a derivative positive right rather than a basic one, and that such a right is accordingly licensed by (20). I shall now argue that the same is true of a woman’s right to abort an unwanted fetus.²⁷

(2) A woman who voluntarily becomes pregnant but later changes her mind has the right to abort her unborn child.

It is sometimes argued that abortion is justified because a fetus is not a person. If that were true, then the right to abortion would simply be a negative right, not a positive one, and would follow from (1) and (20). If a fetus is not a person, but simply a mass of human tissue, like a cancer[182]ous growth, then in aborting it a woman would simply be doing something to her own body, and the right to control one’s own body (so long as one does not invade others’ boundaries in the process) is clearly licensed by (1). If I forcibly interfere with your use of your own body, I am violating your boundaries and may legitimately be restrained by force.

But is it true that fetuses are not persons? I think it is clearly true in the early stages of pregnancy, for the possession of psychological capacities of some sort is essential to personhood. Nothing counts as a person if it lacks psychological capacities; an early fetus lacks neurophysiological structures sophisticated enough to ground psychological capacities; therefore, an early fetus is not a person,²⁸ and so may be killed without violating any rights.²⁹

Nor will it do to protest that an early fetus is at least s potential person, and is deserving of moral concern in virtue of that status. For the claim that an early fetus is potentially a person is ambiguous, and the sense in which the claim is true is different from the sense needed by the opponent of early abortion. The claim might mean that the early fetus is potentiallyidentical with a person, just as I am potentially identical with a grandfather; or it might mean that the early fetus is potentially constitutive of a person, just as clay is potentially constitutive of a person. The claim that the early fetus is potentially identical with a person might make it an object of moral concern (although it is far from obvious that it would do so); but such a claim is simply false. Whatever is a person is essentially a person; the change from person to nonperson or vice versa is not a change that anything could survive.³⁰ So the early fetus cannot be regarded as becoming a person; either we must regard it as going out of existence when the person comes to be (in the same way that an organism goes out of existence when a corpse comes to be), so that an early fetus is replaced by a person, or else we must regard the early fetus as persisting as the underlying matter of the person. In the latter case, an early fetus does turn out to be a potential person, but only in the pickwickian [183] sense of potentially constituting a person; and there is no reason to suppose that something becomes an object of moral concern merely in virtue of constituting a person. The right to abort an early fetus, then, is an unproblematic negative right.

But while this argument serves to justify most abortions (and would certainly justify the use of “morning-after” abortifacients), it does not justify late abortions.³¹ Personhood may not begin at conception, but it does appear to begin before birth; there is no significant difference between the psychological capacities of a newborn infant and those of a late fetus. If we deny that the late fetus is a person, we must also deny that a newborn infant is a person, and therefore justify infanticide. But this conclusion, apart from its morally unpalatable character, also runs afoul of (3); if a newborn infant is not a person, it does not have nay rights, and so may legitimately be abandoned. (The worst thing one could say about it is that it might be a case of littering.) One might respond by claiming that newborn infants can have rights even if they are not persons; but in that case the argument for (2) fails. The claim that late fetuses are not persons counts as an argument for (2) only if personhood is necessary for rights; and in that case (given the similarity between late fetuses and newborn infants) we do not get a view that accommodates both (2) and (3) – which is what we were looking for. So the denial that late fetuses are persons seems like an unpromising strategy.

Let us grant, then, that late fetuses are persons. If that is so, what could justify aborting them? The answer to that question will turn on the answer to a broader question: When, in general, does one have the right to kill a person? Or, more specifically, when does one have the right to kill a person without that person’s consent?³² Well, killing a person without that person’s consent is a boundary-invasion by definition (16), and the general rule governing boundary-invasions was laid down by (20):

(20) Every person has the right not to have her boundaries violated, and also not to have her boundaries invaded unless such invasion is necessary to end some wrongful boundary-invasion of hers, and such invasion is also not disproportionate to the seriousness of her boundary-invasion.

To prohibit abortion is to invade a pregnant woman’s boundary. According to (20), such a prohibition will be illegitimate unless it is necessary to counteract some wrongful boundary-invasion on her part. Now aborting a (late) fetus is certainly a boundary-invasion; but such invasions cannot [184] be prohibited unless they are wrongful, where such wrongfulness is itself defined as the violation of (20). Therefore, to show that a woman has the right to abort an unwanted fetus, it will suffice to show (a) that such an abortion would not involve a boundary-violation, and (b) that such an abortion is necessary to end some wrongful boundary-invasion on the part of the fetus, and is also not disproportionate to the seriousness of the fetus’s boundary-invasion. But by (17), (b) entails (a). So all we need to show is (b): that aborting an unwanted fetus is necessary to end some wrongful boundary-invasion on the part of the fetus, and is also not disproportionate to the seriousness of the fetus’s boundary-invasion.

Here is the quick version of the argument (which I shall go on to defend in more detail): Suppose Miriam is pregnant, and Joshua is her unwanted fetus. Joshua is using Miriam’s body as an incubator; so Joshua is crossing Miriam’s boundary. Miriam wants Joshua out; so Joshua’s boundary-crossing counts as a boundary-invasion. Joshua’s invasion of Miriam’s boundary is not necessary to counteract any invasion on Miriam’s part, so Joshua’s invasion is a boundary-violation. By (20), Joshua has no right to violate Miriam’s boundary; so Joshua’s boundary-violation is a wrongful boundary-invasion. Aborting Joshua is necessary to end his wrongful boundary-invasion of Miriam, and is not disproportionate to the seriousness of the threat; therefore, Miriam has a derivative positive right to abort Joshua.³³ The right to abortion is a special case of the right to self-defense.³⁴

Opponents of the right to abortion find this style of argument repellent. How, they ask, can one treat a fetus as some sort of alien parasite, and pregnancy as a unnatural violation, when pregnancy is the most natural thing in the world?³⁵ Well, sexual intercourse is also the most “natural” thing in the world; but when it is involuntary, it becomes rape. [185] Likewise, when the “natural” process of pregnancy is involuntary, it too becomes an alien intrusion or violation.

Yet if the case of Miriam and Joshua is to justify (2), we must add that Miriam became pregnant voluntarily.³⁶ And if Miriam’s pregnancy is voluntary, then Joshua’s boundary-crossing is not a boundary-violation, and so the argument for Miriam’s right to abort Joshua cannot get off the ground.

We must, however, distinguish between pregnancies that were initially voluntary and pregnancies that are currently voluntary. Suppose that Miriam’s pregnancy began in as voluntary a fashion as one could ask for: she voluntarily engaged in intercourse with the deliberate intention of conceiving a child.³⁷ But now she has changed her mind, and no longer wishes to carry Joshua to term. Her pregnancy was initially voluntary, but now it is involuntary. The same is true of sexual intercourse: a woman who begins intercourse voluntarily but then changes her mind has the right to back out halfway through; if her partner fail to respect this wish, he is guilty of rape, even if he (and she) have been socialized not to think of it that way.³⁸ Consent, to justify, must be sustained.³⁹ [186] When Miriam’s pregnancy was voluntary, Joshua’s occupation of her body was a mere boundary-crossing; but now that Miriam has changed her mind, that occupation has become a boundary-invasion.

Yet what distinguishes Miriam’s case from the case of Stan the airline pilot, or from the case of the initially willing mother who later abandons her baby? Cannot Stan say that he started the flight as a willing pilot, but has since changed his mind, and thus that forcing him to continue piloting would be involuntary servitude? (And likewise for the abandonment case?) It seems that (2) and (3) have come into potential conflict once more. Either we grant Miriam the right to change her mind, in which case it is unclear how we can deny the same right to Stan or to the abandoning mother; or we deny Stan and the abandoning mother the right to change their minds, in which case it is unclear how we can grant that right to Miriam. What could justify giving different answers in the two cases?

Recall that in denying Stan the right to abandon his passengers, and in denying an initially willing mother the right to abandon her child, we appealed to principle (21), which established that such abandonments are killings rather than mere lettings-die. That is, we showed that abandonment in such cases is a boundary-crossing – and therefore, since the victims did not consent, a boundary-invasion. But that was only part of the story. In order to show that Stan’s passengers and the newborn infant have the right not to be abandoned, it was necessary, according to (20), to show that the boundary-invasion would be a wrongful one. This we were able to do by showing that no such boundary-invasion was necessary to counteract any wrongful boundary-invasion on the part of the passengers or of the infant; a boundary-invasion that lacks such an excuse is a boundary-violation, and (20) prohibits all boundary-violations. But in Miriam’s case, killing Joshua is necessary in order to counteract his wrongful invasion of her body,⁴⁰ and so Miram’s act of aborting Joshua is not a boundary-violation. That is the crucial difference between abortion and abandonment.

But although the fact that abandonment is a boundary-violation shows that it is prohibited by (20), the fact that abortion is not a boundary-violation does not show that it is licensed by (20). For (20) also prohibits some boundary-invasions that are not boundary-violations – namely, those boundary-invasions that are disproportionate to the seriousness of the threat they are designed to counteract. (Once again, I do not have the right to blow you away in order to prevent you from treading on my toe, [187] even if no lesser measure would succeed.) In order to show that abortion is licensed by (20), then, we will need to show that Miriam’s killing Joshua is not disproportionate to the seriousness of the threat posed by Joshua to Miriam.

One might try to show this by appealing to the fact that the process of childbirth is (a) life-threatening, and (b) extremely painful.⁴¹ (If the pain involved in childbirth were induced by other means, it would generally be recognized as a form of torture, and a nation that required women to undergo it would be found in violation of Article V of the United Nations Universal Declaration of Human Rights.)⁴² As things stand, then, abortion is not disproportionate to the seriousness of the threat it counteracts, and so is not a wrongful boundary-invasion; we surely have a right to kill in order to avoid being tortured. (It might be objected that killing can never be a proportionate response to any threat short of death. But our concern is with proportionality in moral seriousness, not proportionality in physical effect; to claim that defensive killing can be morally proportionate only to a threat of death is to assume, between aggressive force and defensive force, a moral symmetry difficult to square with [20].) But what if medical science eventually renders childbirth safe and painless? Should abortion still be permitted under those circumstances? A robust defense of (2) should strive to vindicate the right to abortion in general, without relying on historically continent facts about the level of advancement of medical technology. In what follows I thus forgo any appeal to the risk or painfulness of childbirth.

So let us pretend that pregnancy poses no risk to the mother’s life, health, or comfort. If the threat posed by an unwanted pregnancy is not one of pain or physical injury, then how could killing the fetus not be a disproportionate response to the seriousness of the boundary-violation posed by such a pregnancy? How could anything less than a threat of pain or physical injury justify killing one’s assailant in self-defense?

To answer this question, it is useful to consider the analogy of rape. A woman is clearly justified in killing a rapist in self-defense (assuming no lesser measures would be successful). Rape is one of the most profound and traumatizing assaults on one’s personhood that it is possible to inflict; so killing is not a disproportionate response to the seriousness of rape. But a rape need not involve physical injury or pain; if the rape victim is intimidated into failing to resist, then in purely physical terms a rape may be indistinguishable from normal, consensual intercourse. Rape need not be violent in any overt sense.⁴³ Yet it is a rape for all that; for [188] any sexual use of another person’s body without that person’s consent is a rape. What gives a woman the right to kill as rapist in self-defense, then, is not that he threatens her with pain or injury, but that he uses her body in the most deeply intimate and personal way, without her consent (even if she originally consented, then changed her mind). And it is precisely this same fact that gives Miriam the right to kill her unwanted fetus Joshua: not that he threatens her with pain or injury, but that he uses her body⁴⁴ in the most deeply intimate and personal way, without her consent(even if she originally consented).⁴⁵ Hence abortion is not a disproportionate response to the seriousness of the boundary-violation it counteracts. My argument for abortion rights may be expressed, then, by the following syllogism:

(a) One has the right to kill in self-defense if the threat is sufficiently serious.

(b) The threat posed by an unwanted fetus is sufficiently serious.

(c) Therefore, one has the right to kill an unwanted fetus in self-defense.

My analogy between fetuses and rapists will strike many opponents of abortion rights as absurd. Doesn’t this analogy ignore a vitally important difference – namely, that the fetus is innocent? The fetus did not choose to violate its mother’s boundaries; the violation occurred as a result of natural processes over which the fetus, in the nature of the case, could have no control (since these are the same natural processes that produced it).

Yes, this is of course an important difference; but it is not important in the relevant way. An unwanted fetus is an innocent threat, but is a threat nonetheless.⁴⁶ A boundary-violation does not cease to be a boundary-violation just because the boundary-violator was acting involuntarily; nor does such involuntariness transform a profoundly personal intrusion into a minor inconvenience. Proposition (20) therefore licenses the killing of innocent threats in self-defense.⁴⁷ To be sure, considerations of the threat’s innocence or guilt may legitimately affect judgments of the moral [189] proportionality of the response. But when the threat is as personal and intrusive as an unwanted pregnancy, it is difficult to see how the innocence of the fetus could make enough of a difference to justify forcing the mother to quietly endure nine months of what is tantamount to rape. Analogously , even if someone has been involuntarily hypnotized into becoming a literal rapist, his victim still has the right to kill him in self-defense.

Objection: Not if she’s the one who deliberately hypnotized him. Reply: Yes, even then, I think. A woman never has an obligation – or at any rate, never has an enforceable obligation⁴⁸ – to let herself be raped. That’s moral bedrock if anything is. The notion of an enforceable obligation to let one’s body be used by a rapist is a moral obscenity; and the same holds for the notion of an enforceable obligation to let one’s body be used as an incubator by a fetus, even if the mother is responsible for the fetus’s presence there in the first place.⁴⁹ The right to abortion, like the right not to be raped, is inalienable; one cannot legitimately enslave oneself by waiving in advance one’s right to control one’s own body. This follows from (20); since boundary-violations are prohibited, the only way Miriam could alienate her right to her body is by placing her body outside her own boundary, so that someone could use her body as a mere means without using her as a mere means. But unless Miriam has acquired the knack of astral projection, this is impossible.⁵⁰

Yet another disanalogy may be urged between a rapist and an unwanted fetus: Joshua is Miriam’s child, and we have special obligations to our children that we do not have to strangers .⁵¹ Miriam may have a right to deny strangers the use of her body, but it does not follow that she has the right to deny such use to her child, when his very survival depends on being allowed to use her body. The right to an abortion, so this objection runs, is inconsistent with the existence of parental obligations.⁵² [190]

I would hardly deny the importance of parental obligations; but it is generally recognized that adoptive parents have the same obligations as biological parents – and that, conversely, either biological or adoptive parents may divest themselves of their (enforceable) obligations by putting their child up for adoption. Hence a parent’s enforceable obligations appear to stem from his or her voluntary acceptance of the rights and responsibilities of guardianship, and not from a merely biological tie. Hence an appeal to parental obligations in the case of abortion misses the mark, since a woman seeking abortion is ipso facto not a willing guardian.

But even if we were to assume, implausibly, that Miriam has assumed the parental responsibilities of guardianship toward Joshua (e.g., in virtue of having conceived him voluntarily), does it really follow that she has no right of self-defense against him. Until recently, it was widely held (and still is, in some of the world’s more benighted recesses)⁵³ that a woman does not count as having been genuinely “raped” if the assailant is her husband. The “parental obligations” objection to abortion, like the refusal to recognize rape in marriage, appears to stem from a traditional attitude that refuses to acknowledge women as autonomous individuals, and regards their bodies as mere resources to be used by family members.⁵⁴ On this view, a woman is not an independent moral being worthy of respect in her own right, but instead exists only for the sake of her family relationships, and has her moral identity and standing only within that context. Simply by virtue of being a family member, then, a husband or fetus is imagined to possess a rightful claim to sue her body. A woman is only a tenant in her body, not the rightful proprietor.⁵⁵ But this view is wrong, and (1) is one way of expressing the fact that it is wrong. Family relationships are important, but they do not give some people the right to use other people as mere means to their own ends.⁵⁶ [190] What Jefferson said of men is likewise true of women: “[T]he mass of mankind has not been born with saddles on their backs, nor a favored few booted and spurred, ready to ride them legitimately, by the grace of God.”⁵⁷

VIII. Summary

We began with three propositions: that people have a right not to be treated as mere means to the ends of others, that a woman who voluntarily becomes pregnant nevertheless has the right to an abortion, and that a woman who voluntarily gives birth does not have a right to abandon her child until she finds a substitute caretaker. These propositions initially seemed inconsistent, for the prohibition on treating others as mere means appeared to rule out the possibility of positive rights, thus making it impossible to countenance the right to abort or the right not to be abandoned (both of which, it was argued, are positive in form).

But we have seen that the prohibition on treating people as mere means to the ends of others is best understood as ruling out basic positive rights while permitting derivative ones. Since a willing mother is responsible for bringing her child into the world in the first place, she cannot abandon it without violating its negative right not to be killed, and so such a child has a derivative positive right not to be abandoned. A pregnant woman, on the other hand, has a negative right not to have her body invaded, and from this negative right derives a positive right to abort her fetus, so long as doing so is not disproportionate to the seriousness of the threat (as it is not in the case of involuntary pregnancy, or of pregnancy which has become involuntary).Therefore, far from being in conflict, propositions (1), (2), and (3) have been shown to be in harmony with one another, the latter two being plausibly grounded in the first. Insofar as we have reason to accept (1), then, we have reason to accept (2) and (3). Moreover, we have seen that a proper understanding of (1) allows us to embed (2) and (3) in a larger moral respective in which the limits of compulsory altruism are firmly drawn: enforceable rights to the sue or assistance of others may be allowed in to the moral domain only if they are “sponsored” by some negative right. Every putative positive right must find such a sponsor, or perish.

Philosophy, The University of North Carolina at Chapel Hill

[Notes]

* An earlier version of this paper was presented to the Research Triangle Ethics Circle (Chapel Hill, North Carolina), whose participants provided useful and stimulating discussion. I am also grateful to David Schmidtz and Ellen Frankel Paul for helpful written comments.

¹ I take (1) to make a narrower claim than (1*): “Every person has a right to be treated as an end in him or herself.” Proposition (1*) seems to require taking positive action on behalf of the rights-holder, in a way that (1) does not.

² Am I arguing for the truth of these three propositions, or merely for their compatibility? Somewhere in between, actually. I shall not argue for the truth of (1), but I take it to be a plausible moral principle; and I shall argue that that principle, together with what I take to be plausible moral assumptions, yields (2) and (3) as consequences. [167]

³ Discussions of rights sometimes founder on a failure to distinguish these varieties from one another. When Marthe’s mother claims a right to Marthe’s gratitude, she is presumably claiming an AB-right but not a C-right; Marthe ought to be grateful to her mother, but it would be illegitimate to enforce such gratitude at gunpoint. On the other hand, when Hector’s friends tell him that he has no right to complain about the restaurant since he’s the one who insisted they come here, they are denying him an A-right, but as First Amendment devotees they would no doubt grant Hector’s BC-right to complain. Yet again, when Hobbes [168] argues (Leviathan, part I, ch. 14) that in the state of nature everyone has an equal right to everything, he is asserting an AC-right, not a B-right. And so on. (In particular, discussions of abortion often fail to make clear whether it is A-rights or BC-rights that are at issue.) Traditional distinctions between perfect and imperfect rights, or between liberty-rights and claim-rights, capture some of the relevant distinctions, but not, I think, all of them.

⁴ That is, a just political system would refrain from violating these rights itself, and also protect its citizens from having these rights violated by others. In calling these rights “political rights,” I do not mean to suggest that their existence depends on or is an artifact of political institutions or conventions. I mean only that these are the sorts of rights with which a legal system might properly be concerned, whereas the right to gratitude, say, is not.

⁵ That is, political rights are at least BC-rights. A different issue is whether all BC-rights are also ABC-rights. On a liberal conception of politics, the answer will be no, since legal protection for some morally impermissible actions has a basis in one of the central tenets of liberalism, that political institutions should be neutral (at least in a broad range of cases) among their citizens’ competing conceptions of the good. (It does not follow that this liberal neutrality may not itself be founded on a particular conception of the good.) But the liberal doctrine of institutional neutrality is often challenged; for example, a popular slogan among supporters of Pat Robertson’s 1988 presidential campaign was: “People should only have a right to do what is right.” For the purposes of this essay I shall leave it an open question whether BC-rights must also be ABC-rights – though proposition (1) does seem to favor a reply in the negative.

⁶ I here ignore a possible complication, which is that it may be legitimate for some people, but not for others, to enforce the obligation in question. (In particular, the person to whom the obligation is owed might conceivably have different rights of enforcement from those enjoyed by third parties.) [169]

⁷ The lack of such specification can lead to seriously ambiguous rights-claims. For example, is the right to life the right not to be killed? Or the right to be provided with the necessary means of survival? Or the right to be brought into existence? Is the right to work the right to guaranteed employment, or merely the right not to be interfered with in seeking a job? And interfered with how? And so on.

⁸ I am treating rights as all-things-considered rights rather than as prima facie rights. It follows that, on this understanding, conflicts of rights (that is, of BC-rights) can never occur. It may well be that this would justify placing some qualification or exception clause on (1).

⁹ For the purposes of this essay I shall assume without argument that a principled distinction can be drawn between doing and allowing (e.g., between killing and letting die). [170]

¹⁰ I will be appealing to intuitions a lot. My epistemologist says that’s okay. [171]

¹¹ The suggested claim is not that people never have positive obligations to act on behalf of others, or to let themselves be used by others; the claim is rather that such obligations, if they exist, are not legitimately enforceable. On this view, there may be positive B-rights, bur no positive BC-rights. Positive rights are sometimes rejected, not because they violate (1), but because they are allegedly nonuniversalizable; but Andrew Melnyk, “Is There A Formal Argument Against Positive Rights?” Philosophical Studies, vol. 55 (1989), pp. 205-9, has argued persuasively against that view.

¹² By “means” I shall henceforth mean “means to the ends of others.” I leave open the question of whether it is possible to treat someone as a means to his own ends without his consent. If so, then (1) appears to permit paternalism; if not, then it doesn’t. I should think that the intuition behind (1) at least creates a serious presumption against paternalism, but I shall not argue for that claim here. [172]

¹³ One might deny that running someone over counts as a violation of (1), since in running you over I may not be using you as a means (mere or nonmere) to any further end; I may simply be trying to drive through the space you are standing on, and killing you is not a means to, but rather a byproduct of, my pursuit of that goal – a foreseen but not intended consequence. This is true enough if the notion of “using as a mere means” is understood very narrowly; but surely I am forcibly subjecting or subordinating you to my ends, and that is what I take to be the central point of the notion. [173]

¹⁴ The notion of “use” needs to be clarified somewhat. Suppose Professor Kant is so regular in his daily walks that the townspeople set their watches by him. Are the townspeo-[174]ple using Kant as a means (with or without his consent) to their own ends? Does their action lie within Kant’s boundary? Surely not. The townspeople are not doing anything to Kant; they simply accept, and make use of, the information that he distributes gratis. (Recall the fable of the rich man who tried to sue his poor neighbor for enjoying the smells that emanated from the rich man’s kitchen.) Using another person must involve subjecting him to one’s ends, in a way that the Kant case does not. (See previous note.)

¹⁵ I take it that threatening to invade someone’s boundary is itself an invasion of that person’s boundary (since in announcing my intention of using you as a means I am already treating you as the sort of thing it is legitimate to use as a means). Therefore, the right to forestall a threatened invasion would follow from the right to end an actual one. [175]

¹⁶ By “disproportionate” I mean, of course, disproportionate on the side of excess, not of deficiency. [176]

¹⁷ These definitions entail that any use to which the used person consents is not a case of being used as a mere means. This follows from (1), since if it were a violation of my rights for you to cross my boundary even with my consent, then you would have an enforceable obligation not to cross my boundary even with my consent, which means that it would be legitimate to force you not to cross my boundary even with my consent, which in turn means – by definition (17) – that it would be legitimate to violate your boundary, and (1) prohibits boundary-violations if it prohibits anything. This does not entail, however, that all noninvasive boundary-crossings are morally permissible.

¹⁸ Proposition (20) is probably not true without exception as it stands. I am not so extreme a rights fanatic as to deny that minor boundary-violations may occasionally be morally permissible in order to prevent some great evil. But I do not think this qualification affects the central arguments of my essay. [177]

¹⁹ I leave aside the question of whether there are property rights, and if so, what their foundation might be. If, as Aristotle, Locke, Hegel, and Marx thought (in very different ways), one’s property is in some sense an extension of oneself, then (1) will entail the existence of property rights. (For a recent defense of such a view, see Samuel C. Wheeler III, “Natural Property Rights as Body Rights,” Noûs, vol. 16 [1980], pp. 171-93.) But much more remains to be said. [178]

²⁰ Such a view of property rights has an unlikely source in Locke; see his First Treatise of Government, ch. 4, section 42, where welfare rights are treated as derivative positive rights licensed by something like (1).

²¹ It might be objected that (1)’s ban on basic positive rights is toothless, since the case of welfare rights suggests that all basic positive rights can be rewritten as derivative positive rights. But that is a mistake. In order for a positive right to be admitted as derivative and so permissible, it must be possible to identify a specific negative right from which the putative positive right is supposed to arise; and there is no guarantee that such a negative right can be found in every case. From what negative right, for example, could a positive right to health care derive? (I am speaking not of a right to be enabled to purchase health care – as in Medicare, etc., – but of an actual right against some doctor that he or she expend labor in treating the patient – as in the alleged right, defended by J. S. Mill in the penultimate paragraph of Utilitarianism, to “kidnap, and compel to officiate, the only qualified medical practitioner” in order to save a life.) [179]

²² Proposition (3) leaves open the question of how much the mother must do to secure such a caretaker. Must she locate a willing and suitable guardian? Or may she simply leave the infant in a shopping mall, confident that someone will take care of it? It is not my aim in this essay to determine precisely the most that others may ask of her (that is, the most that others may enforce; the mother’s unenforceable obligations may well extend beyond her enforceable ones). The point of (3) is to set at least some lower limits; the mother may not, for example, abandon her infant in a deserted area, or dispose of it in a dumpster.

²³ I do not see any way of generating an enforceable obligation not to abandon a child if that child was born against the mother’s will. An unwilling mother might well have an obligation to care for the child until a substitute caretaker could be found, but that obligation could not be enforced without violating (1) or (20). [180]

²⁴ Cf. Aristotle, Physics, II.3.195a11-14. [181]

²⁵ It may seem odd to speak of “voluntarily giving birth,” when the actual biological process of birth proceeds independently of the mother’s will. But it is clear enough what is meant: if the mother wants to give birth, and has deliberately chosen to do so, in the face of opportunities to terminator her pregnancy, then she gives birth voluntarily. The same remarks apply to voluntary conception. (One can likewise be struck by lightning voluntarily, if one deliberately stays on the roof during a storm, holding up a pitchfork in the hope of getting hit.)

²⁶ I leave aside the question of what responsibilities the father might have to assist her in fulfilling this obligation.

²⁷ As will be obvious, my treatment of the abortion issue is greatly indebted to Judith Jarvis Thomson, “A Defense of Abortion,” Philosophy and Public Affairs, vol. 1 (1971), pp. 47-66. [182]

²⁸ It might be claimed that an early fetus is a person because, as a potential person, it has psychological capacities; it just cannot exercise them yet. But this is to confuse remote with proximate potentialities. Suppose that Pierre is a Francophone, and Zeke is not. Yet Zeke is capable of learning French; Zeke is a potential Francophone. So both Pierre and Zeke have the potentiality to speak French; but that does not make them both Francophones. Pierre’s potentiality to speak French is proximate, while Zeke’s is merely remote. Being a Francophone is defined in terms of the proximate potentiality to speak French; so Pierre is a Francophone while Zeke is not. So even if an early fetus has remote psychological capacities, it is not thereby a person, since being a person is defined in terms of (more) proximate psychological capacities.

²⁹ Aquinas, following Aristotle, regarded the soul as the form of the body, and denied that the fetus could become ensouled until its matter had been sufficiently worked up to enable it to be informed by a human soul; accordingly, he did not regard early abortion as murder. Late abortion was another matter. (See On the Sentences, IV.31.2; On the Politics, VII.12; Summa of Theology, II.II.64.8.2.)

³⁰ While a non-Francophone could become a Francophone, no nonperson could become a person. [183]

³¹ Since I am ultimately going to argue that all abortions are justified, I do not need to worry about fixing the precise temporal cutoff between early and late.

³² Since killing a person with that person’s consent is only a boundary-crossing, not a boundary-invasion, such a killing presumably does not violate (20). [184]

³³ I am not arguing that abortion is morally permissible, but only that a woman has a BC-right to abortion. The view I am defending is thus compatible with any of the following three positions:

(a) A woman has a moral obligation to let an unwanted fetus use her body as an incubator, but it would be illegitimate to enforce that obligation. (I take this to be the position of a majority of the American public.)

(b) A woman has no moral obligation to let an unwanted fetus use her body as an incubator, but it would be morally praiseworthy (albeit supererogatory) for her to do so. (This is Thomson’s position.)

(c) Not only has a woman no moral obligation to let an unwanted fetus use her body as an incubator, but she actually has an obligation (an unenforceable one, of course) not to saddle herself with the physical and psychological burden of an undesired pregnancy and childbirth – an obligation stemming from her duty to care for herself and to make the most of her life. (This is the position to which I incline.)

³⁴ Self-defense is a derivative positive right; my positive right to fight off a mugger derives from my negative right not to be attacked by him in the first place.

³⁵ For this style of argument, see John T. Wilcox, “Nature as Demonic in Thomson’s Defense of Abortion,” New Scholasticism, vol. 63 (1989), pp. 463-84. Since Wilcox’s conception of the “natural” is avowedly statistical (rather than, say, teleological), it is difficult to see how it can bear the normative weight with which he wants to invest it. [185]

³⁶ What makes the difference between voluntary and involuntary pregnancies? Pregnancies that result from rape are clearly involuntary, but pregnancies that result from failed contraception are a more difficult case. Proponents of the right to abortion usually argue that pregnancies that result from failed contraception are clearly involuntary, since steps were taken to avoid them. (Recall Thomson’s “people seeds” in “A Defense of Abortion.”) On the other hand, opponents of the right to abortion usually argue that such pregnancies are voluntary, so long as intercourse was engaged in with full knowledge of the risk (just as actions performed under the influence of an intoxicant are voluntary if the intoxicant was taken voluntarily with full knowledge of the possible results; see Aristotle, Nicomachean Ethics, III.5.1113b21-1114a21). Until recently, I was extremely suspicious of the latter argument; it struck me as analogous to the claim that a rape victim was “asking for it” by walking in a high-crime neighborhood or going on a date. But I have been made somewhat less confident by Wilcox’s observation that if (as I believe) fathers have responsibilities with regard to pregnancies that they cause, even when those pregnancies are the result of failed contraception, then such pregnancies cannot be treated as mere bolts from the blue. It seems that fathers could have such responsibilities only if engaging in intercourse with knowledge of the risks somehow made even pregnancies resulting result from failed contraception voluntary in Aristotle’s broad sense. And if they are voluntary for the father, are they not voluntary for the mother? I need not decide this issue, however, since I shall argue that pregnancies may be terminated whether or not they were (initially) voluntary.

³⁷ It may be objected that, in the context of the overt and covert pressures and attitudes of a male-dominated society, no act of heterosexual intercourse is fully voluntary on the woman’s part, and so no pregnancy resulting from such intercourse could be voluntary either. I think this objection, while insightful, is overstated, and slights women’s capacity to achieve and express genuine autonomy even in the face of an adverse social climate; but for those who accept this objection, it still makes sense to ask whether abortion would be justified in a society that was not male-dominated, and my discussion may be taken as relevant to that question at least.

³⁸ To borrow an example that I believe I heard in a lecture by Claudia Card: If I start to eat a meal and then change my mind halfway through, you cannot use my original intention to eat the entire meal as an excuse to force the rest of the food down my throat. [Post-publication note: Claudia Card, “Rape as a Terrorist Institution,” p. 310; in R. G. Frey and Christopher W. Morris, eds., Violence, Terrorism, and Justice (Cambridge: Cambridge University Press, 1991), pp. 296-319.]

³⁹ Does this undermine the possibility of contractual obligation? No, but that is a long story which I cannot get into here. The short answer is that if I do not do what you paid me to do, I may not legitimately be forced to do so, but I do have a derivative positive obligation to give you back your money (plus, perhaps, another derivative positive obligation to [186] compensate you for taking it under false pretenses). For useful discussion, see Randy E. Barnett, “Contract Remedies and Inalienable Rights,” Social Philosophy & Policy, vol. 4, no. 1 (1986), pp. 179-202.

⁴⁰ If killing Joshua were not necessary – if Miriam could remove him from her body without killing him – the she would, I think, be bound by (20) not to kill him. See Thomson, “A Defense of Abortion,” p. 66. [187]

⁴¹ In most cases anesthetic does not fully counteract the pain, and often cannot be administered early enough anyway.

⁴² “No one shall be subjected to torture or to cruel, inhuman or degrading treatment or punishment.”

⁴³ “The essence of rape isn’t violence, but trespass.” J. Neil Schulman, The Rainbow Cadenza: A Novel in Vistata Form (New York: Avon Books, 1986), p. 190. [188]

⁴⁴ I do not mean to imply that the fetus is a voluntary agent; Joshua may “use” Miriam’s body in the same sense that a Venus’ flytrap “uses” flies, mindlessly subjecting them to its ends. But nothing in the definition of boundary-violations requires them to be voluntary acts.

⁴⁵ The frequently heard claim that women who have abortions are sacrificing human life merely for the sake of their own “convenience” overlooks the profoundly intrusive nature of the fetus’ boundary-violation. Such remarks are on a par with the advice to rape victims to “lie back and enjoy it.”

⁴⁶ If it is wrong to kill an innocent threat, it is difficult to see how the presence of evil thoughts in the threatening agent’s head could suddenly make such a killing legitimate, given that nobody has rightful jurisdiction over the contents of another person’s mind.

⁴⁷ I leave aside the question of whether (20) also licenses the killing of innocent shields. For the distinction between innocent threats and innocent shields, see Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), pp. 34-35. [189]

⁴⁸ I would be willing to defend the stronger thesis, but for the purposes of this essay the weaker claim will suffice.

⁴⁹ It is true, of course, that a woman who deliberately hypnotized a man into attempting rape, and then killed him in self-defense, would be blameworthy for his death; but that is presumably because she violated his rights in the first place by voluntarily placing him in a situation where she would be justified in killing him. There is no analogous rights-violation in the pregnancy case; we surely do not violate a person’s rights merely by bringing him or her into existence.

⁵⁰ If Miriam were to voluntarily cut off or surgically remove some part of her body, she would then be able to alienate her right to that part (as organ donors do), since it would be possible to use the part without using Miriam. But as long as the part is organically attached to Miriam, she and it are a package deal, and her right over the part is inalienable. (One can never place oneself outside one’s own boundary. Hence, the real problem with selling oneself into slavery is that such a contract is fraudulent; the seller offers something that cannot be delivered.)

⁵¹ For this objection, see Wilcox, “Nature as Demonic”; and Francis J. Beckwith, “Rights, Filial Obligations, and Medical Risks,” American Philosophical Association Newsletter on Philosophy and Medicine, vol. 89 (1990), pp. 86-88.

⁵² Beckwith inexplicably calls these filial obligations. Surely a filial obligation is owed by a child to a parent, not the other way around. [190]

⁵³ One of these recesses is the mind of Alaska Senator Paul Fischer, who remarked in 1985: “I don’t know how you can have a sexual act and call it forcible rape in a marriage situation. ... I still believe in the old traditional bond of marriage.” Quoted in Susan Moller Okin, Justice, Gender, and the Family (New York: Basic Books, 1989), p. 42n. Thank you for sharing that with us, Paul.

⁵⁴ In ancient Athens every woman was assigned a kurios, a “boss” – either her husband or a male relative. Today, opponents of abortion may be viewed as assigning to the fetus the status of kurios.

⁵⁵ One is reminded of Herbert Spencer’s insightful remark that opponents of the full and equal rights of women are clinging to an evanescent form of the doctrine that women have no souls. See Herbert Spencer, Social Statics: The Conditions Essential to Human Happiness Specified, and the First of Them Developed (New York: Robert Schalkenbach Foundation, 1970), p. 143.

⁵⁶ Does this essay’s stress on rights and individual autonomy, rather than on relationships of caring and connectedness, express an androcentric perspective inappropriate to the needs and concerns of women, and thus irrelevant to the abortion debate? (See Carol Gilligan, In a Different Voice [Cambridge: Harvard University Press, 1982].) I hope not. To say that people have rights is simply to say that there are some things one should not be allowed to do to others; whoever grants the importance of that fact is committed to granting the importance of rights. Relationships of caring and connectedness are important as well; but if they are not to turn into a self-sacrificial trap, they must operate within a framework of rights. As for individual autonomy, it is precisely this that women have traditionally been [191] denied, with devastating results. Women have been classified as objects, as resources, as property, and have been routinely subordinated to the will of others. A principle that grants every person the right not to be used as a mere means to the ends of others is a feminist principle if anything is. (For a defense of the importance of a “justice perspective” to feminist concerns, see Okin, Justice, Gender, and the Family, p. 15, and indeed the entire book.)

⁵⁷ Letter to Roger C. Weightman, 24 June 1826, in The American Enlightenment: The Shaping of the American Experiment and a Free Society, ed. Adrienne Koch (New York: George Braziller, 1965), p. 372.

Comment on the blog