In Matters of Life and Death, 2d ed., edited by Tom Regan (1986).





  • Undoubted Commonsense Persons
  • Undoubted Nonpersons
  • Hard Cases
  • A Working Criterion of Commonsense Personhood
  1. The Species Criterion
  2. The Modified Species Criterion
  3. The Strict Potentiality Criterion
  4. The Modified or Gradualist Potentiality Criterion
  5. The Actual-Possession Criterion
    • The Killing of Normal Infants
    • The Killing of Radically Deformed Infants
    • Implications for the Problem of Abortion


  • Property Rights over One's Body
  • Self-Defense and Proportionality
  • Self-Defense: The Problem of the Innocent Aggressor
  • Bodily Autonomy: The Example of the Plugged-in Violinist
  • Voluntariness and Responsibility


Abortion can be defined as the deliberate causing of the death of a fetus, either by directly killing it or (more commonly) by causing its expulsion from the womb before it is "viable," that is before it is capable of surviving outside its mother's body. The word 'fetus' in this definition refers to a human offspring at any stage of its prenatal development, from conception to birth. There is a narrower medical sense of the word 'fetus' in which it refers to the unborn entity from roughly the eighth week of pregnancy (when brain waves can first be monitored) until birth, normally at nine months. In this technical sense the word 'fetus' is used to contrast the relatively mature unborn being with its earlier stages when it is called a "zygote" (from conception to implantation a week later in the uterine wall) or an "embryo" (from implantation until the eighth week). I shall use the word 'fetus' as a convenient general term for the unborn at all phases of its development except when the contrast between fetus and embryo or zygote becomes central in the argument.

A number of serious practical questions are raised by the practice of abortion. Two in particular stand out. The first is a question of personal conscience: Under what conditions, if any, is abortion morally justified? The second is a question of public policy: What, if anything, should be done to increase or decrease the opportunities for abortion? Despite its importance, we will be able to consider this second question only partially in the "Postscript" at the end of this chapter.

The question of the moral permissibility of abortion demands that we answer two very hard philosophical questions. The first requires us to consider which traits of the developing fetus are relevant to deciding what morally may and may not be done to it. The general problem discussed in this regard is often called "the problem of the status of the fetus." This question will be discussed in the first part of this essay. The second question that demands our attention can be classified under the general heading "the problem of the conflict of claims." The problem here is that, even granting that the fetus is a person, the fetus is not the only person involved; there are also the mother, the father, possibly other family members, etc. The needs and interests of all the affected parties, not just those of the fetus, seem to have a bearing on the question of the moral permissibility of abortion. Therefore, we must determine whose claims are the strongest, if they happen to clash. This problem is explored in the second part of this essay.


The problem of the status of the fetus can be formulated as follows: At what stage, if any, in the development between conception and birth, do fetuses acquire the characteristic (whatever it may be) that confers on them the appropriate status to be included in the scope of the moral rule against homicide -- the rule "Thou shalt not kill"? Put more tersely: At what stage, if any, of their development do fetuses become people? A variety of familiar answers have been given -- for example, that fetuses become persons "at the moment of conception," at "quickening" (that is, at the moment of their first self-initiated movement in the womb), at viability (that is, when the fetus is able to survive independently outside the mother's womb). Debates about when the fetus becomes a person, however, are premature unless we have first explored what a person is. Answers to the question "When does the fetus become a person?" attempt to draw a boundary line between prepersons and persons; however, even if correctly drawn, a boundary line is not the same thing as a criterion of personhood. Indeed, until we have a criterion, we cannot know for sure whether a given proposed boundary is correct. Let us mean by "a criterion of personhood" a specification of the characteristic (or set of characteristics) that is such that

  1. no being can be a person unless he or she possesses that characteristic,
  2. any being who possesses that characteristic is a person, and
  3. it is precisely that characteristic that directly confers personhood on whoever possesses it.
The person-making characteristic, in brief, is that characteristic (or set of characteristics) that is common and peculiar to all persons and in virtue of which they are persons. "A criterion of personhood," as we shall understand the phrase, is some statement (true or false) of what the person-making characteristic is.

The statement of a mere boundary line of personhood, in contrast, may be correct and useful even though it mentions no person-making characteristic at all. Such a statement may specify only some theoretically superficial characteristic that happens to be invariable (or perhaps only usually and roughly) correlated with a person-making characteristic. For example, even if it is true that all persons can survive outside the mother's womb, it does not follow that being able to survive outside the mother's womb is what makes someone a person or that this is even partially constitutive of what it is to be a person. The "superficial characteristic," then, can be used as a clue, a test, or an indication of the presence of the basic personhood-conferring characteristic and therefore of personhood itself, even though it is in no sense a cause or a constituent of personhood. An analogy might make the point clearer. What makes any chemical compound an acid is a feature of its molecular structure, namely that it contains hydrogen as a positive radical. But it also happens that acids "typically" are soluble in water, sour in taste, and turn blue litmus paper red. The latter three characteristics, then, while neither constitutive nor causative of acidity, can nevertheless be useful and reliable indexes to, or tests of, acidity. The litmus test in particular draws a "boundary" between acids and nonacids. The question now to be addressed is whether a reasonable criterion for personhood can be found, one that enables us to draw an accurate boundary line.


The first step in coming to terms with the concept of a person is to disentangle it from a concept with which it is thoroughly intertwined in most of our minds, that of a human being. In an influential article, the young American social philosopher Mary Anne Warren has pointed out that the term 'human' has two "distinct but not often distinguished senses."1 In what she calls the "moral sense," a being is human provided that it is a "full-fledged member of the moral community," a being possessed (as Jefferson wrote of all "men" -- did he mean men and women?) of inalienable rights to life, liberty, and the pursuit of happiness. For beings to be humans in this sense is precisely for them to be people, and the problem of the "humanity" of the fetus in this sense is that of determining whether the fetus is the sort of being -- a person -- who has such moral rights as the right to life. On the other hand, a being is human in what Warren calls the "genetic sense" provided he or she is a member of the species Homo sapiens, and all we mean in describing someone as a human in the genetic sense is that he or she belongs to that animal species. In this sense, when we say that Jones is a human being, we are making a statement of the same type as when we say that Fido is a dog (canine being). Any fetus conceived by human parents will of course be a human being in this sense, just as any fetus conceived by dogs will of course be canine in the analogous sense.

It is possible to hold, as no doubt many people do, that all human beings in the moral sense (persons) are human beings in the genetic sense (members of Homo sapiens) and vice versa, so that the two classes, while distinct in meaning, nevertheless coincide exactly in reality. In that case all genetically human beings, including fetuses from the moment of conception, have a right to life, the unjustified violation of which is homicide, and no beings who are genetically nonhuman are persons. But it is also possible to hold, as some philosophers do, that some genetically human beings {for example, zygotes and irreversibly comatose "human vegetables") are not human beings in the moral sense (persons), and/or that some persons (for example, God, angels, devils, higher animals, intelligent beings in outer space) are not members of Homo sapiens. Surely it is an open question to be settled, if at all, by argument or discovery, whether the two classes correspond exactly. It is not a question closed in advance by definition or appeals to word usage.


Perhaps the best way to proceed from this point is to stick to the term 'person' and avoid the term 'human' except when we clearly intend the genetic sense, in that way avoiding the ever-present danger of being misunderstood. The term 'person', however, is not without its own ambiguities. The one central ambiguity we should note is that between purely normative (moral or legal) uses of 'person' and purely descriptive (conventional, commonsense) uses of the term.

When moralists or lawyers use the term 'person' in a purely normative way they use it simply to ascribe moral or legal properties -- usually rights or duties or both -- to the beings so denominated. To be a person in the normative sense is to have rights, or rights and duties, or at least to be the sort of being who could have rights and duties without conceptual absurdity. Most writers think that it would be sheer nonsense to speak of the rights or duties of rocks, or blades of grass, or sunbeams, or of historical events or abstract ideas. These objects are thought to be conceptually inappropriate subjects for the attribution of rights or duties. Hence we speak of them as "impersonal entities," the types of beings that are contrasted with objects that can stand in personal relationships to us or make moral claims on us. The higher animals -- our fellow mammalian species in particular -- are borderline cases whose classification as persons or nonpersons has been a matter of controversy. Many of them are fit subjects of right-ascriptions but cannot plausibly be assigned duties or moral responsibilities. These ideas are examined in considerable detail in Peter Singer's essay in the present volume. In any case, when we attribute personhood in a purely normative way to any kind of being, we are attributing such moral qualities as rights or duties but not (necessarily) any observable characteristics of any kind -- for example, having flesh or blood or belonging to a particular species. Lawyers have attributed (legal) personhood even to states and corporations, and their purely normative judgments say nothing about the presence or absence of body, mind, consciousness, color, etc.

In contrast to the purely normative use of the word 'person' we can distinguish a purely empirical or descriptive use. There are certain characteristics that are fixed by a rather firm convention of our language such that the general term for any being who possesses them is 'person.' Thus, to say of some being that he is a person, in this sense, is to convey some information about what the being is like. Neither are attributions of personhood of this kind essentially controversial. If to be a person means to have characteristics a, b, and c, then to say of a being who is known to have a, b, and c that he or she is a person (in the descriptive sense) is no more controversial than to say of an animal known to be a young dog that it is a puppy, or of a person known to be an unmarried man that he is a bachelor. What makes these noncontroversial judgments true are conventions of language that determine what words mean. The conventions are often a bit vague around the edges but they apply clearly enough to central cases. It is in virtue of these reasonably precise linguistic conventions that the word 'person' normally conveys the idea of a definite set of descriptive characteristics. I shall call the idea defined by these characteristics "the commonsense concept of personhood." When we use the word 'person' in this wholly descriptive way we are not attributing rights, duties, eligibility for rights and duties, or any other normative characteristics to the being so described. At most we are attributing characteristics that may be a ground for ascribing rights and duties.

These purely normative and purely descriptive uses of the word 'person' are probably unusual. In most of its uses, the word both describes or classifies someone in a conventionally understood way and ascribes rights, etc., to him. But there is enough looseness or flexibility in usage to leave open the questions of whether the classes of moral and commonsense persons correspond in reality. Although some may think it obvious that all and only commonsense persons are moral persons, that identification of classes does not follow simply as a matter of word usage, and must, therefore, be supported independently by argument. Many learned writers, after all, have maintained that human zygotes and embryos are moral persons despite the fact that they are almost certainly not "commonsense persons." Others have spoken of wicked murderers as "monsters" or "fiends" that can rightly be destroyed like "wild beasts" or eliminated like "rotten apples." This seems to amount to holding that "moral monsters" are commonsense persons who are so wicked (only persons can be wicked) that they have lost their moral personhood, or membership in our moral community. The English jurist Sir William Blackstone (1723-1780) maintained that convicted murderers forfeit their rights to life. If one went further and maintained that moral monsters forfeit all their human rights, then one would be rejecting the view that the classes of moral and commonsense persons exactly coincide, for wicked persons who are answerable for their foul deeds must first of all be persons in the descriptive sense; but as beings without rights, they would not be moral persons.


A criterion of personhood in the descriptive sense would be a specification of those characteristics that are common and peculiar to commonsense persons and in virtue of which they are such persons. They are necessary conditions for commonsense personhood in the sense that no being who lacks any one of them can be a person. They are sufficient conditions in the sense that any being who possesses all of them is a person, whatever he or she may be like in other respects. How shall we formulate this criterion? If this question simply raises a matter of fixed linguistic convention, one might expect it to be easy enough to state the defining characteristics of personhood straight off. Surprisingly, the question is not quite that simple, and no mere dictionary is likely to give us a wholly satisfactory answer. What we must do is to think of the characteristics that come at least implicitly to mind when we hear or use such words as 'person', 'people', and the personal pronouns. We might best proceed by considering three different classes of cases: clear examples of beings whose personhood cannot be doubted, clear examples of beings whose nonpersonhood cannot be doubted, and actual or hypothetical examples of beings whose status is not initially clear. We probably will not be able to come up with a definitive list of characteristics if only because the word 'person' may be somewhat loose, but we should be able to achieve a criterion that is precise enough to permit a definite classification of fetuses.

Undoubted Commonsense Persons

Who are undoubted persons? Consider first your parents, siblings, or friends. What is it about them that makes you so certain that they are persons? "Well, they look like persons," you might say; "They have human faces and bodies." But so do irreversibly comatose human vegetables, and we are, to put it mildly, not so certain that they are persons. "Well then, they are males and females and thus appropriately referred to by our personal pronouns, all of which have gender. We can't refer to any of them by use of the impersonal pronoun 'it', because they have sex; so perhaps being gendered is the test of personhood." Such a reply has superficial plausibility, but is the idea of a "sexless person" logically contradictory? Perhaps any genetically human person will be predominately one sex or the other, but must the same be true of "intelligent beings in outer space," or spirits, gods, and devils?

Let's start again. "What makes me certain that my parents, siblings, and friends are people is that they give evidence of being conscious of the world and of themselves; they have inner emotional lives, just like me; they can understand things and reason about them, make plans, and act; they can communicate with me, argue, negotiate, express themselves, make agreements, honor commitments, and stand in relationships of mutual trust; they have tastes and values of their own; they can be frustrated or fulfilled, pleased or hurt." Now we clearly have the beginnings, at least, of a definitive list of person-making characteristics. In the commonsense way of thinking, persons are those beings who, among other things, are conscious, have a concept and awareness of themselves, are capable of experiencing emotions, can reason and acquire understanding, can plan ahead, can act on their plans, and can feel pleasure and pain.

Undoubted Nonpersons

What of the objects that clearly are not persons? Rocks have none of the above characteristics; neither do flowers and trees; neither (presumably) do snails and earthworms. But perhaps we are wrong about that. Maybe rocks, plants, and lower animals are congeries of lower-level spirits with inner lives and experiences of their own, as primitive men and mystics have often maintained. Very well, that is possible. But if they do have these characteristics, contrary to all appearances, then it would seem natural to think of them as persons too, "contrary to all appearance." In raising the question of their possession of these characteristics at all, we seem to be raising by the same token the question of their commonsense personhood. Mere rocks are quite certainly not crowds of silent spirits, but if, contrary to fact, they are such spirits, then we must think of them as real people, quite peculiarly embodied.

Hard Cases

Now, what about the hard cases? Is God, as traditionally conceived, a kind of nonhuman -- or better, superhuman -- person? Theologians are divided about this, of course, but most ordinary believers think of Him (note the personal pronoun) as conscious of self and world, capable of love and anger, eminently rational, having plans for the world, acting (if only through His creation), capable of communicating with humans, of issuing commands and making covenants, and of being pleased or disappointed in the use to which humans put their free will. To the extent that one believes that God has these various attributes, to that extent does one believe in a personal God, If one believes only in a God who is an unknown and unknowable First Cause of the world, or an obscure but powerful force sustaining the world, or the ultimate energy in the cosmos, then, it seems fair to say, one believes in an impersonal deity.

Now we come to the ultimate thought experiment. Suppose that you are a space explorer whose rocket ship has landed on a planet in a distant galaxy. The planet is inhabited by some very strange objects, so unlike anything you have previously encountered that at first you don't even know whether to classify them as animal, vegetable, mineral, or "none of the above." They are composed of a gelatinous sort of substance much like mucus except that it is held together by no visible membranes or skin, and it continually changes its shape from one sort of amorphous glob to another, sometimes breaking into smaller globs and then coming together again. The objects have no appendages, no joints, no heads or faces. They float mysteriously above the surface of the planet and move about in complex patterns while emitting eerie sounds resembling nothing so much as electronic music. The first thing you will wish to know about these strange objects is whether they are extraterrestrial people, to be respected, greeted, and traded and negotiated with, or mere things or inferior animals to be chopped up, boiled, and used for food and clothing.

Almost certainly the first thing you would do is try to communicate with them by making approaches, gesturing by hand, voice, or radio signals. You might also study the patterns in their movements and sound emissions to see whether they have any of the characteristics of a language. If the beings respond even in a primitive way to early gestures, one might suspect at least that they are beings who are capable of perception and who can be aware of movements and sounds. If some sort of actual communication then follows, you can attribute to them at least the mentality of chimpanzees. If negotiations then follow and agreements are reached, then you can be sure that they are rational beings; and if you learn to interpret signs of worry, distress, anger, alarm, or friendliness, then you can be quite confident that they are indeed people, no matter how inhuman they are in biological respects.

A Working Criterion of Commonsense Personhood

Suppose then that we agree that our rough list captures well the traits that are generally characteristic of commonsense persons. Suppose further (what is not quite as evident) that each trait on the list is necessary for commonsense personhood, that no one trait is by itself sufficient, but that the whole collection of traits is sufficient to confer commonsense personhood on any being that possesses it. Suppose, that is, that consciousness is necessary (no permanently unconscious being can be a person), but that it is not enough. The conscious being must also have a concept of a self and a certain amount of self-awareness. But although each of these last traits is necessary, they are still not enough even in conjunction, since a self-aware, conscious being who was totally incapable of learning or reasoning would not be a person. Hence minimal rationality is also necessary, though not by itself sufficient. And so on through our complete list of person-making characteristics, each one of which, let us suppose, is a necessary condition, and all of which are jointly a sufficient condition of being a person in the commonsense, descriptive sense of 'person'. Let us call our set of characteristics C. Now at last we can pose the most important and controversial question about the status of the fetus: What is the relation, if any, between having C and being a person in the normative (moral) sense, that is, a being who possesses, among other things, a right to life?


It bears repeating at the outset of our discussion of this most important question that formulating criteria of personhood in the purely moral sense is not a scientific question to be settled by empirical evidence, not simply a question of word usage, not simply a matter to be settled by commonsense thought experiments. It is instead an essentially controversial question about the possession of moral rights that cannot be answered in these ways. That is not to say that rational methods of investigation and discussion are not available to us, but only that the methods of reasoning about morals do not often provide conclusive proofs and demonstrations. What rational methods can achieve for us, even if they fall short of producing universal agreement, is to list the various options open to us and the strong and weak points of each of them. Every position has its embarrassments, that is, places where it appears to conflict with moral and commonsense convictions that even its proponents can be presumed to share. To point out these embarrassments for a given position is not necessarily to refute it but rather to measure the costs of holding it to the coherence of one's larger set of beliefs generally. Similarly, each position has its own peculiar advantages, respects in which it coheres uniquely well with deeply entrenched convictions that even its opponents might be expected to share. I shall try in the ensuing discussion to state and illustrate as vividly as I can the advantages and difficulties in all the major positions. Then I shall weigh the cases for and against the various alternatives. For those who disagree with my conclusion, the discussion will serve at least to locate the crucial issues in the controversy over the status of the fetus.

A proposed criterion for moral personhood is a statement of a characteristic (or set of characteristics) that its advocate deems both necessary and (jointly) sufficient for being a person in the moral sense. Such characteristics are not thought of as mere indexes, signs, or "litmus tests" of moral personhood but as more basic traits that actually confer moral personhood on whoever possesses them. All and only those beings having these characteristics have basic moral rights, in particular the right to full and equal protection against homicide. Thus, fetuses must be thought of as having this right if they satisfy a proposed criterion of personhood. The main types of criteria of moral personhood proposed by philosophers can be grouped under one or another of five different headings, which we shall examine in turn. Four of the five proposed criteria refer to possession of C (the traits we have listed as conferring commonsense personhood). One of these four specifies actual possession of C; the other three refer to either actual or potential possession of C. The remaining criterion, which we shall consider briefly first, makes no mention of C at all.

A. The Species Criterion

"All and only members of the biological species Homo sapiens, 'whoever is conceived by human beings,' are moral persons and thus are entitled to full and equal protection by the moral rule against homicide." The major advantage of this view (at least for some) is that it gives powerful support to those who would extend the protection of the rule against homicide to the fetus from the moment of conception. If this criterion is correct, it is not simply because of utilitarian reasons (such that it would usefully increase respect for life in, the community) that we must not abort human zygotes and embryos, but rather because we owe it to these minute entities themselves not to kill them, for as members of the human species they are already possessed of a full right to life equal in strength to that of any adult person.

The species criterion soon encounters serious difficulties. Against the view that membership in the species Homo sapiens is a necessary condition of membership in the class of moral persons, we have the possibility of there being moral persons from other planets who belong to other biological species. Moreover, some human beings -- in particular, those who are irreversibly comatose "vegetables" -- are human beings but doubtfully qualify as moral persons, a fact that casts serious doubt on the view that membership in the species Homo sapiens is a sufficient condition of being a moral person.

The species criterion might be defended against these objections if some persuasive reason could be given why moral personhood is a unique feature of all and only human beings. Aside from an arbitrary claim that this is "just obvious," a position that Singer argues amounts to a pernicious prejudice against nonhuman animals comparable to racism and sexism, the only possible way to defend this claim to uniqueness is by means of some theological argument: for example, that all human beings (including human fetuses) and only human beings (thereby excluding all nonhuman animals and possible beings from other planets) are moral persons because God has made this so.2 Now, if one already believes on faith that God had made it true that all and only humans are moral persons, then of course one has quite conclusive reason for believing that all and only humans are moral persons. But if we leave faith aside and confine our attention to reasons, then we shall have to ask what grounds there are for supposing that "God has made this so," and any reason we might have for doubting that it is so would count equally as a reason against supposing that God made it so. A good reason for doubting that 7 + 5 = 13 is an equally good reason for doubting that God made it to be the case that 7 + 5 = 13; a good reason for doubting that cruelty is morally right is, if anything, a better reason for denying that God made it to be the case that cruelty is right.

B. The Modified Species Criterion

"All and only members of species generally characterized by C, whether the species is Homo sapiens or another and whether or not the particular individual in question happens to possess C, are moral persons entitled to full and equal protection by the moral rule against homicide." This modification is designed to take the sting out of the first objection (above) to the unmodified species criterion. If there are other species or categories of moral persons in the universe, it concedes, then they too have moral rights. Indeed, if there are such, then all of their members are moral persons possessed of such rights, even those individuals who happen themselves to lack C because they are not yet fully developed or because they have been irreparably damaged.

The major difficulty for the modified species criterion is that it requires further explanation why C should determine moral personhood when applied to classes of creatures rather than to individual cases. Why is a permanently unconscious but living body of a human or an extragalactic person (or for that matter, a chimpanzee, if we should decide that that species as a whole is characterized by C) a moral person when it lacks as an individual the characteristics that determine moral personhood? Just because opposable thumbs are a characteristic of Homo sapiens, it does not follow that this or that particular Homo sapiens has opposable thumbs. There appears to be no reason for regarding right-possession any differently, in this regard, from thumb-possession.

C. The Strict Potentiality Criterion

"All and only those creatures who either actually or potentially possess C (that is, who either have C now or would come to have C in the natural course of events) are moral persons now, fully protected by the rule against homicide." This criterion also permits one to draw the line of moral personhood in the human species right at the moment of conception, which will be counted by some as an advantage. It also has the undeniable advantage of immunity from one charge of arbitrariness since it will extend moral personhood to all beings in any species or category who possess C, either actually or potentially. It may also cohere with our psychological attitudes, since it can explain why it is that many people, at least, think of unformed or unpretty fetuses are precious. Zygotes and embryos in particular are treasured not for what they are but for what they are biologically "programmed" to become in the fullness of time: real people fully possessed of C.

The difficulties of this criterion are of two general kinds, those deriving from the obscurity of the concept of "potentiality," which perhaps can be overcome, and the more serious difficulties of answering the charge that merely potential possession of any set of qualifications for a moral status does not logically ensure actual possession of that status. Consider just one of the problems raised by the concept of potentiality itself.3 How, it might be asked, can a mere zygote be a potential person, whereas a mere spermatozoon or a mere unfertilized ovum is not? If the spermatozoon and ovum we are talking about are precisely those that will combine in a few seconds to form a human zygote, why are they not potential zygotes, and thus potential people, now? The defender of the potentiality criterion will reply that it is only at the moment of conception that any being comes into existence with exactly the same chromosomal makeup as the human being that will later emerge from the womb, and it is that chromosomal combination that forms the potential person, not anything that exists before it comes together. The reply is probably a cogent one, but uncertainties about the concept of potentiality might make us hesitate, at first, to accept it, for we might be tempted to think of both the germ cell (spermatozoon or ovum) and the zygote as potentially a particular person, while holding that the differences between their potentials, though large and significant to be sure, are nevertheless differences in degree rather than kind. It would be well to resist that temptation, however, for it could lead us to the view that some of the entities and processes that combined still earlier to form a given spermatozoon were themselves potentially that spermatozoon and hence potentially the person that spermatozoon eventually became, and so on. At the end of that road is the proposition that everything is potentially everything else, and thus the destruction of all utility in the concept of potentiality. It is better to hold this particular line at the zygote.

The remaining difficulty for the strict potentiality criterion is much more serious. It is a logical error, some have charged, to deduce actual rights from merely potential (but not yet actual) qualification for those rights. What follows from potential qualification, it is said, is potential, not actual, rights; what entails actual rights is actual, not potential, qualification. As the Australian philosopher Stanley Benn puts it, "A potential president of the United States is not on that account Commander-in-Chief [of the U.S. Army and Navy]."4 This simple point can be called "the logical point about potentiality." Taken on its own terms, I don't see how it can be answered as an objection to the strict potentiality criterion. It is still open to antiabortionists to argue that merely potential commonsense personhood is a ground for duties we may have toward the potential person. But they cannot argue that it is the ground for the potential person's rights without committing a logical error.

D. The Modified or Gradualist Potentiality Criterion

"Potential possession of C confers not a right, but only a claim, to life, but that claim keeps growing stronger, requiring ever stronger reasons to override it, until the point when C is actually possessed, by which time it has become a full right to life." This modification of the potentiality criterion has one distinct and important advantage. It coheres with the widely shared feeling that the moral seriousness of abortion increases with the age of the fetus. It is extremely difficult to believe on other than very specific theological grounds that a zygote one day after conception is the sort of being that can have any rights at all, much less the whole armory of "human rights" including "the right to life." But it is equally difficult for a great many people to believe that a full-term fetus one day before birth does not have a right to life. Moreover, it is very difficult to find one point in the continuous development of the fetus before which it is utterly without rights and after which it has exactly the same rights as any adult human being. Some rights in postnatal human life can be acquired instantly or suddenly; the rights of citizenship, for example, come into existence at a precise moment in the naturalization proceedings after an oath has been administered and a judicial pronouncement formally produced and certified. Similarly, the rights of husbands and wives come into existence at just that moment when an authorized person utters the words "I hereby pronounce you husband and wife." But the rights of the fetus cannot possibly jump in this fashion from nonbeing to being at some precise moment in pregnancy. The alternative is to think of them as growing steadily and gradually throughout the entire nine-month period until they are virtually "mature" at parturition. There is, in short, a kind of growth in "moral weight" that proceeds in parallel fashion with the physical growth and development of the fetus.

An "immature right" on this view is not to be thought of simply as no right at all, as if in morals a miss were as good as a mile. A better characterization of the unfinished right would be a "weak right," a claim with some moral force proportional to its degree, of development but not yet as much force as a fully matured right. The key word in this account is "claim." Elsewhere I have given an account of the difference between having a right (which I defined as a "valid claim") and having a claim that is not, or not quite, valid. What would the latter be like?

One might accumulate just enough evidence to argue with relevance and cogency that one has a right . . . although one's case might not be overwhelmingly conclusive. The argument might be strong enough to entitle one to a hearing and fair consideration. When one is in this position, it might be said that one "has a claim" that deserves to be weighed carefully. Nevertheless the balance of reasons may turn out to militate against recognition of the claim, so that the claim is not a valid claim or right.5

Now there are various ways in which a claim can fail to be a right. There are many examples, particularly from the law, where all the claims to some property, including some that are relevantly made and worthy of respect, are rejected, simply because none of them is deemed strong enough to qualify as a right. In such cases, a miss truly is as good as a mile. But in other cases, an acknowledged claim of (say) medium strength will be strong enough to be a right unless a stronger claim appears on the scene to override it. For these conflict situations, card games provide a useful analogy. In poker, three of a kind is good enough to win the pot unless one of the other players "makes claim" to the pot with a higher hand, say a flush or a full house. The player who claims the pot with three of a kind "has a claim" to the pot that is overridden by the stronger claim of the player with the full house. The strongest claim presented will, by that fact, constitute a right to take the money. The player who withdrew with a four flush had "no claim at all," but even that person's hand might have established a right to the pot if no stronger claim were in conflict with it.

The analogy applies to the abortion situation in the following way. The game has at least two players, the mother and the fetus, though more can play, and sometimes the father and/or the doctor are involved too. For the first few weeks of its life, the fetus (zygote, embryo) has hardly any claim to life at all, and virtually any reason of the mother's for aborting it will be strong enough to override a claim made in the fetus's behalf. At any stage in the game, any reason the mother might have for aborting will constitute a claim, but as the fetus matures, its claims grow stronger, requiring ever stronger claims to override them. After three months or so, the fact that an abortion would be "convenient" for the mother will not be a strong enough claim, and the fetus's claim to life will defeat it. In that case, the fetus can be said to have a valid claim or right to life in the same sense that the poker player's full house gives him a right to the pot: it is a right in the sense that it is the strongest of the conflicting claims, not in the sense that it is stronger than any conflicting claim that could conceivably come up. By the time the fetus has become a neonate (a newborn child), however, it has a "right to life" of the same kind all people have, and no mere conflicting claim can override it. (Perhaps more accurately, only claims that other human persons make in self-defense to their own lives can ever have an equal strength.)

The modified potentiality criterion has the attractiveness characteristic of compromise theories when fierce ideological quarrels rage between partisans of more extreme views. It shares one fatal flaw, however, with the strict potentiality criterion. Despite its greater flexibility, it cannot evade "the logical point about potentiality." A highly developed fetus is much closer to being a commonsense person with all the developed traits that qualify it for moral personhood than is the mere zygote. But being almost qualified for rights is not the same thing as being partially qualified for rights; nor is it the same thing as being qualified for partial rights, quasi-rights, or weak rights. The advanced fetus is closer to being a person than is the zygote, just as a dog is closer to personhood than a jellyfish, but that is not the same thing as being "more of a person." In 1930, when he was six years old, Jimmy Carter didn't know it, but he was a potential president of the United States. That gave him no claim then, not even a very weak claim, to give commands to the U.S. Army and Navy. Franklin D. Roosevelt in 1930 was only two years away from the presidency, so he was a potential president in a much stronger way (the potentiality was much less remote) than was young Jimmy. Nevertheless, he was not actually president, and he had no more of a claim to the prerogative of the office than did Carter. The analogy to fetuses in different stages of development is of course imperfect. But in both cases it would seem to be invalid to infer the existence of a "weak version of a right" from an "almost qualification" for the full right. In summary, the modified potentiality criterion, insofar as it permits the potential possession of C to be a sufficient condition for the actual possession of claims, and in some cases of rights, is seriously flawed in the same manner as the strict potentiality criterion.

E. The Actual-Possession Criterion

"At any given time (t), all and only those creatures who actually possess C are moral persons at t, whatever species or category they may happen to belong to." This simple and straightforward criterion has a number of conspicuous advantages. We should consider it with respect even before we examine its difficulties if only because the difficulties of its major rivals are so severe. Moreover, it has a certain tidy symmetry about it, since it makes the overlap between commonsense personhood and moral personhood complete -- a total correspondence with no loose ends left over in either direction. There can be no actual commonsense persons who are not actual moral persons, nor can there be any actual moral persons who are not actual commonsense persons. Moral personhood is not established simply by species membership, associations, or potentialities. Instead, it is conferred by the same characteristics (C) that lead us to recognize personhood wherever we find it. It is no accident, no mere coincidence, that we use the moral term 'person' for those beings, and only those beings, who have C. The characteristics that confer commonsense personhood are not arbitrary bases for rights and duties, such as race, sex, or species membership; rather they are the traits that make sense out of rights and duties and without which those moral attributes would have no point or function. It is because people are conscious; have a sense of their personal identities; have plans, goals, and projects; experience emotions; are liable to pains, anxieties, and frustrations; can reason and bargain, and so on -- it is because of these attributes that people have values and interests, desires and expectations of their own, including a stake in their own futures, and a personal well-being of a sort we cannot ascribe to unconscious or nonrational beings. Because of their developed capacities they can assume duties and responsibilities and can have and make claims on one another. Only because of their sense of self, their life plans, their value hierarchies, and their stakes in their own futures can they be ascribed fundamental rights. There is nothing arbitrary about these linkages. For these reasons I am inclined to believe that the actual-possession criterion is the correct one.

Despite these impressive advantages, however, the actual-possession criterion must face a serious difficulty, namely that it implies that small infants (neonates) are not moral persons. There is very little more reason, after all, to attribute C to neonates than to advanced fetuses still in utero. Perhaps during the first few days after birth the infant is conscious and able to feel pain, but it is unlikely that it has a concept of itself or of its future life, that it has plans and goals, that it can think consecutively, and the like. In fact, the whole complex of traits that make up C is not obviously present until the second year of childhood. And that would seem to imply, according to the criterion we are considering, that the deliberate destruction of babies in their first year is no violation of their rights. And that might seem to entail that there is nothing wrong with infanticide (the deliberate killing of infants). But infanticide is wrong. Therefore, critics of the actual-possession criterion have argued that we ought to reject this criterion.

The Killing of Normal Infants

Advocates of the actual-possession criterion have a reply to this objection. Even if infanticide is not the murder of a moral person, they believe, it may yet be wrong and properly forbidden on other grounds. To make this clearer, it is useful to distinguish between

  1. the case of killing a normal healthy infant or an infant whose handicaps are not so serious as to make a worthwhile future life impossible and
  2. the case of killing severely deformed or incurably diseased infants.

Most advocates of the actual-possession criterion take a strong stand against infanticide in the first (the normal) case. It would be seriously wrong for a mother to kill her physically normal infant, they contend, even though such a killing would not violate anyone's right to life. The same reasons that make infanticide in the normal case wrong also justify its prohibition by the criminal law. The moral rule that condemns these killings and the legal rule that renders them punishable are both supported by "utilitarian reasons," that is, considerations of what is called "social utility," "the common good," "the public interest," and the like. Nature has apparently implanted in us an instinctive tenderness toward infants that has proven extremely useful to the species, not only because it leads us to protect our young from death, and thus keep our population up, but also because infants usually grow into adults, and in Benn's words, "if as infants they are not treated with some minimal degree of tenderness and consideration, they will suffer for it later, as persons."6 One might add that when they are adults, others will suffer for it too, at their hands. Spontaneous warmyh and sympathy toward babies then clearly has a great deal of social utility, and insofar as infanticide would tend to weaken that socially valuable response, it is, on utilitarian grounds, morally wrong.

There are other examples of wrongful and properly prohibitable acts that violate no one's rights. It would be wrong, for example, to hack up Grandfather's body after he has died a natural death, and dispose of his remains in the trash can on a cold winter's morning. That would be wrong not because it violates Grandfather's rights; he is dead and no longer has the same sort of rights as the rest of us, and we can make it part of the example that he was not offended while alive at the thought of such posthumous treatment and indeed even consented to it in advance. Somehow acts of this kind if not forbidden would strike at our respect for living human persons (without which organized society would be impossible) in the most keenly threatening way. (It might also be unhygienic and shocking to trash collectors -- less important but equally relevant utilitarian considerations.)

The Killing of Radically Deformed Infants

The general utilitarian reasons that support a rather rigid rule against infanticide in the case of the normal (and not too abnormal) infant might not be sufficiently strong to rule out infanticide (under very special and strict circumstances) when the infant is extremely deformed or diseased. Very likely, a purely utilitarian-based rule against homicide would have exceptive clauses for extremely abnormal neonates. In this respect such rules would differ sharply from rules against infanticide that derive from the ascription to the newborn infant of a full-fledged right to life. If the deformed neonate is a moral person, then he or she is as fully entitled to protection under the rule forbidding homicide as any reader of these words; if the neonate is not a moral person, then in extreme cases there may be a case on balance for killing him. The Partisan of the actual-possession criterion of moral personhood actually takes this consequence of nonpersonhood to be an advantage of his view rather than an embarrassment for it. If his view is correct, then we can destroy hopelessly malformed infants before they grow into moral persons, thus saving them trom a longer life that would be so horrible as to be "not worth living," and this can be done without violating their rights.

Indeed, failure to kill such infants before they reach moral personhood may itself be a violation of their rights, arcorrding to this view. For if we permit such children to grow into moral personhood knowing full well that the conditions for the fulfillment of their most basic future interests have already been destroyed, then we have wronged these persons before they even exist (as persons), and when they become persons, they can claim (or it can be claimed in their behalf) that they have been wronged. I have argued elsewhere that an extension of the idea of a birthright is suggested by this point. If we know that it will never be possible for a fetus or neonate to have that to which he has a birthright and we allow him nevertheless to be born or to survive into personhood, then that fetus or neonate is wronged, and we become a party to the violation of his rights.7

Not just any physical or mental handicaps, of course, can render a life "not worth living." Indeed, as the testimony of some thalidomide babies8 now growing into adulthood shows, it is possible (given exceptional care) to live a valuable life even without arms and legs and full vision. But there may be some extreme cases where deformities are not merely "handicaps" in the pursuit of happiness but guarantees that the pursuit must fail. A brain-damaged, retarded child born deaf, blind, partially paralyzed, and doomed to constant pain might be such a case. Given the powerful general utilitarian case, against infanticide, however, the defender of the "right to die" position must admit that in cases of doubt, the burden of showing that a worthwhile life is impossible rests on the person who would elect a quick and painless death for the infant. And there is almost always some doubt.

Implications for the Problem of Abortion

The implications of the actual-possession criterion for the question of the status of the fetus as a moral person are straightforward. Since the fetus does not actually possess those characteristics (C) that we earlier listed as necessary and sufficient for possessing the right to life, the fetus does not possess that right Given this criterion, therefore, abortion never involves violating a fetus's right to life, and permitting a fetus to be born is never anything we owe it, is never something that is its due.

It does not follow, however, that abortion is never wrong. As we saw earlier, despite the fact that infants fail to meet the actual-possession criterion and thus are not moral persons, reasons can be given, of a utilitarian kind, why it is wrong to kill them, at least if they are not radically deformed. It is possible, therefore, that similar reasons can be given in opposition to aborting fetuses at later stages in their development if they are likely not to be radically deformed when born.

Utilitarian reasons of the sort we have considered are so very important that they might suffice to rule out harsh or destructive treatment of any nonperson whose resemblance or similarity to real persons is very close: not only deceased ex-persons and small babies, but even adult primates and human fetuses in the final trimester of pregnancy. Mr. Justice Blackmun may have had such considerations in mind when in his majority opinion in Roe v. Wade he declared that even though no fetuses are legal persons protected by the law of homicide, nevertheless during the final trimester. "The State in promoting its interest in the potentiality of human life, may if it chooses regulate, and even proscribe, abortion. . . ."9 Whatever interest the State has in "the potentiality of human life" must be derivative from the plain interest it has in preserving and promoting respect for actual human life. It is not only potential persons who merit our derivative respect but all near-persons including higher animals, dead people, infants, and well-developed fetuses, those beings whose similarity to real persons is close enough to render them sacred symbols of the real thing.

In the light of these considerations, it seems that a gradualist approach similar to that discussed earlier is a more plausible solution to the general problem of the moral justifiability of abortion than it is to the narrow problem of the criterion of moral personhood. Even if the fetus as a merely potential person lacks an actual right to life, and even if it would not be homicide therefore to kill it, its potential personhood may yet constitute a reason against killing it that requires an even stronger reason on the other side if abortion is to be justified. If that is so, it is not implausible to suppose that the more advanced the potential for personhood, the more stringent the case against killing. As we have seen, there are reasons relevant to our moral decisions other than considerations of rights, so that sometimes actions can be judged morally wrong even though they violate no one's rights. Killing a fetus, in that case, could be wrong in certain circumstances, even though it violated no rights of the fetus, even though the fetus was not a moral person, even though the act was in no sense a murder.


Killing human beings (homicide) is forbidden both by our criminal law and by the moral rules that are accepted in all civilized communities. If the fetus at any point in its development is a human being, then to kill it at that point is homicide, and if done without excuse or mitigation, murder. But the term 'human being' is subtly ambiguous. The fetus at all stages is obviously human in the genetic sense, but that is not the sense of the term intended in the moral rule against homicide. For a genetically human entity to have a right to life it must be a human being in the sense of a person. But the term 'person' is also ambiguous. In the commonsense descriptive meaning of the term, it refers to any being of any species or category who has certain familiar characteristics, of which consciousness of the world, self-concepts, and the capacity to plan ahead are prominent. In the purely normative (moral or legal), sense, a person is any being who has certain rights and/or duties, whatever his other characteristics. Whether or not abortion is homicide depends on what the correct criterion of moral personhood is.

We considered five leading formulations of the criterion of moral personhood and found that they are all subject to various embarrassments. One formulation in terms of species membership seemed both too broad and too narrow, and in the end dependent on an arbitrary preference for our own species. A more careful formulation escaped the charge of being too restrictive and the charge of arbitrariness but suffered from making the status of an individual derived from his membership in a group rather than from his own intrinsic characteristics. The two formulations in terms of potential possession of the characteristics definitive of commonsense personhood both stumbled on "the logical point about potentiality," that potential qualification for a right does not entail actual possession of that right. The modified or gradualist formulation of the potentiality criterion, however, does have some attractive features, and could be reformulated as a more plausible answer to another question, that about the moral permissibility of abortion. Even if the fetus is not a person and lacks a right to life, ever stronger reasons might be required to justify aborting it as it grows older and more similar to a person.

The weaknesses of the first four proposed criteria of moral personhood create a strong presumption in favor of the remaining one, the "actual-possession" criterion. It is clear that fetuses are not "people" in the ordinary commonsense meaning of that term, hence, according to our final criterion, they are not moral persons either, since this criterion of moral personhood simply adopts the criteria of commonsense personhood. The very grave difficulty of this criterion is that it entails that infants are not people either, during the first few months or more of their lives. That is a genuine difficulty for the theory, but a far greater embarrassment can be avoided. Because there are powerful reasons against infanticide that apply even if the infant is not a moral person, the actual-possession criterion is not subject to the devastating objection that it would morally or legally justify infanticide on demand.


The problem of the status of the fetus is the first and perhaps the most difficult of the questions that must be settled before we can come to a considered view about the moral justifiability of abortion, but its solution does not necessarily resolve all moral perplexities about abortion. Even if we were to grant that the fetus is a moral person and thus has a valid claim to life, it would not follow that abortion is always wrong. For there are other moral persons, in addition to the fetus, whose interests are involved. The woman in whose uterus the fetus abides, in particular, has needs and interests that may well conflict with bringing the fetus to term. Do any of these needs and interests of the woman provide grounds for her having a genuine claim to an abortion and, if they do, which of the two conflicting claims -- the woman's claim to an abortion or the fetus's claim to life -- ought to be respected if they happen to conflict? This is the second major moral question that needs to be examined with all the care we can muster. To do this, one very important assumption must be made -- namely, that the fetus is a moral person and so has a valid claim to life. As we have seen in the previous section, this assumption might very well be unfounded and is unfounded in fact if we accept what appears to be the most reasonable criterion for moral personhood -- namely, the actual-possession criterion. For purposes of the present section, however, we shall assume that the fetus is a moral person; this will enable us to investigate whose claim, the fetus's or the woman's, ought to be honored if both have genuine but conflicting claims.


The right to an abortion that is often claimed on behalf of all women is a discretionary right to be excercised or not, on a given occasion, as the woman sees fit. For that reason it is sometimes called a "right to choose." If a pregnant woman has such a right, then it is up to her, and her alone, whether to bear the child or to have it aborted. She is at liberty to bear it if she choses and at liberty to have it aborted if she chooses. She has no duty to bear it, but neither can she have a duty, imposed from without, to abort it. In respect to the fetus, her choice is sovereign. Correlated with this liberty is a duty of others not to interfere with its exercise and not to withhold the necessary means for its exercise. These duties are owed to her if she has a discretionary right to abortion, and she can claim their discharge as her due.

As a discretionary right, a right to an abortion would resemble the "right to liberty," or the right to move about or travel as one wishes. One is under no obligation to leave or stay home or to go to one destination rather than another, so that it is one's own choice that determines one's movements. But the right to move about at will, like other discretionary rights, is subject to limits. One person's liberty of movement, for example, comes to an end at the boundary of another person's property. The discretionary right to an abortion may be limited in similar ways, so that the statement of a specific right of a particular woman in a definite set of concrete circumstances may need to be qualified by various exceptive clauses -- for example, ", . . may choose to have an abortion except when the fetus is viable." Which exceptive clauses, if any, must be appended to the formulation of the right to an abortion depends on what the basis of this discretionary right is thought to be. For example, if a woman is thought to have a right to an abortion because she has a right to property and because the fetus is said to be her property, then the only exceptions there could be to exercising the right to an abortion would be those that restrict the disposing of one's property. What we must realize, then, is that the alleged right to an abortion cannot be understood in a vacuum; it is a right that can only be understood by reference to the other, more fundamental rights from which it has often been claimed to be derived. Three of these rights and their possible association with the right to an abortion deserve our closest scrutiny. These are

  1. the previously mentioned property rights,
  2. the right to self-defense, and
  3. the right to bodily autonomy.
We shall consider each in its turn.


Property Rights over One's Body

Within very wide limits any person has a right to control the uses of his or her own body. With only rare exceptions, surgeons are required to secure the consent of the patient before operating because the body to be cut open, after all, is the patient's own, and he or she has the chief interest in it, and should therefore have the chief "say" over what is done to it. If we think of a fetus as literally a "part" of a woman's body in the same sense as, say, an organ, or as a mere growth attached to a part of the body on the model of a tumor or a wart, then it would seem to follow that the woman may choose to have it removed if she wishes just as she may refuse to have it removed if she prefers. It is highly implausible, however, to think of a human fetus, even if it does fall short of moral personhood, as no more than a temporary organ or a parasitic growth. A fetus is not a constituent organ of the mother, like her vermiform appendix, but rather an independent entity temporarily growing inside the mother.

It would be still less plausible to derive a maternal right to an abortion from a characterization of the fetus as the property of its mother and thus in the same category as the mother's wristwatch, clothing, or jewelry. One may abandon or destroy one's personal property if one wishes; one's entitlement to do those things is one of the "property rights" that define ownership. But one would think that the father would have equal or near-equal rights of disposal if the fetus were "property." It is not in his body, to be sure, but he contributed as much genetically to its existence as did the mother and might therefore make just as strong (or just as weak) a claim to ownership over it. But neither claim would make very good conceptual sense. If fetuses were property, we would find nothing odd in the notion that they can be bought and sold, rented out, leased, used as collateral on loans, and so on. But no one has ever seriously entertained such suggestions. Finally, we must remember the methodological assumption that we shall make throughout this section, at least for the sake of the argument, that the fetus is a full moral person, with a right to life like yours and mine. On this assumption it would probably be contradictory to think of the fetus as anyone's property, especially if property rights include what we might call "the right of disposal" -- to abandon or destroy as one chooses.

It is more plausible at first sight to claim that the pregnant woman owns not the fetus but the body in which she shelters the fetus. On this analogy, she owns her body (and particularly her womb) in roughly the way an innkeeper owns a hotel or a homeowner her house and garden. These analogies, however, are also defective. To begin with, it is somewhat paradoxical to think of the relation between a person and her body as similar to that of ownership. Is it possible to sell or rent or lease one's body without selling, renting, or leasing oneself? If one's body were one's property the answer would be affirmative, but in fact one's relationship to one's own body is much more intimate than the ownership model suggests. More important for our present purposes, the legal analogies to the rights of innkeepers and householders will not bear scrutiny. One cannot conceive of what it would be like for a fetus to enter into a contract with a woman for the use of her womb for nine months or to fall in arrears in its payments and thus forfeit its right of occupancy. Surely that cannot be the most apt analogy on which to base the woman's abortion rights! Besides, whatever this, that, or the other legal statute may say about the matter, one is no morally entitled, in virtue of one's property rights, to expel a weak and helpless person from one's shelter when that is tantamount to consigning the person to a certain death, and surely one is not entitled to shoot and kill a trespasser who will not, or cannot, leave one's property. In no department of human life does the vindication of property rights justify homicide. The maternal right to an abortion, therefore, cannot be founded on the more basic right to property.

Self-Defense and Proportionality

Except for the most extreme pacifists, moralists agree that killing can be justified if done in self-defense. If, for example, one man (A) is attacked with a lethal weapon by another (B), we think that A has a right to defend himself against B's attack. Sometimes, in fact, we think that A would be justified in killing B if this were the only way for A to defend himself. Now, some of those who urge the maternal right to an abortion believe that this right is associated with the more basic right to self-defense. There are many difficulties standing in the way to rational acceptance of this view. In particular, the innocence and the nonaggressive nature of the fetus need our special attention. We shall turn to these matters shortly. First, though, it is important to realize what reasons would not count as morally good reasons for an abortion if the right to an abortion were supposed to be founded on the more basic right of self-defense.

All parties to the abortion dispute must agree that many women can be harmed if they are required to bring an unwanted fetus to term. Unwanted sexual intercourse imposed on a woman by a rapist can inflict on its victim severe psychological trauma of a sort deemed so serious by the law that a woman is entitled under some rules to use deadly force if necessary to prevent it. Similarly, an unwanted pregnancy in some circumstances can inflict equally severe psychological injury on a woman who is forced to carry her child to birth. There are various familiar examples of such harm. To borrow an example from Judith Thomson, a philosophy professor at the Massachusetts Institute of Technology: A terrified fourteen-year-old high schoolgirl whose pregnancy has been caused by rape has already suffered one severe trauma. If she is now required, over her protests, to carry the child to full term despite her fear, anguish, deep depression, and fancied public mortification, the harmful ramifications may be multiplied a hundredfold. The forty-year-old housewife who has exhausted herself raising a large family in unfavorable economic circumstances while dependent upon an unreliable and unsympathetic husband may find herself, to her horror, pregnant again and rightly feel that if she is forced to give birth to another child, she will forfeit her last opportunity to escape the intolerably squalid conditions of her life. A man must be morally blind not to acknowledge the severe harms that enforced continuance of unwanted pregnancies can inflict on women. An unwanted child need not literally cost the woman her life, but it can effectively ruin her life from her point of view, and it is a useful moral exercise for men to put themselves imaginatively in the woman's place to share that point of view.

At this stage in the argument the antiabortionist has a ready rejoinder. A woman need not keep her child, assume the responsibilities of raising it to adulthood, and forfeit her opportunities for self-fulfillment, he might reply, simply because she forgoes an abortion. She can always put the child up for adoption and be assured in the process that it will find loving foster parents who will give it a good upbringing. All she really has to suffer, the rejoinder concludes, is nine months of minor physical inconvenience. This is an argument that comes easily to the lips of men, but it betrays the grossest sort of masculine insensitivity. In the first place, it is not always true that a woman can have her baby adopted. If she is married, that transaction may require the consent of her husband, and the consent might not be forthcoming. But waiving that point, the possibility of adoption does not give much comfort to the unhappily pregnant woman, for it imposes on her a cruel dilemma and an anguish that far surpasses "minor inconvenience." In effect, she has two choices, both of which are intolerable to her. She can carry the child to term and keep it, thus incurring the very consequences that make her unwilling to remain pregnant, or she can nourish the fetus to full size, go into labor, give birth to her baby, and then have it rudely wrenched away, never to be seen by her again. Let moralistic males imagine what an emotional jolt that must be!

Still, on the scale of harms, mere traumas and frustrations are not exactly equal to death. Few women would choose their own deaths in preference to the harms that may come from producing children. According to a common interpretation of the self-defense rule, however, the harm to be averted by a violent act in self-defense need not be identical in severity to that which is inflicted upon one's assailant but only somehow "proportional'jtojt. Both our prevailing morality and our legal traditions permit the use of lethal force to prevent harms that are less serious than death, so it is plausible to assume that the rule of "proportionality" can be satisfied by something less than equality of harms. But how much less? The late Jane English, a philosopher from the University of North Carolina, offers an answer that, though vague, is in accordance with the moral sentiments of most people when they think of situations other than that involving abortion:

How severe an injury may you inflict in self-defense? In part this depends upon the severity of the injury to be avoided: you may not shoot someone merely to avoid having your clothes torn. This might lead one to the mistaken conclusion that the defense may only equal the threatened injury in severity; that to avoid death you may kill, but to avoid a black eye you may only inflict a black eye or the equivalent. Rather our laws and customs seem to say that you may create an injury somewhat but not enormously greater than the injury to be avoided. To fend off an attack whose outcome would be as serious as rape, a severe beating, or the loss of a finger, you may shoot; to avoid having your clothes torn, you may blacken an eye.10 [Emphasis added.]
Applying English's answer to the abortion case, and assuming that both the fetus and the woman have legitimate claims, we derive the conclusion that killing the "fetal person" would not be justified when it is done merely to prevent relatively trivial harms to the mother's interests. Not all cases of abortion, therefore, can morally be justified, even if there is a maternal right to abortion derived from the more basic right to self-defense.

Self-Defense: The Problem of the Innocent Aggressor

Suppose, however, that the harms that will probably be caused to the mother if the fetus is brought to term are not trivial but serious. Here we have a case where the mother's right to have her important interests respected clashes with the assumed right to life of the fetus. In these circumstances, don't the mother's claims outweigh the fetus's? Doesn't self-defense in these circumstances justify abortion?

There is a serious, previously undiscussed difficulty that calls out for attention. Consider a case where someone aggressively attacks another. The reason we think that, to use English's expression, we may in self-defense create a "somewhat but not enormously greater injury" than would have been caused by the aggressor is because we think of the aggressor as the party who is morally at fault. If he had not launched the aggression in the first place, there should have been no occasion for the use of force. Since the whole episode was the aggressor's fault, his interests should not count for as much as those of the innocent victim. It is a shame that anybody has to be seriously hurt, but if it comes down to an inescapable choice between the innocent party suffering a serious harm or the culpable party suffering a still more serious harm, then the latter is the lesser of the two evils. Aggressors of course, for all their guilt, remain human beings, and consequently they do not forfeit all their human rights in launching an attack. We still may not kill them to prevent them from stealing $10. But their culpability does cost them their right to equal consideration; we may kill them to prevent them from causing serious harm.

But now suppose that the party who threatens us, even though he is the aggressor who initiates the whole episode, is not morally at fault. Suppose the person cannot act otherwise in the circumstances and thus cannot justly be held morally responsible. For example, he was temporarily (or permanently) insane, or it was a case of mistaken identity (he mistook you for a former Gestapo agent to whom you bear a striking resemblance), or someone had drugged the person's breakfast cereal and his behavior was influenced by the drug. George Fletcher, a Columbia law professor, provides a vivid illustration of the probleem in what he calls "the case of the psychotic aggressor":

Imagine that your companion in an elevator goes berserk and attacks you with a knife. There is no escape: the only way to avoid serious bodily harm or even death is to kill him. The assailant acts purposively in the sense that his means further his aggressive end . . . [but] he does act in a frenzy or a fut . . . [and] it is clear that his conduct is non-responsible. If he were brought to trial for his attack, he would have a valid defense of insanity.11
The general problem, as lawyers would put it, is "whether self-defense applies against an excused but unjustified aggression."12 To justify an act is to show that it was the right thing to do in the circumstances; to excuse an act is to show that although it was unjustified, the actor didn't mean it or couldn't help it, that it was not, properly speaking, his doing at all. In the "excused but unjustified aggression" we have a more plausible model for the application of self-defense to the problem of abortion, for the fetus is surely innocent (not because of insanity but because of immaturity, and because it did not choose to threaten its mother -- it did not "ask to be born").

Upon reflection, most of us would agree, I think, that one would be justified in killing even an innocent aggressor if that seemed necessary to save one's own life or to prevent one from suffering serious bodily injury. Surely we would not judge harshly the slightly built lady who shoots the armed stranger who goes berserk in the elevator. If we were in her shoes, we too would protect ourselves at all costs to the assailant, just as we would against wild animals, runaway trucks, or impersonal forces of nature. But while the berserk assailant, as well as those persons mentioned in the last paragraph, all are innocent -- are not morally responsible for what they do -- they all are assailants, and in this respect they differ in a quite fundamental respect from the fetus. For the fetus is not only innocent but also not an aggressor. It didn't start the trouble in any fashion. Thus, it would seem that while we are justified in killing an innocent assailant if this is the only way to prevent him from killing us, it does not follow that we are similarly justified in killing a fetal person, since, unlike the innocent aggressor, the fetus is not an aggressor at all.

Judith Thomson has challenged this argument. She presents the following farfetched but coherent hypothetical example:

Aggressor is driving his tank at you. But he has taken care to arrange that a baby is strapped to the front of the tank so that if you use your anti-tank gun, you will not only kill Aggressor, you will kill the baby. Now Aggressor, admittedly, is in the process of trying to kill you; but that baby isn't. Yet you can presumably go ahead and use the gun, even though this involves killing the baby as well as Aggressor.13
The baby in this example is not only "innocent" but also the "innocent shield of a threat."14 Still it is hard to quarrel with Thomson's judgment that you may (not that you should) take the baby's life if necessary to save your own, that it is morally permissible, even if it is not morally obligatory, to do so. After all, you are (by hypothesis) perfectly innocent too. This example makes a better analogy to the abortion situation than any we have considered thus far, but there are still significant dissimilarities. Unless the fetus is the product of rape, it cannot conceivably be the shield of some third-party aggressor. There is simply no interpersonal "aggression" involved at all in the normal pregnancy. There may nevertheless be a genuine threat to the well-being of the mother, and if that threat is to her very life, then perhaps she does have a right to kill it, if necessary, in self-defense. At any rate, if the threatened victim in Thomson's tank example is justified in killing the innocent shield, then the pregnant woman threatened with similar harm is similarly entitled. But all that this would establish is that abortion is justified only if it is probably required to save the mother's life. So not only could we not use the self-defense argument to justify abortion for trivial reasons, as was argued earlier; it appears that the only reason that authorizes its use is the one that cites the fact that the mother will probably die if the fetus is not aborted.

Bodily Autonomy: The Example of the Plugged-in Violinist

The trouble with the use of self-defense as a model for abortion rights is that none of the examples of self-defense makes an exact analogy to the abortion situation. The examples that come closest to providing models for justified abortion are the "innocent aggressor cases" and these would apply, as we have seen, only to abortions that are necessary to prevent death to the mother. Even these examples do not fit the abortion case exactly, since the fetus is in no way itself an aggressor, culpable or innocent, but is at most a "nonaggressive, nonculpable threat," in some respects like an innocent shield.15 And the more we change the examples to bring them closer to the situation of the fetus, the less clear is their resemblance to the central models of self-defense. Once we are allowed to protect ourselves (and especially to protect interests less weighty than self-preservation) at the expense of nonaggressive innocents, it becomes difficult to distinguish the latter from innocent bystanders whom we kill as means to our own good, and that, in turn, begins to look like unvarnished murder. The killing of an innocent person simply because his continued existence in the circumstances would make the killer's life miserable is a homicide that cannot be justified. It is not self-defense to kill your boss because he makes your work life intolerable and you are unable to find another job, or to kill your spouse because he or she nags you to the point of extreme misery and will not agree to a divorce,16 or (closer to the point) to kill your shipwrecked fellow passenger in the lifeboat because there are provisions sufficient for only one to survive and he claims half of them, or to kill your innocent rival for a position or a prize because you can win only if he is out of the running. In all these cases the victim is either innocent or relatively innocent and in no way a direct aggressor.

Partly because of deficiencies in the hypothetical examples of self-defense, Thomson invented a different sort of example intended at once to be a much closer analogy to the abortion situation and also such that the killing can be seen to be morally justified for reasons less compelling than defense of the killer's very life:

You wake up in the morning and find yourself back to back in bed with an unconscious violinist. A famous unconscious violinist. He has been found to have a fatal kidney ailment, and the Society of Music Lovers has canvassed all the available medical records and found that you alone have the right blood type to help. They have therefore kidnapped you, and last night the violinist's circulatory system was plugged into yours, so that your kidneys can be used to extract poisons from his blood as well as your own. The director of the hospital now tells you, "Look, we're sorry the Society of Music Lovers did this to you -- we would never have permitted it if we had known. But still they did it, and the violinist now is plugged into you. To unplug you would be to kill him. But never mind, it's only for nine months. By then he will have recovered from his ailment, and can safely be unplugged from you." Is it morally incumbent on you to accede to this situation? No doubt it would be very nice of you if you did, a great kindness. But do you have to accede to it? . . . What if the director . . . says . . . "Granted you have a right to decide what happens in and to your body, but a person's right to life outweighs your right to decide what happens in and to your body. So you cannot . . . be unplugged from him." I imagine you would regard this as outrageous . . . .17

Suppose that you defy the director on your own, and exercise your control over your own body by unplugging the unconscious violinist, thereby causing his death. This would be to kill in defense of an interest far less important than self-preservation or the prevention of serious, injury to oneself. And it would be to kill an innocent nonaggressor, indeed a victim who remains unconscious throughout the entire period during which he is a threat. We have, therefore, an example which -- if it works -- offers far more encouragement to the proabortion position than the model of self-defense does. We must now pose two questions: (1) Would you in fact be morally justified in unplugging the violinist? and (2) How close an analogy does this bizarre example make to the abortion situation?

There is no way to argue conclusively that unplugging the violinist would be morally justified. Thomson can only make the picture as vividly persuasive as possible and then appeal to her reader's intuitions. It is not an easy case, and neither an affirmative nor a negative judgment will seem self-evident to everyone. Still the verdict for justification seems as strong as in some of the other examples of killing innocent threats, and some additional considerations can be brought to bear in its support. There is, after all, a clear "intuition" in support of a basic right "to decide what happens in and to one's own body," even though the limits of that right are lost in a fog of controversy. So unless there is some stronger competing claim, anyone has a right to refuse to consent to surgery or to enforced attachment to a machine. Or indeed to an unconscious violinist. But what of the competing claim in this example, the violinist's right to life? That is another basic right that is vague around the edges.

In its noncontroversial core, the right to life is a right not to be killed directly (except under very special circumstances) and to be rescued from impending death whenever this can be done without unreasonable sacrifice. But as Warren has pointed out, one person's right to life does not impose a correlative duty on another person to do "whatever is necessary to keep him alive."18 And though we all have general duties to come to the assistance of strangers in peril, we cannot be forced to make enormous sacrifices or to run unreasonably high risks to keep people alive when we stand in no special relationship to them, like "father" or "lifeguard." The wife of the violinist perhaps would have a duty to stay plugged to him (if that would help) for nine months; but the random stranger has no such duty at all. So there is good reason to grant Thomson her claim that a stranger would have a right to unplug the violinist from herself.

But how close an analogy after all is this to the normal case of pregnancy? Several differences come immediately to mind. In the normal case of pregnancy, the woman is not confined to her bed for nine months but can continue to work and function efficiently in the world until the final trimester at least. This difference, however, is of doubtful significance, since Thomson's argument is not based on a right to the protection of one's interest in efficient mobility but rather on a right to decide on the uses of one's own body, which is quite another thing. Another difference is that the mother and her fetus are not exactly "random strangers" in the same sense that the woman and the violinist are. Again the relationship between mother and fetus seems to be in a class by itself. If the person who needs to use the woman's body for nine months in order to survive is her mother, father, sister, brother, son, daughter, or close friend, then the relationship would seem close enough to establish a special obligation in the woman to permit that use. If the needy person is a total stranger, then that obligation is missing. The fetus no doubt stands somewhere between these two extremes, but it is at least as close to the "special relationship" end of the spectrum as to the "total stranger" end.

The most important difference, however, between the violinist case and the normal pregnancy is that in the former the woman had absolutely nothing to do with creating the situation from which she wishes to escape. She bears no responsibility whatever for being in a state of "plugged-in-ness" with the violinist. As many commentators have pointed out, this makes Thomson's analogy fit at most one very special class of pregnancies, namely those imposed upon a woman entirely against her will, as in rape. In the "normal case" of pregnancy, the voluntary action of the woman herself (knowingly consenting to sexual intercourse) has something to do with her becoming pregnant. So once again, we find that a proabortion argument fails to establish an unrestricted moral right to abortion. Just as self-defense justifies abortion at most in the case where it is necessary to save the mother's life, the Thomson defense justifies abortion only when the woman shares no responsibility for her pregnancy, as, for example, when it has been caused by rape (force or fraud).

Voluntariness and Responsibility

If we continue the line of reasoning suggested by our criticism of the violinist example, we will soon reach a general principle, namely, that whether or not a woman has a duty to continue her pregnancy depends, at least in part, on how responsible she is for being pregnant in the first place, that is, on the extent to which her pregnancy is the consequence of her own voluntary actions. This formula, in turn, seems to be an application of a still more general moral principle, one that imposes duties on one party to rescue or support another, even a stranger and even when that requires great personal sacrifice or risk, to the degree that the first party, through his own voluntary actions or omissions, was responsible for the second party's dependence on him. A late-arriving bystander at the seaside has no duty to risk life or limb to save a drowning swimmer. If, however, the swimmer is in danger only because the bystander erroneously informed him that there was no danger, then the bystander has a duty to make some effort at rescue (though not a suicidal one), dangerous as it may be. If the swimmer is in the water only because the "bystander" has pushed him out of a boat, however, then the bystander has a duty to attempt rescue at any cost to personal safety,19 since the bystander's own voluntary action was the whole cause of the swimmer's plight

Since the voluntariness of an action or omission is a matter of degree, so is the responsibility that stems from it, as is the stringency of the duty that derives from that responsibility. The duty to continue a pregnancy, then, will be stronger (other things being equal) in the case where the pregnancy was entered into in a fully voluntary way than it will be in the case that fits the violinist model, where the pregnancy is totally involuntary. But in between these two extremes is a whole range of cases where moral judgments are more difficult to make. We can sketch the whole spectrum as follows:

  1. Pregnancy caused by rape (totally involuntary).
  2. Pregnancy caused by contraceptive failure, where the fault is entirely that of the manufacturer or pharmaceutical company.
  3. Pregnancy caused by contraceptive failure within the advertised 1 percent margin of error (no one's fault).
  4. Pregnancy caused by the negligence of the woman (or the man, or both). They are careless in the use of the contraceptive or else fail to use it at all, being unaware of a large risk that they ought to have been aware of.
  5. Pregnancy caused by the recklessness of the woman (or the man, or both). They think of the risk but get swept along by passion and consciously disregard it.
  6. Pregnancy caused by intercourse between partners who are genuinely indifferent at the time whether or not pregnancy results.
  7. Pregnancy caused by the deliberate decision of the parties to produce it (completely voluntary).

There would be a somewhat hollow ring to the claim in case 7 that one has no obligation to continue one's bodily support for a moral person whose dependence on that support one has deliberately caused. That would be like denying that one has a duty to save the drowning swimmer that one has just pushed out of the boat. The case for cessation of bodily support is hardly any stronger in 6 and 5 than in 7. Perhaps it is misleading to say of the negligence case (4) that the pregnancy is only partially involuntary, or involuntary "to a degree," since the parents did not intentionally produce or run the risk of producing a fetus. But there is no need to haggle over that terminological question. Whether wholly or partially involuntary, the actions of the parents in the circumstances were faulty and the pregnancy resulted from the fault (negIigence), so they are to a substantial degree responsible (to blame) for it. It was within their power to be more careful or knowledgeable, and yet they were careless or avoidably ignorant. So they cannot plead, in the manner of the lady plugged to the violinist, that they had no control over their condition whatever. In failing to exercise due care, they were doing something else and doing it "to a degree voluntarily." In these cases -- 4, 5, 6, and 7 -- the woman and her partner are therefore responsible for the pregnancy, and on the analogy with the case of the drowning swimmer who was pushed from the boat, they have a duty not to kill the fetus or permit it to die.

Cases 2 and 3 are more perplexing. In case 2, where the fault was entirely that of the manufacturer, the woman is no more responsible for being pregnant than in case 1 where she is the unwilling victim of a rape. In neither case did she choose to become pregnant. In neither case was she reckless or negligent in respect to the possibility of becoming pregnant. So if she has no duty to continue to provide bodily support for the dependent fetus in the rape case, then equally she has no duty in the other case. To be sure, there is always some risk of pregnancy whenever there is intercourse, no matter how careful the partners are. There may be only 1 chance in 10,000 that a contraceptive pill has an undetectable flaw, but there is no chance whatever of pregnancy without intercourse. The woman in case 2, then, would seem to have some responsibility, even if vanishingly small, for her pregnancy. She could have been even more careful by abstaining from sex altogether. But notice that much the same sort of thing could be said of the rape victim. By staying home in a locked building patrolled round the clock by armed guards, she could have reduced the chances of bodily assault from, say, 1 in 50,000 to effectively nil. By staying off the dangerous streets, she would have been much more careful than she was in respect to the risk of rape. But surely that does not entitle us to say that she was "partially responsible" for the rape that made her pregnant. When a person takes all the precautions that she can reasonably be expected to take against a certain outcome, then that outcome cannot fairly be described as her responsibility. So in case 2, where the negligence of the manufacturer of the contraceptive is the cause of the pregnancy, the woman cannot be held responsible for her condition, and that ground for ascribing to her a duty not to abort is not present.

Case 3 brings us very close to the borderline. The couple in this example do not choose to have a baby and, indeed, they take strong precautions against pregnancy. Still they know that there is a 1 percent danger and they deliberately chose to run that risk anyway. As a result, a woman becomes pregnant against her will. Does she then have a right to abandon to a certain death a newly formed moral person who is even less responsible for his dependence on her than she is? When one looks at the problem in this way from the perspective of the fetus to whom we have suppositively ascribed full moral rights, it becomes doubtful that the pregnant woman's very minimal responsibility for her plight can permit her to abandon a being who has no responsibility for it whatever. She ran a very small risk, but the fetus ran no risk at all. Nevertheless, this is a borderline case for the following reason. If we extend to this case the rule we applied to case 2, then we might be entitled to say that the woman is no more responsible than the fetus for the pregnancy. To reach that conclusion we have to judge the 1 percent chance of pregnancy to be a reasonable risk for a woman to run in the circumstances. That appraisal itself is a disguised moral judgment of pivotal importance, and yet it is very difficult to know how to go about establishing it. Nevertheless, if it is correct, then the woman is, for all practical purposes, relieved of her responsibility for the pregnancy just as she is in cases 1 and 2, and in that event the fetus's "right to life" does not entail a duty on her part to make extreme sacrifices.


Assuming that the fetus is a moral person, under what conditions, if any, is abortion justifiable homicide? If the woman's right to an abortion is derived from her rieht to own property in her body (which is not very plausible), then abortion is never justifiable homicide. Property rights simply can't support that much moral weight. If the right is derived from self-defense, then it justifies abortion at most when necessary to save the woman from death. That is because the fetus, while sometimes a threat to the interests of the woman, is an innocent and (in a sense) nonaggressive threat. The doctrine of proportionality, which permits a person to use a degree of force in self-defense that is likely to cause the assailant harm greater (within reasonable limits) than the harm the assailant would otherwise cause the victim, has application only to the case where the assailant is culpable. One can kill an "innocent threat" in order to save one's life but not to save one's pocketbook. The right of bodily autonomy (to decide what is to be done in and to one's own body) is a much solider base for the right to abortion than either the right to property or the right to self-defense, since it permits one to kill innocent persons by depriving them of one's "life-support system," even when they are threats to interests substantially less important than self-preservation. But this justification is probably available at most to victims of rape, or contraceptive failure caused by the negligence of other parties, to the risk of which the woman has not consented. That narrow restriction on the use of this defense stems from the requirement, internal to it, that the woman be in no way responsible for her pregnancy.

It does not follow automatically that because the victim of a homicide was "innocent," the killing cannot have been justified. But abortion can plausibly be construed as justifiable homicide only on the basis of inexact analogies, and then only (1) to save the mother from the most extreme harm or else (2) to save the mother from a lesser harm when the pregnancy was the result of the wrongful acts of others for which the woman had no responsibility. Another possibility that was only suggested here is (3) when it can be claimed for a defective or diseased fetus that it has a right not to be born. These narrow restrictions on the right of the woman to an abortion will not satisfy many, people in the proabortion camp. But if the assumption of the moral personhood of the fetus is false, as was argued in the first part of this essay, then the woman's right to bodily autonomy will normally prevail, and abortions at all but the later stages, at least, and for the most common reasons, at least, are morally permissible.


Since this article was written there have been numerous legislative attempts to undermine the constitutional right to an abortion declared by the United States Supreme Court in Roe v. Wade (1973). Perhaps the most prominent of these efforts was the bill, S. 158, introduced in the Senate in 1982 by Senator Jesse Helms, which resolves that "human life shall be deemed to exist from concention." The Congress, had it passed this bill, would have created a legal status for human fetuses and embryos that presumably would imply their possession of various constitutional rights, including "the right to life." Among the other intended legal consequences of this new status is that the killing of a fetus at any stage of its development would be a kind of criminal homicide.

Whatever difficulties might attend a legislative or constitutional rule endorsing some particular point in mid or late pregnancy as the starting point of moral personhood (and they, too, would be considerable), the consequences of the criterion endorsed by S. 158, or any other rule dating personhood from "the moment of conception," are intolerably paradoxical. Consider how this essay has treated the question: Is the single-cell zygote a moment after conception already a full-fledged human being? Depending on how we interpret the question, Senator Helms's affirmative answer is either self-evidently, but trivially and irrelevantly, true, or else absurd. The one-celled conceptus, the product of a human spermatozoan and a human ovum, is, of course, itself human. No scientist who counted its chromosomes could possibly classify it as feline, canine, or equine. The species to which it belongs is obviously the human species. Moreover, unless or until it aborts along the way, it is clearly a living thing as opposed to an inert piece of dead tissue. Therefore, the newly conceived "zygote" (the word for the fertilized egg before it becomes a hollow sphere of cells or "blastocyst" on the sixth day) is -- in the most trivial and obvious sense -- a form of "human life." There would appear to be nothing to quarrel about in this innocent claim -- except, of course, whether the zygote is already "human life" in the sense that is morally significant, namely that of a person with rights, the legal status Senator Helms wishes to establish.

To show that living human zygotes have rights, it is not enough to point out that they are living things associated with the human species. One must also show that they are people, in all morally relevant respects like you and me. Fertilized human eggs, of course, are potential people. That is to say, that they will develop into actual people if all goes well in the normal course of pregnancy. But that obvious truth is not sufficient to establish their present rights. To be a potential human person, as we have seen, is to be only a potential bearer of human rights; actual rights must await actual personhood. To infer actual possession of rights from future qualification is a mistake that seems peculiar to the abortion controversy; we are rarely tempted to make it in other contexts. A twelve-year-old American child, for example, is a potential voter in American elections in the sense that he will have an actual right to vote six years from now if all goes well in the natural course of his adolescence. But we are not tempted to admit him to the voting booths now simply on the grounds that he is a "potential voter" who will be qualified later.

The relevant question, then, is whether human zygotes, a few minutes after conception, are already actual people. Senator Helms's affirmative answer to this question conflicts violently with common sense. The only people who are likely to agree with it are those who are prepared to abandon common sense or those who confuse this question with the obvious but irrelevant questions about life and species membership that we have already dismissed. The one-celled speck of protoplasm, or small cluster of cells, of which we speak has none of the characteristics we normally have in mind when we speak of persons. No one believes that this tiny entity is already conscious of itself and the world, capable of sensation and emotion, able to understand and reason, remember and anticipate, make plans and act, be pleased or frustrated or hurt. Nor do we take seriously the logical consequences of "deeming" human embryos in their earliest stages to be people. In particular our attitudes toward spontaneous miscarriage ("natural abortion") show that we do not regard embryos as full-fledged moral persons. We may be disappointed when a pregnancy miscarries because our desire to produce a child has been at least temporarily frustrated, but we do not grieve for the embryo's sake. Funeral rites are not performed for tiny clusters of cells; baptism and extreme unction are not given upon the arrival of a tardy menstrual period over menses that may have contained an embryo; names and "conception dates" are not recorded; death certificates are not required. Why would we be so casually negligent if we believed with Senator Helms that a real person has died?

More importantly, if zygotes and embryos are actual people, why do we not make a monumental effort to "rescue" the millions who are bound to perish each year from natural causes? If there are a dozen trapped coal miners in a caved-in mine or a million persons threatened by starvation in a foreign famine, we are prepared to spend millions of dollars to save them because they are, after all, undeniably fellow human beings. Why then do we not budget millions more for research toward the discovery of a drug that would prevent spontaneous abortions? Embryologists have estimated that only 58 percent of fertilized ova survive until implantation (seven days after conception) and that the spontaneous abortion rate after that stage is from 10 to 15 percent. If we left that many miners and farmers to die each year without rescue efforts, we would be very callous indeed. But if we did save all the fetuses with some wonder drug, then, according to one embryologist, instead of a population in which approximately 2 percent surfer from relatively minor congenital defects, we would have a population in which 10 to 20 percent would be abnormal "and most of the defects would be gross and incapacitating." (Malcolm Potts, Biology and Ethics, The Academic Press, 1969). Does Senator Helms intend to prepare us to face this consequence, or is he content to let millions of salvageable "human lives" perish?

Other absurd consequences of the redefinition in the "human life statute" have been more widely publicized. Women who use contraceptives that make the uterine wall unreceptive to implantation, for example, would themselves be murderers since they deliberately cause the death of fertilized human ova. Government could take vigilant steps to protect unborn persons, and many of these would involve intrusions into women's private affairs -- requiring monthly pregnancy tests, for example, to determine whether any unborn persons exist in her womb, or requiring registration of all suspected pregnancies. If we really took seriously the view that fertilized eggs are people with the full panoply of human rights, these absurd practices would not seem implausible.

It seems to me that a case can be made for taking a human life statute that dates the origin of personhood at conception to be an "establishment" of religious doctrine. The argument runs as follows. For reasons given above, it is quite contrary to common sense to claim that a newly fertilized human ovum is already an actual person. Employing the term 'person' in the normal fashion, no one thinks of a fertilized egg in that way. The only arguments that have been advanced to the conclusion that fertilized eggs are people, common sense notwithstanding, are arguments with theological premises. These premises are part of large theological and philosophical systems that are very much worthy of respect indeed, but they can neither be established nor refuted without critical discussion of the whole systems of which they form a part. In fact, many conscientious persons reject them, often in favor of doctrines stemming from rival theological systems; so for the state to endorse the personhood of newly fertilized ova would be for the state to embrace one set of controversial theological tenets rather than others, in effect to enforce the teaching of some churches against those of other churches (and nonchurches), and to back up this enforcement with severe criminal penalties. The state plays this constitutionally prohibited role when it officially affirms a doctrine that is opposed to common sense and understanding and whose only proposed arguments proceed from theological premises. This case, it seems to me, is a good one even if there is reason, as there might be, for affirming the personhood of fetuses in the second or third trimester of pregnancy.


1. Mary Anne Warren, "On the Moral and Legal Status of Abortion," The Monist 57 (1973), pp. 43-61. Reprinted in J. Feinberg and H. Gross (eds.), Liberty: Selected Readings, pp. 133-143. The quotation is from the latter source, p. 138.

2. See Paul Ramsey, "The Morality of Abortion," in D. H. Labby (ed.), Life or Death: Ethics and Options (Seattle and London: University of Washington Press, 1968), pp. 60-93.

3. These problems are discussed in more detail in Joel Feinberg, "The Rights of Animals and Future Generations" (Appendix: The Paradoxes of Potentiality), in W. T. Blackstone (ed.), Philosophy and Environmental Crisis (Athens, Ga.: University of Georgia Press, 1974), pp. 67-68.

4. Stanley I. Benn, "Abortion, Infanticide, and Respect for Persons," in J. Feinberg (ed.), The Problem of Abortion (Belmont, Calif.: Wadsworth, 1973), p. 102.

5. Joel Feinberg, Social Philosophy (Englewood Cliffs, N.J.: Prentice-Hall, 1973), p. 66.

6. Benn, op. cit., p. 102.

7. Joel Feinberg, "Is There a Right to Be Born?" in James Rachels (ed.), Understanding Moral Philosophy (Encino, Calif.: Dickenson, 1976), pp. 353-354.

8. Thalidomide is the trade name of a potent tranquilizer once manufactured in Europe but never permitted in the United States. In the late 1950s, thousands of deformed babies were born to European women who had taken thalidomide during pregnancy.

9. From Mr. Justice Blackmun's opinion in Roe v. Wade 410 U.S. 113 (1973).

10. Jane English, "Abortion and the Concept of a Person," Canadian Journal of Philosophy 5 (1975), p. 242.

11. George Fletcher, "Proportionality and the Psychotic Aggressor: A Vignette in the Comparative Criminal Law Theory," Israel Law Review 8 (1973), p. 376.

12. Fletcher, loc. cit.

13. Judith Jarvis Thomson, "Self-Defense and Rights," The Lindley Lecture, 1976 (Lawrence, Kans.: University of Kansas Philosophy Department, 1977).

14. The term comes from Robert Nozick, Anarchy, State, and Utopia (New York: Basic Books, 1974), p. 35.

15. Even when self-defense is acceptable as a defense to homicide in the case of forced killings of nonaggressive innocents, that may be because it is understood in those cases to be an excuse or a mitigation rather than a justification. If a criminal terrorist from a fortified position throws a bomb at my feet and I can escape its explosion only by quickly throwing it in the direction of a baby buggy whose infant occupant is enjoying a nap, perhaps I can be excused for saving my life by taking the baby's, perhaps the duress under which I acted mitigates my guilt, perhaps the law ought not to be too severe with me. But it is not convincing to argue that I was entirely justified in what I did because I was acting in self-defense. But the problem is a difficult one, and the case may be borderline.

16. See Ludwig Lewisohn's remarkable novel, The Case of Mr. Crump (New York: Farrar, Straus, 1947).

17. Judith Jarvis Thomson, "A Defense of Abortion," Philosophy and Public Affairs 1 (1971), pp. 48-49.

18. Warren, op. cit., p. 135.

19. The examples are Sissela Bok's. See her article "Ethical Problems of Abortion," Hastings Center Studies 2 (1974), p. 35.


Further readings are mentioned below in connection with the section (§) of the essay to which they are most pertinent.

§3. A very helpful collection of articles on the concept of a person is The Identity of Persons, edited by Amelie O. Rorty (Berkeley, Los Angeles, London: University of California Press, 1976). In particular, the editor's introduction and postscript and Daniel Dennett's "Conditions of Personhood" are especially useful. A very penetrating and original analysis of commonsense personhood (not in the Rorty collection) is Harry Frankfurt's "Freedom of the Will and the Concept of a Person," Journal of Philosophy 68 (1971), reprinted in J. Feinberg (ed.), Reason and Responsibility, 4th ed. (Belmont, Calif.: Wadsworth, 1978). For articles that defend particular accounts of personhood in the context of the abortion problem, see especially Mary Anne Warren, "The Moral and Legal Status of Abortion," The Monist 57 (1973), and Joseph Fletcher, "Indicators of Humanhood: A Tentative Profile of Man," The Hastings Center Report 2 (1972). See also Lawrence C. Becker's "Human Being: The Boundaries of the Concept," Philosophy and Public Affairs 4 (1975).

§4. The most influential recent defenses of a species criterion are probably those of the Catholic legal scholar John T. Noonan, Jr., and the Protestant theologian Paul Ramsey. See Noonan's "Abortion and the Catholic Church: A Summary History," Natural Law Forum 12 (1967), and his rejoinder to critics, "How to Argue About Abortion" (New York: Ad Hoc Committee in Defense of Life, Inc., 1974). Ramsey's views are well stated in "The Morality of Abortion," in Life or Death: Ethics and Options, edited by D. H. Labby (Seattle: University of Washington Press, 1968).

A nicely nuanced defense of a kind of synthesis of the strict potentiality and species principles can be found in Philip E. Devine's book, The Ethics of Homicide (Ithaca, N.Y.: Cornell University Press, 1978). Although Devine's conclusions differ from those of this essay, the methodology is very similar to that employed here.

A thorough presentation of a modified potentiality criterion can be found in Daniel Callahan's learned and thoroughly moderate book, Abortion: Law, Choice, and Morality (London and New York: Macmillan, 1970). See especially chapters 10 and 11.

The most uncompromising defense of an actual-possession criterion is that of Michael Tooley, "Abortion and Infanticide," Philosophy and Public Affairs 2 (1972). Tooley defends not only abortion but, under certain circumstances, infanticide as well. His view on the latter question is criticized on utilitarian grounds by Stanley Benn and Jane English in the articles mentioned in the text of this essay.

§7. An eloquent account of the serious psychological harms that can be done to women by enforced pregnancies can be found in Lorenne M. G. Clark's "Reply to Professor Sumner," Canadian Journal of Philosophy 4 (1974). Baruch Brody argues forcefully that "self-defense" cannot justify abortion in his Abortion and the Sanctity of Human Life: A Philosophical View (Cambridge, Mass., and London: M.I.T. Press, 1975). An authoritative and very accessible discussion of the application of "self-defense" and other justifications to borderline cases of justifiable homicides of all kinds is Sanford Kadish's "Respect for Life and Regard for Rights in the Criminal Law," California Law Review 64 (1976), reprinted in S. F. Barker (ed.), Respect for Life in Medicine, Philosophy, and the Law (Baltimore and London: Johns Hopkins University Press, 1976).

Judith Thomson's use of the "plugged-in violinist" example is sharply criticized by John Finnis in his "The Rights and Wrongs of Abortion," in Philosophy and Public Affairs 2 (Winter 1973). Thomson's rejoinder to Finnis, "Rights and Deaths," is included in the same issue. Both articles can be found in The Rights and Wrongs of Abortion, edited by M. Cohen, T. Nagel, and T. Scanlon (Princeton, N.J.: Princeton University Press, 1973). Another useful anthology with more recent articles is The Problem of Abortion, 2nd ed., edited by Joel Feinberg (Belmont, Calif.: Wadsworth, 1983).