Cloning,killing, and identity78 Cloning,killing, andidentity entity or, as it maybe, series ofentities resulting fromnucleartransfer at anystage in the develop-mental process initiated

Journal ofMedical Ethics 1999;25:77-86

Cloning, killing, and identityJeff McMahan University of Illinois at Urbana, USA

AbstractOne potentially valuable use of cloning is to provide asource of tissues or organs for transplantation. Themost important objection to this use of cloning is thata human clone would be the sort of entity that itwould be seriously wrong to kill. I argue that entitiesof the sort that you and I essentially are do not beginto exist until around the seventh month offetalgestation. Therefore to kill a clone prior to that wouldnot be to kill someone like you or me but would beonly to prevent one of us from existing. And evenafter one of us begins to exist, the objections to killingit remain comparatively weak until its psychologicalcapacities reach a certain level of maturation. Theseclaims support the permissibility of killing a cloneduring the early stages of its development in order touse its organs for transplantation.(journal ofMedical Ethics 1999;25:77-86)Keywords: Cloning; killing; identity; ethics; fetus

The owners of an aging dog have recently donatedUS$2.3 million to researchers in order to make aclone of their pet. This and many other proposeduses of cloning are not only frivolous but are alsopremised on erroneous assumptions about iden-tity. The dog owners presumably believe that theclone will be their original dog, resurrected; forone does not pay $2.3 million in order to have adog that is merely very similar to another. But it isconfused to suppose that a clone would be identi-cal to-that is, one and the same entity as-theindividual from whom its genes were taken.There are, however, potential uses of cloning

that are quite important. When, for example, itbecomes possible to clone a human being, cloningcould be used to save people's lives by providing asource of tissues or organs for transplantation. Aperson needing a transplant could have him- orher- self cloned and then have the required tissueor organ extracted from the clone. The problem ofestablishing tissue compatibility would not arise.The prospect of using cloning for this purpose

has been opposed on moral grounds. Some of theobjections are general and apply to all instances ofcloning. These objections would apply even if theextraction of tissue from a clone would becompatible with the clone's survival and develop-

ment into a mature human being. My concern,however, is with an objection that applies only ifthe extraction of tissues or organs would be fatalto the clone. The objection is that the clone wouldbe the kind of entity that it would be seriouslywrong to kill. Certainly if a clone of a normalhuman being develops normally, there will besome point at which it is seriously wrong to kill it.If, for example, the clone had to grow to be twoyears old in order for the needed organ to becomesufficiently developed for transplantation, therewould be no question of killing the child in orderto take the organ.Many critics contend that the clone is from its

inception the kind of entity that it is seriouslywrong to kill. Opposition to cloning for thepurpose of organ procurement and opposition toabortion may thus have overlapping sources. Inboth cases, however, the opposition may rest on aconfusion about identity-not the crude mistakeof supposing a clone to be identical with the indi-vidual cloned but a subtler mistake in supposingthe clone in its earliest stages to be identical withthe person into whom it might eventually develop.The cloning procedure is in principle quite

simple, though there are formidable technicalproblems in implementing it. To produce a cloneof a particular human being, one would extractthe nucleus from one of that individual's cells andinsert it into a human egg cell from which theoriginal nucleus had been removed. The resultinghybrid cell would contain the individual's com-plete DNA code. But it would also be totipotent:it would have the potential to develop into amature human organism. To realise this potential,it would have to be subjected to electrical pulses inorder to initiate the process of cell division andthen be implanted in a natural or artificial uterus.The process of nuclear transfer parallels the

process of ordinary fertilisation. The difference isthat, rather than blending the incomplete geneticcodes oftwo gametes, cloning involves the transferof a complete genetic code into the germ plasm ofan egg cell. But the product of each process is, ineffect, the same: a single-celled zygote.

I will continue to refer to the product of a suc-cessful instance of nuclear transfer as "the clone".This is intended as a neutral designation for the

on July 19, 2020 by guest. Protected by copyright.

http://jme.bm

j.com/

J Med E

thics: first published as 10.1136/jme.25.2.77 on 1 A

pril 1999. Dow

nloaded from

http://jme.bmj.com/

78 Cloning, killing, and identity

entity or, as it may be, series of entities resultingfrom nuclear transfer at any stage in the develop-mental process initiated by this event. It isintended to avoid begging important questionsabout the nature or status ofwhatever is present atany point in the process: a cell or cluster of cells, ahuman organism, a person, or whatever.The most important question, for our purposes,

is at what point in the process initiated by nucleartransfer the clone would achieve a moral statussufficient to make it morally objectionable to killit. As I noted, some believe that the clone wouldhave this status from the moment the nucleartransfer was completed. Others believe that itwould not achieve this status until later. Whethercloning for the purpose of procuring transplant-able tissue is permissible may depend on which ofthese views is correct.What might determine whether the status of the

clone is such as to make it seriously objectionableto kill it? One relevant consideration is whether, atany given point between nuclear transfer andbirth, the clone would be identical with-that is,one and the same individual as-the person intowhom it might later develop. This depends onwhether, at that point, the clone would be thesame kind of entity that you and I essentiallyare-that is, whether it would be, as I will say, oneof us.

Imagine that you were produced by cloning.One morally significant point in the process initi-ated by nuclear transfer in your case would havebeen the point at which you began to exist.Suppose, for the sake of argument, that you beganto exist at some point after nuclear transfer butbefore birth. Between the moment of nucleartransfer and the time at which you began to exist,what I am calling "the clone" would not have beenyou. Suppose, for example, that you began to existat time t. If someone had killed the cloned organ-ism from which you developed prior to t, thiswould not have killed you but would merely haveprevented you from existing. If, by contrast, theclone had been killed after t, this would haveinvolved killing you. In general we recognise asharp moral distinction between acts that merelyprevent a person from coming into existence andacts that kill a person-for example, between con-traception and infanticide. For an act that merelyprevents someone from existing need not be worsefor any actual person. But an act that killssomeone is worse for that individual provided thatthe life he would otherwise have had would havebeen worth living. For the life that the individualwill have if he is killed will be substantially lessgood than the life he would have had if he had notbeen killed.

So this could be one reason why the status ofthe clone might be such that it would not be seri-ously objectionable to kill it: the clone, in its earlystages, might not be anyone might not, that is, bean entity of our kind, one of us. To kill it might notbe to kill someone like you or me; it might insteadbe only to prevent one of us from existing.The answer to the question whether the clone

would, at a certain point in its development, beone of us would not, however, necessarily settlethe question whether it would be permissible tokill it at that point. There are several possibilities.The two most obvious are the ones we have justnoted.(1) At a certain point in its development, the

clone would be one of us-that is, would beidentical with the person into whom it mightdevelop. Therefore it would be seriouslyobjectionable to kill it at that point.

(2) At a certain point in its development, theclone would not be one of us; therefore itwould not be seriously objectionable to kill itat that point.

Two other options are:(3) At a certain point in its development, the

clone would be one of us. But, because itwould lack some morally significant propertyor properties, it would not be seriously objec-tionable to kill it at that point.

(4) At a certain point in its development, theclone would not be one of us. But its moralstatus would nevertheless be such that itwould be seriously objectionable to kill it.

According to the third of these options, theremight be times in my life when it would not beseriously wrong for others to kill me-evenwithout my consent. In the public debate aboutabortion, for example, it is often claimed that it ispermissible to kill the fetus because it is not a per-son. This is not necessarily meant to imply thatthe fetus is not the same individual as the personinto whom it might develop; it might instead meanthat, when each of us was a fetus, he or she lackedcertain properties (for example, self-consciousness) that are necessary for being a per-son, and that, when we lacked those properties,the moral objections to killing us were muchweaker, or failed to apply altogether. Differentexplanations might be given for this. It might sim-ply be that, as non-persons, our moral status waslower. Alternatively, we might have lacked somenecessary condition for the possession of rights.Or it might be that before we became sentient wewere incapable of having interests, so that ourbeing killed could not have been against our inter-ests before that point. I will later defend a differentexplanation.


http://jme.bm

j.com/

J Med E


pril 1999. Dow

nloaded from

http://jme.bmj.com/

McMahan 79

According to the fourth option, even if the clonein its early stages would not be identical with theperson into whom it might develop, it mightnevertheless be an entity of a sort that it would beseriously objectionable to kill. It might, for exam-ple, have an intrinsic value that is independent ofits being one of us. Or it might merit protection onthe ground that it has the potential to become (inthe sense of giving rise to the existence of) one ofus.

Again, my principal concern is whether theclone, during the early phase of its existence,would be identical with the person into whom itmight develop-despite the fact that the answerwill not decisively establish the permissibility orimpermissibility of killing the clone at that stage.At a minimum, the answer will determine whetherto kill a clone during that early period would be tokill someone like you or me or whether it would beto prevent someone from ever existing.The question: "When in the process initiated by

nuclear transfer does someone like you or mebegin to exist?" does not appear to raise any issuesdifferent from those raised by the question:"When in the process initiated by the fertilisationof a human egg does someone like you or mebegin to exist?" The two processes run paralleleven though the event that initiates each process-nucleation in the germ plasm of the egg cell-isdifferent in the two cases. Hence we need addressonly one question: "When do we-that is, beingsof our kind-begin to exist?" If the answer formost of us is "at conception", we will have to givea different answer in the case of people producedby cloning. In these cases, our answer would pre-sumably be "when nuclear transfer is completed".Otherwise the answer in the case of peopleproduced by cloning should be the same as theanswer in the case of people produced throughfertilisation.

In order to determine when we begin to exist, itis helpful to understand what sort of being weessentially are. To say that we are essentially a cer-tain kind of being is to say that we could not ceaseto be beings of that kind without also ceasing toexist. The two most common views are that we aresouls, or non-material substances, and that we arehuman organisms. If I am essentially an incorpo-real soul, I began to exist when this soul was cre-ated. If I am essentially a human organism, Ibegan to exist when this organism-myorganism-began to exist. These two views arealso the only views that are compatible with thebelief that we begin to exist at conception or, inthe case of a clone, at the moment of nucleartransfer. For there is really no other kind of thingthat might be supposed to be present immediately

after conception (or nuclear transfer). If we areneither souls nor human organisms, we must havebegun to exist at some point later than concep-tion.Suppose we are essentially souls. If the soul

appears at conception, it is reasonable to supposethat it is also present upon the completion ofnuclear transfer. And if it is present at either event,it seems that the soul can exist without the capac-ity for consciousness, for there appears to benothing present in the womb in the early stages ofpregnancy that has the capacity for consciousness.The soul, it seems, must await the development ofthe brain in order for consciousness to becomepossible. If, however, consciousness depends onthe functioning of the brain, there is no reason tosuppose that the soul is capable of consciousnessafter the brain has ceased to function at death. Ifthe soul continues to exist after death, we shouldconclude that we will be unconscious throughoutthe afterlife, just as we were between conceptionand the point at which our brains began to gener-ate consciousness. Since this is contrary to whatmost believers in the soul suppose, we should per-haps revise our assumptions.An alternative conception of the soul is that the

soul is, as Descartes said, a thinking thing: itessentially has the capacity for consciousness-that is, it cannot exist without the capacity forconsciousness. It is the soul, not the brain, thatthinks and feels and perceives. If we conceive ofthe soul in this way, there are three possibilities.One is that we begin to exist at conception and areactually conscious from that point on. We are,from conception to some point late in pregnancy,like locked-in patients who later suffer retroactiveamnesia, so that we cannot now remember ourconscious life during that period. This is ex-tremely implausible. The second possibility is thatwe begin to exist at conception and have thecapacity for consciousness at that point, but can-not exercise our capacity, just as adults who aretemporarily comatose have the capacity forconsciousness but are prevented from exercisingit. This too seems entirely ad hoc. Moreover, thisunderstanding of our early life may also haveunappealing implications for the afterlife. If theexplanation of why the soul is unable to exerciseits capacity for consciousness in the periodfollowing conception is related to the absence ofbrain function, then we should expect that thesoul will also be unable to exercise that capacity inthe afterlife, when brain functions have ceased.The only reason to accept that there is anything

present at conception that has the capacity forconsciousness is that this is implied by our twoassumptions: that the soul essentially has the


http://jme.bm

j.com/

J Med E


pril 1999. Dow

nloaded from

http://jme.bmj.com/


capacity for consciousness and that it is present at

conception. But the fact that these assumptionshave this implication seems more obviously to bea reason to reject one of the assumptions. If weaccept that the soul essentially has the capacity forconsciousness, it seems more reasonable to

conclude that the soul does not begin to exist untilthere is independent evidence of conscious life inthe womb. According to this third view, even

though we are souls, we do not begin to exist untilat least five months after conception, when there isbehavioural evidence of consciousness. Thus tokill a clone before that point would not be to killone of us; it would instead be to prevent one of usfrom existing. (It would, of course, be to killsomething-namely an organism not yet animatedby a soul; but what that organism's moral statuswould be I leave open at this point.)A view similar to this was held by Aquinas and

prevailed throughout much of the Western worlduntil about the middle of the nineteenth century.According to this view, the fetus does not come tolife, or "quicken," until "ensoulment" occurs-

that is, when it is endowed with a soul-and it wascommonly held that this does occur until wellafter conception. The pregnant woman's ability tofeel the fetus moving was regarded as sufficientevidence that it had come to life. We still refer tothe moment when fetal movement is first detectedas "quickening," though we no longer understandthis to be the moment when the fetus becomesalive.

It is, however, implausible to suppose that we

are souls, conceived as non-material substancesthat essentially have the capacity for conscious-ness. Most of the objections to the idea that we are

souls derive from our increasing understanding ofthe dependence of consciousness and cognitionon the operations of the brain. Many are familiarand I will not rehearse them here, though I havedefended some of the stronger objectionselsewhere.' Here I can only assert that, if the claimthat it is wrong to kill a clone at any point afternuclear transfer depends on the idea that the cloneis a soul, that claim is in a precarious position.Turn now to the other common view ofwhat we

essentially are: that we are biological organisms ofthe human species. If we are human organisms,did we begin to exist at fertilisation? If so, a personproduced through cloning would begin to exist atthe completion of nuclear transfer.Some have argued that the possibility that the

human embryo may divide to form monozygotictwins shows that there cannot be a human organ-ism present until fourteen days after fertilisation,when the possibility oftwinning has ceased. For, ifwe suppose that the embryo is an organism and

twinning does occur, there is no explanation ofwhat happens to the organism. It makes no senseto suppose that the original organism survives asonly one of the twins, for the relation that it bearsto each twin is exactly the same. Nor is it coherentto suppose that the original organism is identicalwith both twins, for that would imply that the twinsare identical to each other-that is, that they arenot distinct but are one thing. Finally, it would beodd to claim that the original organism has died,for it leaves "no earthly remains".2 So, it isclaimed, it is best to accept that, prior to twinning,there was no organism there at all but merely a cellor collection of cells.There is a reply to this argument. If an organism

begins to exist at fertilisation, we can say that, iftwinning occurs, that organism simply ceases toexist. Ceasing to exist through division is not thesame kind of event as death and does not leavedead remains behind. Thus, when an amoebadivides, it ceases to exist though nothing actuallydies.

This reply commits us to the view that, whenmonozygotic twinning occurs, a human organismceases to exist. One may be reluctant to accept thisimplication if one believes that it is always a seri-ous misfortune when an innocent human beingceases to exist prematurely; for it does not seemthat twinning involves anything regrettable. Forreasons that I will later indicate, I reject the furtherassumptions necessary for the conclusion thattwinning is regrettable; hence I would findnothing embarrassing in accepting that a humanorganism ceases to exist whenever monozygotictwinning occurs.

Reflection on the phenomenon of twinningseems to show that, even if most of us are organ-isms that come into existence at fertilisation,monozygotic twins are different. The earliest theycould begin to exist is when the original embryodivides to form two new embryos.

Granting that monozygotic twins may be differ-ent, is it plausible to suppose that the rest of usbegin to exist at fertilisation? There are two inter-pretations of the processes that occur during thefortnight following conception, before cell differ-entiation begins, the "primitive streak" is formed,and twinning ceases to be possible. According toone understanding, the successive cell divisionsare all events in the history of a single individual.When the initial single-celled zygote divides, itbecomes, or continues to exist as, a two-celledentity. The two-celled entity is simply the succes-sor state of the initial one-celled entity. Similarly,further divisions that continue to expand thenumber of cells are simply further phases in thegrowth and development of that individual. When


http://jme.bm

j.com/

J Med E


pril 1999. Dow

nloaded from

http://jme.bmj.com/

McMahan 81

each cell divides, it ceases to exist, but theindividual that is constituted by the various cellscontinues to exist through each successive expan-sion of its constituent matter.According to the alternative interpretation,

when the initial single-celled zygote divides, thereis nothing that continues to exist. Just as, in twin-ning, when the zygote divides it ceases to exist, soin the initial division the single cell ceases to existand is supplanted by its two qualitatively identicaldaughter cells. Similarly, when each of the daugh-ter cells divides, it ceases to exist and is replacedby its own two qualitatively identical daughtercells. Again, there is nothing-no individual-thatpersists through these transformations. Onlywhen the cells begin to be differentiated, to takeon specialised functions, and to be organisedtogether in an integrated way do they togetherconstitute a further individual: an organism.

If the first of these interpretations is correct, it isplausible to suppose that the zygote formed atconception is a human organism. If, furthermore,we are human organisms, then to kill any humanembryo-even one formed via nucleartransfer-is to kill one of us. If, by contrast, thesecond interpretation is correct, a human organ-ism does not begin to exist until at least fourteendays after conception. If this is right, to kill ahuman embryo during the first fourteen days afterfertilisation is not to kill one of us but to preventone of us from existing.Which interpretation is correct? The funda-

mental difference is that the first holds that theclustered, proliferating cells together constitute afurther individual that is distinct though not sepa-rable from them, while the second interpretationdenies this. The case for the first interpretationdepends on the observation that the cells are bun-dled together within a single membrane (the"zona pellucida") and eventually begin to take onspecialised tasks in the constitution of the organ-ism and its prenatal environment. Unlike a seriesof amoebas, where division produces daughtercells that may wander off to lead quite independ-ent lives, the embryonic cells and their progenyform a discrete unit. That unit, it is claimed, is theorganism.The case for the second interpretation is that,

during the first two weeks after conception, thecells are only loosely grouped within the zona pel-lucida. They are independent and uncoordinatedand, at least until the eight-cell stage, each is"totipotent"-that is, capable, if separated fromthe others, of developing into, or giving rise to, acomplete adult organism. It is this lack of integra-tion among the cells that suggests that they do nottogether constitute an organism.

Consider, as an analogy, an island on whichthere are people. Suppose these people are entirelyunrelated: each came to the island independently,lives in solitude with no communication or coop-eration with the others, and is even unaware of theexistence of most of the others. It seems clear thatthese individuals do not together constitute anindividual of any substantial sort. If, by contrast,various relations obtain among them-if, forexample, they are related genealogically, speak thesame language, follow the same customs, cooper-ate together in complex ways, and so on-then itis plausible to suppose that they together consti-tute a distinct individual: a nation, for example.

If the zygotic cells are relevantly like the inhab-itants of the island in the scenario in which theyare unrelated, it seems that the cells do not consti-tute an individual and, a fortiori, are not anorganism. If, instead, they are relevantly like thecomplexly related inhabitants, it is more plausibleto regard them as together constituting an organ-ism.

It seems to me more plausible to regard the cellsduring the first fourteen days as insufficientlyclosely related to constitute an organism. Yet Iwould not claim to be right about this; for it seemsthat there is no deep truth here to be discovered.3Consider again the analogy with nations. A nationis nothing but a set of people related to oneanother in certain ways. We accept that there canbe borderline cases in which it is indeterminatewhether a group of people are sufficiently relatedin the relevant ways to constitute a nation. Thecase is the same with organisms. An organism isnothing more than a collection of cells related invarious complex ways. Thus we should accept thatthere can be borderline cases in which it isindeterminate whether a collection of cells consti-tutes an organism.

If there were no instances of indeterminacy,there would have to be a precise point in the proc-ess of cell division initiated by fertilisation atwhich an incremental increase in the unifyingrelations among the cells would make thedifference between there being an organism andthere not being one. That is surely implausible.Instead, it seems that, if an organism does notbegin to exist at fertilisation, there is no precisepoint at which it does begin to exist. There is abrief period following fertilisation in which noorganism exists, followed by a brief period duringwhich it is indeterminate whether the proliferatingcells constitute an organism. Eventually the cellsare sufficiently unified that it is clear that theyform an organism.The problem is that our concept of an organism

is not precise enough to allow us to identify an


http://jme.bm

j.com/

J Med E


pril 1999. Dow

nloaded from

http://jme.bmj.com/


exact point at which cellular organisation hasbecome sufficiently integrated that an organismhas begun to exist where previously there was onlya collection of cells. At a certain period during theprocess initiated by fertilisation, we could knoweverything that is happening at the cellular levelwithout this being sufficient to determine whetheror not the cells together constitute an organism.

If this is right that the issue is merely aconceptual one-it is hard to see how anything ofdeep and intrinsic moral significance coulddepend on it. Yet, if we are organisms and it mat-ters morally whether to kill an embryo is to kill oneof us, then it may matter greatly whether theembryo is an organism. I will suggest a response tothis problem later in the paper.

In summary, although it is not unreasonable tobelieve that an organism begins to exist atfertilisation, it is more plausible to suppose that ahuman organism does not begin to exist untilabout fourteen days after fertilisation. If an organ-ism does not exist until a fortnight afterconception, and ifwe are organisms, then to kill anembryo prior to that point would not be to kill oneof us but would merely be to prevent one of usfrom existing. If this is right, it removes a majorobjection to killing a cloned embryo during thefortnight following nuclear transfer. And there arevaluable therapeutic uses to which embryonic tis-sues could be put. But these uses are limited. Ifweare organisms that begin to exist around fourteendays after fertilisation, and if it is seriously objec-tionable to kill one of us for the purpose ofextracting transplantable tissues, then the killingof cloned embryos will have to be confined to thefortnight following nuclear transfer and the use ofcloning as a source of transplantable tissue will beseverely restricted. In order for it to be acceptableto kill a clone at a later stage of development, oneof two claims must be true: either we are notessentially human organisms or the moral objec-tions to killing people like you and me do notapply, or apply only very weakly, in the very earlystages of our existence. I will argue that both theseclaims are true.There are various arguments against the view

that we are essentially human organisms. I willpresent, in abbreviated form, two that seem mostpersuasive. The first appeals to an example drawnfrom science fiction. Suppose that one's entirebrain were extracted from one's body andtransplanted into the cranium of one's identicaltwin, whose own brain had just been removed.Most people believe that one would continue toexist in what was formerly the body of one's twin.One would then occupy, and animate, a differentorganism. There are various things that might

happen to one's original organism: it might be leftto die; it might receive a brainstem transplant,thereby remaining alive though incapable ofsupporting consciousness; its functions might besustained indefinitely with the aid of artificial life-support systems; or it might have the whole ofsomeone else's brain transplanted into it, therebynot only remaining alive but also becoming host toan entirely different person. Whichever course ofevents we imagine happening, one's originalorganism would continue to exist (as a corpse, ina persistent vegetative state, or whatever) on onebed in the operating theatre while one would one-self regain consciousness in a different organismon another bed. There would then be two separateand distinct entities: oneself and one's originalorganism. But, if one would not then be identicalwith that organism, it follows that one never wasidentical with it. Generalising, we can concludethat none of us is identical with his or her physicalorganism.Some will object to this argument on the

ground that our intuitions about cases drawn fromscience fiction are unreliable. The second argu-ment cannot be rejected on that ground. Thisargument appeals to the phenomenon known as"dicephalus," in which a human embryo dividesincompletely, resulting in twins conjoined belowthe neck. In dicephalic twinning, as in other formsof twinning, it is clear that there are two people. Inone contemporary instance, Abigail and BrittanyHensel present a spectacle of two heads sproutingfrom a single torso; yet no one doubts that they areseparate and distinct little girls.4 Each has her ownprivate mental life and her own character, eachfeels sensations only on her own side of the body,and each has exclusive control over the limbs onher side. But, although Abigail and Brittany aretwo different persons, there seems to be only oneorganism between them. If so, then neither girl isidentical with that organism. For they cannot bothbe identical with the organism, as that wouldimply that they were identical with each other.Nor is it plausible to suppose that one of them isidentical with the organism while the other issome sort of parasite hosted by the organism.(Even if this understanding were plausible, itwould implicitly concede that there is at least oneperson who is not an organism.) So it seems thatwe should accept that neither is identical with theorganism they share. But, if dicephalic twins arenot human organisms, this strongly suggests thatnone of us is an organism. For, despite theiranomalous physical condition, there is no reasonto suppose that dicephalic twins are fundamen-tally different types of being from the rest of us.


http://jme.bm

j.com/

J Med E


pril 1999. Dow

nloaded from

http://jme.bmj.com/

McMahan 83

Those who hold that we are essentially humanorganisms seem to have only two options. One isto claim that, because dicephalic twins constitutea single human organism, that organism can be atmost one person-a person with a divided mind,or perhaps a person with two minds. Thisresponse denies that the Hensel twins are differentpersons. It treats them as analogous to the severedhemispheres in a commissurotomy patient: dis-tinct centres of consciousness but together consti-tutive of a single person. This seems obviouslyunacceptable.The other option is to claim that dicephalic

twins are actually two distinct though overlappingorganisms. This understanding is entirely compel-ling in the case of conjoined twins who are onlysuperficially melded (and therefore potentiallyseparable), as well as in cases in which there is onlya limited sharing of organs or parts but extensiveduplication of others. It is substantially less plau-sible, however, when, there is only very limitedduplication of organs. Although the Hensel twinshave two hearts and two stomachs, they otherwiseshare a single set of organs wrapped in a singleskin, sustained by a single coordinated system ofmetabolism, served by a single integrated circula-tory system, protected by a single immune system,and so on. These systems and the processes theysustain together constitute a single biological life,despite the fact that various aspects of this life arejointly governed by two brains. It is a singlebiological life that supports the existence of twodistinct persons. (Recorded cases of dicephalusshow varying degrees of duplication and it seemspossible that there could be an even purer casethan that ofthe Hensel twins in which there wouldbe virtually no duplication of organs below theneck.)These arguments strongly challenge the idea

that one is the same thing as one's organism. If,however, we are neither souls nor humanorganisms, what are we? The most plausible view,I believe, is that we are essentially embodiedminds. According to this view, I began to existwhen the brain in this body-my body-firstacquired the capacity to support consciousness.My continued existence has depended on thephysical and functional continuity of enough ofthis brain to be capable of continuing to supportconsciousness. I will cease to exist when thoseareas of my brain in which consciousness isrealised irreversibly lose the capacity to generateconsciousness.There is no space to argue for this view here,

though I have sought to defend it elsewhere.' If itis correct, we do not begin to exist until approxi-mately 28 to 30 weeks after fertilisation-

assuming that current estimates of when in thecourse of fetal development consciousness be-comes possible are roughly accurate. This, ofcourse, is well after our organisms begin to exist.But, if one is not identical with one's organism, itis perfectly possible that one began to exist after itdid, and equally possible that one may cease toexist before it does, or even before it dies.A case can be made for the permissibility of

killing a developing human organism prior to thepoint at which one of us begins to exist in associ-ation with it. For this would not be to kill one of usbut only to prevent one of us from existing. As Iacknowledged earlier, of course, there might bemoral objections to killing a developing humanorganism even if this would not be to kill someonelike you or me. Again, there is no space to considerthis possibility in detail. Here I can only note thatI do not believe that there can be seriousobjections. There is no more reason to supposethat an unoccupied organism has interests orrights than there is to suppose that a plant does. Itdoes not even have any significant potential. (Itmay have the potential to cause a person to exist,just as a sperm and an egg do; but, again like thesperm and egg, it does not have the potential tobecome a person, in the sense of "becoming" thatpresupposes identity. For, if my earlier argumentsare right, persons are never identical with mereorganisms.) It may, admittedly, have intrinsicvalue by virtue of the marvellously intricateorganisation of its myriad parts; but it should haveno more value of this sort than a comparablycomplex non-human organism.

If it would be permissible to kill a humanorganism prior to the point at which that organismwould support the existence of one of us, killing aclone in order to obtain transplantable tissuewould be permissible up to about the seventhmonth of pregnancy. This would allow for a mod-erate degree of differentiation and maturation oftissues and organs, making cloning a potentiallymuch richer source of transplantable tissue than itwould be if a clone could permissibly be killedonly during the fortnight following nuclear trans-fer. A further possibility, of course, would be tointervene to prevent the clone from ever develop-ing the capacity for consciousness. One might, forexample, deliberately create an anencephalicclone by suppressing the action of the genesresponsible for the development of the cerebralhemispheres. In so doing, one would prevent thecloned organism from ever giving rise to the exist-ence of someone like you or me. Because it wouldbe a permanently unoccupied organism, it mightpermissibly be allowed to develop even furtherbefore its organs were extracted for transplanta-


http://jme.bm

j.com/

J Med E


pril 1999. Dow

nloaded from

http://jme.bmj.com/


tion. There are various objections to this proposal,among which is the claim that it is wrong to createor tamper with the nature of a human being with-out appropriate regard for the good of that humanbeing. It is, however, questionable whether thissort of consideration applies to the creation ofmere organisms that is, organisms that neversupport the existence of a mind (or self or person).Again, there is no space to pursue this here.

I have suggested that killing a cloned organismmight be permissible up to the point at which oneof us begins to exist in association with it, on theground that this would be merely to prevent one ofus from existing. But there is a general objectionto this kind of claim. Suppose we can identify thepoint, subsequent to conception, at which we-orentities of our kind begin to exist. Call this pointt. The suggestion is that, prior to t, there is no onefor whom it is worse if a cloned organism is killed.But, after t, there is an individual present that, if itlives, will become (that is, will be identical with) aperson like you or me. To kill that individualwould be worse for it. For its life as a whole if it iskilled will be a worse life than the longer life itwould have if it were not killed. Indeed, if the bad-ness of death varies with the magnitude of the losssuffered by the victim, death immediately afterone has begun to exist must be the worst deathone can suffer; for one thereby loses the whole ofa human life rather than only the later stages of alife. This suggests, however, that while to kill aclone immediately prior to t would be innocuous,to kill a clone just after t would be to inflict one ofthe gravest harms that it is possible for us to suf-fer. Yet, intuitively, there does not seem to be anypoint at which killing a fetus (cloned or otherwise)ceases to be innocuous and suddenly becomesmurderous instead. Fetal development is asmoothly continuous process; it seems impossibleto identify any event in this process that is signifi-cant enough to make the difference between per-missible killing and murder. Intuitively, the degreeto which killing is morally objectionable seems toincrease slowly and gradually, in a way that corre-sponds to the process of fetal maturation.How can we explain and justify this common

intuition? I believe the best explanation lies inunderstanding the significance ofthe way in whichthe fetus would be related to itself in the future. Tosee this, we must begin by noting that each of ushas a special sort of concern for his or her ownfuture. We anticipate our own experiences in away that is different from our attitudes to otherpeople's future experiences. As one writer ob-serves, "we all know the difference betweenfearing our own pain and fearing the pain ofsomeone else. The difference here consists not in

degree-I may care more about the pain of mybeloved than about my own-but in kind".5 Let uscall this special kind of concern about the future"egoistic concern".

It has always been assumed that the basis foregoistic concern is identity that the mere factthat someone will be me gives me a reason to carein an egoistic way about what will happen to thatperson. This assumption has, however, been chal-lenged by Derek Parfit.6 His argument appeals toa hypothetical case-the case of "Division" inwhich a person divides into two people. Suppose,for example, that the cerebral hemispheres of aperson, A, are divided and separately transplantedinto the bodies of his identical triplet brothers,whose cerebrums have been destroyed andremoved but whose brainstems and bodies areintact and functional. Following the transplantoperations, two people, B and C, will come toconsciousness. Each will have most of A's memo-ries, beliefs, values, and character traits, and eachwill believe himself to be A. But because there aretwo such people, they cannot both be A, for thatwould imply that they are identical with eachother (that is, that they are one and the same per-son). Parfit contends that the logic of identityforces us to conclude that neither is A-that A hasceased to exist. Notice, however, that if only one ofthese people were to exist, (if, for example, one ofthe transplant operations had failed), most of usbelieve that that person would be A. For the rela-tion between A and B, for example, contains allthat is necessary for identity. (That relation is notsufficient for identity because it can hold betweenA and more than one person.) Notice, too, that ifonly B were going to exist, A would have reason tocare in an egoistic way about B. The basis for ego-istic concern is present in the relation between Aand B. And this fact (unlike the claim about iden-tity) is unaffected by the additional presence of C.Indeed, if the basis for egoistic concern is presentin the relation between A and B, it must also bepresent in the relation between A and C, for thetwo relations are substantially the same. Thus Ahas reason, prior to the operations, to beegoistically concerned about both B and C; but hewill be neither. Hence identity cannot be the basisfor egoistic concern.What is the rational ground of egoistic concern

about the future, if it is not identity? In practice,each of us is egoistically concerned only abouthim- or her- self. Unless this tendency is radicallymisguided, the relations that ground egoistic con-cern must coincide with identity in most cases.The best explanation for this is that the basis foregoistic concern at least includes those relationsthat are constitutive of personal identity in the


http://jme.bm

j.com/

J Med E


pril 1999. Dow

nloaded from

http://jme.bmj.com/

McMahan 85

normal case. According to the view I cited earlier,the relations that are constitutive of our identityover time are physical and functional continuity ofcertain parts of the brain. I believe that thepresence of these relations is a sufficient conditionfor the rationality of at least a minimal degree ofegoistic concern about the future. But I followParfit in believing that another component of thebasis of egoistic concern is psychological connected-ness. A person at an earlier time and a person at alater time are psychologically connected if thereare direct connections of memory, belief, desire,etc between them-for example, if B rememberssomething that A experienced. Psychological con-nectedness is a matter of degree: generally thereare more direct psychological connections overshorter periods than over longer periods. And thedegree of connectedness also varies with thedegree of development of an individual's psycho-logical capacities: thus there are more psychologi-cal connections between a person now andhimself a month ago than there are between a dognow and itself a month ago. I follow Parfit inbelieving that, if the relations that ground egoisticconcern are a matter of degree, then egoistic con-cern may also rationally be a matter of degree.Thus, if psychological connectedness is at leastpart of the basis of egoistic concern, an individualmay rationally have a weaker degree of egoisticconcern about himself in the future when thedegree to which he would be psychologically con-nected to himself in the future would bediminished.

If identity is not the basis for egoistic concern, itis possible that egoistic concern can be rational inthe absence of identity. In Division, for example, Amay rationally be concerned in an egoistic wayabout both B and C. It is also possible, though Ihave doubts about this, that there could beidentity but no basis for egoistic concern. A lessradical possibility is that just suggested: that therecould be identity but only a weak basis for egoisticconcern. This, I will suggest, is true in the case ofthe relation between a fetus and the person itmight later become.One problem, of course, is that a fetus lacks the

capacity to be concerned about itself in the future.This requires the introduction of a new concept.Let us say that an individual has an egoistic interestin some future possibility if it is rational for thatindividual to be egoistically concerned about thatpossibility or if it would be rational for thatindividual to care in an egoistic way about thatpossibility if the individual could conceive of it. Itis important to note that an egoistic interest is dif-ferent from an interest as ordinarily understood,for the latter presupposes identity whereas the

former does not. One's interests correspond towhat is better or worse for oneself, taking accountof one's life as a whole. But one's egoistic interestsmay in principle range outside of one's own life (asin the case of Division); or they may be weakerthan one's interests, if the relations that groundegoistic concern are weak.The claim I wish to make can now be

articulated by saying that, even though deathwould be much worse for a fetus than continuingto live (and therefore might be said to be greatlyagainst its interests), its egoistic interest incontinuing to live is comparatively weak. For thestrength of an individual's egoistic interest in a setof future goods is a function both of the amount ofgood he can anticipate in an egoistic way and ofthe strength of the relations that make his egoisticconcern for those goods rational. In the case of thefetus, the amount ofgood that its future life wouldcontain if it were not to die is very great. But atleast some of the relations that bind the fetus toitself in the future in the way that makes egoisticconcern rational are present only to a very weakdegree. For there would be virtually no psycho-logical connections between the fetus and the per-son it might later become. It is the fact that thefetus is thus psychologically cut off from itself inthe future-so that its future is, in a figurativesense, relevantly like someone else's future ratherthan its own-that it matters less, from the fetus'spresent point of view, whether it lives to have thatfuture or not.

If this is right, it provides an explanation ofwhythere is no point at which killing a fetus immedi-ately shifts from being innocuous to being gravelywrong. In the early stages ofpregnancy, before oneof us begins to exist in association with the devel-oping organism, there is no one who would beaffected for the worse if the organism were killed.Thus the killing of a clone in the early stages of itsexistence can be regarded as largely innocuous.After one of us begins to exist, however, there issomeone for whom death would be worse, so thatto kill that individual would be to harm it. Yet,because of the primitive nature of the fetus's men-tal life, it is psychologically cut off from its ownfuture and thus may have only a weak egoisticinterest in continuing to live. Because of this, themoral objection to killing it may be comparativelyslight. As the fetus matures, however, its mentalcapacities are increasingly enhanced, therebyestablishing the potential for more extensivepsychological connections with itself in the future.As its mental life becomes richer and moreunified, the strength of its egoistic interest in con-tinuing to live increases, and the moral objectionto killing it becomes correspondingly stronger. In


http://jme.bm

j.com/

J Med E


pril 1999. Dow

nloaded from

http://jme.bmj.com/


short, it is the fact that the fetus's egoistic interestin continuing to live increases in strength as itspsychological capacities mature that explains whykilling becomes more morally serious as fetaldevelopment continues.These claims apply even if we are essentially

human organisms that begin to exist at concep-tion. For, if psychological connectedness is part ofthe basis for rational egoistic concern about thefuture, we would, at conception, have only a weakegoistic interest in continuing to live, for at thatpoint we have no psychological connections withourselves in the future. This would explain why,even if we are organisms, early spontaneous abor-tions are not tragedies. It would also explain whymost of us do not regard abortion as morallycomparable to the killing of a child or adult.The fact that we may reach similar conclusions

whether we assume that we are embodied mindsor human organisms illustrates an importantpoint. This is that, if identity is not the basis foregoistic concern, our understanding of what weare and when we begin to exist has substantiallyless practical significance.The skeptical reader may wonder why our

moral concern should be guided by an individual'segoistic interests rather than by what would bebetter or worse for that individual. I haveconceded that, if we consider the fetus's life as awhole, it is clear that death would be much worsefor it than continuing to exist. Why should not thisfact rather than the fact that the fetus's egoisticinterest in continuing to live is weak-determinethe morality of killing the fetus? This is a difficultquestion, but there are several points that may bebriefly made in reply. An individual's interests andhis egoistic interests normally coincide. It is onlyrarely, in cases in which the basis of an individual'segoistic concern about the future is unusuallyweak, that they diverge. Where they diverge, theegoistic interest seems prudentially and morallymore significant. In the case of the fetus, forexample, it is the egoistic interest that explains oursense that its death matters less than the death ofan older person, which in turn supports the viewthat killing the fetus would be less objectionable.Division is another case in which an individual'sinterests diverge from his egoistic interests. Heretoo the egoistic interest seems to capture what ismore important. It is against A's interests todivide. If he divides, he will cease to exist; therewill be no further good in his life. Yet it may beegoistically rational for him to want to divide. If hedivides, he will be connected in the way thatgrounds egoistic concern to two future lives ratherthan one.

I do not claim, however, that killing is wrongonly because, and only to the extent that, itfrustrates the victim's egoistic interest in continu-ing to live. Where the killing of persons isconcerned, the frustration of the victim's egoisticinterests is only one factor that contributes tomaking killing wrong. There is another dimensionto the morality of killing that is concerned with therespect that we owe to other persons as our moralequals. It is doubtful, however-though I cannotdefend this claim here that these further consid-erations apply in the case of individuals, such asfetuses, that lack self-consciousness andautonomy.7 If this is right, the moral objection tokilling a fetus should be commensurate, otherthings being equal, with the strength of the fetus'segoistic interest in continuing to live.

In summary, I suggest that the killing of a clonein the early stages of its development is permissi-ble on the ground that there is as yet no one therewho would be harmed by being killed. Later, afterone of us begins to exist in association with thecloned organism, there is someone who would beaffected for the worse by being killed. But themoral objection to killing that individual is initiallyrelatively feeble owing to the extreme weakness ofthe individual's egoistic interest in continuing tolive. Killing becomes increasingly objectionable asthe individual's psychological capacities develop,but may remain permissible for some time(perhaps until around the time of birth) providedthat there is a sufficiently significant end thatmight thereby be achieved. Saving a person's life issuch an end.8

AcknowledgementI am grateful to David McCarthy and JulianSavulescu for comments on this paper.

J7eff McMahan is Associate Professor of Philosophy,University of Illinois at Urbana, USA.

References1 McMahan J. Killing at the margins of life. New York: Oxford

University Press, forthcoming: ch 1.2 Kuhse H, Singer P. Individuals, humans and persons. In: SingerP et al, eds. Embryo experimentation. Cambridge: CambridgeUniversity Press, 1990: 67.

3 The following remarks parallel one of Derek Parfit's argumentsfor a reductionist view of persons. See Parfit P. Reasons and per-sons. Oxford: Oxford University Press, 1984:211-12.

4 Miller K. Together forever. Life 1996 Apr: 44-56.5 Schechtman M. The constitution of selves. Ithaca: Cornell

University Press, 1996:52.6 See reference 3:ch 12.7 See McMahan J. Killing and equality. Utilitas 1995;7:1-29.8 This article was stimulated by my reading a provocative

unpublished paper by Julian Savulescu. The article also drawsheavily on ideas developed in more detail in Killing at the mar-gins of life: see reference 1.


http://jme.bm

j.com/

J Med E


pril 1999. Dow

nloaded from

http://jme.bmj.com/

Documents

Cloning,killing, and identity78 Cloning,killing, andidentity entity or, as it maybe, series ofentities resulting fromnucleartransfer at anystage in the develop-mental process initiated