CHAPTER FOUR: POSITIVE AND NEGATIVE GENETIC INTERVENTIONS

I. Old Distinctions in New Clothes

Positive and Negative Eugenic Goals for Populations

Earlier in this century, the eugenics movement was more concerned with the genetic quality or "health" of populations than with the health and welfare of individuals. Indeed, as noted in Chapter Two, some in that movement were concerned that keeping "unfit" individuals healthy might have dysgenic effects. This focus on populations rather than individuals is part of what makes the movement and its goals seem so threatening.

We also saw that the eugenics movement had two main goals. Its negative goal was the reduction in "dysgenic" effects or burdens on the gene pool by eliminating genetic diseases, disorders, disabilities, and other "defects." Negative eugenics aimed to improve the health and performance of the population by preventing reproduction of its least healthy and least capable members. This goal required severe restrictions on reproductive rights, for those with "defects" had to be kept from reproducing, if necessary through the involuntary sterilization of "mental defectives."

The goal of "positive eugenics" was to improve the health and performance of the population by increasing the rate of reproduction of those harboring its best traits and capabilities. Pursuing this goal also involved modifying traditional reproductive practices, though usually through voluntary measures. For example, as also noted in Chapter Two, to inspire the right marriage choices, there were competitions at county fairs aimed at displaying the best human "breeding stock."

The boundary between positive and negative eugenics did not mark a generally accepted moral boundary for proponents of the eugenics movement, though some attributed moral significance to it. Rather, it reflected two aspects of the same goal of improving a population. The distinction between them rested crucially on drawing a further line between what was considered sub- or abnormal or defective and what was considered normal or even superior. The scope of negative eugenics--that is, what counted as defects to be eliminated--clearly depended on what conception of normal or even superior phenotype and genotype was used. To the extent that "defects" were limited to clear, paradigmatic cases of disease or disability, negative eugenics could be portrayed as a form of disease prevention--though the infringements of reproductive rights committed in its name are disturbing and unacceptable regardless of how benign the categorization of disease happened to be.

A highly idealized or perfectionist view of superior or normal traits would mean that a trait we ordinarily take to be normal would count as "defective." Elimination of these traits would then become the legitimate target of negative eugenics. To the extent that racial stereotypes of "higher" and "lower" races or "socially superior" and "inferior types" were involved in defining normal and defective traits, negative eugenics risked becoming--and in the United States and Germany, dramatically became--racist or even genocidal.

Many of the most serious abuses were actually committed in the name of negative rather than positive eugenics. The seemingly benign goal of disease prevention was transformed into the goal of people prevention (or even elimination), at least for those groups of people harboring "defective" traits. Today, we are at least as horrified at some of the ways the distinction between normal and defective was drawn as we are at the infringements of reproductive rights involved in the pursuit of eugenic goals. We are appalled at the ways in which bad science was harnessed to serve discriminatary attitudes toward race, class, and disability.

Positive and Negative Interventions and the Health and Welfare of Individuals

Today a distinction is commonly drawn between negative and positive "genetic interventions." As noted in Chapter One, we use this latter term in a very broad sense to include somatic or germline uses of modern genetic technologies and pharmacological applications of those technologies, as well as uses of genetic technology in screening aimed at family planning. As in its earlier use, the negative/positive distinction depends on our being able to distinguish disease, disorder, impairment or disability from normal traits or capabilities.

In many contemporary discussions, the negative/positive distinction is used to draw a fundamental moral boundary. There is a presumption that negative genetic interventions--ranging from screening and selective abortions to somatic cell and (more problematically) germline replacement therapy--are morally permissible, whereas positive interventions are morally impermissible or at least highly problematic (Anderson 1990, 1980, 1985, 1988; cf. Glover 1988; Kitcher 1995).

The explanation for this presumption, which is surprising in light of the serious abuses committed in the name of negative eugenics, is that negative genetic interventions are in general no different from other medical treatments, whether preventive or curative, of disease or impairment, whereas positive genetic interventions are aimed at the morally problematic enhancement of normal traits and capabilities. The implication is that there is little that is morally problematic about treating disease but there may be much moral controversy about which normal traits, if any, should be enhanced. Obviously, if this negative/positive distinction translates into the distinction between permissible and impermissible uses of genetic interventions, it would be of great importance for public policy, and that explains the interest in this issue and our focus on it in this chapter.

It is important not to confuse the negative/positive distinction with the difference between somatic and germline replacement therapies--that is, between replacing genetic material in the somatic cells of an individual (affecting the expression of some trait solely in that individual) and replacing genetic material in germ cells (affecting the expression of traits for the individual resulting from those germ cells and his or her descendants).

The somatic-germline distinction, so prominent in the literature on genetic technologies, is often used to draw a similar line between morally permissible and morally impermissible or suspect interventions (Council For Responsible Genetics 1993). But it cuts across the distinction between negative and positive genetic engineering. A somatic replacement intervention could be pursued, had we the technology, to correct a defect or to enhance an otherwise normal capability. The same is true for germline replacements. Any systematic differences in risks and uncertainties, in the relation between risks and benefits, or in the prospects for obtaining informed consent of the affected party that support different moral conclusions about the permissibility of somatic and germline interventions will be independent of the distinction between negative and positive interventions (Kitcher 1995), which is our concern here.

Having noted that the shadow of the negative/positive eugenics distinction falls darkly on the negative/positive genetic intervention distinction, it is important to emphasize some decisive differences in the social context (discussed in more detail in Chapter Two). One important difference is that the primary concerns of those interested in either somatic or germline gene replacement today is not the "health" or "capability" of populations or the "quality" of the gene pool, but the health and well-being of individuals and their descendants. In addition, there is a greatly enhanced respect for and legal recognition of individual reproductive freedom, as well as for the requirement that voluntary, informed consent or its legitimate proxy be obtained for all medical (and experimental) interventions.

Contemporary developments in genetics also take place on the heels of a highly successful disabilities rights movement. Activists have educated the public about people with disabilities, established strong legal protections for them, and moved the public to broader acceptance of the "diversity" they represent. Finally, the contemporary positive/negative intervention distinction is drawn in a period when we are still highly sensitive to, and frightened by, the historically recent slide from eugenics to genocide.

Together these differences in the social context do much to dispel the shadow cast by the eugenics movement. Nevertheless, we are still far from a worldwide climate in which racist ideology is no longer a threat. We continue to witness "ethnic cleansing" and genocide in parts of Europe and Africa; we see active racist movements even in liberal democracies promoting ideology that explicitly invokes Nazi ideas with favor. Echoes of racism underlie the appeal of many of the current "hot" social reform topics in the United States--affirmative action, welfare reform, crime, and gun control. Much but not all has changed. We leave these comments on the social context, however, to concentrate on the moral issues surrounding the positive/negative distinction and its uses.

Moral Boundaries and the Positive/Negative Distinction

The question remains, Can the negative/positive intervention distinction be drawn as cleanly as its proponents imply, and with anything like the moral and public policy implications they suggest? We believe the distinction and the moral implications drawn from it deserve more careful examination for two reasons.

First, the treatment/enhancement distinction on which the contrast between negative and positive intervention rests is used for a very different moral purpose in most medical and medical insurance contexts. Specifically, it is currently used to draw a line between services that are obligatory to provide to others--for example, in private or public insurance schemes--and those that are not. But a line between what is obligatory and nonobligatory in insurance schemes is not to be confused with a line between what it is permissible and impermissible for anyone to do with genetic interventions. Obviously, what is obligatory in an insurance scheme is permissible, and what is impermissible for anyone to do cannot be obligatory in insurance schemes. But what is nonobligatory in insurance may be either permissible or impermissible for individuals to do, and what is permissible for individuals to do may be either obligatory or nonobligatory in insurance. So, even given the distinction between treatments and enhancements, we shall have to consider whether this gives us any guidance on the boundary between permissible and impermissible.

Second, the treatment/enhancement distinction itself has been sharply criticized even in its current use for reasons that must be taken seriously. To decide whether the negative/positive genetic intervention distinction has any public policy implications--that is, whether it helps us draw any moral boundaries--we will have to consider more carefully the treatment/enhancement distinction on which it rests. In the remainder of this chapter, we shall consider various objections to the treatment/enhancement distinction: that it is difficult to draw, that it does not give us the boundary between what is obligatory and nonobligatory to provide in medical insurance schemes, and that it leaves us with hard cases that make the distinction seem arbitrary.

To resolve some of these issues, we will have to discuss alternative accounts of our obligations to assist others with medical services, including those based on genetic technologies. Specifically, we shall consider the suggestion discussed in Chapter Three, by one of the more expansive interpretations of the level playing field conception of equal opportunity and by resource egalitarianism, that we drop the distinction between treatments and enhancements altogether in favor of more directly assessing whether an intervention is required because it equalizes opportunities.

On that view, which we dubbed the brute luck view (following Scanlon), equality of opportunity requires eliminating all disadvantaging deficits in capabilities for which an individual is not responsible, whether or not they are the result of disease or impairment or merely the result of bad luck in a natural lottery for (otherwise normal) capabilities. A supporting intuition for this expansive account of equal opportunity is the claim that contingencies in the natural lottery are as morally arbitrary as those in the social lottery, so that if equal opportunity is supposed to correct for social contingencies, it also ought to correct for natural ones that produce unchosen disadvantages.

In this chapter, we explore and try to motivate further an alternative, more restrictive approach to the notion of equal opportunity and its implications for our obligations in health care. Described briefly in Chapter Three as the social structural view, this conception of justice supports a qualified and limited defense of the treatment/enhancement distinction. By drawing out the implications of the social structural view, we will make much more concrete the rather abstract examination of justice in Chapter Three and bring it into closer contact with basic issues of social policy.

Before trying to defend the distinction between positive and negative intervention or a theoretical perspective that uses it, we first reinforce some of the theoretical doubts about it raised in the previous chapter. We do so by focusing on a range of "hard cases" drawn from actual medical practice, in which use of the distinction sometimes seems especially problematic, even though it is part of a widely accepted insurance practice. We then reverse course, defending on both policy and theoretical grounds a more limited goal of maintaining "normal functioning" through health care rather than the more expansive goals that might be supported by the brute luck view. Nevertheless, this approach, though different in its theoretical conception from the more expansive brute luck conception of equal opportunity, comes very close to it in its practical implications. This convergence in practice, which was briefly noted in Chapter Three, is of considerable interest itself.

The limited defense offered here of the treatment/enhancement distinction means it cannot provide a clear or unequivocal guide to the moral boundaries between what is obligatory and nonobligatory to provide in insurance or between what it is permissible and impermissible for individuals to do. Neverthless, it has a useful, if modest, bearing on each, especially the former. In the remainder of this chapter, we are concerned more with negative genetic interventions and the obligatory/nonobligatory boundary. In Chapter Six, we explore further the general issue of improving our offspring through genetic interventions, both positive and negative.

II. Treatment Versus Enhancement: Wide Use, Hard Cases, Strong Criticism

Insurance Coverage and "Medical Necessity"

As noted, the treatment/enhancement distinction draws a line between services or interventions meant to prevent or cure (or otherwise ameliorate) conditions viewed as diseases or impairments and the interventions that improve a condition viewed as a normal function or feature of members of our species. Glover (OTA 199X) comments that, like night and day, the distinction may pose boundary problems, but the line it draws is nevertheless useful. We will have to assess whether he is underestimating the challenges to the distinction. Still, the line drawn here is widely appealed to in medical practice and medical insurance contexts, as well as in our everyday thinking about the medical services we do and should assist people in obtaining.

The treatment/enhancement distinction is closely related to the concept of "medical necessity" that appears in legislation regulating public insurance in both the United States and Canada, and in private insurance contracts. Medically necessary services are those that effectively treat physical or mental disease and impairment or ameliorate conditions deriving from them (Daniels and Sabin 1991, Sabin and Daniels 1994). They may effectively produce benefits for other conditions, but those do not count as medically necessary.

For example, insurance coverage is provided by public and private schemes for growth hormone treatment for children projected to be very short, provided that there is an underlying disease condition, such as a diagnosable growth hormone deficiency or Turner's Syndrome. If there is no underlying disease condition, insurers do not cover the treatment for children whose parents simply want them to be taller, regardless of how short they will be. Similarly, insurers will generally reimburse--and in some states, like Massachusetts, they are mandated to reimburse--reconstructive breast surgery following mastectomy or trauma. But they do not reimburse "cosmetic" surgery, however strongly a woman may feel that her life will be improved if her breasts are made larger or smaller.

The same distinction plays a role in coverage for mental health therapies. Consider the following cases illustrating coverage policy within the Harvard Community Health Plan (HCHP) prior to 1994, then a staff-model HMO that served over 550,000 people in New England (the cases are drawn from Sabin and Daniels 1994). An adult patient with a history of bipolar disorder had been stabilized on lithium for some years. He remained shy, however, and was referred to an out-of-plan group therapy situation, from which he clearly benefited over a period of several years. In its original benefit structure, this long-term treatment could not have been covered by HCHP. HCHP revised its benefit structure, allowing an "extended benefit" that would cover protracted therapy of this sort without extensive co-pays provided treatment was for a serious condition. But does treatment of shyness count as treatment of a serious disorder? The psychiatrist managing the Shy Bipolar's case believed that the shyness was the result of the onset of the bipolar disorder; had the disorder not interfered with the adolescent development of this man, who was normally outgoing before its onset, he would probably have been more outgoing. Consequently, the therapist reasoned, the "extended benefit" should be given. Had the shyness not been "diagnosed" as the result of the bipolar disorder, then even if it were comparably serious, there would have been no eligibility for an extended benefit.

The factor underlying this reasoning did not have to do with the degree of suffering involved in being shy; the reasoning depended on the etiology or explanation of the shyness. The point is illustrated by another case in which clinicians distinguish between treatment of illness and enhancement of well-being. An intelligent, professionally successful married father of two children sought treatment because of severe unhappiness associated with marital distress. His wife suffered from a serious mental illness that made her very difficult to live with. The Unhappy Husband was committed to maintaining the marriage. A V code diagnosis ["Conditions not attributable to a mental disorder that are a focus of treatment" (Diagnostic and Statistical Manual III-R:359)] (marital problem) was made. In 26 highly productive sessions of psychotherapy, the man was able to clarify some of the pertinent dynamic issues in his marriage, and developed a number of adaptive strategies for lessening his distress. The patient wished that his treatment would be covered by insurance, but he agreed that he was not suffering from an illness and that it was fair to expect him to pay.

The Unhappy Husband was probably suffering more than many of the HMO members being treated for illnesses, and psychotherapy definitely enhanced his well-being. What possible rationale could there be for not covering his treatment? The clinician's decision hinged on the question of what the Unhappy Husband is suffering from. By the criteria set forth in DSM-III-R, the man did not have an illness. His suffering arose from the fact that although his wife's unchanging condition caused him great pain, his values precluded divorce. The clinician believed that under the prevailing agreements that govern insurance, individuals like the Unhappy Husband should be responsible for some or all of the cost of ameliorating the unhappiness associated with an unfortunate existential situation. Paradoxically, if the Unhappy Husband expressed his suffering through somatic symptoms, and presented to an internist rather than a mental health clinician, insurance would typically cover medical investigation and treatment, which would probably be less effective but costlier than psychotherapy. A 1989 survey of medium and large firms showed that only 2 percent of insured employees have coverage for outpatient mental health services equivalent to other medical services (Scheidemandel 1993:44).

Treatment/Enhancement and Moral Hazard

Some clinicians and some members of the public balk at the line being drawn here. They may be inclined to say that if there is "suffering" involved, as there clearly might be for short children, shy adults, or husbands opposed to divorce from ill wives, then we should relieve the suffering with medical interventions funded through insurance. Nevertheless, there is a very good reason why insurers prefer to insist on a diagnosis of disease or impairment as an eligibility condition for reimbursement for beneficial services, and why they do not simply agree to reimburse anything that is effective at producing a benefit, at least as perceived by the patient or even the therapist. Without the requirement for a disease diagnosis, insurers would be exposed to what is termed "moral hazard."

Moral hazard refers to the modification of behavior that individuals make in light of the incentives provided by insurance coverage--in particular, behavioral changes that make them eligible for benefits they would not otherwise be entitled to have. For example, someone with extensive fire insurance might seek to make a payoff more likely by setting a fire (here the moral hazard leads to arson and fraud) or by failing to take reasonable precautions against one. Obviously, insurers do not reimburse people for fire damage that is the result of arson the insuree arranges or to gross negligence. Similarly, there is no insurance market for reimbursement for the costs of speeding tickets.

If individuals could define their otherwise normal condition as one that involved "suffering" or one that imposed unacceptable disadvantage, given their expectations, then insurance against such suffering or disadvantage would encourage extensive moral hazard. An insurance market against certain risks can function only if the risks are measurable actuarially and the market does not create conditions (such as moral hazard) that make the risk unmeasurable.

Treatments and the Limits of Obligations

This point about insurance is related to a deeper fact about when we feel obligated to assist others, including when we feel obligated to reduce or mitigate the effects of an inequality that arises among us. Consider the case of the Plain Hero who feels very dissatisfied with his appearance: his face is normal but hardly handsome. He feels he might be more successful in seeking companionship or in "presenting" himself in business relationships if his face or hair better matched some social model of handsomeness. He may have developed elaborate explanations for his failures that excuse himself of any responsibility and place the blame on his appearance and the "superficiality" of others so affected by it. The solution does not lie in changing how he thinks or behaves with prospective social or business partners but in how he appears. Are we obliged to relieve his suffering by providing him with the means to obtain cosmetic surgey?

Some may feel so obliged, but most would not. They would complain that the problem here lies in the beliefs, attitudes, and excuses harbored by our Plain Hero. It is his responsibility to rethink his goals and his means of achieving them. If he is unable to do so, it may indicate some deeper psychological problem, and the treatment he may need is a mental health intervention, not cosmetic surgery. In part, what we resist here is the idea that an individual can develop very expensive tastes or preferences, for example for being one of "the beautiful people," and that others owe him the chance to satisfy those preferences. We resist being held hostage in our obligations to assist others (or to reduce inequality) by expensive tastes. If the Plain Hero wants to invest his own resources in removing what he sees as the obstacle to his success, then let him. We owe him a new face no more than we owe him a Porsche, which he might just as justifiably think would open new doors to him.

The Plain Hero, of course, is to be distinguished from someone who has been disfigured by congenital deformity, disease, or trauma, for whom reconstructive plastic surgery meets a medical need (Daniels 1985:31n9; MacGreggor 1979; de Beaufort, et al 1996). Obviously, there will be gray areas in which it is unclear whether plastic surgery meets a medical need or simply a cosmetic preference.

This concern about social hijacking by extravagant preferences cannot be all there is to the matter. The Unhappy Husband in the example discussed earlier has not adopted extravagant tastes. His aversion to abandoning his ill wife may be the result of some unshakeable moral conviction, perhaps one inculcated in his religious upbringing, or it may be the result of a moral conviction he affirms and takes responsibility for--one that is in fact admirable, as loyalty often is, despite its consequences for him. Still, in general we do not owe it to people to compensate them for all sacrifice and suffering they incur when they live up to what they see as their obligations, though sometimes--for example, with veterans--we may think compensation is appropriate. We are often unhappy in carrying out our moral duties, yet we do not generally think that others owe us something for doing so--that is, other than praise or recognition or moral support. We no doubt owe moral support to the Unhappy Husband in a way we do not owe such support to the Plain Hero.

As a result of this discussion, we might conclude that the treatment/enhancement distinction derives some support from the general fact that disease and impairment are conditions that we are generally not responsible for (leaving aside the issue of self-induced disease, for the moment).[1] In addition, they are conditions that we can generally agree involve some sort of objectively specifiable burden or harm. For example, the harm might be an impairment of the range of opportunities open to us (see Daniels 1985) because of the reduced functioning or diminished capabilities that result from the disease or impairment. We might also think the "pain and suffering" involved as a result of these conditions can be recognized as typical features of the human condition. In contrast, we may want to dismiss the pain of the Plain Hero as an "exaggerated" and quite atypical response to the human condition, saying "There need have been no real suffering there, only a change of his attitude."

Hard Cases and the Expansion of Obligations

Unfortunately, if we look more closely at certain difficult cases, we are more hard pressed to attribute to the treatment/enhancement distinction the weight it is given medical practice, including insurance schemes. For the sake of vividness, let us put names on the faces of the growth hormone treatment cases.

Johnny is a short 11-year-old boy with documented GH (Growth Hormone) deficiency resulting from a brain tumor. His parents are of average height. His predicted adult height without GH treatment is approximately 160 cm (5 feet 3 inches).

Billy is a short 11-year-old boy with normal GH secretion according to current testing methods. However, his parents are extremely short, and he has a predicted adult height of 160 cm (5 feet 3 inches) (Allen and Fost 1990:117).

These cases make the distinction seem arbitrary for several reasons. First, Johnny and Billy will suffer disadvantage equally if they are not treated. There is no reason to think the difference in the underlying causes of their shortness will lead people treat them in ways that make one happier or more advantaged than the other. Second, although Johnny is short because of dysfunction whereas Billy is short because of his (normal) genotype, both are short through no choice or fault of their own. The shortness is in both cases the result of a biological, "natural lottery." Both thus seem to suffer undeserved disadvantages. Third, Billy's preference for greater height, just like Johnny's, is a preference that most people hold; it is not peculiar, idiosyncratic, or extravagant. Indeed, it is a response to a social prejudice, "heightism." The prejudice is what we should condemn, not the fact that they both form an "expensive taste" in reaction to it.

If we return to the case of the Shy Bipolar, we could raise exactly the same points in comparing him to an equally shy but otherwise normal person, the Shy Normal. They will suffer the disadvantages of shyness equally. Both are shy through no fault of their own--assuming normal shyness is a feature that is significantly determined by temperament or by exposure to early learning situations that a person did not choose to be in. And most people would prefer to be more outgoing and to enjoy the relationships we think come with such a posture toward others--the preference to change from being shy is not idiosyncratic or extravagant.

Cases like these raise several questions: Does the concept of disease underlying the treatment/enhancement distinction force us to treat relevantly similar cases in dissimilar ways? Are we violating the old Aristotelian requirement that justice requires treating like cases similarly? Is dissimilar treatment unfair or unjust? Any defense of the moral use to which we put the treatment/enhancement distinction in medical and insurance contexts must respond to this concern. These cases make more concrete the worries raised in Chapter Three that there is something "morally arbitrary" about addressing the disadvantages produced by disease and impairment and not addressing those imposed by disadvantageous--but normal--allotments of capabilities or talents and skills.

The Microstructure of the Normal, and Moral Arbitrariness

Before responding to these questions, we would like to deepen the sense that there may be something morally arbitrary about the use to which the treatment/enhancement distinction is put, and perhaps even about the distinction itself. To do so we appeal, at least hypothetically, to something we may learn from the Human Genome Project and from the greater knowledge we get about how genes regulate growth.

Suppose we learn that some particular pattern of genes explains the extreme shortness of Billy, the child who did not seem to be growth-hormone-deficient. Suppose we learn that some particular genes Billy has make some receptors to growth hormones slightly less responsive than the genes that would lead someone to be of average or above-average height; perhaps there are fewer such receptors, or perhaps they shut down earlier than in those whose genotypes generally make them taller, or perhaps they slow down production of growth hormones sooner. We learn, that is, just which "losing numbers" in the natural lottery placed Billy in the bottom few percent of the normal distribution for height. Suppose, further, that we also identify a gene that disposes Johnny to develop the brain tumor that caused his growth hormone deficiency. We now have traced both Johnny and Billy's shortness back to specific genes. Billy's genes work directly to make him short; Johnny's work indirectly to do so, through causing a tumor that disrupts hormone production. Why does having one set of genes give Johnny a claim on social resources necessary for growth hormone treatment but Billy no such claim?

Of course, this story really adds nothing new. We already knew that Johnny and Billy's troubles were rooted in their biology. Adding the genetic details only makes things seem more vivid. Still, if we can identify the specific genes that contribute to Billy's shortness, we may be more tempted to think of them as "bad" genes: they lead to Billy's unhappiness or disadvantage in a "heightist" world. We may be more tempted to think of them very much on the model of genetic defects or diseases, especially if they work through mechanisms that have some analogy to pathological defects. We will be tempted, that is, to medicalize what we have hitherto considered normal.

There may be an interesting analogy here to the differentiation into "learning disabilities" of what had earlier been thought of as the "not so bright" end of the school performance trait. In neuropsychology, students whose performances are at variance with their measured intelligence and whose performance discrepancies are not explained by poor teaching are considered likely to have some specific learning disability. The general working hypothesis is that there is a specific information processing problem or deficit--a departure from or impairment of normal information processing--that probably has some neurological basis.[2]

The analogy is to our finding specific growth hormone deficiencies in a person who is otherwise constituted (genotypically) to be taller (in a normal range of environments). In the case of learning disabilities, we disentangle people who were simply grouped as "dull or ineffective students" at the low end of a "normal distribution" of school performance into those with specific disabilities and those who just remain the unanalyzed "dull" students. What happens, however, if we find out that being taller or normally intelligent is nothing more than having more of (an appropriate set of) the relevant microstructures/processes? Someone who is very short or dull is then missing more of them. Why is this not seen as just as much a "defect" as having a learning disability or growth hormone deficiency?

The hard cases thus pose this question: What justifies us treating the normal but "bad" or disadvantageous genes differently from genes that lead to growth hormone deficiency or to receptor insensitivity to growth hormone (or to learning disabilities)? If we can remedy the effects of these genes with growth hormone treatment or other treatments, including genetic tampering, we might think it quite arbitrary to maintain the treatment/enhancement distinction.

Two Objections to the Treatment/Enhancement Distinction

The hard cases we have been considering, especially in light of the speculation about what we might learn from the Human Genome Project and related genetic research, really raise two quite distinct kinds of objections to the treatment/enhancement distinction. One is that the distinction, even assuming we can draw a persuasive line between the treatment of disease or impairment and the enhancement of otherwise normal traits, does not have the moral import that is commonly attributed to it--for example, in our insurance practices. Some nondisease conditions seem to oblige us to provide assistance to people for the very kinds of reasons that some diseases or impairments do. If so, the treatment/enhancement distinction does not map onto the boundary between morally obligatory and nonobligatory services, as the defender of our medical insurance practices might have hoped.

The second objection challenges the basis on which the treatment/enhancement distinction is itself drawn. By implication, this objection then challenges whether we can use the treatement/enhancement distinction to draw further moral distinctions without some kind of circularity. On this view, it is not because there is something biologically distinctive about Johnny's condition as opposed to Billy's that led us to describe Johnny as having a disease and Billy not (although Johnny does have a tumor and Billy does not). Rather, our "social construction" of disease draws on a set of values that happens to have singled out Johnny rather than Billy in this way. But if we come to see that the same value that leads us to consider Johnny's tumor as a disease condition (suppose it is the concern we have to assist people whose conditions put them at a certain kind of disadvantage through no fault of their own) also applies to Billy's condition, then we should reconstruct our view of disease to include Billy's condition.

According to this line of thinking, it is our norms and values that define what counts as disease, not merely biologically based characteristics of persons, and the arbitrariness in these hard cases comes from inconsistently applying our values. Pointing to the line between treatment and enhancement is not, then, pointing to a biologically drawn line; rather, it is an indirect way of referring to valuations we make. We cannot point to such a line as the grounds or basis for drawing moral boundaries since we are only pointing to a value-laden boundary we have constructed.

This objection echoes a worry we expressed in commenting on the positive/negative eugenics distinction: judgments about eugenicists made about "defects" reflected value judgments about what counted as normal or superior traits. Aimed at the treatment/enhancement distinction more generally, the objection is that we are being offered an apparently "natural" baseline between disease (and impairment) and the biologically normal, when there really is none. The effect is that, disingenuously or not, we disguise hidden moral judgments that really form the basis of the moral boundary we purport to derive.

Are there reasonable replies to this objection? Can the treatment/enhancement distinction justifiably do at least a significant portion of the work we commonly rely on it to do?

III. A Limited Defense of the Treatment/Enhancement Distinction and Its Circumscribed Use

Treatment/Enhancement and the Obligatory/Nonobligatory Boundary

No reasonable defense of the treatment/enhancement distinction is possible if we expect too much of it. Specifically, we should not expect that distinction to translate without qualifications into the boundary between obligatory and nonobligatory services, even if it justifiably plays an important role in medical insurance coverage decisions. There are good reasons why we are not obliged to provide all and only treatments as oppposed to enhancements. For the sake of argument in much of this section, we shall assume we can draw a plausible line between the treatment of disease or impairment and the enhancement of otherwise normal traits, and we shall revisit the assumption at an appropriate point.

There are two basic reasons why the treatment/enhancement distinction does not coincide with the boundary between obligatory and nonobligatory services. First, resources will be too limited to meet all our needs for the treatment of disease or impairment. Justice then requires that we meet the most important needs first, leaving people to fend for themselves in meeting less important medical needs. In other words, the class of beneficial treatments is broader than the class of services we are obliged to provide, given reasonable resource constraints. Being a treatment is thus not a sufficient condition for our being obliged to provide a service to people. Still, we might believe that being a reasonably effective treatment for a disease or impairment is still a necessary condition, an eligibility condition, for something being included in an insurance package or being thought of as a service we are obliged to provide.

The second reason for limiting the role of appeal to the treatment/enhancement distinction also rules out the unqualified claim that it provides a boundary between the obligatory and the nonobligatory. Society may--indeed does--have certain moral or legal obligations to offer medical services that do not involve the treatment of disease. For example, we would argue (were this the occasion) that society's obligations to respect the equality of women compel it to make abortion a covered service in a national benefit package, just as it is already a covered service in most existing private insurance in the United States. But the reason for including it has nothing to do with treating a disease or impairment, since an unwanted pregnancy is not a disease or impairment but rather the result of normal functioning. (President Clinton's proposed Health Security Act [1994] disguised the issue by including nontherapeutic abortions as "pregnancy related services.")

Indeed, the reasons many have for excluding abortion also have nothing to do with the distinction between treatment and enhancement: some people are opposed to permitting abortions even when continued pregnancy involves a threat to the health or life of the mother and is therefore "therapeutic" or "medically necessary" abortion. If we are right that nontherapeutic abortion services should be included in standard benefit packages, because of concerns about the equality of women, then treatment of disease and impairment does not capture the class of services society is obliged to provide once we consider all of our obligations. As we shall see in what follows, sometimes concerns about equal opportunity, whether construed more narrowly, as in this chapter, or more expansively, as in the brute luck view described in Chapter Three, may oblige us to provide some genetic interventions even when they are not treatments of disease.

The Primary Rationale for Medical Obligations

It may still be the case that the primary rationale for claiming that society is obligated to provide people with a medical service is that it meets an important need for the treatment of disease or impairment.[3] This may be the reason people typically agree on and justifiably cite when they think about the moral importance of health care services. If there is a plausible defense of the treatment/enhancement distinction, it will be in the limited role pointed to here: Our primary justification for considering a health care service to be something society is obliged to offer people is that it is a reasonably effective treatment for a disease or impairment (and resource constraints permit treating this condition). Other reasons may broaden societal obligations, but the primary justification gives us the core.

This limited use of the treatment/enhancement distinction is elaborated in Daniels's (1985) account of justice and health care, as described in Chapter Three. Disease and impairment, both physical and mental, are construed as adverse departures from or impairments of species-typical normal functional organization or "normal functioning," for short (see endnote 2). The biomedical sciences for humans, like the veterinary sciences for animals, study both the variation in the functional organization typical for our species and the departures from normal functioning that we call disease and impairment (Boorse 1975, 1976, 1977; Kitcher 1995).[4] The line between disease and impairment and normal functioning is thus drawn in the relatively objective and nonevaluative context provided by the biomedical sciences, broadly construed. What counts as a disease or impairment from the perspective of these sciences is largely free from controversy in the broad range of cases.

Of course, sometimes value judgments, including prejudices, as well as errors intrude, and we get examples of conditions or behaviors that are improperly classified as disease or impairment, such as the disease of masturbation (Engelhardt 1974) or homosexuality (Bayer 1981). But just as whales are not fishes, though they were long classified as such, so too these conditions are not diseases,[5] even if complex social conditions and attitudes contributed to their being viewed as such.

We are not, however, just interested in categorizing disease and impairment. Rather, we have an important interest in their effects, since they often cause pain and suffering, shorten life, and quite generally impair in varying degrees the range of opportunities open to us. On Daniels's view, the central moral importance, for purposes of justice, of treating disease and impairment with effective health care services (construed broadly to include public health and environmental measures, as well as personal medical services) derives from the way that protecting normal functioning contributes to protecting opportunity. Specifically, by keeping people close to normal functioning, health care preserves for people the ability to participate in political, social, and economic life. It sustains them as fully participating citizens, as "normal competitors" in all spheres of social life.

By maintaining normal functioning, health care protects an individual's fair share of the normal range of opportunities (or plans of life) reasonable people would choose in a given society. That individual fair share, however, is defined by reference to the individual's talents and skills, suitably protected against mis- or underdevelopment as a result of unfair social practices. Accordingly, a principle assuring fair equality of opportunity should be extended to govern the design of health care systems. (As we saw in Chapter Three, Rawls [1971] simplifies his theory by assuming all are fully functional; Rawls [1993] endorses Daniels's approach.)

By keeping people functioning as close to normally as possible, within resource limits, we discharge our obligations to protect (a suitably extended principle of) fair equality of opportunity. Daniels's extension of "fair" equality of opportunity to health care broadens the level playing field to include socially correctible departures from normal functioning. In doing so, it corrects both for some natural and some socially induced disadvantages.[6] His account, however, stops short of levelling the field further to include the redistribution of otherwise normal but competitively disadvantageous capabilities (as the "brute luck" view does). We return to consider arguments in favor of this stopping point in the next few sections, after adding an important detail to this account.

The relative moral importance of treating different diseases and impairments can in part be judged by reference to their impact on the range of opportunities open to us. Because this range of opportunities is itself socially relative, being affected by technology, education, wealth, and other cultural factors, judgments about the relative importance of treating different diseases and impairments will have some social variability. (See our discussion in Chapter Seven on the morality of inclusion as it applies to disabilities.) For example, dyslexia is a cognitive impairment in any society, but it is disability and hence important to treat only in literate societies (Daniels 1985). Within a society, relative to its normal range of opportunities, some diseases and impairments are more important to treat than others, and this will affect our decisions about which treatments to offer when we cannot provide all the ones people need.

The appeal to a principle assuring fair equality of opportunity is an attempt to explain why we attribute special moral importance to treating disease and disability. If the particular interpretation of equal opportunity is acceptable, it explains the primary rationale for the provision of health care services. Aiming to maintain normal functioning makes a limited, but crucial, contribution to protecting fair equality of opportunity.[7] It leaves considerable room for social relativity about the importance of some treatments compared with others, but much less about what counts as a disease. Still, it is important to keep in mind that the treatment/enhancement distinction does not specify the boundary between obligatory and nonobligatory medical services. Some obligations derive from considerations beyond the primary rationale, and the primary rationale includes a respect for reasonable resource constraints.

Hard Cases and Expansive Views of Medical Obligations

If we return to the hard cases we considered in the last section, we see that we are pulled away from the primary rationale in two directions. In both the growth hormone (Johnny and Billy) and Shy Bipolar cases, we were concerned that people who suffered equal disadvantage from being short or shy were not being treated the same way. If the therapy could be counted as treating a disease (growth hormone deficiency in Johnny's case, or the effects of bipolar condition on personality development), then reimbursement for the therapy was assured. But if the therapy counted as enhancement of an otherwise normal condition (shortness for Billy, or shyness for an otherwise normal person), then reimbursement was not available.

According to the fair equality of opportunity account of justice and health care (that is, the "social structural" view), we are concerned about disease and impairment because of their impact on opportunity. But here Billy suffers as much loss of opportunity as Johnny; so too does the Normal Shy person as compared to the Shy Bipolar. We must face this crucial question: Why should we put so much emphasis on the distinction between treatment and enhancement rather than calculate the effect on opportunity directly? The pull here is to equalize capabilities, or at least to reduce disadvantages that result from less-than-equal capabilities, regardless of the role of disease or impairment. This pull thus concretely reflects the instability noted in Chapter Three regarding positions that pursue only partially the level playing field interpretation of equal opportunity. The instability seems to imply that we should either agressively and completely pursue the level playing field strategy by moving toward a brute luck or equal resources view, abandoning such partial approximations as the treatment/enhancement distinction, or we should abandon the distinction altogether.

The other pull is visible in the case of the Unhappy Husband who is clearly unhappy and finds therapy effective. Moreover, he is not unhappy because he has cultivated exotic tastes-- indeed, we think of his compunctions about abandoning his wife as admirable (although we do not ordinarily think we should compensate people for any sacrifices they make in order to meet their moral commitments). He has not in any obvious sense "chosen" to adopt idiosyncratic or selfish or self-serving attitudes or preferences that are the source of his unhappiness. (One qualification is needed here: His religious or moral compunctions about divorce and abandonment of his sick wife may be the result of early childrearing, but we think of these as values that are reaffirmed--and thus "chosen" when we are adults.) Why should such effective therapy not be reimburseable? Some clinicians--call them "expansive" clinicians--think that such a person should be eligible for reimbursed treatment. They think that health care should have the goal of removing sources of unhappiness from which we suffer through no fault or choice of our own. Health care should give people an opportunity to be happy that is just like the opportunity enjoyed by others.

The three positions described here--the primary rationale and the two pulls from other directions--reflect three distinct interpretations of our egalitarian concerns that have been developed in the philosophical literature. More specifically, each can be read as a different gloss on the scope of our concerns about equality of opportunity. If we think that concerns about equal opportunity should guide the design of a health care system, then each leads us to a different view of our obligations to assist others through health care services, including genetic interventions. The primary rationale retains an appeal to the treatment/enhancement distinction, but the other two positions abandon it. Each needs further explanation, and ultimately we shall have to decide whether the alternative positions give us sufficient reason to abandon or modify the primary rationale.

Three Philosophical Models of The Relationship Between Equal Opportunity and the Goals of Health Care

The Normal Function Model

The primary rationale appealed to by Daniels rests on what we shall call a "normal functioning" interpretation of the requirements of what was referred to in Chapter Three as the social structural view, which Rawls calls fair equality of opportunity (Rawls 1971).

Historically, our understanding of the requirements of equality of opportunity has evolved, as noted in Chapter Three. We have come to oppose allowing certain human traits--originally race, religion, ethnicity, and class origin; later gender, age, disability, and sexual orientation--to serve as the basis for assigning people to jobs or offices. Rather, these traits are seen as "morally irrelevant," and we believe people should be judged by their capabilities to perform in jobs, offices, or educational settings. (This is the "nondiscrimination" conception of equal opportunity described in Chapter Three).

But we also recognize that past unjust social practices may have distorted the development of peoples' talents and skills, and that we may have to compensate people for the effects of those practices--for example, by special programs aimed at correcting for those distorting effects. Otherwise, expecting people to compete on the basis of the talents and skills they actually have may be expecting them to compete on an unfair basis. This was the idea behind the federal program Operation Head Start, one of the most successful compensatory education programs attempted in the United States, and, more generally, it is the idea behind universal public education.

Whereas the nondiscrimination view of equal opportunity simply requires us to eliminate reference to morally irrelevant traits in determining access to desirable social positions, "fair" equality of opportunity (the "social structural" version of the level playing field view) requires us to correct where we can for the mis- or underdevelopment of talents and skills--that is, for the effects of the social lottery, at least where these are influenced by past injustices.

The normal functioning interpretation of the requirements of equality of opportunity--whether in Rawls's (1971, 1993) general account of distributive justice or in Daniels's (1985) extension of it to health care--assumes that there is a background inequality in the distribution of capabilities, an effect of the natural lottery. Equal opportunity then does not require assuring truly equal chances of success, which could result only if we eliminated this inequality in the distribution of capabilities. The levelling of the playing field goes only so far, on the normal functioning view.

Applied to health care and construed as an account of the goals of health care, the normal functioning interpretation of the fair equality of opportunity principle thus ascribes to health care the relatively modest and limited task of keeping people functioning as close to normally as possible. In effect, health care, like compensatory education programs, aims to produce "normal competitors." These are not equal competitors; quite generally, some will suffer the disadvantage that comes from some normal but not optimal or even average capabilities. An unequal distribution of capabilities is left intact, once the distorting effects of past social practices and treatable disease and impairment are addressed.

Why should we take the natural distribution of talents and skills as a baseline, as the normal functioning view does? If we can redistribute or modify some of those disadvantaging talents and skills, why not alter the baseline? This was the key question raised in Chapter Three and we see it echoed in the pull of the hard cases. If medical technology--whether genetic or not--allows us to redefine the baseline and produce more equal chances of success, are not we committed to doing so wherever we can if we believe in levelling the playing field?

In Chapter Three we noted some reasons for levelling the field more completely. Here we develop further some reasons for abiding by the normal functioning (or "social structural") view. As noted earlier, these views are farther apart in theory than in practice, but for the moment we want to clarify further the theoretical issues.

A fundamental point to note is that our egalitarian concerns in general, and our concerns about equal opportunity in particular, form only part of our concerns about what justice requires. A theory of justice in general, or of justice for health care in particular, must combine concerns about equality with those about liberty. And both these must be reconciled with considerations about efficiency and the allocation of resources. Even if the fundamental intuition underlying our concerns about equality of opportunity pushed us toward thinking that we were obliged to take some steps to redistribute capabilities such as talents and skills, rather than treating their "natural" distribution as a baseline, we must reconcile the pull of that concern with conflicting goals regarding liberty and efficiency. The presumption in favor of modifying the baseline, coming from the egalitarian pull of a concern about equality in opportunity, can be defeated by other key components of our concerns about justice. If, for example, we were only concerned about equality of opportunity, we might be indifferent between "levelling down" by reducing the greater capabilities of the better off and "levelling up" by improving the capabilities of the worst off. But considerations of both liberty and efficiency rule out levelling down.

In Rawls's (1971) theory, this reconciliation of equality with liberty and efficiency takes place through the choice of principles that deliberators would make in the Original Position--Rawls's social contract situation. Suppose that contractors knew that it was sometimes possible to alter the natural distribution of talents and skills and that doing so might make it possible to better promote equality of opportunity. They would still have to solve the more general problem of distributive justice posed by the fact that, in general, some unequal distribution of talents and skills can and must be taken as a baseline.

For example, Rawls is quite explicit that environmental factors, including culture, family influence, and individual responsiveness to educational and compensatory educational measures, will unavoidably lead to some ineliminable inequality in the distribution of talents and skills. In the general case, it may be better for deliberators, even for those who anticipate they may turn out to be worse off with regard to marketable talents and skills, to mitigate the effects of inequalities by redistributions of other important goods than to insist on what may turn out to be a highly inefficient "equalizing" of the distribution of natural talents and skills (or even a more modest elimination of obviously disadvantaging traits). And this general case is what yields Rawls's integration of efficiency and equality concerns through a division of labor between principles. The Difference Principle, which allows inequalities only if they maximally benefit the worst off, provides maximal mitigation of the consequences of the natural lottery, at least in the general case. Insisting quite generally on more direct and radical forms of seeking equality through modifying traits might in fact make those seeking modification worse off than they would otherwise be. Rawls assumes that deliberators in his Original Position would make just such a reconciliation of competing concerns, requiring that the system as a whole can be made to work to the advantage even of those worst off with regard to marketable talents and skills. In specific instances, however, where we can identify a highly efficient intervention to eliminate traits that are obviously disadvantaging, then there is nothing in the the theory to block this course of action, and much to support it, especially since it is may be clearly superior in all ways to mere mitigation.[8]

Interpreted in this way, there is no reason to think that Rawls's account, or Daniels's extension of it to health care, would rule out sometimes being obliged to use genetic technologies to alter the distribution of talents and skills. The Rawlsian reconciliation between equality and efficiency, even if it captures the general case, will not justify treating the natural baseline as if it is an uncrossable boundary. In some special cases, where there might be a particularly good likelihood of inexpensively and safely reducing clear disadvantages, there may well be adequate reason to go beyond our standard notion of equal opportunity to permit some enhancement in the name of eliminating obvious disadvantages.

Even with such case-by-case exceptions, however, the normal functioning conception of equal opportunity retains a central feature of the conception of equal opportunity that is part of our public culture--namely, that not all unchosen competitive disadvantages are unfair and require elimination or compensation. As standardly conceived, we do not think competition is automatically unfair--or that opportunities are automatically unequal--just because some people are naturally endowed with talents or skills that others cannot match. The normal functioning view of equal opportunity, even with the exceptional extensions allowed here, retains the core idea that has longstanding in our democratic culture--namely, that a competition for jobs or offices is fair if it tests people for their possession of the relevant capabilities, provided society has not unfairly distorted them. (We return to this point about democratic culture in a later section of this chapter.)

This point about fair competition draws support from other contexts as well. In athletic competitions, we may sometimes set up categories, such as weight ranges in boxing or wrestling or gender classifications in many sports, to recognize some obvious sources of biological difference that might make competition unfair. We recognize, however, that some people are naturally faster or stronger in ways that no training regimen can correct for, and we do not view such competition as intrinsically unfair. In any case, what is made unlikely by Rawls's account of fair equality of opportunity or by Daniels's normal functioning account is that an individual who finds himself or herself at a competitive disadvantage because of an otherwise normal capability thereby has a prima facie claim, based on equality of opportunity, to assistance to rectify that (perceived) deficit.[9] In what follows, we try to clarify this point by examining two alternatives to the normal functioning view. In Chapter Three, we considered a much more demanding and expansive interpretation of the requirement of equality of opportunity--at least in theory--than the normal functioning account. To explore that approach more thorougly, we examine two variations that appear in the philosophical literature, both of which abandon the treatment/enhancement distinction. The "equal opportunity for welfare" view explicitly appeals to a principle similar to the "brute luck" view, calling for compensation whenever a disadvantage or loss of opportunity for happiness comes about through no individual fault. The "equal capabilities view" actually appeals to a different philosophical idea, the notion of positive freedom--that is, the freedom to do or be whatever we want. We begin with the equal capabilities view.

The Equal Capabilities Model

In a series of articles and books, Sen (1980, 1990, 1992) argues that the object of our egalitarian concerns is equality in what we can do or be.[10] Think of all the things we can do or be as forming a complex set of our actual capabilities. (Sen talks of an "n-tuple of doings and beings" as forming our "capability sets.") To the extent that we are concerned about equality, we should think not simply about the resources we have available to us, but our ability to transform those resources into actual capabilities, since variations in individuals will lead to different capabilities for the same inputs of resources. (On Sen's view, as on Rawls's, we are also not concerned about producing experiential states, like happiness or pleasure.) Sen views this account as an explication of the idea of "positive freedom," but it is not implausible to view it as an account of equality of opportunity. We have equality of opportunity when our capability sets are equal. (If capabilities are not equal because some people have developed certain ones rather than others they could have developed, we may ignore that inequality; we will simply talk about "capabilities" in what follows.[11])

If we could simply translate this claim about equality of opportunity into an account of the goals of health care--which we shall argue we cannot--then those goals would be dramatically expanded from the limits set by the standard rationale discussed earlier. Health care would now have as its task the goal of reconstructing people in ways that make their capabilities more equal. With this expansion, there would simply be no point in drawing a line between treatments and enhancements. Society would be just as much obliged to enhance the capabilities of those who function normally but with less-than-equal capabilities as to keep people functioning as close to normally as possible. (One perhaps disturbing wrinkle in this approach is that we may not be under an obligation to correct for the effects of disease and capability where they work to increase equality in capabilities--for example, by diminishing the capabilities of those endowed with superior capabilities.)

To see why we cannot so directly use the equal capabilities approach to expand the goals of health care, we must examine Sen's account more carefully. A crucial point is one already made: a theory of justice requires integrating concerns for equality with concerns for liberty and efficiency. Sen's claim that our egalitarian concerns really focus on capabilities, not resources or experiential states, or the claim that we explicate the goal of equal opportunity as the goal of achieving equality in capability sets, does not tell us what justice requires us to do. Sen (1992) comments that we must reconcile our concerns about efficiency with our concerns about equality in order to find out what justice requires. In other words, justice might not require us to pursue equality of capabilities so directly after all, if the cost in terms of efficiency were too great. We might instead be required to permit some inequalities and act to mitigate their effects. Sen does not discuss how this reconciliation of justice and efficiency would take place in a theory of justice.

Rawls, as we have argued, does attempt a particular reconciliation: we should take some natural distribution of talents and skills as a baseline (generally, this is the normal distribution). Although those with the least marketable capabilities will have lower prospects in life than those with more marketable capabilities, Rawls mitigates the effects of this basic, residual inequality by requiring that inequalities in primary social goods like wealth and income be constrained so that they work to the advantage of those with the worst prospects in life. In this way, those with more marketable talents and skills must harness their advantages to maximizing the prospects of those who are worst off with regard to talents and skills. In effect, Rawls divides responsibility for meeting our egalitarian concerns between two different principles of justice. The principle governing equality of opportunity leaves the normal distribution of capabilities in place, but the principle governing overall inequality in prospects in life mitigates the effects of doing so.

If health care should be governed by a principle governing fair equality of opportunity, then, for the general case, it too may leave the normal distribution of capabilities in place, concerning itself only with keeping people functioning as close to normally as possible. It would then be necessary to rely on other principles of justice governing inequalities in life prospects to mitigate the effects of this pragmatic decision. The fact that Rawls appeals to the "moral arbitrariness" of the natural lottery for capabilities does not mean that the only reasonable way to address this problem within a comprehensive theory is to devote extensive resources toward equalizing capabilities. It may be that we can do better, even by those worst off with regard to capabilities, by leaving the distribution of capabilities mostly in place and mitigating the effects of doing so in other ways--at least that is his rationale.

This dramatic restriction on the scope of what we can pursue in the name of equal opportunity--as compared to Sen's account--may not be what Sen has in mind, but then we need some clear idea of what restrictions are compatible with the overall demands of justice. (The reader is also reminded that Rawls's rationale for the general case does not preclude a more restrictive, case-by-case assessment of claims that eliminating some obviously disadvantaging [but normal] traits is justifiable on grounds that modestly extend the concept of equal opportunity.)

Even leaving aside the competing claims of liberty, efficiency, and equality, the equal capabilities model does not strictly speaking involve a pursuit of equal capability sets. Sen offers an approach to ranking differences in capabilities. How important a particular capability is will depend on the system of values--the plan of life or conception of the good--adopted by an individual. John may rank capability set A as better than capability set B, but Jane may make the opposite judgment from her conception of the good. We may find some cases in which all can agree that set C is worse than sets A and B, but we may get no rankings for a broad range of capability sets.

In fact, we are most likely to find that the clear-cut cases in which a set (say C) is ranked lower than others will be cases in which there is a significant departure from normal functioning--a disease or impairment that has a significant impact on capabilities and thus on opportunities. In those cases, Sen's account will tend to agree with the standard model for thinking about our obligations to assist others with medical interventions, including genetic ones. But for a broad range of differences in capability sets, there may be "incommensurability" in the sense that these sets are ranked differently by people with different conceptions of what is good in life. Because of this, our commitment to pursuing equality of capabilities is seriously limited, since for a substantial range of capabilities, there is no common basis for determining what increase the equality of those sets.

There is a biological point that bears on this incommensurability and on the treatment/enhancement distinction more generally.[12] Consider the accompanying figure, which reports on the results of an experiment testing how different genotypes from natural populations survive in a variety of environments that vary in temperature.

<