On the very idea of genetic justice: why Farrelly’s pluralistic prioritarianism cannot tackle genetic complexity1 Short title: “On the very idea of genetic justice” Author: Michele Loi Mailing address: Università Vita-Salute San Raffaele, Facoltà di Filosofia Via Olgettina 58, I-20132 Milano (Italy) Email: loi.michele@hsr.it Office phone: (+39) 0226435871 Mobile: (+39) 3200273272 Contributor’s profile: Michele Loi, Ph.D. (Political Theory), is postdoctoral research fellow at CeSEP, Centro Studi di Etica Pubblica, at Università Vita-Salute San Raffaele, Italy. He is working on a book on distributive justice and genetic technology. Fundings This research has benefited from funding by FIRB – Founds for Investment in Basic Research – by the Italian Ministry of Scientific and Technologic Research. 1 The author wishes to thank Oliver Feeney, Darryl Gunson, Massimo Reichlin, Allen Buchanan, Vilhjálmur Árnason, and Shlomi Segall for useful suggestions on previous drafts of this paper. The author wishes to thank all visiting scientists and personnel of the Fondation Brocher for intellectual and material support when researching on the topic of this paper. 1 TEXT Innovations in science and technology are often the source of public concern, but few have generated debates as intense and at the same time with such a popular fascination as those surrounding genetic technologies. Unequal access to pre-implantation diagnosis could give some individuals the opportunity to select children with more advantageous predispositions.1 Genetic therapies and enhancements (improvements of healthy characteristics) involving changes in germline cells (egg or sperm cells, or their precursors) could lead to inheritable genetic advantage, and then, to an accumulation of natural advantage along lineages.2 Unequal access to genetic technology could contribute to an increase in existing social inequalities, by intensifying congenital inequalities, which would lead to significant competitive disadvantages for the socially and economically disadvantaged members of society. However, more inequality is not necessarily unjust. As many contemporary egalitarians are happy to admit, some inequalities might not be objectionable if, through trickle down or incentive effects, they allow further improvements of the conditions of the worst off. This general consideration is relevant because it is misleading to frame access to therapies and enhancements as a zero-sum game, in which one person’s win is another’s loss.3 That would be the case if the only goods produced by cooperation were positional: goods like being tall, whose value increases the higher one is in the hierarchy of people having access to that good.4 On the contrary both therapies and enhancements, like education, contribute to creating resources which governments could redistribute.5 To judge whether an unequal distribution of a good is unjust, one must know where the inequality comes from. For instance: is it produced by a scheme of cooperation which is in the long-term benefits of the least advantaged, or by a scheme with a tendency to worsen the expectations of the least favored group? An additional complication is that different conceptions of justice measure injustice by focusing on different types of inequalities. When justice is said to obtain in virtue of how the good 2 X is distributed, X may be called the currency or distribuenda of the conception of justice in question. It is useful to distinguish theories in terms of the currency in which social institutions operate. John Rawls’s theory of justice as fairness, for example, evaluates justice on the basis of how social institutions distribute social primary goods. These are, by definition, all-purpose instrumental goods whose distribution is directly controlled by social institutions.6 (Revisions of this formula in more recent formulations of the idea of justice as fairness change nothing for our discussion).7 Natural primary goods are also, by definition, general purpose goods; one might think that health, vigor, intelligence and imagination are such goods. They are not regarded to be the currency of distributive justice, because, according to Rawls, they are only influenced, not directly controlled, by them.8 Rawls’s choice of social primary goods as the sole currency of justice has been criticized for several reasons. It has been argued, for instance, that it does not directly justify special compensations for disabled and chronically ill individuals. If one considers natural primary goods, welfare9, or Sen’s capabilities10, instead, one can justify them straightforwardly, on the basis of the observation that most disabled individuals obtain fewer primary goods than most healthy people, or that they are less able to convert external resources into welfare or capabilities . For that reason, some philosophers claim that conceiving justice as a function of the distribution of social primary goods currency cannot account for widespread moral convictions, such as the alleged one concerning special compensation for disabilities.11 Other philosophers think that justice involves the rectification of inequalities that are due to ‘brute luck’, that is to say, to factors individuals cannot be held responsible for.12 They are suspicious of the distinction between natural and social primary goods, because it is orthogonal to the distinction between brute luck and responsibility. A person’s initial social circumstances, for instance, are socially produced and are certainly a matter of brute luck for the person who benefits from them. Health is perhaps describable as a natural (biological) good, but bad health can result from either brute luck, for instance genetic predispositions, of from choices analogous to deliberate gambles, for instance winter sports, or both. 3 The debate about access to gene-therapy and genetic screening suggests a different criticism of the social primary good currency, namely that the boundary between what counts as natural and what counts as social is blurred.13 This critique is made more vivid by recent advances in biotechnology which suggest that one might be able, in the future, to alter the distribution of initial natural characteristics to an unprecedented degree. Among the philosophers who have argued in this guise, Colin Farrelly’s argument deserves special attention, since he is among the few to draw a precise normative implication from it. He has argued that: Rawls's general conception of justice stipulates that only the SPG [social primary goods] are to be equally distributed, unless an unequal distribution of any, or all, of these values is to everyone's advantage. But what about the NPG [natural primary goods]? […] [Genetic] technologies will make it possible for the distribution of NPG to be directly (though not totally) under our control. Advances in screening and diagnostic technology, gene therapy, and genetic enhancement mean that the goods to be distributed by principles of social justice will be radically changed (78).14 As we shall see, Farrelly concludes that access to genetic technology ought to be regulated by a principle governing the distribution of genes relevant to the distribution of natural primary goods: the genetic difference principle.15 This argument appears to be independent from the arguments coming from utilitarians, capability theorists and luck egalitarians, cited above, because it does not rely on the same premises and moral intuitions.16 This paper has the following structure: in section one I explain why Farrelly correctly avoids turning justice into a function of the actual distribution of natural primary goods. In the second section, I criticize the lax genetic difference principle, which turns justice into a function of the distribution of genetic endowments (which affect expectations of natural primary goods). I argue that the second approach provides a poor guide for policy when interactions among genetic and environmental factors are significantly complex. 4 1. Natural primary goods Rawls rejects the inclusion of natural primary goods in the currency of justice by arguing that these are not controlled directly by social institutions. Does the possibility of controlling genes make any difference? To answer this question, one must specify what direct control involves and why it is morally relevant. To clarify matters, it is useful to introduce the classical biological distinction between genotype and phenotype. Following Lewontin, we can define the genotype as the inherited set of DNA molecules present in the fertilized egg and the phenotype as that “set of traits that characterizes the whole organism at each stage of its life” which develops in virtue of “complex […] processes that occur not in a vacuum, but in an impinging world of environmental circumstances and developmental accidents” (165).17 Consider the natural primary goods listed above. They are phenotypical traits and, as such, influenced by both genetic and environmental circumstances. The distribution of health, vigor, intelligence, and imagination, is not directly controlled, only influenced, by the initial genetic distribution. Even if individuals begin life with identical genomes, they will be influenced by different developmental accidents. Even if a population consists of identical twins, they will sooner or later become free to choose (to some degree) the environmental influences affecting their development. The problem with controlling distributive shares of health, vigor, intelligence and imagination, directly, is that, in order to realize a specific pattern of distribution, the state must engage in continuous redistribution, thus interfering with people’s lives in an especially burdensome way. Developmental accidents and personal choices cause a departure from the distributive target specified by the principle in question. For instance, if the principle prescribes equality of health, vigor, intelligence and imagination, inequalities due to accidents or individual choices must be corrected or compensated for, in order to reestablish equality. It seems more 5 difficult and morally problematic for the state to control natural, rather than social primary goods, in this way, as I shall explain shortly. To achieve a target distribution of wealth and income, the state taxes individual assets; if taxes are not delivered, it may expropriate some properties. These assets are things, features of the environment, external to the person. On the contrary, qualities like health, intelligence, vigor and imagination are essentially connected to physical and mental properties intrinsic to the person. Controlling (rather than influencing) a person’s assets of these goods might require controlling amounts of physical exercise, diet and rest, or even modifying the body through biological interventions, in the most challenging cases. Few would accept extending the authority of the state over the individual in this way. The point is that, in contrast to wealth, health, vigor, intelligence and imagination cannot be transferred from those who have more to those who have less of them, without violating the right to physical and mental integrity of the person, by which I mean the (moral) right of self-ownership of mind and body, one of the basic liberties recognized by any state calling itself “liberal”.18 These goods are in that sense, what economists call “indivisible goods’. These assets cannot be divided, but their distribution can be affected: taxes on cigarettes and subsidies on sport can affect a person’s tendency to build assets of such goods. Legitimate social institutions can influence the distribution of these assets but not control them. If justice, on the contrary, required realizing patterns of natural primary goods, it would require direct control over the distribution of such goods, leading to excessively intrusive governmental policies. This leads to moving from distributing natural primary goods such as health or intelligence, to distributing opportunities for these goods; that is to say, ensuring that the playing field be appropriately leveled in relation to chances of obtaining such goods, by setting up institutions accordingly. Many social goods, in particular educational opportunities and access to health care, contribute to determine a person’s lifetime opportunities not only for social, but also for natural primary goods, like health and intelligence. It follows that social opportunities for natural primary 6 goods – social factors affecting the expectations of natural primary goods – can be controlled (modified, redistributed, etc) by controlling the distribution of external goods, like wealth and legal entitlements. This strategy avoids the problem of personal integrity. Are there natural goods which, like wealth and unlike intelligence, can be directly controlled (e.g. added, reduced, transferred) without violating personal integrity? Genetic endowments may satisfy this description in certain conditions. If a very young embryo in a test tube is not regarded as a person, intervening on the genes of this embryo (or selecting a different embryo for implantation) cannot be regarded as a violation of that future’s person personal integrity. Let us concede this possibility, for the sake of argument. This reasoning leads to and justifies Farrelly’s approach. He explicitly acknowledges Rawls’s point that society cannot control the distribution of natural primary goods directly, and concludes that institutions should seek to “fairly distribute those genes that are most important for the [natural primary goods]”(80).19 In the next section, I shall argue that this – more promising – strategy is unsatisfactory, too. 2. Genes Farrelly’s “genetic difference principle”, henceforth (GDP) states that inequalities in the distribution of genes important to the NPG [natural primary goods] are to be arranged so that they are to the greatest benefit of the least advantaged (81).20 To apply this principle, one has to be able to identify the worst off group and compare different distributions of genetic endowments in terms of the expected benefit of its members. A just distribution of genes is one (rationally) expected to maximize the natural primary goods holdings of those with lesser primary goods. The principle does not assume genetic determinism: genes do not determine individual levels of health, vigour, intelligence and imagination, but affect the probabilities that specific levels be reached. Farrelly’s view is not egalitarian, but prioritarian in inspiration. Prioritarianism is characterized by Farrelly as the idea that “it is morally more important to benefit the people who are worse off”(48).21 7 It is more instructive (and intuitive) to imagine the principle applied to single-gene traits. Farrelly points out that: there are over 6000 known single-gene disorders, which occur in 1 out of every 200 births. A single-gene disease is a disease caused by a single malfunctioning allele. Such disease typically can develop in practically all usual environments (47).22 Suppose that we are dealing with a genetic disease for which no easy and inexpensive environmental fix exist. Huntington’s disease, for instance, leads to severe cognitive impairment and paralysis after age 40. The mutation which is responsible for it obviously reduces life-time expectations of natural primary goods. In a world in which no single-gene condition worse than Huntington’s disease exists, the GDP would prescribe adopting institutions that increase the chance that individuals with Huntington’s disease will obtain gene-therapy. Yet even in this hypothetical scenario, the principle is problematic because it would justify channelling too many social resources into genetic technology. For, genetic inequalities are not the only inequalities which deserve attention from the point of view of justice. As Farrelly points out: many of those who are worse off in terms of their expected lifetime acquisition of natural primary good are worse off, not because of their genetic endowments, but because they lack access to adequate housing, basic healthcare, or long-term economic security (50).23 Farrelly imagines a government which must choose between improving public education and investing in public genetic enhancements. He imagines a case in which […] if the government had pursued improving education they would have improved the life prospects of many other people who are not genetically disadvantaged. […] Even those who have both favourable genes and social environment could benefit from quality public education (W26).24 As Farrelly acknowledges, the stringent interpretation of the genetic difference principle leads to unreasonable policies, based on the implausible idea that “a small gain to the genetically 8 disadvantaged is more important than a great loss to both the socially disadvantaged and [the overall] advantaged” (W26).25 This leads to what Farrelly calls the “problem of weight”. It is not a specific problem for prioritarian theories of genetic justice. It is equally a problem for egalitarian theories, and, indeed, for any conception of justice where the distribution of genes is regarded as one of justice’s distinctive goals. In the context of egalitarianism, it can be defined as the problem of: […] balancing the desire for achieving genetic equality with the desire for achieving other kinds of equality (e.g., wealth and income) and other values (e.g., utility and freedom) (W24).26 Farrelly claims to endorse a form of pluralistic prioritarianism. This is, as defined by Farrelly himself, a kind of prioritarianism which takes into due account the two complications cited above: the importance of values other than justice and the existence of non-genetic forms of advantage.27 Farrelly claims that the problem of weigh can be solved by moving away from the strict interpretation of the GDP. He endorses the lax genetic difference principle, which says that when selecting institutions affecting the distribution of genes, society should aim to ensure the “greatest reasonable benefit to the least advantaged”(W26).28. The qualification “reasonable”, “permits a number of diverse considerations” – such as general utility and freedom – “to come to the fore”(W26).29 Moreover, pluralistic prioritarianism must take into account the so-called problem of genetic complexity. Farrelly notices that […] The most prevalent diseases – like heart disease, cancer, and diabetes – are not caused by a single malfunctioning allele. These more common diseases are multifactorial diseases. Their development depends on a variety of factors beyond our genetic constitutions. Environmental factors like diet and lifestyle (e.g. exercise, smoking, stress levels, etc.) often play a more important role in determining our risk of 9 developing multifactorial diseases. […] Our genes do play a role [...] But environmental factors play a much greater role (47).30 Farrelly claims that “to track genetic complexity seriously, justice theorist ought to embrace pluralistic prioritarianism”(48).31 Yet Farrelly’s “pluralistic prioritarianism” has another feature Farrelly does not elaborate on: it is conceived as the result of an intuitive weighing of different particularistic claims of justice. To understand this “particularistic” aspect of pluralistic prioritarianism, it is useful to contrast it with an instance of what I shall label “monistic” prioritarianism: Arneson’s “responsibility catering” welfare-prioritarianism.32 Like pluralistic prioritarianism, monistic prioritarianism “is egalitarian in tilting in favor of those who are badly off”(343).33 But in contrast to the pluralistic version “priority is assigned to aiding an individual in virtue of how badly his life is going, as measured by an objective scale of well-being”(343).34 Wellbeing here provides the currency of justice, in the sense that the moral weight of removing genetic or environmental forms of deprivation is established by converting all the relevant factors into a unified welfare metrics. Suppose that the well-being of people is affected only by the distribution of four goods: germ-line gene therapy, health-care, access to education, and wealth. While each factor adds something to well-being, there is no need to invoke distinct principles for dealing with each factor. Rather, the weight and importance of each depends on its contribution to distributing wellbeing. It is not necessary to calculate separately which educational arrangement would maximize well-being for the worst off, etc. Rather,: entire social assets are evaluated to see which combination of genetic modification, health-care, access to education and wealth is most likely to benefit the worst off.35 By contrast, Farrelly’s prioritarianism includes (explicitly) a principle of genetic justice and (implicitly) norms of “non-genetic” justice; that is to say, different norms exert competing demands on the social system, with no pre-established order of priority.36 No general formula is provided for weighing them, while it is recognized that, for philosophy to be relevant for policy, it must achieve 1 0 some weighing.37 This analysis is supported by analyzing three features of Farrelly’s approach: first, the claim that we need: an account of genetic justice, that is, an account of what constitutes a fair distribution of genetic endowments that influence our expected lifetime acquisition of natural primary goods (46).38 Second, notice that according to both the strict and the lax version of the genetic difference principle, genetic justice can be realized without modifying the distribution of other goods: the genetic difference principle never requires distributing goods beside genetic endowments. Third, one finds the idea of overall justice as the outcome of a reasonable weighing of heterogeneous moral claims: justice all things considered results from finding the appropriate balancing between particularistic claims of justice: claims of justice involving different particular goods. If that is true, it makes sense to distinguish all-things-considered justice, which is what follows from such weighing, and genetic justice in the strict sense, which is what the genetic difference principle would require, if it didn’t have to be weighed and compromised against equally legitimate moral claims of justice and other values. When Farrelly claims that we need an account of genetic justice, I take him to mean that we need an account of what a fair distribution of genetic endowments would have to be apart from the requirements of other particular aspects of justice. For otherwise, what justifies invoking the genetic difference principle, a principle for dealing specifically with genetic endowments? It might be objected that, in section 1, I have already provided the answer to the question I’m asking. I argued that principles dealing with the distribution of natural primary goods, directly, involve a violation of personal integrity. This would be correct if the only available options were a) distributing actual natural primary goods; b) distributing genetic endowments “that influence our expected lifetime acquisition of natural primary goods”(46).39 There are, however, many other options, for instance: (c) distributing opportunities, by modifying social factors affecting (statistically definable) expectations of natural primary goods. This is not the same as (b), because 1 1 the distribution of opportunities is affected not only by genetic endowments, but also by a person’s initial position in the social system. For instance, in most existing societies, parental socioeconomic status is a strong predictor of a person’s future educational level and income;40 which are, in turn, strong predictors of health outcomes.41 In the light of this analysis, one could object that Farrelly provides no justification for the claim that we need an account of genetic justice understood in this way. What, in his theory, explains the special moral importance attached to genetic endowments? What justifies appealing to a prioritarian principle whose scope is so specific? I shall not press such questions further. Rather, I shall point out that, when there are significant interactions between genetic and environmental factors which can be changed politically, Farrelly’s particularistic approach fails to justify searching for better policies (better in prioritarian perspective), thus contradicting its own prioritarian rationale. Let us consider the distribution of the genes responsible for predispositions of diseases which respond to both environmental and genetic factors. One might be convinced on the basis of epidemiological evidence that “the major preventable environmental causes of illness and death are tobacco use, unhealthy diet, physical inactivity, excess alcohol use, infections, trauma, and exposure to environmental toxins”(600).42 Merikangas and Risch conclude that, for multi-factorial disorders amenable to environmental modifications Gene hunting for disorders that appear to be highly amenable to environmental modification, such as type 2 diabetes, AIDS, alcohol dependence, and nicotine dependence, would have lower [public health] priority, even though genes may be involved in their etiology (601).43 Khoury et al object that this way of thinking “could perpetuate the false competition between nature and nurture”(800).44 Their point is not merely speculative. Some known environmental causes of disease, like smoking and drugs, also cause addiction, which affects the success rate of interventions at the population level, like information campaigns. Addiction is, in turn, influenced 1 2 by genetic predispositions, which could perhaps be better dealt with through advances in pharmacologic and behavioural interventions, derived from genomics research.45 A similar critique applies to Farrelly’s approach. His concrete proposal involves: Defin[ing] the category of the genetically “least advantaged” as those individuals whose genetic constitutions place them below half of the median for the expected lifetime acquisition of natural primary goods (49).46 On the basis of this criterion, people with a genetic predisposition which makes it especially difficult to overcome drug addiction could never be among the genetically worst off in the following hypothetical scenario. Let us suppose that the gene is spread uniformly across different social groups, so that 80% of the people with that genotype have, for social reasons, little chances of developing an addiction to begin with. However, 20% of the population with the predisposition in question belongs to a socially disadvantaged group with high chances of drug abuse. Among the members of this social group, those with the unfavorable genetic predisposition could fall below half of the median for the expected lifetime acquisition of natural primary goods within that (already disadvantaged) group. People from unfavorable social backgrounds with a genetic predisposition to hold on to their addictions could represent, within society, the “socio-genetic” group with the lowest expectations of natural primary goods. (Even if having the genetic predisposition in question, in and by itself, does not place the individual below half of the median for the expected lifetime acquisition of natural primary goods.) So if it were the case that genetherapy could remove the genetic cause of disadvantage among those from disadvantageous social background, by increasing their chances of quitting addiction, committed prioritarians should perhaps favor the policy in question. Notice that this is a policy addressing the distribution of genetic endowments, yet Farrelly’s principle cannot be used to justify it. I shall conclude the argument against Farrelly’s particularistic idea of genetic justice by arguing that the lax genetic difference principle fails to deal adequately with cases where 1 3 interactions between genes and environment involve what have been called “interactive predispositions”. These are cases in which, by definition: The presence of a genetic difference between various groups both increases and decreases the probability of individuals from one group, in comparison to individuals from the other group(s), developing a particular phenotypic trait depending on the environmental conditions experienced (35).47 What is in question here is a particular type of interaction between genes and environments (G x E interaction) in which features of the environment not only affect the probability for individuals with a certain genotype to develop the phenotype in question, but involve a change of rank: a case where different genetic groups respond differently to the same array of environments and that difference in phenotypic response is so extreme that the higher ranking group in one environment becomes the lower ranking group in a different environment (32).48 Again, this is not a mere theoretical possibility, but it has already been found in a genetic predisposition to sociopathy,49 as reported by James Tabery: In a widely acclaimed study from 2002, Avshalom Caspi, Terrie Moffitt, and their colleagues found a case of GXE for a gene controlling neuroenzymatic activity (lowvs. high monoamine oxidase A [MAOA] and the development of antisocial personality disorder (ASPD). ASP is the clinical term for sociopaty, defined by a pervasive pattern of disregard for and violation of the rights of others. MAOA is a metabolic enzyme that inactivates neurotransmitters, and deficienzies in the enzyme have been associated with aggression. In the Capsi-Moffitt study, childhood maltreatment consisted in physical and sexual abuse. However, other experiences also counted as childhood maltreatment: harsh discipline, neglect, multiple changes in the primary caregiver, and poor motherchild interactions. […] (28).50 The special ethical and philosophical significance of this case, that Tabery’s article elucidates, is that “the MAOA study by Caspi and Moffitt is […] an instance of GxE resulting in a change of 1 4 rank”(32).51 In other words, one allele (version of the gene) correlates to better outcomes if degree of childhood maltreatment is high; the other correlates to better outcomes if degree of childhood maltreatment is low. Cases of GxE interaction involving a change of rank create a special problem for Farrelly’s approach, as I shall show by discussing a hypothetical example. In order to simplify the analysis of the case at hand, I shall imagine a society made up by genetically identical individuals, except for the two genes P and R, that are roughly equally distributed across the population. The two genes confer an interactive predisposition with an inversion of rank: gene P is a predictor of higher IQ in poor learning environments and gene R is a predictor of higher IQ in rich learning environments. Suppose that parents could buy gene R in an unregulated “genetic supermarket” for their children. It would be rational for benevolent wealthy parents to confer that genetic endowment on their children, because they could safely predict a rich learning environment for their children. Assume, for the sake of argument, that this group of individuals with higher I.Q. contributes little to raising the expectations of natural or social primary good of those who are worst off in social and economic terms. That might be the case if, for instance, a significant proportion of high I.Q. individuals undertakes careers in art, philosophy, and social science which have many intrinsic merits, but cause little, if any, trickle down effects. Let us also assume that the above described market for genetic enhancements contributes to raising the expectations of natural primary goods (intelligence, above all) of people with a favorable starting position in society, while making it increasingly difficult for the unenhanced to compete for social positions which contribute to accumulate assets of such good, such as, let us suppose, positions in art or education. If so, the enhancement market reduces the average expectations of intelligence for citizens from disadvantaged social backgrounds, in absolute terms. By raising inequality in the distribution of natural primary good (intelligence), said market exacerbates the sort of inequalities with which Farrelly’s principle is concerned, without any corresponding benefit for those who are worst off socially and economically. 1 5 Farrelly’s approach has problems dealing with cases like this one. I shall consider two cases: one in which roughly half of the population is from a socially and economically privileged background and obtains a rich educational environment, and one in which only a minority of the population enjoys this privileged condition. In the first scenario, 50% of children enjoy a rich educational environment. In this case, the lax genetic difference principle does not single out P or R as all things considered better genes, in terms of expectations of natural primary goods. For, after all, R is not correlated with better outcomes than P overall, given that 50% of the R-gene population grows up in disadvantaged scenario, where R is a predictor of I.Q. disadvantage. It follows that citizens with the R gene cannot be classified as the genetically advantaged and that the reverse is also true. While the genetic supermarket affects the distribution of expectations of natural primary goods across differently advantaged social classes, it does not affect the distribution expectations across differently advantaged genetic classes, for in this scenario all genetic classes are equally well off. Hence it could not be prohibited by invoking the lax genetic difference principle. Nor can this principle be invoked to justify a tax-subsidy scheme, where taxes on genetic enhancements are used to boost education for the socially least well-off citizens. This follows from at least two considerations: first, the redistributive policy in question cannot be said to improve the expectations of the genetically worst off: socially disadvantaged citizens are not ipso facto worst off genetically. Second, principles of genetic justice are not intended for educational policy: in Farrelly’s pluralistic framework the requirements of genetic justice are meant to be balanced against those of educational justice, which means that they cannot possibly be meant to include them.52 Let us now consider the second scenario, in which only a minority of the population (say 10%) grows up in a rich and stimulating learning environment. There are two ways of analyzing the condition of “being born with the R-gene” in this scenario. One option is the quasi-utilitarian one of regarding the R-gene as the “worst” genetic predisposition, because it correlates with the worst outcome in the most widespread environmental circumstances (those involving poor learning 1 6 conditions). The number of cases in which R correlates with better expectations is significantly lower than the number of cases in which R correlates to worse expectations, because a substantial majority of the population enjoys poor learning environments, in which the P-gene is advantageous. Under this interpretation (objectionable, as we shall see), individuals who have inherited the R gene represent the genetically worst-off group. It turns out that when privileged parents buy genetherapy interventions for their socially privileged children, they maximize the natural primary goods expectations of the genetically worst off group. At the individual level, every wealthy couple who, through germ-line gene therapy, replaces the P gene with the R gene in their child improves the lot of R-gene children as group: thanks to the enhancement market, a larger proportion of children with the R gene will be born in socially and economically advantaged circumstances and have better expectations of I.Q, than otherwise. The market therefore improves the (average) expectations of natural primary goods of the genetically “worst off” group and therefore the genetic difference principle is satisfied. It is remarkable that, in this case, prioritarian genetic justice can be achieved by changing the genes of socially advantaged children, but not by improving the social conditions of the socially disadvantaged children with the R gene. Changing the genes of socially advantaged children from the P to the R allele raises the expectation of individuals who are born in favorable social circumstances and, because in those environmental conditions the R-gene becomes relatively advantageous, in favorable natural circumstances too. The set of individuals whose expectations are thus improved, defined by both genetic and social characteristics, have the highest expectations of natural primary goods, overall. What is questionable, therefore, is that the prioritarian policy justified by the genetic difference principle justifies improving the conditions of a group which is better off in terms of expectations of natural primary goods, under such (objectionable) classification of advantageous and disadvantageous genetic predispositions. It could be objected that it is a mistake, conceptually speaking, to treat an interactive predisposition as an ordinary one. When a genetic predisposition is interactive, neither of the two 1 7 traits should be considered better or worse “as such”, i.e. as the predisposition attached to the worst expectations, or the one to “modify” or “screen away”.53 If this point is granted, the problem with genetic justice is that it fails to address inequalities arising in virtue of combinations of genetic and environmental factors where the two interact with inversions of rank. In the above described case, a market of genetic enhancements leads to a distribution of the R allele, which contributes to an increase in inequality of natural primary goods. Yet, from the standpoint of the genetic difference principle, this inequality neither augments, nor reduces genetic inequality, which is, by definition, an inequality among groups defined by their genetic endowments (not by membership to a social group).54 For all we know, many genetic predispositions could involve inversions of rank. If principles of genetic justice cannot be applied to such cases, what are they useful for? While it promotes a more unequal society in terms of natural primary goods, a free market of R-gene enhancements in the second scenario is not unjust from the point of view of genetic justice. Saying “an unregulated market of genetic-enhancements is a worse policy, for the worst off, than one in which the enhancement market is taxed, and the tax used for improving education for members of socially and economically disadvantaged class” represents a piece of sound moral reasoning which cannot be framed in terms of the “weighing” idea. It can only be applied to the choice whether the worst group in terms of genetic determinants of natural primary goods, or the worst group in terms of social determinants of social primary goods, is to be privileged by policy, on a tight budget constraint. The framework can avoid such counterintuitive consequences only if one claims that interactive predispositions fall outside its intended scope, but (depending on how the biological facts turn out) that might amount to conceding the practical irrelevance of the enterprise. §3 Conclusions. It should be clear at this point, that the problem of genetic complexity challenges an essential feature of Farrelly’s idea of pluralistic prioritarianism: taking genetic endowments as the 1 8 currency of a distinct principle of justice, despite the fact that the value of genetic endowments cannot be assessed independently of environmental circumstances. In this perspective, a redistribution of the social determinants of advantage (education) cannot be considered a requirement of genetic justice. What appears as an intuitive and elegant framework, however, makes no sense in the light of the way in which genes and environments interact. The objection against pluralistic prioritarianism is that it does not require the “ministry of genetic justice” and the “ministry of education” to talk to each other. It is plausible to think that this result can be extended to non-prioritarian frameworks which require balancing the claims of different “particularistic” principle, having to do with genes and other goods. The conclusion is not that a conception of social justice ought to be monistic, that is to say, a function of the realization of a single principle or of the distribution of a single good, such as Arneson’s responsibility catering (welfarist) prioritarianism. Rather, the point is that the cut between genetic goods and other goods makes no sense in the light of the importance of potential interaction between them. Farrelly’s framework of pluralistic prioritarianism is unable to cope with the fact of genetic complexity. One ought to abandon not only Farrelly’s genetic difference principle, but the very idea that regulating access to genomics involves an account of genetic justice, that is, an account of what would constitute a fair distribution of genetic endowments, describable as a distinct ideal, independent from the social bases of inequality. Notes 1. Buchanan AE, Brock DW, Daniels N, Wickler D. From Chance to Choice: Genetics and Justice. Cambridge, UK: Cambridge University Press; 2000:96-97. 2. Mehlman MJ. The law of above averages: leveling the new genetic enhancement playing field. Iowa Law Review 2000 ;85(2):51793. 3. Buchanan A. Enhancement and the ethics of development. Kennedy Institute of Ethics Journal 2008;18(1):1-34 at pp. 7-10. 4. Brighouse H, Swift A. Equality, priority, and positional goods. Ethics 2006;116(3):471-497. 5. See note 3, Buchanan 2008:7-10. 6. Rawls J. A Theory of Justice, 2nd ed. Cambridge, MA: Harvard University Press; 1999: 54-55. The moral importance of generalpurpose means, in contrast to narrow-range instrumental goods, is explained by the fact it would be rational to desire more of them 1 9 whatever one’s ends in life, for a large range of human ends. Thus their goodness is alleged to by explained by a “thin theory of the good” valid for every rational person (see Rawls 1999: 347-350). 7. In Political Liberalism, for example, primary goods are regarded as things citizens need, if they see themselves as free and equal persons. See Rawls J. Political Liberalism, exp. ed. New York: Columbia University Press; 1996:187-190. The inclusion of leisure time and freedom from physical pain as additional primary goods (in Rawls 1996:181-182), is only contemplated as a possibility but never definitively endorsed. 8. See note 6, Rawls 1999:54. 9. Arrow KJ. Some ordinalist-utilitarian notes on Rawls’s theory of justice. Journal of Philosophy 1973;70(9):245–263. 10. Sen A. Equality of what? In: Choice, Wefare and Measurement. Cambridge, MA: Cambridge University Press; 1982:353-369. 11. These problems were central to the debate concerning the “currency of egalitarian justice”. See note 10, Sen 1982; Daniels N. Equality of what: welfare, resources, or capabilities? Philosophy and Phenomenological Research 1990;50:273-296; Arneson RJ. Equality and equality of opportunity for welfare. Philosophical Studies 1989;56(1):77-93. 12. Cohen GA. On the currency of egalitarian justice. Ethics 1989;99(4):906-944. 13. See note 1, Buchanan, Brock, Daniels, Wickler 2000:83. 14. Farrelly C. Genes and social justice: a Rawlsian reply to Moore. Bioethics 2002;16:72-83 at p. 78. 15. See note 14, Farrelly 2002:81; cf. Farrelly C. The Genetic Difference Principle. American Journal of Bioethics 2004;4:W21-W28 at p. W26. 16. Notice that the analogies and relationships between certain genetic characteristics or natural abilities and general-purpose means (primary goods) is used by different philosophers to make different points. See for instance, note 1, Buchanan, Brock, Daniels, Wickler 2000:167-168; Allhoff F. Germ-line genetic enhancement and Rawlsian primary goods. Kennedy Institute of Ethics Journal 2005;15(1):39-56. 17. Lewontin RC. Biological determinism. Tanner Lectures on Human Values 1983;4:147-183 at p. 165. 18. This is the view of John Rawls, who claims that “even if an equal distribution of natural assets seemed more in keeping with the equality of free persons, the question of redistributing these assets (were this conceivable) does not arise, since it is incompatible with the integrity of the person”. See note 7, Rawls 1996:283. 19 . See note 14, Farrelly 2002:80. 20. See note 14, Farrelly 2002:81. 21. Farrelly C. Genetic justice must track genetic complexity. Cambridge Quarterly of Healthcare Ethics 2008;17(01):45-53 at p. 48. 22. See note 15, Farrelly 2004:47. 23. See note 21, Farrelly 2008:50. 24. See note 15, Farrelly 2004:W26. 25. See note 15, Farrelly 2004:W26. 26. See note 15, Farrelly 2004:W24. 27. See note 21, Farrelly 2008:48-50. 28. See note 15, Farrelly 2004:W26. 2 0 29. See note 15, Farrelly 2004:W26. 30. See note 21, Farrelly 2008:47. 31. See note 21, Farrelly 2008:48. 32. Arneson RJ. Luck egalitarianism and prioritarianism. Ethics 2000 Gen;110(2):339-349. 33. See note 32, Arneson 2000:343. 34. See note 32, Arneson 2000:343. 35. Another feature of Arneson's approach is that it attaches a different moral weight to “altering a state of affairs in a way that makes someone better off or worse off” according to a person’s “degree of responsibility [...] for her present condition” (344), hence the qualification of “responsibility sensitive” in its label. See note 32, Arneson 2000. 36 See note 15, Farrelly 2004:W26. 37. Farrelly has stated it very clearly that his ambition is not to provide a set of serially ordered principles of justice, like those of Rawls’s theory. Serial ordering is only one possible way in which trade-offs between principles could be defined (one could do it, in theory, by providing a set of logical and mathematical functions). Yet, Farrelly does not seem to aim to provide any comprehensive theory of justice, one in which the problem of evaluating trade-offs is explicitly dealt with. He claims that the lax genetic difference principle “is designed so that it can be balanced against the demands of other principles of justice” (see note 15, Farrelly 2004:W26), via a reasonable genetic intervention model. This is offered as a procedural solution (a fair deliberative procedure) through which different stake-holders contribute to determine the trade-off between the cost and benefits of restricting reproductive freedom for the sake of other values. See Farrelly C. Preimplantation genetic diagnosis, reproductive freedom, and deliberative democracy. Journal of Medicine and Philosophy 2009;34(2):135-154. By analogy, Farrelly could propose a procedural solution for the problem of assessing interactions between different goods. So far, however, this goal seems to lie outside the horizon of his preoccupations. 38. See note 21, Farrelly 2008:46. 39. See note 21, Farrelly 2008:46. 40. Duncan G, Kalil A, Mayer SE, Tepper R, Payne M. The apple does not fall far from the tree. In: Bowles S, ed. Unequal Chances: Family Background and Economic Success. Princeton, N.J.: Princeton University Press; 2005:23-79. 41. Marmot M. Social differentials in health within and between populations. Daedalus 1994;123(4):197. See also Marmot M. The Status Syndrome: How Social Standing Affects Our Health and Longevity, 1st ed. New York: Times Books; 2004. 42. Merikangas KR, Risch N. Genomic priorities and public health. Science 2003;302(5645):599-601 at p. 600. 43. See note 42, Merikangas, Risch: 601. 44. Khoury MJ, Davis R, Gwinn M, Lindegren ML, Yoon P. Do we need genomic research for the prevention of common diseases with environmental causes? American Journal of Epidemiology 2005;161(9):799 -805 at p. 800. 45 . Berrettini W, Bierut L, Crowley TJ, Cubells JF, Frascella J, Gelernter J, et al. Setting priorities for genomic research. Science 2004;304(5676):1445-1447. 46. See note 21, Farrelly 2008:49. 47. Tabery J. From a genetic predisposition to an interactive predisposition: rethinking the ethical implications of screening for geneenvironment interactions. Journal of Medicine and Philosophy 2009;34(1):27-48 at p. 35. 2 1 48. See note 47, Tabery 2009:32. 49. Caspi A, McClay J, Moffitt TE, Mill J, Martin J, Craig IW, et al. Role of genotype in the cycle of violence in maltreated children. Science 2002;297(5582):851-854. 50. See note 47, Tabery 2009:28. 51. See note 47, Tabery 2009:32. 52. It could be objected that in the scenario in question, improving educational conditions for half of the population qualifies as a goal of justice independently from duties of genetic justice and that, for that reason, the scenario in question could never occur in a fully just society. This might be true in ideal theory, but not in the context of non-ideal theory, for which Farrelly’s principles are intended. E.g. when international monetary institutions require a balanced budget, educational justice might have to be compromised with other legitimate goals of justice. 53. See note 47, Tabery 2009:41-43. 54. See note 21, Farrelly 2008:49. 2 2