Published online by Cambridge University Press: 06 February 2004
The emerging phenomenon of genetic paternity testing shows how good science and useful social reform can run off the rails. Genetic paternity testing enables us to sort out, in a transparent and decisive way, the age-old but traditionally never-quite-answerable question of whether a child is genetically related to the husband of the child's mother. Given the impossibility of settling this question for certain, British and American law has long held that a biological relationship must almost always be assumed to exist. According to what is known as the “marital presumption” or “presumption of legitimacy,” a child born to a woman within a marital relationship is assumed to be the biological child of the woman's husband unless he was absent, impotent, or sterile. In other words, if paternity was not a physical impossibility for the husband, there was a nearly irrebuttable presumption that he was the father of the child. The husband was locked into the role of fatherhood.I am grateful to the participants of Genetic Ties and the Future of the Family, a research project run conjointly by The Hastings Center and the Institute for Bioethics, Health Policy, and Law at the University of Louisville School of Medicine. Discussions held in the course of this project have influenced this paper in various ways. I am especially grateful to Mary Anderlik for detailed comments. An earlier version of this paper was presented at the 2002 annual meeting of the American Society of Bioethics and Humanities and at a colloquium at Oxford University on Genetic Technologies and the Family, and the paper has benefited from comments offered on each of those occasions. Funding for Genetic Ties and the Future of the Family is provided by the National Institutes of Health (grant #HG02485).
The emerging phenomenon of genetic paternity testing shows how good science and useful social reform can run off the rails. Genetic paternity testing enables us to sort out, in a transparent and decisive way, the age-old but traditionally never-quite-answerable question of whether a child is genetically related to the husband of the child's mother. Given the impossibility of settling this question for certain, British and American law has long held that a biological relationship must almost always be assumed to exist. According to what is known as the “marital presumption” or “presumption of legitimacy,” a child born to a woman within a marital relationship is assumed to be the biological child of the woman's husband unless he was absent, impotent, or sterile. In other words, if paternity was not a physical impossibility for the husband, there was a nearly irrebuttable presumption that he was the father of the child.1
Michael H. v. Gerald D., 491 U.S. 110, 124 (1989) (citing Nicholas H. Adulturine Bastardy 1836;1:9–10).
Some now argue that this presumption is outmoded.2
Lewin T. In genetic testing for paternity, law often lags behind science. New York Times 2001 Mar 11:A1.
See note 2, Lewin 2001. See also: Stanley A. So, who's your daddy? in DNA tests, TV finds elixir to raise ratings. New York Times 2002 Mar 19:C1. The media coverage is also discussed at length in: Nelkin D. Paternity palaver in the media: selling identity tests. Unpublished manuscript, Genetic Ties and the Future of the Family, University of Louisville Center for Bioethics and The Hastings Center.
See, for example: http://www.dnacenter.com; http://www.genetree.com; and http://www.betagenetics.com.
Anderlik M. Disestablishment suits: daddy no more? Unpublished manuscript, Genetic Ties and the Future of the Family, University of Louisville Center for Bioethics and The Hastings Center. See: 1999 Ohio H.B. 242 (2000 Jul 27).
Egerton B. DNA tests don't let dads off the hook: man supports sons not biologically his. Dallas Morning News 1999 Oct 31:A1.
The pressing conceptual issue behind paternity testing is the nature of parenthood. Paternity testing encourages us to suppose that a parental relationship to a child is fundamentally a genetic relationship, or at least necessarily includes a genetic relationship. Paternity testing also encourages us to suppose that a parental relationship can be reassessed years or decades after the parties in the relationship thought it was established.
Insofar as paternity testing enables people to reopen the question of parenthood, it enjoys some support from a broad and ongoing social transformation. In the mid-twentieth century, the underlying commitment behind much of family law was preservation of the social order. Family law enforced social conventions about the familial roles people filled—as parent, husband, wife, or child. Increasingly, as the legal historian Michael Grossberg has argued, this commitment gave way to a “commitment to individual choice and private ordering” (p. 7).7
Grossberg M. How to give the present a past? family law in the United States 1950–2000. In: Katz SN, Eekelaar J, MacLean M, eds. Cross Currents: Family Law and Policy in the United States and England. New York: Oxford University Press; 2000:3–29.
See note 7, Grossberg 2000:4–12.
In almost every case, this broad reform is desperately needed and still unfinished, but I will argue against unrestricted genetic paternity testing. The central conceptual issue raised by paternity testing is the nature of parenthood. My general conclusion—I will not set out policy options in detail—is that society may reasonably constrain individuals' use of and access to DNA-based paternity testing—at least if those individuals have had the opportunity to develop relationships with the children and they seek the testing to reassess the existing family relationships. I approach this issue by setting out the sorts of cases that involve paternity testing and commenting along the way on the range of issues involved in them.
Sometimes, paternity testing is an inadvertent side effect of medical tests. In one of the better publicized legal cases involving claims about biological paternity, for example, the father discovered that he was not the biological father of a child he was helping raise when he learned that the child had cystic fibrosis but that he was not a carrier for the disease.9
See note 6, Egerton 1999.
In other cases, paternity testing is intentional. Commentators frequently organize these cases into rough categories according to what the testing is intended to establish. There are four primary categories. In one, genetic paternity testing is sought by a man to terminate parental responsibilities.10
See, for example: Paternity of Cheryl, 746 N.E. 2d. 488 (Mass. 2001); NPA v. WBA, 380 S.E. 2d. 178 (Va. Ct. App. 1989); Dews v. Dews, 632 A. 2d. 1160 (D.C. 1993).
In a second and very common type of case, paternity testing can be sought by the child's mother to impose parental responsibilities—chiefly, the responsibility to help support the child financially. She may also be under pressure from the state to locate the genetic father, because if no man is located to provide child support payments, the financial burden may shift to the state itself. Alternatively, the state itself may require the testing. Federal welfare policy over the last several decades has given increasing support to the use of paternity testing to enforce child support, culminating in the Personal Responsibility and Work Opportunity Act of 1996, which required states to give child-support enforcement agencies authority to order genetic testing.11
Anderlik MR, Rothstein MA. DNA-based identity testing and the future of the family: a research agenda. American Journal of Law and Medicine 2002;28:215–32, at 218.
The other two kinds of cases involving possible parents concern the establishment of parental rights. Paternity testing can be sought by a man to assert parental rights, possibly over against the mother, or possibly over against a presumptive father.12
See, for example: Michael H. v. Gerald D., 491 U.S. 110 (1989); Rodney F. v. Karen M., 71 Cal. Rptr. 2d. 399, 401 (Cal. Ct. App. 1998).
See, for example: Cavanaugh v. deBaudiniere, 493 N.W. 2d. 197, 197 (Neb. Ct. App. 1992); Marriage of K.E.V., 883 P. 2d. 1246 (Mont. 1994).
To these four categories must be added an assortment of other types of cases in which genetic paternity testing might be sought by people other than the presumptive or putative parents. Paternity testing can be sought indirectly by other relatives, through tests to determine whether someone is a genetic grandparent, full sibling, aunt, or uncle.14
William Carl Langston v. Alice L. Riffe; William Carl Langston v. Sharon Locklear; Danielle R. v. Tyrone W., 359 Md. 396; 754 A. 2d. 389 (Md. 2000). Testing for such uses is also promoted; see http://www.dnagenetesting.com.
Individuals' efforts to gain control over their familial relationships are illustrated in these categories to various degrees. The most straightforward instances are those cases in the first and fourth categories, in which paternity testing is associated with divorce proceedings. Indeed, it is likely that, in some instances, the motivation to procure testing is not genuinely to discover whether a parental relationship exists but to bring about an outcome that the individual wants anyway, whatever the biological facts may be.
In all of these types of cases, questions about the man's relationship to the child tend to get mixed up with questions about his relationship to his wife. In particular, we are likely to be distracted by the thought that the child is the result of a sexual betrayal. If we are, then we will translate decisions about the man's relationship with the child into contests between adults about their relationship with each other, and specifically about how they will assign parental rights and responsibilities with respect to a now somewhat-invisible child. But sexual betrayal is not, after all, itself a feature of the relationship between a man and the child. It is a feature of his relationship with his wife. Even if the parental relationship is construed merely as a biological phenomenon, the marital sexual betrayal can be only connected to the facts of parentage.
Because our starting typology organizes cases specifically in terms of types of contests between adults, it does not raise the question that we most need to attend to: How should we understand the relationship between the man and the child? What, in other words, is “fatherhood”?
The main premises of the argument for paternity testing are that parenthood is a biological phenomenon and that determinations of parentage should therefore adhere to the biological facts. For most of the course of Western history, a biological relationship has indeed been accepted as the overriding criterion of fatherhood. Further, this claim has often been cast as a matter of natural law. In an older and particularly stark form of this position, a child was held to belong, as property, to the man who had sired it—in nearly the way that any artifact would presumptively belong to the person who manufactured it—unless of course that man himself was the property of another.15
See: Grossberg M. Governing the Hearth: Law and Family in Nineteenth-Century America. Chapel Hill: University of North Carolina Press; 1985; Murray T. The Worth of a Child. Berkeley: University of California Press; 1996.
Pufendorf S. On the Law of Nature and Nations. Oldfather C, Oldfather W, trans. 1934:ch. 2, § 4.
A recent variant of this position draws from contemporary views about the genes' role as the “blueprint” for the person. True family bonds are encoded directly in genetic relationships. The genes so thoroughly control a person's nature, according to this position, that two people who are genetically related to each other will immediately feel a personal bond between them. In Kaja Finkler's study of adopted children who have searched for their genetic parents, a young woman explains that “with your own you instinctively feel a kinship connection that passes through…. It's a memory code in the genes.”17
Finkler K. Experiencing the New Genetics: Family and Kinship on the Medical Frontier. Philadelphia: University of Pennsylvania Press; 2000:127. Emphasis in the original.
Lifton BJ. Journey of the Adopted Self: A Quest for Wholeness. New York: Basic Books; 1994, quoted in: Nelkin D. The DNA Mystique: The Gene as a Cultural Icon. San Francisco: Freeman; 1995:66.
This reductive and deterministic view of the genes' contribution to a person's character is increasingly popular in the media, as anyone who spends much time in grocery store check-out lines can attest to, but genetic scientists now doubt that the genes exert the kind of control that this latter picture imputes to them. Human behavioral traits—those aspects of a person's character on which an interpersonal bond would depend—are thought to depend on extremely complex interactions both among genes themselves and among genes and the physical and social aspects of environment.19
For an extended discussion of this view and criticism of opposing views, see: Turkheimer E. Heritability and biological explanation. Psychological Review 1998;105:782–91.
Even Pufendorf's more moderate explanation for why a genetic relationship is the natural basis for a parental relationship has come into question, however. There are no studies of the impact on children's welfare of a discovery of misattributed paternity—much less of the impact of undiscovered misattributed paternity!—but studies of analogous groups are reassuring. The closest analog is provided by children conceived through sperm donation, as they are genetically related to their mother but not to their rearing father and they have a social relationship with both beginning at conception. In her review of studies on children's psychological development, Susan Golombok concludes that these children tend to do as well as children conceived naturally, and she tentatively infers that a genetic relationship to the rearing father provides no apparent benefit to a child.20
Golombok S. Parenting: What Really Counts? London: Routledge; 2000:36.
Adopted children offer another analog, though a more distant one, as they are genetically related to neither parent and their social relationship begins sometime after birth. Here, the news is cautionary at first blush: behavioral problems are more likely to be diagnosed in adopted children than in children who are genetically related to their parents.21
See note 20, Golombok 2000:27–8.
See note 20, Golombok 2000:28–9.
There might, of course, be times when it is indeed in a child's interest to remain with its genetic parents. But if we do not suppose that children belong to those who brought them into the world, nor that children should be raised by their biological parents as a matter of natural law, nor that human nature is so determined by the genes that genetically related people will have a personal relationship by a kind of physical necessity, then we will recognize that the parental relationship, like all human relationships, is first and foremost a psychological and social phenomenon and may or may not be entwined with an underlying biological relationship. This view is sometimes known as the functional or psychological account of parenthood. It is described by Joseph Goldstein, Anna Freud, and Albert Solnit, in an influential work on child development, as follows: “Whether any adult becomes the psychological parent of a child is based on day-to-day interaction, companionship, and shared experiences. The role can be filled either by a biological parent or by an adoptive or by any other caring adult.”23
Goldstein J, Freud A, Solnit A. Beyond the Best Interests of the Child. New York: Free Press; 1973:19.
See note 15, Murray 1996:57ff.
The notion of rearing parenthood has ancient roots, traceable at least to Roman conventions and law concerning parent-child relationships.25
See note 15, Murray 1996:47ff. Murray draws from: Boswell J. The Kindness of Strangers: The Abandonment of Children in Western Europe from Late Antiquity to the Renaissance. New York: Pantheon; 1988.
Dolgin JL. Choice, tradition, and the new genetics: the fragmentation of the ideology of family. Connecticut Law Review 2000;32:529.
To refine the typology of cases with which the discussion began, we need to know whether the man whose paternity is in question is the child's rearing father. We need to know whether he has been supporting and raising a child in the way that fathers do and whether the child has been looking to him as to a father. An individual who has been successfully filling this role is the best candidate, on the available evidence, to continue to support the child, to uphold and advance the child's interests. Furthermore, if parenthood is a fundamentally social and psychological phenomenon, then the central value of parental relationship is precisely that it upholds and advances the child's interests and, secondarily, the interests of society and of the adults whose lives are invested in that child—that is, of the rearing parents. The child's interests are foremost among these three because of the asymmetry of the parental relationship: the child, who is uniquely dependent on the adults raising him, simply has more at stake in the parental relationship than do the adults.27
For an extended discussion of the relevance of these three interests to parenthood, see: Blustein J. Parents and Children. New York: Oxford University Press; 1982:139ff.
Given these considerations, the results of paternity testing will be rendered moot in some kinds of cases. First, it will often be appropriate to allow a man who has an established parental relationship with a child to be completely released from it. Indeed, this appears to be a reasonable presumptive rule.28
The presumptive rules I identify are treated similarly by: Blustein J. Ethical issues in DNA-based paternity testing. Unpublished manuscript, Genetic Ties and the Future of the Family, University of Louisville Center for Bioethics and The Hastings Center.
Second, it will often be inappropriate to deprive him of the parental relationship—the goal of those cases in the third and fourth categories, in which a woman or another man seeks paternity testing to reassign parental rights. This appears to generate another presumptive rule. The parental rights ride along with the relationship and properly belong, we may presume, to whatever person has been building and perpetuating that relationship. It will probably be in the interests of the child that the relationship continue. And it will probably be in the interests of the man, who has by now invested himself psychologically in the relationship.
I have argued in favor of viewing parenthood as a psychological instead of a genetic relationship, and decisions about paternity as psychosocial rather than genetic questions. But there must be something to say on behalf of genetic relationships. If a genetic relationship were of no importance at all, then there would be no good objection to adopting a policy that involved randomly redistributing babies in the hospital. But such a proposal is unsettling, if not appalling. Plainly, a genetic relationship can also have some importance in parenthood.
One way in which the genetic relationship is important is that it is a feature of a causal relationship. Typically, an adult is the genetic parent of a child only when the adult has knowingly acted in such a way that a child could be created. We often hold each other responsible for the foreseeable consequences of our actions, and given our current social mechanisms for rearing and supporting children, it is reasonable to hold people accountable when through their actions they bring a person into existence. Here we have another presumptive rule: in the context of our current social institutions for raising children, one way of undertaking a responsibility to help cover the material costs of raising a child is to have a child.
Few will argue with this rule. The trick is to delimit it. Fathers' rights advocates, for example, tend to accept it but then use it to argue, first, that genetic fathers have not only parental responsibilities toward their child but rights to the child and, second, that men who are shown not to be genetic fathers of a child have neither rights nor responsibilities toward a child. Evidence that is sufficient to “convict” a man of parenthood ought by parity of reasoning to be sufficient to uphold a man's claims to parenthood or to “exonerate” him of it. The argument is attractively simple, but the response is by now familiar. If parenthood is (like all other human relationships) first and foremost a social and psychological phenomenon—in this case a rearing relationship—then either genetic testing never establishes parenthood at all or it establishes only one biological feature that is commonly part of the social relationship. Or, we might say, it establishes one, narrow, biological kind of parenthood—not a relationship between persons, but one between organisms. However we put it, what establishes the biological and causal relationship can in no way be assumed to establish a personal relationship (and its attendant rights) nor to disestablish a personal relationship (and its attendant responsibilities).
The genetic relationship is, of course, commonly also one aspect of the social and psychological relationship, and this gives us much more to say about its importance. A genetic relationship can often offer a variety of social and psychological benefits. The most prosaic of these is that a transparent biological relationship gives children easier access to their genetic relations' medical histories, and so better knowledge of their own medical risks. The child also may gain access to various social benefits, ranging from material benefits that are conditional on genetic relatedness to less tangible social and psychological benefits, such as those enjoyed by people who can prove that they are descendants of Thomas Jefferson.
A variety of other possible benefits are conditional in some measure on contemporary views about parenthood and biology. For example, again given contemporary views about parenthood and biology, a biological relationship sometimes contributes toward a psychological relationship. It can provide one way in which the range of experiences that the child and its parents share can be extended, the story that the child and its parents tell about themselves lengthened and deepened. It means that the caring and nurturing begin before the child has even quite appeared on the scene. It gives parents the very earliest possible experiences one can have of another person—the prenatal stirrings and sonographic images that we search (however foolishly) for clues to the child's emerging personality—and it enables parents and children to draw connections (however groundlessly) between a child's traits and those of parents and other relatives.
Knowledge about the biological relationship can also touch on the child's and the adults' sense of self. Partly by discovering shared traits that may be traceable to shared genetic factors, partly because of common sentiments about the value of genetic relatedness, a genetic relationship perhaps sometimes helps children develop a sense of identity. Adopted children who do not know the identity of their genetic relatives occasionally claim that they lack a sense of their own identity, and they sometimes launch exhaustive searches to locate their genetic parents for no other reason than to see where they fit into the genetic web of humankind, believing that this understanding will give them a better sense of who they are.29
See note 17, Finkler 2000:117ff.
For all of these reasons and for others besides, a genetic parental relationship can be one of the things that gives meaning to life. For the parent, it can be meaningful to participate in the project of physically bringing a child into the world. For the child, it can be meaningful to feel that one's physical existence is owed partly to your parents' own meaning-making. Being situated in this way in another person's story can feel good.
Many of these points are suggested by James Lindemann Nelson's concept of the “genetic narrative” view of the family. Explains Nelson:
The role of genetic connections in our lives can be seen, not as “gene calling to gene,” but rather as a part of our interest in perceiving the connections between our lives and the lives of others—connections which add depth and richness to the continuing story in which we participate, and which can therefore be referred to as narrative connections. (p. 81)30
Nelson JL. Genetic narratives: biology, stories, and the definition of the family. Health Matrix 1992;2:71–83.
These claims about the benefits of a genetic relationship are all either limited or provisional. They do not generate a manifesto for the Western family. They do not establish that parenting that involves a biological relationship is intrinsically better or more meaningful than other sorts of parental relationships, which may offer their own special benefits. There are other ways of obtaining knowledge of one's genetically acquired medical risks than by being raised by one's genetic parents, nor does being raised by one's genetic parents guarantee that one knows those risks. The meaning, the sense of identity, and the richness of shared life-story to which genetic parenthood can contribute are themselves partly conditional on Western views of the family, not intrinsic features of genetic parenthood, and they would be more or less canceled out if the genetic parents did not seize the moment at the child's birth to establish a rearing relationship.
The psychosocial benefits also do not establish that men who are genetically related to their children will tend to be better rearing parents. That last claim is already undermined by studies of children conceived with donor-provided sperm, but even if it were true, if we regard parenthood as a social and psychological relationship, we would tend to look first at the existing relationships. If one man already is functioning well as a parent, we would tend to favor preserving that relationship, particularly because disrupting a parent-child relationship appears to be detrimental to a child's health.
Thus these considerations would count for very little in most cases involving genetic paternity testing, particularly when the testing is sought to test the paternity of a man who has already been raising a child as his own for some time.
If these observations are plausible, then it would be appropriate to impose a variety of constraints on genetic testing when it is sought for purposes of reassessing and reordering existing parental relationships. These constraints could take different forms.31
A much more thorough discussion of the current law and policy options is found in: Rothstein M. Translating values and interests into the law of parentage determination. Unpublished manuscript, Genetic Ties and the Future of the Family, University of Louisville Center for Bioethics and The Hastings Center.
First, constraints might bear on how the testing is used. When the results of genetic testing are intended for use in legal paternity cases, for example, it would be appropriate to limit their weight and perhaps even to prevent them from being introduced as evidence in the first place. They will have weight chiefly in proceedings initiated by the mother to force a biological father to provide at least financial support to a child. In hearings to decide whether a rearing father has ongoing parental rights and responsibilities toward a child, they should typically either have very little weight or not be admissible as evidence.
Because of the traditional strength of the marital presumption, nearly all jurisdictions barred the results of genetic tests in paternity cases in which a presumptive father sought the test to challenge his paternity—at least until the recent wave of state legislation that makes it easier to use genetic testing, either by loosening the evidentiary restrictions on genetic paternity testing or by increasing the weight that knowledge of genetic relations should be given.32
See note 5, Anderlik; 1999 Ohio H.B. 242 (2000 Jul 27).
Oddly, even some of those who object to the new wave of testing-friendly legislation adhere to the genetic view. The Uniform Parentage Act of 2000, a model law developed by the National Conference of Commissioners on Uniform State Laws, bars proceedings to adjudicate parentage if the child is more than 2 years old and there is a presumed father—a man who is married to the mother and who either cohabited or had sex with the mother around the time of the child's conception.33
National Conference of Commissioners on Uniform State Laws. Uniform Parentage Act, 2001 Jan 5 [UPA 2000]. The law also prohibits proceedings involving acknowledged or adjudicated fathers.
See note 33, UPA 2000, sect. 608.
Roberts P. Biology and Beyond: The Case for Passage of the New Uniform Parentage Act. Washington, D.C.: Center for Law and Social Policy; 2000:16ff.
Glennon T. Somebody's child: evaluating the erosion of the marital presumption of paternity. West Virginia Law Review 2000;102:547–605, at 569.
A less conceptual muddle arises in a second appropriate constraint on genetic paternity. In addition to constraining how people use genetic testing, we may constrain how testing is provided to them. One kind of appropriate constraint on the provision of genetic testing is obvious: the laboratories that offer genetic paternity tests may and probably should be regulated to ensure that testing is accurate and results dependable, even if the regulation means that the tests are more expensive or are offered by fewer laboratories. Given the level of certainty possible with genetic testing, results should almost never be overturned by further testing. To permit inaccurate testing is, of course, to compound the disruption that testing causes.
We can also reasonably limit access to testing by requiring that both legal parents (or the legal parent, if there is only one) grant consent to paternity testing. In most cases, the legal parents will be the rearing parents as well as the presumptive genetic parents. If the parental relationship is understood as a rearing relationship, then an existing rearing relationship should not be undermined secretly by someone else—whether a spouse, another family member, or a putative biological father outside the marriage.
There are some other, firmer ways of limiting access to genetic paternity testing. At the very end of the scale, we could consider prohibiting genetic testing altogether in certain categories of cases. Or we could make genetic testing permissible only when there is a court order for it. However, given the medical and sometimes social value of knowing who one's genetic relations are, this measure seems unwise. Further, whereas paternity testing that is intended to reassess parent-child relationships merits constraint, testing at birth, with the goal of initially sorting out parent-child relationships, presents a somewhat different set of concerns. The argument of this paper provides reasons for permitting such testing, even for encouraging it in those cases in which the parents themselves raise questions about genetic paternity: It would help children acquire the medical benefits of knowing their genetic ancestry. It could give biological parents an opportunity to “seize the moment” and establish a strong rearing relationship. Although it would disrupt some families that otherwise would never have discovered misattributed paternity, it might help avoid the disruption of later disputes. Given the rising importance of genetics in medicine, it may well be that an ever larger percentage of cases of misattributed paternity will be uncovered inadvertently anyway.
What argues most strongly against ever encouraging testing at birth is just that it would cultivate the very confusions that draw people to testing later on. If we could someday demystify the genetic relationship, recognize it for what it is, and understand both its appeal and its limits, we might even encourage testing at birth as a general social policy. But that day is not at hand.