maternal-infant bonding: the politics of falling in love with your child

Maternal-Infant Bonding: The Politics of Falling in Love with Your ChildAuthor(s): William Ray ArneySource: Feminist Studies, Vol. 6, No. 3 (Autumn, 1980), pp. 547-570Published by: Feminist Studies, Inc.Stable URL: http://www.jstor.org/stable/3177482 .

Accessed: 10/06/2014 22:19

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Feminist Studies, Inc. is collaborating with JSTOR to digitize, preserve and extend access to Feminist Studies.

http://www.jstor.org

This content downloaded from 194.29.185.22 on Tue, 10 Jun 2014 22:19:20 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=femstudies

http://www.jstor.org/stable/3177482?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


MATERNAL-INFANT BONDING: THE POLITICS OF FALLING IN LOVE

WITH YOUR CHILD

WILLIAM RAY ARNEY

According to an independent midwife, "Bonding is that process [through] which the mother and the child come together and the child realizes that that's the mother it was born out of and the mother makes the outward realization that that is her child."' It is a process which, some argue, has both social and biological bases. Extant literature is unclear about the definition of bonding or about the nature of the "attachment" process, but significant sectors of diverse groups-doctors, midwives, social workers, and feminists-have embraced the notion that bonding does exist. There is widespread agreement that the effects of bonding, or its absence, are profound and long lasting. Indeed, some people feel that the evidence concerning the effects of bonding is so substantial, so unequivocal, that policy can be based on it. Already there are recommendations for change in hospital policies on childbirth and newborn care, on childcare in the family, and for changes in national social policy.

Much of the research on bonding is methodologically flawed, yet bonding theory has been widely accepted despite its demon- strable scientific problems. One must ask how flawed science can slip through the critical screens of the medical profession, which presents itself as scientific, at least in part. The answer is that "scientific" production of knowledge based on bonding theory is, in fact, pseudoscience used to reinforce and defend social institutions that are suffering attacks on their legitimacy. Bonding theory is being used to reform hospital practices, social policies involving children, childcare, and family life just when these institutions required reconstruction for their continued existence. One can understand the widespread acceptance of bonding theory and its rapid development in obstetrics, only by concentrating on the political uses of the theory, rather than on the scientific activ- ity the concept has spawned.

Feminist Studies 6, no. 3 (Fall 1980). © 1980 by Feminist Studies, Inc.



548 William Ray Arney

THE RESEARCH AND LOGIC OF THE MATERNAL-INFANT BONDING LITERATURE

Evidence that suggests bonding occurs and that it has important consequences derives from two sources: ethological research and quasi-experimental and experimental human research. A logic of inference is used to derive more profound implications about bonding from the data.

Ethology. Within the ethological tradition, there are three major kinds of studies: inference from evidence on birth practices in a wide variety of species, separation studies in which the effects of postnatal separation of mother and infant are studied, and adoption studies in which the researcher tries to get an animal to adopt some other animal's offspring after the newborn has been separated from members of its own species for varying lengths of time.

Species whose perinatal behavior has been studied intensively include cats, rats, goats, sheep, and various primates. Reviews of this literature can be found in Marshall Klaus and John Kennell's major summary of their work, Maternal-lnfant Bonding.. The Im- pact of Early Separation or Loss on Family Development, and in Edward 0. Wilson's Sociobiology. The New Synthesis.2 The results of these studies show that parental behavior toward offspring is species-specific, but, it is argued, "different species have evolved comparable caretaking behaviors to meet similar needs of newborns." Regardless of the social behavior peculiar to each species "individualized, enduring bonds develop between mother and infant."3

There is great variation in birth practices across species, and bonding researchers might be criticized for concentrating too much on commonalities instead of studying this interspecific diversity. But bonding theorists would argue that their cross-species comparative research is merely suggestive, not definitive, evidence for the existence of bonding. Their case for the existence of "bonds" in animals rests on separation and adoption studies.

Separation of mother and infant in various species is effected by removing a newborn from its mother after birth, or by surgically delivering the baby. Dependent variables in separation studies, those variables thought to be influenced by the experimental manipulation, include "acceptance" of the young (nursing, re- trieving young moved a short distance from the mother, etc.) and species-specific maternal behaviors such as nest building in rats. Different species have different reactions to separation. Goats whose babies are taken from them in the first five minutes of life will reject their young. Goats left with their young for five minutes after birth and then separated will reaccept their



William Ray Arney 549

offspring after several hours. Rats too are thought to have a "critical period," the time after birth when attachment is thought to occur, but it appears to be longer and much less well defined than in goats. From these and other separation studies, attachment theorists conclude:

Separation of a newborn or young animal from its mother therefore signifi- cantly alters maternal behavior. The sooner after birth the separation occurs, the stronger are the effects. For each species there seems to be a specific length of separation that can be endured. If separation extends beyond this sensitive period, the effects on mothering behavior during this breeding cycle are often drastic and irreversible.4

Adoption studies tend to confirm the existence of a sensitive period, but add one qualification: the setting in which one tries to effect an adoption and the behavior of the infant both affect maternal response. This qualification suggests the importance of mother-infant interaction and the importance of the environment on "attachment."

Investigation of the physiological bases of attachment behavior has paralleled studies based on observation of behavior studies. Spurred on by developments in endocrinology that demonstrate the impact of social stimuli on hormonal secretion and subsequent behavior,5 investigators have tried to show how the presence or absence of a newborn affects the hormonal status of a new mother. The results of this work, according to Klaus and Kennell, show that

biological mechanisms are primarily responsible for a mother's receptivity to a young at the time of birth but quickly subside afterward. Within the sensitive period, maternal behavior quickly disappears if young are not present to elicit and maintain it.... Because of her physiological state after parturition, a mother is sensitized to the behavior cues of her newborn and begins to respond to them. The infant, in turn, responds to maternal behavior, and patterns of interaction quickly develop that establish the bond between mother and infant, preventing her from abandoning him.6

Studies of Humans. Quasi-experimental bonding research on humans has involved babies separated from their parents by "accident," by war, or by death of the parents. In fact, for "quasi-experimental evidence," one can often read "anecdotal evidence." That is, the evidence comes from studies comparing separated infants to notions of normal development instead of to a control group of nonseparated infants as an experiment would require.

Researchers in the 1930s and 1940s studied children who were institutionalized and claimed that deviant modes of mothering




led to long-term disturbances of personal-social relations, altered perceptual-motor functioning, and promoted intellectual retarda- tion.7 Subsequent research questioned the long-term impact of institutionalization,8 however, and pointed out that the complex- ity of maternal-infant interaction made it difficult to isolate those variables related to observed developmental disorders. Further- more, the detrimental effects of institutionalization were found to be "apparently reversible if the infant is placed in a home or given supplementary caretaker attention and stimulation before he or she is past six months of age."9

The findings of institutionalization studies and other anecdotal information on stimulation and attention were not extraordinarily important in themselves. They are historically important because of the impetus they gave to recent experimental research on bonding.

Ethical considerations prevent prolonged, indefinite separation of mothers and infants for experimental purposes. Therefore, the usual independent, or explanatory, variables in human separation studies are timing and duration of maternal-infant contact after birth. In the now classic study by Marshall Klaus et al.,10 a study typical of the genre, twenty-eight mothers were divided into an "'extended contact group" and a control group. Each member of the control group saw her baby for a short time soon after birth, had brief contact again at six to twelve hours after birth for iden- tification purposes, and then visited her child for twenty to thirty minutes every four hours for feedings. This was standard hospital procedure for postpartum mother-infant contact. Members of the extended group "were given their nude babies ... for one hour within the first three hours after birth, and also five extra hours of contact each afternoon of the three days after delivery."1' Thus, the difference in contact totaled sixteen hours over the first three postpartum days.

Dependent variables, the phenomena to be explained, in separation studies of this kind are numerous. The studies reported in Klaus and Kennell's book take data on dependent variables from three sources: an interview with mothers when their babies were one month old, maternal behavior during physical examinations of babies at one month and one year of age, and films of feeding at one month and one year. The studies report statistically significant associations between variations in the independent and dependent variables. For example, differences in indexes of maternal behavior at home and during a physical exam were found. Also, according to analyses of films taken during feedings, the extended contact group engaged in en face behavior (alignment of mother's




face with baby's face) and fondling to a greater extent than did the control group. Subsequent studies on subsets of the original twenty-eight subjects showed differences in maternal linguistic behavior toward their infants at two years of age and in tests of infants' language ability and IQ at five years of age.l2 The conclusion from this extended series of studies is that "just sixteen extra hours of contact within the first three days of life affect maternal behavior for one year and possibly longer, and they offer support for the hypothesis of a maternal sensitive period soon after birth."13

The work by Klaus and Kennell is representative of human experimental research on bonding. Put simply, one manipulates maternal- infant contact and then searches for associated differences in maternal behavior or in child development. Other studies have adopted variations on this theme using different schedules of maternal- infant contact and different dependent variables. But the conclusions are clear, at least to bonding proponents: "On the basis of [the] evidence, we strongly believe an essential principle of attachment is that there is a sensitive period in the first minutes and hours after an infant's birth which is optimal for parent-infant attachment."14

METHODOLOGICAL CRITIQUE

In the introduction to an extensive review of the research on childcare in the family, Alison Clarke-Stewart makes this sweep- ing, general, and accurate description of research on child development: "The field is, in fact, one beset by methodological problems-arising at all stages of the research endeavor: design, conception of variables, sampling, recording, compiling, analyzing, and interpreting data."'5 To appreciate the full force of this criticism for the bonding literature, let us accept bonding research, for the moment, on the terms in which it is presented. Because bonding research is offered as scientific, we shall examine it from that perspective and see whether it merits its own claim. This section shows that from its own perspective, the research is flawed and that one cannot come to the conclusions bonding researchers would have us draw from their work. Having shown that the work is hopelessly flawed and, furthermore, that it would be impossible to conduct research that could conceivably lead to the conclusions of bonding research, we will be forced to ask how and why such flawed science has become so scientifically acceptable. The answer, as a subsequent section shows, is that bonding theory serves political ends.

The first problem with the field is that there is no agreement on




what attachment or bonding is. There are definitions, of course. Klaus and Kennell say: "An 'attachment' can be defined as a unique relationship between two people that is specific and endures through time."16 This is consistent with conceptions of attachment existing before they started their research. Over the past few years, however, the vagueness of this definition has been questioned and alternatives to it proposed.17 For example, Miriam Rosenthal ob- jects to viewing attachment as a bond that exists between two people and that has a significant influence on the development of only one member of the dyad. Instead, it is argued, attachment should be viewed simply as one aspect of mother-infant interaction. As she says, "Attachment is not a 'thing' . . . but rather a characteristic of some patterns of interaction between mother and infant .. ."18 Others have argued that if one views the child as an "active partici- pant in a social network," research on a child's social behavior and development will be more fruitful than if one focuses narrowly on "attachment" behavior and its consequences.19

Asserting a lack of conceptual consensus is not a particularly damaging criticism, but in this case as in others, imprecision in the definition of concepts establishes the basis for serious methodological errors. Lack of consensus permits, or perhaps requires, that multiple indicators of a concept be used. Each indicator brings with it its own difficulties of measurement.20 To reduce the importance of difficulties with any single indicator of attachment, researchers generally search for correlative differences in several indicators-fondling, kissing, soothing, proximity-seeking behavior, en face behavior, attempts to breastfeed, laughing or singing to an infant, subjective feelings about separation, and so on-in the same study of attachment. Separate tests of statistical significance for each indicator often are performed. This practice, however, biases results in favor of the researcher because each addi- tional test increases the probability of finding a "statistically significant" test simply by chance.

The multiple indicator criticism is a common one in social research, but it is particularly severe in this field because there is a high degree of inconsistency in results and because replication of findings is relatively rare. In a critical review of the psychological literature on attachment John Masters and Henry Wellman21 conclude "correlations which are significant in one sample often fail to research significance in an independent sample and . . .the tem- poral stability, cross-situational consistency, and cross-behavioral consistency of attachment behaviors are not great." In fact, of the preeminent approach to the study of attachment, they write that "correlational analysis of human infant attachment behaviors does




not provide substantial support for the concept of attachment as a psychological state or central motive state."22 That is, there is little empirical validity for the concept thought to underlie the many indicators used in attachment research. A field lacking conceptual clarity permits a low incidence of replication. Developing new and improved measures of attachment instead of working with old ones becomes acceptable scholarship when there is no agreement on how concepts should be measured in the first place. New measures abound, but there are few attempts to check the reliability of reported results.

Yet there are even more serious errors of method to be found in the attachment literature. In a paper published in Child Develop- ment, A. D. Leifer et al.23 report many complicated tests of significance on many indicators of attachment. Some tests support a hypothesis to the effect that differences in behavior are caused by extended contact, some do not. In a kind of scholarly post script- um, Leifer et al. add:

There are other differences between the groups that are not reflected in the observational data presented here and are probably outcomes of their separation experience. For instance, within the sample of 26 separated and 23 contact mothers of prematures who participated in any phase of the study, two relinquished custody of their infants sometime after hospital discharge. Both of these mothers were in the separated group. Also, there have been six instances of divorce among the parents in this study, five of them in the separated group.... Finally, there were four mothers who attempted to breast-feed their infants. Two of these mothers were in the contact group and two in the separated group; one mother in each group was a primipara and the other a multipara who had successfully breast-fed at least one previous child. The only mother to succeed was the multiparous contact mother.24

They conclude:

These actions (divorce and relinquishing custody) represent severe disturbance of normal maternal and marital behavior.... The existence of these cases, while few in number, suggests that early separation of mother and infant may seriously disrupt normal maternal behavior.25

The ideology of "normality" used to arrive at such a conclusion is, in itself, disturbing. More to the point of the present discussion, however, is the fact that this is a blatant example of searching out data to bolster a theory when the original investigation is less than totally successful in doing so.

In addition to problems of conceptualization and measurement, there are serious problems with the design of separation studies.




There are difficulties in obtaining comparable experimental and control groups, and there are problems in experimental treatments being confounded with other variables.

Investigators make much of the fact that experimental and control groups are "matched" in separation studies. Matching controls the effects of variables, other than the experimental treatment itself, that might influence attachment behavior. In studies reported to date, experimental and control groups are, in fact, "matched" on variables that might influence results prior to the beginning of each study. The problem is that there is no assurance that groups remain comparable as a study proceeds, and thus, there is no assurance that the effects of life events on long-term results of attachment or nonattachment have been controlled. But such assurance is necessary when differences in groups up to five years from experimental assignment are associated with the experimental manipulation. Klaus et al.,26 for example, randomly divided subjects into control and experimental groups based on their day of delivery, sex and weight of infants, and socioeconomic status. But does knowing that the groups were matched on these variables allow the inference that differences at one month, one, two, and five years are attributable only to the experimental manipulation of sixteen extra hours of maternal-infant contact in the first three days of life? To make such an inference the life courses of experimental and control group members prior to and after delivery would have to be similar. For example, did equal proportions of experimental and control groups have prenatal education courses? What were parents' prenatal expectations concerning their babies? What pro- portion of the pregnancies were planned? If we are to take the studies of Charles Kempe and Ray Helfer27 seriously, mothers' first reactions to their babies would have to be the same. After discharge, to what kind of home environment did the mothers return? Was there continuing employment for the breadwinner(s)? Were there pregnancies during the five-year follow-up period? What were the reactions to them? What impact did subsequent pregnancies and births have on households? These questions concerning post-discharge life course are especially crucial because attachment studies are often conducted on people drawn from the unmarried ranks of lower social strata where life courses are likely to change frequently and rapidly.28 In short, the reader is asked to assume out of existence potentially significant experimental "noise" because the members of experimental and control groups have comparable social statuses at the time of birth. Simple randomization of a small number of subjects does not adequately eliminate the plausible hypothesis that it was life course events, and not the experimental manipulations, that caused the observed differences.




Even if the problems of life course comparability could be atten- uated by conducting studies with large samples, the problem of confounded experimental variables would remain. Bonding researchers claim that extended contact between mother and infant is the only difference between experimental and control groups. This is not so. Contact and interpersonal behavior between health professionals and subjects differ also. Experimental and control groups were clearly treated differently by the doctors and nurses who facilitated contact between mothers and infants in one group and denied contact in another. It is likely, moreover, that these health professionals influenced maternal behavior toward infants because they affect the patients' conception of their own health status and of their babies' health status. This observation is especially important because one indicator of attachment is maternal behavior during a physician's examination of the mother's infant. Mothers in extended contact groups show a greater propensity to stand near the examining table and to soothe their babies during the exam than do mothers in control groups. But it is conceivable that experimental group mothers might presume that they need not surrender their children to health professionals and remain in the background during an exam. Such a presumption could be drawn directly from their experiences with health professionals soon after birth, experiences that are associated with experimental treatment. Similarly, en face or fondling and soothing behavior may be unintentionally or intentionally encouraged by staff members during the extended contact period. Indeed, "attachment behavior" encouraged in early contact can take on the character of "normal behavior" because health professionals have taken upon themselves and are generally accorded by patients the task of deciding just what "normal perinatal maternal behavior" is.

The effects of extended contact are inseparable from the effects of other variables that the experimental situation introduces. It would be impossible, moreover, to design appropriate experiments for attributing behavioral differences to separation. It is impossible, that is, to conduct the scientifically required "critical test," for what would constitute a critical test? If agreement on the meaning of attachment could be reached, experiments similar to those conducted by Harlow on monkeys would be required. One group of children would have to be reared normally, and one group would have to be reared with surrogate mothers. Development would have to be monitored for many years, certainly through adoles- cence. To control for genetic endowment, it would be best if identical twins could be used, one going into the experimental group, the other into the control group. Why be so demanding?




The answer, from attachment theorists themselves, is that there is a wide range of behaviors that can lead to a successful "bond." In Every Child's Birthright.' In Defense of Mothering, Selma Fraiberg writes that

there are an infinite number of normal variations in patterns of mothering and great diversity in the mode of communication between baby and mother. Any of a vast number of variations in the pattern can be accommodated in the human baby's development and still ensure that a human bond will be achieved.29

Yet the literature leads us to believe that there is a period very early in life during which proper behavior is essential and during which improper behavior can have dire consequences. The test described above would meet the scientific obligations incumbent on attachment theorists, but it is unreasonable to demand that such a test be conducted. It is impossible, therefore, to determine the prolonged effects of early parent-infant contact.

THE USES OF BONDING LITERATURE

Tucked away in disciplinary boundaries, protected by a thick layer of scientific doubt, and properly critiqued, bonding theory and research might seem relatively benign. But knowledge of any sort is a resource available for use beyond the walls of the academy, and bonding theory very quickly found its way out of the labora- tory and into medicine and social policy. Scholars and scientists developed knowledge which was then used to foster some rather remarkable claims about human behavior.

Some authors, for example, argue from animals to humans using evolution as the basis for their logic. Evolutionary argumentation suggests that because other animals bond, humans probably do, too, because we have acquired everything animals have and more. Some authors have even used a teleological evolutionism to argue from animals to human beings. Alice Rossi, in her article "A Bio- social Perspective on Parenting," argues from an evolutionary perspective and claims that the division of labor among humans is genetically influenced.30 Rossi believes that through processes of adaptation and selection, females and males have developed different innate behaviors: "Biologically males have only one innate orientation, a sexual one that draws them to women, while women have two such orientations, a sexual one toward men and a reproductive one toward the young."31 Added to the evolutionary argument is a kind of biological imperative which is based on the relative immaturity of human infants. There is "even greater




need for close bonding of the human infant to its mother than there is in other species."32

Fraiberg takes evolutionary arguments one step further than Rossi by adding intentionality to evolution.

There are ancient traditions which have bound the baby and his parents together from the first hours of life, traditions deeply rooted in our biological heritage. It was not known until our time why the traditions existed or whether in fact the human family should be bound to them. But the evidence that now emerges from a large body of scientific work is incontrovertible: the traditions themselves were "intended" to insure the love bonds between baby and his parents.33

Fraiberg also uses the semantic device of "striking parallels" to provide the connective tissue that permits her to base normative statements about human mothering on ethological research. She says that "there are some striking parallels between [studies of attachment in animals] and our own studies of normal development and of certain aberrant patterns in early childhood."34 Frai- berg simply asks the reader to make the leap from animal to human behavior and accept the possibility that ethological findings apply directly to human behavior.

If, along with Fraiberg, we make the leap from animals to humans, what are we asked to conclude? For Fraiberg, a clinician, the answer is clear and profound:

The pathways that lead from infant love to the love of maturity can be out- lined.... Love of a partner and sensual pleasure experienced with that partner begin in infancy, and progress to a culminating experience, "falling in love," the finding of a permanent partner, the achievement of sexual fulfillment.

In every act of love in mature life, there is a prologue which originated in the first year of life....35

And if one does not have the prologue to "falling in love"? If one fails to bond? "In the case of a child who has been deprived of human partners in the formative years he may lack inhibitions of aggressive impulses or [experience] extraordinary problems in the regulation of his aggression."36 One can succumb to the "diseases of nonattachment" which may "give rise to a broad range of disordered personalities."37

In personal encounters with such an individual [one who suffers from the diseases of nonattachment] there is an almost perceptible feeling of interven- ing space, of remoteness, of "no connection." ... There is no joy, no grief, no guilt, and no remorse.... Many of these people strike us as singularly humorless.... A very large number of them have settled inconspicuously in the disordered landscape of a slum, or a carnie show, or underworld enterprises... .38




This is what happens to the child who fails to bond. The parent who remains unbonded comes under surveillance because of a propensity to child abuse. Several studies suggest that "separation is indeed one factor resulting in physical abuse of pre-term infants."39

The logic we are asked to follow from research to these grim prognoses is as flawed as the research itself. Marshall Sahlins has pointed out that evolutionary arguments about biological necessity and genetic intentionality are a peculiar perversion of the theory of evolution and natural selection. He says that

the traditional understanding of "natural selection" has been progressively assimilated to the theory of social action characteristic of the competitive marketplace.... Selection successively became synonymous with optimiza- tion or maximization of individual genotypes. ... In the structure of evolutionary argumentation, selection takes the role of a means of the organism's end.40

Argumentation about genetic influences on social behavior, in its modern form, is always presented with the following kind of equi- vocation or qualification, this one taken from Rossi's views on parenting:

A biosocial perspective does not argue that there is a genetic determinism of what men can do compared to women; rather it suggests that the biological contributions shape what is learned, and that there are differences in the ease with which the sexes can learn certain things.41

But given the insurmountable problems of inquiry mentioned previously we must hold, with Sahlins, that the biosocial view of parenting is just a method for introducing teleology into adaptation where the developers of the theory of natural selection never intended it to be.42 This kind of argument permits policymakers to claim legitimacy for their projects by insisting that they are simply allowing the means-ends chains, which are embedded in our genes, to be realized. This kind of argument obscures the choices-the political choices-being made for all of us for it prevents scrutiny and discussion.

The "striking parallels" argument used by Fraiberg is simply a form of argumentation from heights of authority or of argumentation by conviction and persuasion. As such, it violates one of the central principles of scientific inquiry. It violates Robert Merton's "principle of organized skepticism" that work be judged only pro- visionally, and Andre Cournand's "doubt of certitude" or "ques- tioning what is asserted authoritatively."43 Fraiberg's book vir- tually begins with a phrase that should raise the suspicions of the scientifically minded: "But the evidence that now emerges from




a large body of scientific work is incontrovertible."44 In fact, the entire scientific enterprise rests on the capacity for controversy, on the capacity for controvertibility. The search for results which might, potentially, undermine bonding theory, and indeed the willingness to recognize and pursue anomalous results when they arise, are conspicuously absent in the attachment literature. Every- one has adopted a deceptively uncritical stance toward bonding research, and we can now ask, "Why?"

THE POLITICS OF FALLING IN LOVE WITH YOUR CHILD

The scientific method is an idealized set of rules for the creation of a particular kind of knowledge. If one follows the rules, the knowledge acquires a privileged status in most sectors of the con- temporary world. Occasionally, however, a body of knowledge is accorded scientific privilege even though the method by which it was created merely imitated the scientific method and even though it clearly violated rules and modes of reasoning generally accepted by scientists as the only appropriate rules and modes of reasoning. This has happened with bonding theory. Why? Put too simply, but reasonably accurately, it has happened because bonding theory is appealing; it is socially and politically useful. It is so appealing and so useful that violations of rules for the creation of knowledge-rules accepted by scientific subcultures which have acquired powerful, socially strategic positions in the modern world-have been ignored. Bonding theory is being used to reform various social institutions and social practices of long standing that have been objectified, reified, and have seemingly acquired a life of their own. Without reform, these institutions would succumb to the challenges presently mounted against them.

What follows, of course, should not be construed as a "conspir- acy theory" of doctors against society or doctors against women. Although it is clear from their own statements that attachment theorists have extra-scientific interests,45 those interests derive mainly from their proponents' institutional placement in society, rather than from conspiratorial aims. I argue that attachment theory and research that supports it are reasonable responses by responsible people to threats to the institutional order of society, to the network of social control.46 In fact, bonding theory or a functional equivalent of it is, I think, the only institutionally permissible response available to those in positions of responsibility.

Hospital Practices. Klaus and Kennell, for example, wish to change hospital obstetrical practices:




All these clinical reports suggest that the events occurring during the first hours after birth have special significance for the mother. Nursery practices in the modern hospital in the United States do not generally acknowledge this, and, instead, separate mother and infant immediately after birth.... If there were convincing evidence of an early sensitive period in the human being, major changes in hospital care would be necessary....47

Why has an interest in changing hospital practice emerged and why at this particular time?

Is it possible that before the 1970s women expressed no interest in a different form of childbirth and doctors, therefore, had no impetus to change their practices? The research of Richard Wertz and Dorothy Wertz suggests not.48 Birth moved into the hospital around the turn of the century, and thirty to forty years later the first expressions of disenchantment with hospital birth began to appear in the popular press. "Natural childbirth," with its empha- sis on ecstasy, psycho-sexual experience, safety, and a return to nature, caught the public interest in the 1940s, and the number of proponents grew steadily throughout the next three decades.

By the 1970s the thrust for natural childbirth, which had been a loosely organized cultural movement among middle class women, aimed at enhanc- ing their experience at birth, acquired a social and political cast; women of all classes began to organize, to educate one another, and to try to change or avoid the professional and institutional structures that exerted such dom- inance over birth.49

Thus, challenges to hospital practices in childbirth began in earnest thirty-five to forty years ago.

Why did the profession not respond affirmatively? Put simply, it could not. Klaus and Kennell show50 that hospital practices the profession wished to change-early separation of mother and infant and little contact after birth-developed out of "scientific" concern over the introduction of infectious diseases into the newborn nursery. Practices that have their own scientific basis cannot simply be changed in a profession the very existence of which is dependent on the autonomy that science, or at least a facade of scientism, accords it.51 The only way for medicine to change previously adopted practices and, at the same time, maintain political autonomy and control over its social domain is to present convincing scientific evidence a) that previous scientists were wrong, b) that matters have changed sufficiently to contravene the practices suggested by previous scientists, or c) that the risks of previous practices are outweighed by the benefits of new practices. Condition c can occur only when it is shown that more about the underlying physiological processes is known now than previously, that is,




when it is shown that there are more dimensions to the problem on which to assess risks and benefits than existed previously. Investigations showing that risks of infection caused by parents visiting their children are not high permit change because they meet criterion a, while research into bonding permits change because of criterion c. Because of bonding theory, investigators can claim that we know more, in a scientific sense, about the mothering process than we ever knew before. Therefore, change to accom- modate new knowledge becomes permissible when it was not permissible previously.

Change might have been permissible because of the new research, but something that is permissible is not absolutely necessary. The impetus for change came not from research, but from women who threatened obstetricians' livelihoods by taking their obstetrical business elsewhere. In a Wall Street Journal article on the rise of "birthing rooms" and the relaxation of hospital birth rituals and rules, David Stewart, executive director of the National Associa- tion of Parents and Professionals for Safe Alternatives in Childbirth, is quoted as saying: "Hospitals that don't set up a birthing room will go out of the baby business because of the competition. And there's no question that consumer pressure is bringing this about."52 That is, there is economic pressure for a change.

Simply changing in response to women's demands, however, would be impossible from the obstetrical profession's point of view. Women are demanding that birth be treated differently, that it be treated less medically and more like a natural, normal phenomenon. Doctors' control of birth has rested, from the earliest days of the profession, on conceptions of pregnancy and childbirth which treat them as diseases, illnesses, and threats to health. Women's demands for natural birth threaten the very basis of a doctor's work, the ideology on which doctors' control of birth is based. Acceding to demands for natural childbirth is tantamount to giving up those conceptions and ideologies on which one's control, one's work, one's project rest. As pregnancy became natural once again, therefore, it was essential that there be new "diseases" for the doctor to "treat," potentially pathological procesess in which she or he could intervene.

Fraiberg's "diseases of nonattachment" fill the space left by the "naturalization" of childbirth. It is on the basis of this new set of diseases that the doctor's right of intervention in childbirth is preserved: "By understanding the genesis of maternal and attachment behavior, perhaps we can better envision interventions that will foster change in those cases where such is desirable for mother and infant."53 The doctor's job is no longer simply to "deliver" a labor-




ing mother, but now includes "Helping Mothers to Love Their Babies," the title of an editorial in the British Medical Jotrnal.54 By expanding the medical domain to include the social relations of parents and infants, medicine preserves institutionalized rela- tionships between physicians and patients. Hospital practices are reformed to protect the interests that they embody.

Social Policy. The "diseases of nonattachment" are also used as justification for criticizing and reforming social policies involving children. Lack of bonding in early infancy allegedly contri- butes to all sorts of social woes, as Fraiberg states:

These bondless men, women, and children constitute one of the largest aberrant populations in the world today, contributing far beyond their numbers to social disease and disorder. These are people who are unable to fulfill the most ordinary human obligations in work, in friendship, in marriage, and in childrearing.... Where there are no human attachments there can be no con- science. As a consequence, the hollow men and women contribute very largely to the criminal population. It is this group, too, that produces a particular kind of criminal, whose crimes, whether they be petty or atrocious, are always characterized by indifference. The potential for violence and destructive acts is far greater among these bondless men and women; the absence of human bonds leave a free "unbound" aggression to pursue its erratic course.55

Prescriptions for change in the judiciary system, in childcare practices, and in social welfare policy follow naturally from this analysis. Fraiberg asks courts to recognize the drastic implications of "unbinding" children through custody decisions. She suggests that children be assigned advocates in court because "the supporting staff of social workers and psychologists have neither specialized professional training nor the vocational commitment to children which qualifies them as advisors to the court."56 "Child Care Industries, Inc." daycare is issued a "blanket indictment" for providing "anonymous sitters for small children." And it is claimed that public welfare policy regarding children compels choices only among bad alternatives. Bonding theory, therefore, gives social reformers license to prescribe changes to eradicate the "diseases of nonattachment."

But criticism and proposals for reform are only one side of the phenomenon. Bonding theory also provides social institutions a species of relief. The judiciary, for example, is presently under attack for alleged subjectivism and for being fraught with loop- holes that any good lawyer can use. Scientific theory can be used by the courts to create a facade of rationality, objectivity, and fairness based on science. Specifically, bonding theory can be used to assure the public that action taken is in the scientifically




demonstrated "best interests of the child." Appealing to a theory, based on "incontrovertible evidence," is an effective weapon against continued attacks on the system.

This is not the first time scientists have presumed that they have uncovered data which "if properly interpreted, carry their own pre- scription for prevention"57 of general social ills. In fact, the language used by attachment theorists is remarkably similar to that used for over fifty years in the IQ controversy. All the characteristics of the literature under review here were present in the IQ literature years ago-the claim that criminality and mental defectiveness result from one's early development or biological heritage, the claim that social dysfunctionality also results from it, the presumption that admittedly limited evidence is definitive and provides a ground for policy, the concern over "cures," and the presumption that the responsibility for treatment devolves onto the state or its agencies.

In fact, the bonding literature has all the characteristics of a pseudoscience-"a sustained process of false persuasion transacted by simulation or distortion of scientific inquiry and hypothesis testing"-a term Jeffrey Blum uses to characterize the psychomet- ric movement of the early twentieth century.58 Viewing bonding theory as a pseudoscience helps explain its appeal and the paucity of criticism directed toward it. A pseudoscience garners no popular criticism because it "serves as a kind of wish fulfillment, enabling people to discover what they would like to believe."59 It garners no scientific criticism because, according to Blum, "scientists generally do not like polemics, and when conflict occurs they usually try to minimize its importance."60 Bonding provides the basis upon which a judiciary that "Divides the Living Child" (one of Fraiberg's chapter titles) between parents in custody suits may be reformed so that the institution is essentially preserved while only its more disagreeable aspects are eliminated. Bonding also fulfills the desires of people who wish we could do without "Child Care Industries, Inc.," but who are reluctant to say anything against a segment of our society which presumably frees people from the social requirement that children be attended constantly by one parent or the other. And bonding does this on the basis of science. Bonding theory, then, is being used in direct ways to alter social institutions that desperately need real reform. But bonding theory is being used in a more diffuse way, as well, and it is to this more subtle application of the theory that we now turn.

Social Order. In 1916, a paper in the American Journal of Sociology described how major social institutions and aspects of social life-law, art, public opinion, education, and other "devices"-were used to get women to bear and raise children.




Leta Hollingsworth argued that the "social guardians" manipulated social life to increase the birthrate to "secure order, and insure that individuals will act in such a way as to promote the interests of the groups, as those interests are conceived by those who form the rad- iant points of social control."61 In early twentieth-century Amer- ica it was in the national interest to produce children, particularly children from families of means. Women did bear and raise children despite the fact that doing so was "painful, dangerous to life, and involve[d] years of exacting labor and self-sacrifice" and despite the fact that research of the day showed "there [was] no evidence to show that a maternal instinct exists in women of such all con- suming strength and fervor as to impel them voluntarily to seek the pain, danger, and exacting labor involved in maintaining a high birthrate."62 Professor Hollingsworth, in a style appropriate to the sociology of her day, concluded:

The time is coming, and is indeed almost at hand, when all the most intelligent women of the community, who are the most desirable child-bearers, will become conscious of the methods of social control. The type of normality will be questioned; the laws will be repealed and changed; enlightenment will pre- vail; belief will be seen to rest upon dogmas; illusion will fade away and give place to clearness of view; the bugaboos will lose their power to frighten. How will "the social guardians" induce women to bear a surplus population when all these cheap, effective methods no longer work?63

Over sixty years later we see bonding theory, a social-psychological- biological theory, a resurrected if somewhat modified and scientifi- cized version of "maternal instinct," being used to keep women in their place.

Social interests are different today but the social tactics are rather similar. A high birthrate is no longer crucial but continued subordination of women and other groups is crucial to the mainten- nance of the status quo. Bonding theory lends legitimacy to the notion that women are the only appropriate attendants for children. It is an ideological justification for tying women to their own children and keeping them in the home, thereby justifying a social order based on patriarchy.

The social interests served by bonding research are revealed by the methodological biases of the work.64 Consider, for example, some measures of attachment used in bonding research. When their babies were one month old, mothers were asked two questions about their behavior and took the sum of their scores as an indication of successful bonding.65 The first question was: "When baby cries, has been fed, diapers are dry, what do you do?" The respondent received no points if she "always let him cry it out,"




one point if she "tended to let him cry it out," two points if she "tended to pick him up," and the maximum three points if she "always pick[ed] him up." The second question was, "Have you gone out since infant born[sic] ? How did you feel?" The respondent scored zero for going out and feeling good about it, one point for going out and thinking about her child, two points for going out and worrying about her child, and three points for not going out or going out and thinking constantly about her child. The bias is clear. Mothers are better if they stay home and concern themselves entirely with their infants.

Also, the literature on bonding, and child development generally, is characterized by a high degree of "father absence."66 Rossi, for example, does not discuss the potential importance of nonbiolog- ical or nonfemale parents. And Klaus and Kennell focus to a very great degree on maternal-infant bonding. Fraiberg is perhaps the most liberal in terms of recognizing the importance of human attachments, but her review is constrained by the "father absent" character of the literature on which her argument is based. "Fath- er absence" is conspicuous in the face of Clarke-Stewart's general finding that:

it seems likely that in the first six months, at least, fathers and mothers have a parallel influence on their infant's behavior, the relative extent of their influence being determined by the amount and quality of interaction each has with the infant.67

Thus, it does not seem farfetched to assert that bonding theory is prejudiced against women interested in pursuing a life in which children are not the reason for being or the exclusive focus of attention.

Bonding theory, like so many other ideologies posed as social theory, turns social issues into individuals' problems. All of the social ills that concern bonding theorists are reconstrued by bonding theory as the problems of individuals-women-not bonding to their babies. Women are singled out by calling attention to the possible biological bases of bonding and by arguing that it is only women who possess the biological constitution for solving our social problems. Social order can be maintained, it seems, by acquiescing to our biological heritage. Even social theories which propose new forms of social life must be reconsidered in light of our biological makeup. Rossi writes that

the particular version of egalitarianism underlying sociological research on, and advocacy of, "variant" marriage and family forms is inadequate and mis- leading because it neglects some fundamental human characteristics rooted in our biological heritage. Unless these biosocial factors are confronted,




allowed for, and, if desired, compensated for, the current press toward sexual equality in marriage and the work place and shared child rearing may show the same episodic history that so many social experiments have demonstrated in the past.68

Rossi does not argue outright for the maintenance of the status quo, but she does issue a strong warning that those interested in change will have to attend to the biological dimensions of social behavior.

If society wishes to create shared parental roles, it must either accept the high probability that the mother-infant relationship will continue to have greater emotional depth than the father-infant relationship, or institutional- ize the means for providing men with compensatory exposure and training in infant and child care in order to close the gap produced by the physiological experiences of pregnancy, birth, and nursing.69

As one critique has argued persuasively, this is simply another attempt to relocate sex role behavior in the individuality of biological differences.70 This is happening despite the impossibility of making the attributions bonding theory makes and despite the fact that the women's movement and feminist scholarship have rather firmly located sex role behavior in history and society rather than in biology.

CONCLUSION

Fraiberg begins her book with a disrespectful look at developmental psychology through the eyes of her grandmother. She allows that grandmothers knew, in some nonscientific way, everything about childrearing that is now being rediscovered through science. She says: "While my grandmother and I would see eye to eye on a number of major issues in infant psychology we would run into problems on rules of evidence. "If I know," her argument would go, "why do I have to prove it?"71

The way this question is put is deceptive, but it provides a point of departure for specifying precisely the course that the bonding literature has taken. No one has ever asked mothers, or grandmothers, to prove that what they know is correct. In fact, mothers have never been given that chance. Other people, respectable scientists, have developed a theory, investigated a phenomenon, and come to conclusions about the nature of human behavior. The theories and conclusions of scientists are then proffered as what is known about a mother's relationship with her child. The scientist's knowledge is accorded a privileged status and is used to prevent women's knowledge from being heard and considered.




Women are silenced, not through the cunning and deceit of doctors and scientists, but through social institutions bolstered by social scientific theory.

Bonding theory urges us to look inward to our biology for solu- tions to major social ills. Our prototype for order is to be the order apparent in the biological world. Rossi's "radical" proposal for society makes this clear. She wants "a society more attuned to the natural environment, in touch with, and respectful of the rhythm of our own body processes."72 But to suggest that we orient ourselves to our biology is a political act. It is a decision to accept an ideology that provides an individualistic rationaliza- tion for the social problems of crime and violence, and for keeping women by the hearth. Orientation to biology is also disrespectful of the subjective expressions of an individual concerning the way she feels about and experiences herself, her child, and her intimate relations generally. Adoption of the biosocial view of the family presents a "significant retreat from the experimental mode" of social interaction.73 In the words of Marshall Sahlins, in order to embrace the sociobiological paradigm that underlies bonding theory, "We should have to abandon all understanding of the human world as meaningfully constituted, and so the one best hope of knowing ourselves."74

The alternative, as Sahlins implies, is commitment to the notion that humans give meaning and significance to the world, that we continually create and recreate the social world we live in. We live in a world that consists largely of "ideologically frozen relations of dependence that can in principal be transformed."75 Females of the species gestate and deliver children. Humans have, through history, attached meaning to that biological phenomenon and used the phenomenon as the basis for a social order in which women are dependent on and inferior to men. In principle, though, people can act collectively to change the situation. As bonding theory has developed, with its methodological flaws and ideological biases, it is an impediment to change and should be recognized as such. Otherwise, its popularization will negate the process that the scholarship and social criticism of feminism have provided.

NOTES

Thanks to David Armstrong, Joan Smith, and reviewers for Feminist Studies for help- ful comments on earlier drafts of this essay.

This paper was presented at the Medical Sociology Conference, York, England, September 26, 1979.




1 Vermont Educational Television, "New Odds for an Old Gamble?" 1978. Quote is from an independent midwife practicing in Vermont.

2Marshall H. Klaus and John H. Kennell, Maternal-Infant Bonding: The Impact of Early Separation or Loss on Family Development (St. Louis: Mosby, 1976); Edward O. Wilson, Sociobiology: The New Synthesis (Cambridge: Harvard University Press, Bel- knap Press, 1975).

Klaus and Kennell, Maternal-Infant Bonding, p. 22. 4Ibid., p. 28. 5Alice Rossi, "A Biosocial Perspective on Parenting," Daedalus (Spring 1977): 1-31. 6Klaus and Kennell, Maternal-Infant Bonding, p. 37. 7John Bowlby, "Nature of a Child's Tie to His Mother," International Journal of

Psychoanalysis 39 (1958): 350-73; John Bowlby, Attachment and Loss, vol. 1 (New York: Basic Books, 1969); Arthur J. Brodbeck and Orvis C. Irwin, "The Speech Behav- ior of Infants Without Families," Child Development 17 (1946): 145-56; Anna Freud and Dorothy Burlingham, Infants Without Families (New York: International Univer- sities Press, 1944); and Rene A. Spitz, "Hospitalism," Psychoanalytic Study of the Child 2(1946): 113-17.

8Leon J. Yarrow, "Maternal Deprivation: Toward an Empirical and Conceptual Re- evaluation," Psychological Bulletin 58 (1961): 45 9 -90.

9 Alison Clarke-Stewart, Child Care in the Family: A Review of Research and Some Propositions for Policy (New York: Academic Press, 1977), p. 12.

10M. H. Klaus, R. Jerauld, N. C. Kreger, W. McAlpine, M. Steffa, and J. H. Kennell, "Maternal Attachment: Importance of the First Post-Partum Days," New England Journal of Medicine 286 (1972): 460-63.

Ibid., p. 461. 12Norma M. Ringler, John H. Kennell, Robert Jarvella, Billie J. Navojosky, and Mar-

shall H. Klaus, "Mother-to-Child Speech at 2 Years: Effects of Early Postnatal Contact," Journal of Pediatrics 86 (1975): 141-44; Norma M. Ringler, Mary Ann Trause, and Marshall H. Klaus, "Mother's Speech to Her Two Year Old, Its Effect on Speech and Language Comprehension at 5 Years," Pediatric Research 10 (1976): 307.

13Klaus and Kennell, Maternal-Infant Bonding, p. 59. 14Ibid., p. 65-66. I

Clarke-Stewart, Child Care in the Family, p. 6. 16Klaus and Kennell, Maternal-Infant Bonding, p. 2. 17Leslie Jordon Cohen, "The Operational Definition of Human Attachment," Psycho-

logical Bulletin 8 (1974): 207-17; Jacob L. Gewirtz, "Attachment, Dependence, and Distinction in Terms of Stimulus Control," in Attachment and Dependency, ed. Jacob L. Gewirtz (Washington: Winston, 1972); Michael E. Lamb, "A Defense of the Concept of Attachment," Human Development 17 (1974): 376-85; John C. Masters, and Henry M. Wellman, "The Study of Human Attachment: A Procedural Change," Psychological Bulletin 81 (1974): 218-37; Miriam K. Rosenthal, "Attachment and Mother-Infant Interaction: Some Research Impasses and a Suggested Change in Orientation," Journal

of Child Psychology and Psychiatry 14 (1973): 201-207; Tannis MacBeth Williams, "Infant Development and Supplemental Care: A Comparative Review of Basic and Applied Research," and Marsha Weinraub, Jeanne Brooks, and Michael Lewis, "The Social Network: A Reconsideration of the Concept of Attachment," both in Human

Development 20 (1977): 1-47. 8 Rosenthal, "Attachment and Mother-Infant Interaction," pp. 201-2.

19Weinraub, Brooks, and Lewis, "Social Network." 20Andrew Whiten, "Assessing the Effects of Perinatal Events on the Success of

Mother-Infant Relationship," in Studies in Mother-Infant Interaction, ed. H. R. Schaffer (London: Academic Press, 1977), pp. 403-25.

21 Masters and Wellman, "Study of Human Attachment," p. 218.




22Ibid., p. 228. 23A. D. Leifer, P. H. Leiderman, C. R. Barnett, and J. A. Williams, "Effects of Mother-

Infant Separation on Maternal Attachment Behavior," Child Development 43 (1972): 1203-18.

24Ibid., pp. 1213-14 (italics mine). 25Ibid., p. 1214. 26Klaus, Jerauld, "Maternal Attachment." 27Charles Henry Kempe, and Ray E. Heifer, Helping the Battered Child and His Fam-

ily (Philadelphia: J. B. Lippincott, 1972); Ray E. Heifer, "High-Risk Signals in the Pre- natal Setting, and in the Post-Partum Period With Follow-ups Recommendations," personal correspondence with Dr. Helen Krell, 1976.

28See Judith Bernal Dunn and M. P. M. Richards, "Observations on the Relationship Between Mother and Baby in the Neonatal Period," in Studies in Mother-Infant Inter-

action, ed. H. R. Schaffer (London: Academic Press, 1977), pp. 427-55, for this criticism and an exception..

29Selma Fraiberg, Every Child's Birthright: In Defense of Mothering (New York: Basic Books, 1977), p. 56 (italics mine).

30Rossi, "Biosocial Perspective," pp. 3-4. 31Ibid., p. 5. 32Ibid., p. 5 (italics mine). 33Fraiberg, Every Child's Birthright, p. 3. 34Ibid. (italics mine). 35Ibid., p. 31-32. 36Ibid., p. 34. 37Ibid., p. 49. 38Ibid., p. 47. 39Robert W. ten Bensel and Charles L. Paxon, "Clinical Notes: Child Abuse Following

Early Postpartum Separation," Journal ofPediatrics 90 (1977): 490-91; Margaret A.

Lynch, "Ill-health and Child Abuse," Lancet 2 (1975): 317-19; and Avroy A. Fanaroff, John H. Kennell, and Marshall H. Klaus, "Follow up of Low Birth Weight Infants: The Predictive Value of Maternal Visiting Patterns," Pediatrics 49 (1972): 287-90.

40Marshall Sahlins, The Use and Abuse of Biology: An Anthropological Critique of Sociobiology (Ann Arbor: University of Michigan Press, 1976), pp. xiv-xv.

41 Rossi, "Biosocial Perspective," p. 4.

42Sahlins, Use and Abuse of Biology, p. 81. See also the critique by Jill S. Quadagno, "Paradigms in Evolutionary Theory: The Sociobiological Model of Natural Selection," American Sociological Review 44 (1979): 100-109.

43Robert K. Merton, "The Normative Structure of Science," in The Sociology of Science, ed. Robert K. Merton (Chicago: University of Chicago Press, 1973); Andre Cournand and Michael Meyer, "The Code of the Scientist and its Relationship to Ethics," Science 198 (1977): 699-705; and Andre F. Cournand and Harriet Zucker- man, "The Code of Science: Analysis and Some Reflections on its Future," Studium Generale 23 (1970): 941-62.

4Fraiberg, Every Child's Birthright, p. 3 (italics mine). 45See Abraham Kaplan, The Conduct of Inquiry: Methodology for Behavioral Science

(San Francisco: Chandler, 1964), for a philosophical discussion of how the extra-scientific interests influence all aspects of social science inquiry. More recent publications discussing this issue include Alfred McLung Lee, Sociology for Whom? (New York: Oxford University Press, 1978); and Barry Barnes, Interests and the Growth of Knowl- edge (London: Routledge and Kegan Paul, 1977).

46Peter L. Berger, and Thomas Luckman, The Social Construction of Reality: A Treatise in the Sociology of Knowledge (New York: Doubleday, 1966), is still the best discussion of this issue.




47Klaus and Kennell, Maternal-Infant Bonding, p. 53. 48Richard M. Wertz and Dorothy C. Wertz, Lying-In: A History of Childbirth in

America (New York: Free Press, 1977). 49Ibid., p. 179. 50Klaus and Kennell, Maternal-Infant Bonding, pp. 6ff. 51 Jeffrey Lionel Berlant, Profession and Monopoly: A Study of Medicine in the

United States and Great Britain (Berkeley: University of California Press, 1975). 52Joann S. Lublin, "The Birthing Room: More Hospitals Offer Maternity Facilities

That Feel Like Home," Wall Street Journal 193 (15 February 1979): 1, 22. 53Klaus and Kennell, Maternal-Infant Bonding, p. 41. 54

Leader, "Helping mothers to love their babies," British Medical Journal, September 3, 1977.

55 Fraiberg, Every Child's Birthright, p. 62. 56Ibid., p. 76. 57Ibid., p. 62. 58Jeffrey M. Blum, Pseudoscience and Mental Ability: The Origins and Fallacies of

the IQ Controversy (New York: Monthly Review Press, 1978). 59Ibid., p. 146. 60Ibid., p. 156. 61 Leta S. Hollingsworth, "Social Devices for Impelling Women to Bear and Rear Chil-

dren," American Journal of Sociology 22 (1916): 19-29, p. 21 (italics in original). 62Ibid., pp. 20-21. 63Ibid., p. 29. 64Bias is different from error. Research can be well done and yet biased. Methodol-

ogical errors were discussed earlier. Errors are important in this context because they force one to ask why rules of inquiry were violated. Bias points one in the direction of answers to this crucial question.

65Klaus, "Maternal Attachment." 66Wini Breines, Margaret Cerullo, and Judith Stacey, "Social Biology, Family Studies,

and Antifeminist Backlash," Feminist Studies 4, no. 1 (February 1978): 43-67.

67Clarke-Stewart, Child Care in the Family, p. 21. 68Rossi, "Biosocial Perspective," p. 2. 69Ibid., p. 18. 70Breines, Cerullo, and Stacey, "Social Biology." 71 Fraiberg, Every Child's Birthright, p. 6. 72Rossi, "Biosocial Perspective," p. 25. 73Breines, Cerullo, and Stacey, "Social Biology," p. 43. 74Sahlins, Use and Abuse of Biology, p. 107. 75 Jurgen Habermas, Knowledge and Human Interests (Boston: Beacon Press, 1971),

p. 310. This discussion of the social uses of theory draws on the excellent appendix in this book.



maternal-infant bonding: the politics of falling in love with your child

Documents