American Psychologist

Deconstructing the Essential Father

Louise B. Silverstein and Carl F. Auerbach
Yeshiva University
American Psychologist June, 1999 Vol. 54, No. 6, 397-407

Neoconservative social scientists have claimed that fathers are essential to positive child development and that responsible fathering is most likely to occur within the context of heterosexual marriage. This perspective is generating a range of governmental initiatives designed to provide social support preferences to fathers over mothers and to heterosexual married couples over alternative familv forms. The authors propose that the neoconservative position is an incorrect or oversimplified interpretation of empirical research. Using a wide range of cross-species, cross-cultural, and social science research, the authors argue that neither mothers nor fathers are essential to child development and that responsible fathering can occur within a varietv of family structures. The authors conclude with alternative recommendations for encouraging responsible fathering that do not discriminate against mothers and diverse family forms.

In the past two decades, there has been an explosion of research on fathers (see Booth & Crouter, 1998; Lamb, 1997, Phares, 1996, for recent reviews). There is now a broad consensus that fathers are important contributors to both normal and abnormal child outcomes. Infants and toddlers can be as attached to fathers as they are to mothers. In addition, even when fathers are not physically present, they may play an important role in their children's psychological lives. Other important issues about fathers and families remain controversial. For example, scholars continue to debate the extent to which patemal involvement has increased over the past 20 years (Pleck, 1997). Similarly, researchers are only beginning to study the ways that fathering identities vary across subcultures (Auerbach, Silverstein, & Zizi, 1997; Bowman & Forman, 1998; Roopnarine, Snell-White, & Riegraf, 1993), and the effects of divorce on fathers and their children are not yet clearly understood (Hetherington, Bridges, & Insabella, 1998).

Overall, this explosion of research on fathering has increased the complexity of scholarly thinking about parenting and child development. However, one group of social scientists (e.g., Biller & Kimpton, 1997; Blankenhorn, 1995; Popenoe, 1996) has emerged that is offering a more simplistic view of the role of fathers in families. These neoconservative social scientists have replaced the earlier "essentializing" of mothers (Bowlby, 1951) with a claim about the essential importance of fathers. These authors have proposed that the roots of a wide range of social problems (i.e., child poverty, urban decay, societal violence, teenage pregnancy, and poor school performance) can be traced to the absence of fathers in the lives of their children. Biller and Kimpton (1997, p. 147) have even used the term "paternal deprivation" in a manner parallel to Bowlby's (1951) concept of maternal deprivation. In our view, the essentialist framework represents a dramatic oversimplification of the complex relations between father presence and social problems.

We characterize this perspective as essentialist because it assumes that the biologically different reproductive functions of men and women automatically construct essential differences in parenting behaviors. The essentialist perspective defines mothering and fathering as distinct social roles that are not interchangeable. Marriage is seen as the social institution within which responsible fathering and positive child adjustment are most likelv to occur. Fathers are understood as having a unique and essential role to play in child development, especially for boys who need a male role model to establish a masculine gender identity (see Table I for a definition of the essentialist perspective).

Our research experience has led us to conceptualize fathering in a way that is very different from the neoconservative perspective. Over the past six years, we have studied the fathering identities of men who are actively involved with their children (Auerbach & Silverstein, 1997, Auerbach et a]., 1997; Silverstein, 1996: Silverstein. Auerbach, Grieco, & Dunket, in press; Silverstein & Phares, 1996; Silverstein & Quartironi, 1996). To date, approximately 200 men from 10 different subcultures within U.S. society have participated in this qualitative research. Our research participants include Haitian Christian fathers; Promise Keeper fathers; gay fathers; Latino fathers; White, nongay divorced fathers; Modern Orthodox Jewish fathers; and Greek grandfathers.

DA*DI noted Father-absence in the
March, 1998 Frontspiece for the new
APA Journal of Family Psychology
In contrast to the neoconservative perspective, our data on gay fathering couples have convinced us that neither a mother nor a father is essential. Similarly, our research with divorced, never-married, and remarried fathers has taught us that a wide variety of family structures can support positive child outcomes. We have concluded that children need at least one responsible, caretaking adult who

Editor's note. Cheryl B. Travis served as action editor for this article.

Author's note. Louise B. Silverstein and Carl F. Auerbach. Ferkauf Graduate School of Psychology, Yeshiva University.
Correspondence concerning this article should be addressed to Louise B. Silverstein, 99 Clinton Street, Brooklyn, NY 11201. Electronic mail mav be sent to

June 1999 - American Psychologist
Copyright 1999 by the American Psychological Association, Inc. 0003-066X/99/S2.00 Vol 54, No. 6. 397-407


has a positive emotional connection to them and with whom they have a consistent relationship. Because of the emotional and practical stress involved in child rearing, a family structure that includes more than one such adult is more likely to contribute to positive child outcomes. Neither the sex of the adult(s) nor the biological relationship to the child has emerged as a significant variable in predicting positive development. One, none, or both of those adults could be a father (or mother). We have found that the stability of the emotional connection and the predictability of the caretaking relationship are the significant variables that predict positive child adjustment.

We agree with the neoconservative perspective that it is preferable for responsible fathers (and mothers) to be actively involved with their children. We share the concem that many men in U.S. society do not have a feeling of emotional connection or a sense of responsibility toward their children. However, we do not believe that the data support the conclusion that fathers are essential to child well-being and that heterosexual marriage is the social context in which responsible fathering is most likely to occur.

Many social scientists believe that it is possible to draw a sharp distinction between scientific fact and political values. From our perspective, science is always structured by values, both in the research questions that are generated and in the interpretation of data. For example, if one considers the heterosexual nuclear family to be the optimal family structure for child development, then one is likely to design research that looks for negative consequences associated with growing up in a gay or lesbian parented family. If, in contrast, one assumes that gay and lesbian parents can create a positive family context. then one is likely to initiate research that investigates the strengths of children raised in these families.

The essentialist theoretical framework has alreadv generated a series of social policy initiatives. For example, a 1998 congressional seminar recommended a series of revisions to the tax code that would reward couples who marry, and end taxes altogether for married couples with three or more children (Wetzstein. 1998). Other federal legislation has emerged with a similar emphasis on the advantages of marriage. The 1996 welfare reform law begins by stating, "Marriage is the foundation of a success-

Table 1
The Essentialist Paradigm

Essentialist Belief


Biological sex differences construct
gender differences in parenting

The biological experiences of pregnancy and lactation generate a strong, instinctual drive in women to nurture. In the absence of these experiences, men do not have an instinctual drive to nurture infants and children.

The civilizing effects of marriage Because a man's contribution to reproduction is limited to the moment of conception, active and consistent parenting on the part of men is universally difficult to achieve.

The best way to ensure that men will consistently provide for and nurture young children is to provide a social structure in which men can be assured of paternity (i.e., the traditional nuclear family). Without the social institution of marriage, men are likely to impregnate as many women as possible, without behaving responsibly toward their offspring.

The importance of a male role model

If men can be induced to take care of young children, their unique, masculine contribution significantly improves developmental outcomes for children. This is especially true for boys who need a male role model to achieve psychologically healthy masculine gender identity.

398 _________________________________ June 1999, American Psychologist

ful society" (Welfare Reform Act, 1996, p. I 10). Similarly, a housing project in Hartford, Connecticut now provides economic supports to married couples and special opportunities for job training to men (but not to women) who live with their families (LaRossa, 1997). In 1997, Louisiana passed a Covenant Marriage Act (1997) that declared marriage a lifelong relationship and stipulated more stringent requirements for separation and divorce.

The social policy emerging out of the neoconservative framework is of grave concern to us because it discriminates against cohabiting couples, single mothers, and gay and lesbian parents. The purpose of the current article is to present a body of empirical data that illustrates the inaccuracv of the neoconservative argument. Throughout our discussion, we focus on the work of Blankenhorn (1995) and Popenoe (1996) because they have been most influential in structuring both public debate and social policy (Haygood, 1997; Samuelson, 1996).

Specific aspects of the neoconservative paradigm have been critiqued elsewhere. For example, McLoyd (1998) has pointed out that families without fathers are likely to be poor, and it is the negative effects of poverty, rather than the absence of a father, that lead to negative developmental outcomes. Similarly, Hetherington et al. (1998) have made the point that divorce does not always have negative consequences for children.

However, the neoconservative argument as a whole has not been deconstructed. Thus, it tends to be absorbed in a monolithic fashion, buttressed by unconscious gender ideology and traditional cultural values. Therefore, we think that a systematic counterargument is necessary. We cite research indicating that parenting roles are interchangeable, that neither mothers nor fathers are unique or essential, and that the significant variables in predicting father involvement are economic, rather than marital. We also offer an alternative framework for encouraging responsible fathering.

We acknowledge that our reading of the scientific literature supports our political agenda. Our goal is to generate public policy initiatives that support men in their fathering role, without discriminating against women and same-sex couples. We are also interested in encouraging public policy that supports the legitimacy of diverse family structures, rather than policv that privileges the two-parent, heterosexual, married familv.

We also realize that some of the research we cite to support our perspective will turn out to be incorrect. Harawav (1989) pointed out that as research paradigms evolve to reflect diverse gender, ethnic, class, and culturral perspectives, much of the established body of "scientific fact" has turned out to be science fiction. Fischhoff (1990) identified two options for psychologists in the public arena: helping the public define their best interests or manipulating the public to serve the interests of policvmakers. Thus, despite the fact that new data will inevitably prove some aspects of our argument wrong, we hope that by stimulating scholarly debate, we will contribute to the process by which the public more accurately defines its best interests.

We begin by presenting cross-species and crosscultural data that contradict the claim that parenting behaviors are constructed by biological differences. We argue that parenting involves a series of caregiving functions that have developed as adaptive strategies to specific bioecological contexts. These caregiving functions can be performed by parenting figures of either sex, whether or not they are biologically related to the child.

We then review the research on marriage and divorce. This body of data suggests that the poor psychological adjustment observed in some children in divorcing families is caused by the disruption of the child's entire life circumstances, rather than simply by the dissolution of the marriage or the absence of a father. We present data illustrating that emotionally connected, actively nurturing, and responsible fathering can occur within a variety of family structures.

Finally, we examine why the neoconservative perspective has been so widely accepted within popular culture. We speculate that the appeal of neoconservative ideology is related to two social trends: a genuine concem about children and a backlash against the gay rights and feminist movements. We then offer social policy recommendations that support men in their fathering role, without discriminating against women and same-sex couples.

The Essentialist Position

Biological Sex Differences Construct Gender Differences in Parenting

One of the cornerstones of the essentialist position is that biological differences in reproduction construct gender differences in parenting behaviors. This theoretical framework proposes that the biological experiences of pregnancy and lactation generate a strong instinctual drive in women to nurture. This perspective assumes that men do not have an instinctual drive to nurture infants and children.

399 __________________________________June, 1999 American Psychologist

The neoconservative perspective relies heavily on evolutionary psychology to support this argument. Evolutionary psychologists cite Trivers's (1972) sexual-conflict-of-interest hypothesis to explain sex differences in mating strategies. Trivers's hypothesis states that all other things being equal, male animals will maximize their evolutionary fitness by impregnating as many females as possible, while investing very little in the rearing of any individual offspring. Female mammals, in contrast, invest a great deal of physiological energy in pregnancy and lactation and thus are motivated to invest a corresponding amount of time and energy in parenting.

Trivers's (1972) hypothesis accurately predicts behavior in many mammalian species. However, Smuts and Gubemick (1992) have shown this hypothesis to he inaccurate in predicting male involvement with infants among nonhuman primates. Unfortunately, Smuts and Gubernick's critique of the relevance of Trivers's hypothesis for primate behavior has not been integrated into evolutionary psychological theory.

Evolutionary psychology has recently gained prominence within psychology and other social sciences (e.g., Archer, 1996. Buss, 1995). Because the formal academic training of most social scientists does not include crossspecies research and evolutionary theory, many social scientists have accepted evolutionary psychologists' use of Trivers's (1972) hypothesis in relation to primate behavior. However, many scholars within the natural science community have been critical of evolutionary psychology (see, e.g., the more than 20 negative commentaries on Thomhill & Thomhill's, 1992, article or Gould's, 1997, critique of evolutionary psychology).

Blankenhorn (1995) and Popenoe (1996), like many social scientists, have incorrectly assumed that Trivers's (1972) theory is true of all primates and is universally applicable across many different ecological contexts. However, all other things have generally not been equal over the course of evolutionary history. As bioecological contexts change, so do fathering behaviors, especially among male primates.

Marmosets are an extreme example of primates who live in a bioecological context that requires males to become primary caretakers (Smuts & Gubemick, 1992). Because marmosets always have twins, female marmosets must nurse two infants simultaneously. This generates nutritional pressure for the mother to spend all of her time and energy feeding herself. Therefore. the father most commonly performs all parenting behaviors. Thus, these animals do not conform to Trivers's (1972) hypothesis about the universality of nonnurturing male primates. Male marmosets behave like full-time mothers.

Marmosets illustrate how, within a particular bioecological context, optimal child outcomes can be achieved with fathers as primary caretakers and limited parenting involvement by mothers. Human examples of this proposition include single fathers (Greif & DeMaris, 1990), two-parent families in which the father is the primary caretaker (Pruett, 1989), and families headed by gay fathers (Patterson & Chan, 1997).

Another cornerstone of the essentialist position is that the traditional division of labor characteristic of Western, industrialized societies has been true throughout human evolutionary history. Popenoe (1996) stated that our hominid ancestors "had a strong division of labor in which males did most of the hunting and females did most of the gathering" (p. 167). Zihlman (1997), in contrast, has pointed out that for most of our evolutionary history, human societies were nomadic.

This bioecological context required both men and women to travel long distances, hunt, gather food, and care for older children and other members of their community. Similarly, in contemporary foraging and horticultural societies, women perform the same range of tasks as men do and add infant care to their other responsibilities. Cross-cultural research illustrates that women are capable of traveling long distances, carrying heavy loads, and participating in hunting. Thus, the assertion that a rigid sexual division of labor existed over most of our evolutionary history is not supported either by what is known about human society in prehistory or by contemporary preagricultural cultures.

The neoconservative perspective has also assumed that providing has been a universal male role. Yet Nsamenang (1992) pointed out that in many West African rural cultures, tradition places the sole responsibility for providing food on mothers. Similarly. in hunting-gathering cultures, women typically provide 60% of a faniily's nutritional requirements (Zihlman, 1997). Thus, in most preindustrial cultures, fathers have never been sole providers, and in some cultures they do not participate at all in the provider role.

The neoconservative perspective has also claimed that mothers are more "natural" caregivers than fathers. Yet, more than a decade ago, Lamb (1987) reported that research on mothers and fathers during the newborn period yielded no differences in parenting behaviors. Neither mothers nor fathers were natural parents. Because mothers tended to spend so much more time with their infants, they became much more familiar with their biological rhythms, visual and behavioral cues, and so forth. Therefore, when observations were repeated after a year, mothers appeared as much more competent caregivers than fathers. Many subsequent studies have shown that when fathers assume the primary caretaking role. they are as competent and as sensitive as mothers (Lamb, 1997).

In summary, the neoconservative position is simply wrong about the biological basis of observed differences in parenting behaviors. Cross-species and cross-cultural data indicate that fathering can vary from a high level of involvement to a total lack of involvement. Given these wide variations in paternal behaviors, it is more accurate to conclude that both men and women have the same biological potential for nurturing and that the sexual division of labor in any culture is defined by the requirements of that culture's specific bioecological context.

Marriage Matters

The neoconservative perspective has argued that without a biological basis for nurturing in men, the best way to

400 __________________________________ June, 1999 American Psychologist

ensure that men will behave responsibly toward their offspring is to provide a social structure in which men can be assured of patemity (i.e., the traditional nuclear family).

Nonhuman primate behavior. This point of view is based on a corollary of Trivers's (1972) sexual conflict-of-interest hypothesis, the paternity hypothesis. Trivers reasoned that without paternity certainty, male primates would not risk investing time and energy in another male's offspring, thereby decreasing their own evolutionarv fitness.

However, Smuts and Gubernick (1992) have demonstrated that Trivers's (1972) paternity hypothesis is not generally predictive of fathering behavior among nonhuman primates. If paternity certainty were the most significant variable. then male primates should show greater paternal involvement in species where several females live with only one breeding male. In species where several males and several females live together (and therefore multiple mating opportunities make patemity uncertain), males should have lower paternal involvement.

The paternity hypothesis does correctly predict male care of infants in most monogamous species. In most monogamous mating pairs, males provide a high level of care. However, the paternity hypothesis does not accurately predict male care in other primate social groupings. With the exception of mountain gorillas, males in one-male groups (where paternity is certain) show less paternal involvement than males in multimale groups (where paternity is uncertain).

Smuts and Gubemick (1992) found that the amount of time and energy males invest in nurturing and protecting infants varies depending on the mutual benefits that males and females have to offer each other within a particular bioecological context. These authors offered an alternative hypothesis, the "reciprocity hypothesis," to account for variations in male care of infants. The reciprocity hypothesis predicts that male care of infants will be low when either males or females have few benefits to exchange. The probability of high male care of infants increases when females have substantial benefits to offer males (e.g., when females can offer to mate more frequently with specific males or can provide males with political alliances that enhance their status within the male dominance hierarchy).

Smuts and Gubernick (1992) found that male care of infants is lower in one-male groups because this system of reciprocal benefits does not exist. Each female has no alternative except to mate with the single male, whether or not he cares for her infants. Because she has no other mating possibilities, she cannot offer preferential mating opportunities in exchange for infant care. Similarly, in a one-male group, the breeding male does not have to compete for a place within a male dominance hierarchy. He is the only male in the group. Therefore, females cannot offer political assistance to enhance his dominance ranking. Because females lack benefits to offer males in exchange for infant care, male involvement, in contrast to what would be predicted by the patemity hypothesis, is low in one-male-groups.

Overall, a very large body of animal research points to the importance of an array of variables, which we refer to as bioecological context, in determining parenting behaviors. Low levels of infant care do not characterize all male primates, nor is biological paternity the most significant variable in increasing the probability of high male involvement. Other feminist anthropologists and sociobiologists have similarly deconstructed Trivers's (1972) theory (e.g., Gowaty, 1997; Hrdy, 1997). In contrast to Trivers's emphasis on universal sex differences and the relative fixity of behaviors, these feminist researchers have pointed to the overlap of behaviors between the sexes and the relative flexibility of complex human behaviors. Unfortunately, this feminist scholarship has not been integrated into most social science literature.

Human primate behavior. Smuts and Gubernick (1992) have made a strong case for the power of the reciprocity hypothesis to predict male involvement among nonhuman primates. However, does their hypothesis predict human primate behavior? We argue that the reciprocity hypothesis does predict male involvement among human primates.

In cultures where women have significant resources to offer men in exchange for child care, paternal involvement should be higher than in cultures where women have fewer resources. In line with this prediction, paternal involvement in the United States, Sweden, and Australia is higher than in more traditional cultures, such as Italy and Spain, where women's workforce participation is less widespread (Blossfeld, 1995). Similarly, Haas (1993) reported that a survey of more than 300 Swedish families indicated that fathers participated more in child care if their partners made as much or more money than they did.

Erikson and Gecas (1991) have provided examples of how paternal involvement varies on the basis of the benefits men have to exchange. These authors pointed out that the least amount of father involvement in U.S. society has been observed in two groups of fathers: poor, unmarried teenage fathers and upper-class fathers in traditional nuclear families. Teen dads in U.S. society are often undereducated and underemployed. Therefore, they cannot make a meaningful contribution to the economic security of their children. Poor teen fathers do not have meaningful benefits to offer their child's mother. As the reciprocity hypothesis would predict, these fathers are often minimally involved in the lives of their children.

Continued at APA-2