Chapter Six—
Fathers and Difference

Whereas Chodorow suggests that women's mothering lies behind women's being more relational than men, I stress the way in which fathers, not mothers, have influenced the specifically "sexual" differences between women and men. Fathers, I maintain, have more to do with how men and women relate to each other in heterosexual contexts than mothers do. In order to see how Chodorow's and my positions relate to each other, it helps to understand that Chodorow is focusing on what psychoanalytic theory calls the preoedipal period and the mother, and I am concerned with the oedipal period and the father.^[1]

Chodorow argues that mothers differentiate between their male and female children. I argue that fathers differentiate more than mothers between their male and female children and in ways different from the mothers greater bonding with daughters. Fathers, I claim, influence the specifically heterosexual aspects of interaction more than mothers do, and these adult heterosexual interactions are characterized by male dominance. By heterosexual interaction, I mean all interaction that takes place between males and females, but, beyond "cross-gender interaction," I also imply that these relationships have a sexual or potentially sexual component and that they also tend to be male-dominated in adulthood. Heterosexual relations are not inherently male-dominated, but they tend to become so for psychological and structural reasons. Mothers may indeed be part of the reason for women's more relational personalities, but that does not determine in itself whether a woman would hold the power in any given relationship. Fathers, as family heads, I argue, are central to the genesis of male dominance

in heterosexual relationships and to women's tendency to seek to marry males who are superior to them. That is, fathers train women to be wives whether they intend to or not.

My interest in the hypothesis that men have a tendency to stress sex differences (I use the term sex here to imply especially the heterosexual aspects of gender difference) arose in connection with my interest in the father-daughter relationship. At the time I was not "seeing" male dominance nor male bonding but rather was concerned with investigating parental influences on expressive and instrumental orientations in college women. To my surprise at the time, the differences I found among women had more to do with their relationships to their fathers than to their mothers. I then discovered by canvassing the literature concerning parental influences on gender-related phenomena that fathers seemed to behave more differently toward sons and daughters than mothers did and were responded to by sons and daughters more differently than were mothers. The facts had not been recognized nor emphasized by researchers, in part because there was no theory that could explain them or into which they could be assimilated.

In 1963, I wrote an article, "Sex Role Learning in the Nuclear Family," in which I attempted to make sense of the findings concerning fathers by using the reciprocal role interaction framework (see Chapter 5). My argument was that children of both genders first learned expressiveness in interaction with the mother in a reciprocal relationship whose main content was mutual "love." I then argued that fathers entered the child's purview somewhat later and introduced differentiation between boys and girls by further reinforcing an expressive orientation in the daughter and encouraging a less expressive (at the time I called it instrumental) orientation in the son. Although the daughters interaction with her father is expressive, it also mimics the husband-wife relationship in which the daughter is the subordinate appreciator of the father. The son is also subordinate to the father, but in this relationship the son is being trained not so much to admire the father as to be like the father, in the sense of taking the attitude of the father toward projects, problems, and females themselves. I also pointed out that mothers' interaction with their sons is not the mirror image of the father-daughter relationship. A mother does not play wife to her son to the extent that the father plays husband to daughter.

In 1963, I described the above facts as the learning of "appropri-

ate sex roles," but I now see that these same facts are intimately related to male dominance and learning one's place in a male-dominant society. From this point of view the father-daughter relationship becomes problematic for me as a feminist, because it reinforces male dominance. The mother-son relationship works against this dominance, however, because the mother is dominant when the son is a child. Thus in a male-dominant society, the male's learning to be masculine by interacting with his mother introduces a contradiction, since the mother-son relationship reverses the dominance situation that characterizes adult heterosexual relationships. Although mothers and sons may sometimes interact in a way that mimics an adult heterosexual relationship, the mother-son relationship is clearly not the prototype of such a relationship.^[2]

The above observations have a parallel in mother-son incest being both less frequent and more tabooed than father-daughter incest. In fact, the social psychologist Serge Moscovici argues that because the mother-son relationship involves female dominance, the fundamental reason for the incest taboo attached to that relationship is to destroy the power of the mother and install male dominance. In essence, he says that if sons were allowed to marry their mothers, women, not men, would be the dominant group. Moscovici contends that the only true incest taboo is that between mother and son and that the fear of—and, I might add, fascination with—mother-son incest is inspired by the fear of female dominance.^[3] From a developmental standpoint, male shows of dominance over women as wives is a method of minimizing males' incapacity to dominate women as mothers.

Father-daughter incest, in contrast, is the acting out of male dominance. Although daughters are not allowed to marry their fathers, they come much closer to it than sons come to marrying their mothers. Not only does father-daughter incest occur much more frequently than mother-son incest, but also the connotations of being a daddy's girl are far more socially acceptable than the connotations of being a mama's boy. A daddy's girl is likely to be considered cute and "feminine"; a mama's boy is seen as dependent, and definitely not masculine. These connections are basic to the argument that masculinity and femininity in contemporary society are defined in terms of dominance and nondominance in crossgender relationships. Moreover, in the father-daughter relation-

ship, the dominance aspects of adult heterosexual relationships are exaggerated, because the father is both male and a generation older than the daughter.

In sum, I am suggesting that adult heterosexual relationships are structured in terms of male dominance, if other contravening sources of power, such as social class, are not operating. This male dominance is fostered within nuclear families by the difference between the father-daughter relationship and the father-son relationship. Compared to the father-child relationship, the mother-child relationship is relatively undifferentiated by gender. Chodorow may be right about differing emotional ties between mothers and sons and mothers and daughters, but the evidence I present indicates that father-son and father-daughter ties differ much more. The description of research that follows will first simply document in various ways the general proposition that fathers are the focus of a certain type of differentiation. I will then show that the type of differentiation and male dominance that fathers foster has to do with the specifically sexual aspects of gender difference.

Examining the Father-Differentiating Hypothesis

The evidence has continued to accumulate since I brought it together in 1963 that fathers differentiate between sons and daughters more than mothers do. Although mothers clearly respond differently to their male and female children, the tendency to respond differently is more marked in fathers. Moreover, male and female children respond more differently to fathers than they do to mothers. The greater differentiation that fathers make between their sons and daughters and the greater differentiation sons and daughters make in their responses to fathers has been found in experimental and observational studies and in unconscious and conscious behavior and statements. These findings also seem to hold regardless of the age of the child. Although the associations are rarely strong and are not always found, the persistence of the trend in various types of studies suggests that they reflect an underlying phenomenon, namely, male dominance and male control over female sexuality.

This basic phenomenon is overlaid by other factors. First, in the

middle class in the United States at least, there is a remarkable lack of differential treatment of children of either gender by either parent, and also few behavioral differences between males and females have been found.^[4] Certainly compared to other societies, children in the United States are treated remarkably similarly and do not differ greatly from each other on the basis of gender, at least with respect to the variables ordinarily measured. Also in the larger picture, fathers not only have much less contact with children than mothers do, but also their style of interacting with children differs from that of mothers regardless of the child's gender. Fathers are more likely to engage in play activities than in routine caretaking, and they interact more briefly and intensely and physically with their children than do mothers. It is within these larger trends that the greater tendency for fathers to differentiate between genders occurs.

The greater differentiation on the part of fathers occurs even among parents who believe in "nonsexist" child-rearing and who conscientiously try to minimize gender-differentiating interactions. Moreover, this more-differentiating behavior of fathers cannot be explained by the fathers' lesser experience than mothers with early child care. Even when fathers are experienced primary caretakers, the differentiating behaviors continue.^[5] These findings fit with the commonsense perception that men and women are different, and they suggest that there may be some unconscious basis for some of these tendencies.

Generally, the gender-differentiating behavior of fathers takes two forms: withdrawal from girls (or preference for boys) and a differentiated style of interaction with girls. Fathers explicitly state that they feel more responsibility toward a male child than toward a female child. Generally, fathers want to have male offspring, and in the event of divorce they are more likely to maintain frequent contact with their boys than with their girls.^[6] Generally, fathers pay more attention to male children than to female children, both rewarding and punishing boys more than they do girls. For example, Margolin and Patterson, in a careful observational study of fourteen families with a boy and a girl, showed that although mothers did not differ in their responses to their sons and daughters, fathers provided almost twice as many positive responses to their sons as

to their daughters.^[7] Boys receive more punishment from both mothers and fathers than girls do, but fathers are more likely to physically punish boys than girls. Mothers tend to use indirect or more psychological methods on both genders.^[8]

Studies concerning infants find that fathers touch and vocalize to their newborn sons more frequently and contingently than to their newborn daughters and that fathers' relative withdrawal from daughters becomes much more pronounced during the second year of life, when fathers become twice as active in interacting with their sons as with their daughters. Mothers do not withdraw from sons nor step up their interaction with daughters in a complementary fashion.^[9]

A recent study of adult-child verbal interaction, using trained child confederates in order to control for the influence of children themselves, found that women did not respond differently on the basis of the child's gender in either an unstructured or a structured situation. Men, in contrast, made a substantial differentiation on the basis of gender; they spoke more often and longer with boys than with girls in the unstructured situation. In the structured situation men spoke more to both boys and girls than women did.^[10] Here the father's actions might be interpreted as modeling for and being a fellow male with the boy and paying less heed to the girl.

Going beyond the question of amount of interaction, studies that have included both fathers and mothers and sons and daughters show that fathers both behave differently and reinforce gender-stereotypical behaviors more than mothers do. Fathers have more gender-stereotyped attitudes toward newborns than mothers do, and fathers are more likely to encourage gender-stereotyped toys and play behaviors.^[11]

Fathers, more than mothers, interact in a different manner with boys and girls and "come across" to them differently. Alfred Heilbrun of Emory University found that brothers and sisters had differing reactions to the same father, with girls perceiving him as being more nurturant than boys saw him to be. Sons tended to describe their fathers in terms of adjectives depicting domination and control, such as aggressive, assertive, autocratic, confident, dominant, forceful, outspoken, and strong. Sons also use adjectives associating their fathers with goal attainment, such as deliberate,

enterprising, foresighted, industrious, inventive, and shrewd. Daughters, by contrast, described their fathers as forgiving, modest, praising, sensitive, talkative, and warm. Brothers and sisters mostly agreed with each other in their perceptions of their mothers, though, and the disagreements that showed up had no consistent pattern.^[12] Heilbrun's findings strongly suggest that boys associate fathers with dominance and instrumental qualities and that girls see fathers as expressive and nondemanding toward them.

Jeanne Block has reported that, in each of four separate studies of parents of children of four different age groups, greater paternal differentiation in child-rearing orientations is consistently found. Block has also showed that in teaching situations fathers interact differently with sons and daughters. With their sons, fathers set higher standards and place greater emphasis on solving the cognitive problem at hand; with their daughters, fathers are more likely to focus on the pleasures of interacting together. With daughters, fathers are less likely to explain principles and patterns or to show the way to solutions, and they are more likely to treat mistakes as nothing to worry about. A study on parents of toddler girls found that fathers like to have toddler girls around but are less likely than mothers to let them "help." Fathers are especially likely to comfort daughters and to try to protect them from experiencing failure.^[13]

On the basis of findings such as these, it is not surprising that the factor of "paternal nurturance" has been found to be significantly associated with cognitive competence for boys but not for girls. Norma Radin, who reports this, suggests that perhaps the father's presence stimulates cognitive growth in girls only if the relationship is not too loving.^[14] My feeling is that even if fathers were to emphasize the cognitive aspects of a learning situation with daughters, there would likely be some sort of metacommunication transmitted that refers to the father's perceptions of difference. I am thinking particularly of a scene in a film designed to encourage fathers to stimulate their daughters' intellectual capacities. In this scene the father is showing a leaf to his toddler daughter while he is holding her as if she were made of china and were likely to break. He was also speaking to her in a whispery voice that would seem quite strange if the child had been a toddler boy. At a somewhat

later age the father may undercut cognitive learning with a flirtatious attitude toward the daughter. One of the earliest studies to discuss father differentiating phenomena noted that fathers of young girls were more likely to flirt with them than were mothers to flirt with young sons.^[15]

The father is also likely to be giving messages about gender to his son, namely, that it is important for males to be cognitively competent and to compete effectively. A male student of mine once commented with some bitterness that his father was highly critical of his moves when he played chess with him, whereas he was very appreciative of his sister's least sign of intelligence. With the son, the father said in effect, "That will do for a start"; with the daughter, the father said, "You are so cute." It is in ways such as these that fathers encourage a kind of male-oriented "darling girl" orientation in daughters and a power-competence orientation in sons. Mothers with sons do not make differentiations symmetrical to fathers'.

In a systematic observational study of some seventy-eight parent-child dyads in Central Mexico, Phyllis Bronstein found that mothers did not differ in their behaviors toward girls and boys but that fathers did distinctly differentiate along the lines discussed above.^[16] She suggests that this

may play an important role in the socialization of their children for traditional sex roles. Boys were listened to and shown how to do things, which would seem to convey the message that what they have to say is important and that they are capable of mastering new skills. Girls, on the other hand, were treated especially gently, and at the same time, with a lack of fall attention and an imposing of opinions and values. The gentle treatment would seem to convey the message that they are fragile and docile. . . . The inattention and imposing of opinion would seem to communicate the view that what females have to say is less valuable than what males have to say, so that females need to be told more what to think and do, and can more readily be interrupted or ignored. Thus in this sample of Mexican families, very different messages were being transmitted to girls and boys—about their roles, their temperaments, their thinking, and their expected behavior—and fathers were the main transmittors of those messages.
                                                                                                                                   (p. 1001)

These data are important in suggesting that at least in patriarchal cultures where fathers, as opposed to brothers or other male kin, head families, the behavior of fathers is similar.

Other evidence comes from a study by William McBroom of unmarried university women's acceptance of stereotyped expectations for the domestic role of married women. Using a scale containing such items as "a woman's place is in the home" and "a woman should yield to her husband regarding her own employment," McBroom found that women who rejected such items had poor childhood relationships with their fathers and that the women who endorsed such expectations had good relationships with their fathers. Good childhood relationships with mothers had no such traditionalizing effects.^[17]

Fathers also seem to be less alarmed about aggression in sons than about their sons' not fitting in with their male peers. In a study of forty sets of middle-class parents of children of both genders, Margaret Bacon and Richard Ashmore report that mothers ranked hostile aggressive behaviors highest on a continuum of behaviors they would consider to be problems in both sons and daughters; fathers ranked poor peer adjustment as an even greater problem than aggression in boys. Fathers consider such hypothetical items as "being pushed around," "getting stomachaches before school," "being left out of games," and "just sitting around" to be more serious problems than being hostile and aggressive. The authors suggest that the fathers might have perceived poor peer relations as indicating "feminine" or "sissy" behavior in boys. Here again is some evidence that fathers are concerned about boys being masculine and that fathers mediate between boys and the male peer group.^[18]

When one considers how many studies have been done on mother-child interaction and how few of these find differences in the way mothers interact with male and female children, and how few studies have been done on father- or adult male-child interactions and how many of these do find differences, one must suspect that fathers' differentiating attitudes will become increasingly apparent as more research on parenting includes both fathers and mothers.

A number of other studies seem to suggest in one way or another that girls who are close to their fathers are more, rather than less,

"feminine" and that boys who are close to their fathers are more, rather than less, "masculine." In these studies, the measures of masculinity and femininity are highly problematic and often obscure the maternal heterosexual distinction that I am making. That is, one cannot tell whether the qualities being measured are those related to mothering or to being a dependent wife and a subordinate in heterosexual relations. Generally, the masculinity-femininity tests used were self-administered and consisted of questions describing behavioral preferences or attitudes that men and women tend to consistently answer differently, but the reasons for the differences are not always clear. Nevertheless, the studies are worth examining.

Perhaps the most frequently used masculinity-femininity test is the Gough Femininity Scale (Fe Scale) of the California Personality Inventory. This same inventory also has a Socialization Scale (So Scale), which differentiates to some extent between delinquents and nondelinquents and measures the degree to which individuals can view themselves from the point of view of the other. Jack Block, Anna Von der Lippe, and Jeanne Block examined the association between adult subjects' scores on these two scales and variables related to parents that were collected when the subjects were children. The results suggest that the mother plays a more prominent role in general socialization and that the father is more involved with "femininity," which on this test does not mean mothering but may mean being passive and attractive. In this study the prior variables that distinguished women who ranked high on the femininity scale from those who ranked low on the socialization scale were their own physical attractiveness, a rejecting mother who did not emphasize values of tenderness and love in relationships, and a seductive father. Since the mothers of the highly feminine girls were rejecting and inadequate, one might presume the girls' "femininity" did not come from them, but since the fathers were seductive, this seductiveness may have been the source of a passive, male-oriented (and nonmaternal) femininity. The other finding of interest was that girls who were ranked low on the femininity scale and high on the socialization scale were, first of all, the best adjusted group.^[19] Femininity has been found in many other studies not to be related to adjustment. Less "feminine" women had affectionate fathers who were ranked high on marital adjustment, and who, presumably, were less inclined to play husband-

wife games with their daughters. Fathers who play these games are powerfully reinforcing male dominance, because their own objective power position over the daughter is great.

There is considerable evidence that both boys and girls do use the relationship with their father, or with adult males, as a lever to help them out of their early dependency relationship with their mother. Father "identification" involving both imitation and attachment increases with age and intelligence for children of both genders.^[20] Helene Deutsch uses clinical findings to make the same point in The Psychology of Women: "As the child with gradually increasing intensity turns away from the mother and childhood dependencies in favor of active adjustment to reality, this reality is represented more and more by the father —and this is true of both boys and girls" (vol. 1, p. 243). Significantly enough, Deutsch follows this by saying, "Contrary to our previous views, the girl's first turn toward the father has an active, not a passive character, and her passive attitude is only a secondary development" (p. 252). What seems to happen is that the girl turns to her father to help her out of her dependency on her mother, and within the fatherdaughter relationship the girl may indeed achieve a measure of independence from the mother, but she is at the same time likely to become dependent on the father and to overvalue men.

Whereas most fathers do seem to "traditionalize" daughters, there is considerable research support for the view that many, though by no means the majority of, female high achievers were close to their fathers and had their fathers' support and encouragement.^[21] But this closeness and support may be a mixed blessing at best if it perpetuates women's achieving simply to please a man or to do credit to the judgment of a man. Being a daddy's girl can interfere with female autonomy. Susan Contratto writes of seeing many high-achieving middle-class women patients in therapy who in one way or another find it extremely difficult to be assertive with men.^[22] Daddy's girls may be very assertive with men if the men approve of what they are being assertive about and back them up. Such women, however, have trouble asserting, or even knowing, their own needs. This lack of autonomy is the direct product of the nuclear family situation in which the father-daughter relationship can all too easily train the daughter, even though she may be highly "accomplished," in wifely deference.

Fathers and Differentiated Sexual Interaction

In the research described so far, the nature of the gender differentiation that fathers seem to foster is not always clear. Generally, fathers seem to reinforce passivity, dependence, and attractiveness in girls. They do this with their more permissive and nurturant attitude toward girls and their more demanding stance toward boys. The lenience and nurturance toward daughters found to be characteristic of middle-class fathers may not be true, however, of working-class fathers. Several studies in the United States have found working-class fathers to be strict and punitive toward their daughters.^[23] This contradiction may be more apparent than real, however.

If we take a sufficiently general perspective, class variations are not contradictory but reflect in different ways the father's awareness of the daughter's sexuality. Fathers may respond directly to their daughters' sexuality or they may suppress this response and focus on protecting the daughter from outsiders. Both responses, of course, can and do occur at once. The father may reward and reinforce the heterosexual aspects of his daughters femininity both directly by interacting with her as an "interested" male and indirectly by protecting her from outside males.

Anne Parsons's description of the position of the daughter in a Southern Italian family reflects both these aspects.^[24] "The most fully institutionalized masculine role in Southern Italy, one that is defined positively and not by rebellion, is that of the protection of the honor of the women who are tabooed" (p. 386). "Thus, the South Italian girl does not appear as inhibited or naive for precisely the reason that even though carefully kept away from outside men, she has in a great many indirect ways been treated as a sexual object by her father (and brothers or other male relatives) both at puberty and during the oedipal crisis" (p. 387).

The lack of symmetry between mother and son in these matters is so obvious that it is seldom pointed out. Mothers do not protect sons from outside women nor do they treat their sons as sexual objects—certainly not so clearly and not in the same way that men protect and sexualize their daughters. The insight that feminism brings to these findings is that they are very much related to male

dominance and the sexual control of women by men. Thus paternal concern with gender differentiation, however benevolent it may seem, tends to reproduce and maintain the system in which men treat women as sex objects for themselves and protect them from outside men.

Data from the United States indicate that fathers are more likely than mothers to support a double standard of sexual behavior for their male and female children; fathers are more tolerant of premarital sexual activity for boys (sometimes positively enjoining it on them) than for girls. A large-scale study on a probability sample of parents in the Cleveland, Ohio, area, showed that fathers are more likely than mothers to want the son to receive one message about sexuality and the daughter another. Although many fathers were more liberal than mothers about what erotic practices, such as oral sex, might be considered desirable, they were more concerned about controlling or curtailing their daughter's sexual activity in general than mothers were.^[25] In a study of occupational choices of high school students, a considerable number of girls answered the question, what would your father not want you to be, by saying their fathers would not want them to be a prostitute; few girls mentioned prostitution as an undesirable job their mother might think of.^[26] The father's disapproval of his daughter being a prostitute reflects his control over her sexuality and his view that she should not make herself accessible to "outside" men.

In her book Like Father, Like Daughter, Suzanne Fields frequently speaks of fathers as being their daughter's "protectors." What these fathers largely seem to be protecting their daughters from is other men. They worry about their adolescent daughters and want to protect them because they know what their own male peers were like in adolescence. As fathers and husbands, men protect their domestic domain and their women from other men. They "understand" the male-peer-group mentality that makes women need protection from men. These facts are well known, but few seem to notice that "the problem" does not lie with the behavior of women but with the behavior of men.

Most fathers also protect themselves from any potential sexual temptation that their daughters might present. Fathers' withdrawal from girls, which I described at the beginning of this discussion, may well be a way of avoiding their sexuality. It is significant that

the father's marked withdrawal, which Lamb and Lamb observed during the girl's second year of life, coincides with girls often beginning to show awareness of their genitalia and to masturbate in the second year. The father's avoidance of his daughter at older ages may also be related to his awareness of her sexuality. There is much evidence that males' overall style of caretaking consistently differs from females'. This difference exists even if the male is a primary caretaker. Generally, fathers "play" with children, and mothers "nurture" them. The context of play itself might act as a deterrent to sexualizing the relationship. Clearly there is much we do not know about the relationship between eroticism and child care and how or why it might differ between female and male caretakers. Patterns of sexual avoidance within the family is an important area for future research.

In this culture at least, outside the family nudity among males is a symbol of locker-room or peer group camaraderie, and nudity between males and females is seen by many (perhaps especially males) as being related to sexual contact, specifically intercourse. Kissing between males has generally been associated in this culture with homosexuality. Thus one may suppose that fathers' avoidance of kissing boys staves off homosexual threat with sons, and that fathers' avoidance of nudity with daughters defends against heterosexual threat with them. These findings fit with earlier research on touch that showed that men and women touch and are touched by their mothers and their friends of the other gender in equivalent degree but differ in their touching relationship to their fathers. A study on physical affection between parents and children finds that fathers display considerably more physical affection toward daughters than they do toward sons, while mothers express physical affection equally toward both boys and girls. Here fathers differentiate by denying physical affection to boys. This denial may contribute to males' tendency to distance themselves from nonsexual physical closeness. It may also contribute to the general tendency for males to be "distanced."^[27]

In general, daughter-father sexual discomfort appears to be stronger than son-mother sexual discomfort. David Finkelhor found only one significant difference between boys and girls at the age of twelve with regard to what would cause them embarrassment with their mothers but several differences in what would cause embar-

rassment with their fathers. Girls report more often than boys that they would have been embarrassed to be seen naked or in their underwear by their father, to tell him a dirty joke, or to tell him about a sexual experience. Boys and girls did not differ from each other, however, with regard to the mother except that boys would have been more embarrassed than girls to be in the bathroom with her.^[28]

In line with this, another study found that fathers are more likely to knock before entering their daughter's bedroom than their son's. Mothers knock less often than fathers and make little distinction between sons and daughters in whether or not they knock.^[29]

With regard to "sex education," in the sense of explaining such things as intercourse and masturbation, middle-class fathers do not ordinarily give information to their children. The heart-to-heart talk a boy and his father are supposed to have about sex is indeed a myth. Certainly one major reason for a father's difficulties in talking to his son about sex may stem from the cross-pressures of his wanting his son to be a "real man" in terms of the standards of the male peer group and his wanting his son to be a "good person" in terms of more general humanistic and maternal standards. Does he tell his son to "get all he can" or that "women are human beings too"? He may say a little bit of both. Kanin, in his comparison of potential rapists with nonrapists, found that rapists' peers had encouraged sexual aggression more than their fathers, although rapists' fathers had discouraged it less than the fathers of the control group.^[30]

All this suggests that fathers, more than mothers, see their daughters as potential sex objects for themselves or for other men. As fathers, they generally do not make their daughters into sex objects directly, but there are elements of flirtation and mock courtship in the relationship. One of the clearest expressions of this comes from an early study of upper-status parents of nursery school children. Fathers reacted with pleasure to their daughters being "little flirts" and knowing how to flatter them.^[31]

The tendency in men to fall into a romantic relationship with their daughters gains specific support from a recent study by Diane Ehrensaft of five families in which the father was actively involved in child care along with the mother. Ehrensaft tells us that the fathers seem to be "in love with" their child and the mothers merely to "love" their child. Moreover, these heavily involved fathers pre-

fer girls. Ehrensaft is careful to point out that these fathers do not sexualize the relationship, but clearly they do romanticize it. For example, she quotes one father as saying about his daughter, "I'm absolutely in love with her. Just passionately in love with her. On occasion that's almost frightening."^[32]

Ehrensaft reports that one father admitted to using the child as a wife substitute in saying "I still have a lot of passion in me that I enjoy having Renee (daughter) available to share with me. And who Joan (mother) is, at times that's not available to me. That kind of romance. Renee and I have the metaphor of what it means to have a cozy place" (p. 330). Their outings seem like dates. "I go out with Renee. Go dancing together. Renee and I get all dressed up and go out for dinner" (p. 333). Ehrensaft does not find that mothers feel the same about their sons as fathers feel about their daughters. She reports that a mother "would talk about her attachment, her endearing feelings, her sense of closeness to her child, but never in the discourse of a love affair" (pp. 326–27). The mothers tend to think of parenting in terms of responsibility and display a certain ambivalent distaste for having too much responsibility. These women deliberately sought a man who would share the responsibility of child care, but the men did not turn out to be second mothers; instead, they became their children's lovers, albeit without overt sexual activity. The fathers, Ehrensaft found, took less responsibility than the mothers. For example, the fathers were more willing to take the child out to be cared for when sick than to stay home from work to take care of the child themselves. They were less concerned about diapers that were long overdue for a change or making sure the child was dressed appropriately for the occasion or the weather.

Ehrensaft's findings graphically support exactly what I mean when I say that the father-daughter relationship is more nearly a paradigm for adult heterosexual relationships than the mother-son relationship. A mother could not take her son out dancing without somehow threatening his masculinity, but a father finds it pleasurable and not at all anomalous to go out dancing with his daughter. From the father's standpoint, the father-daughter relationship offers a nice resolution to whatever fears he may have of being taken over by women as mothers. He is clearly in the superior position in this intimate relationship, and all threats of being overpowered by

a dominant female are removed. Father-daughter intimacy does not pose the same threat to male power that mother-son intimacy does. In the father-daughter relationship, the father is in control; in the mother-son relationship, the mother is in control.

In line with Ehrensaft's observations, Dorothy Burlingham reported earlier that when fantasizing about a daughter, expectant fathers focused on her romantic love life. with the objects of her love clearly representing themselves. The fathers wanted their daughters to respond to their loving feelings. Burlingham noted that fathers' fantasies about sons, however, focused on the boy learning from the father and being strong, capable, and competent like the father's self-ideal.^[33]

The fathers whom Ehrensaft studied had been nontraditional males when they were growing up, not tied in closely with a male peer group. Being nontraditional perhaps made it possible for them to take on the role of mothering in the first place, but the kinds of gratification they obtained were not those of mothers but of lovers. Having a lover instead of a mother as a young child could be a threat to the daughter's development as an adult woman and gives one pause about the "equal parenting" solution to male dominance. Mothers simply do not seem to see children as lovers in the way these fathers did. Mothers love their children without the romance.

Fathers and Male-Dominated Heterosexuality

Typically in this society one does not inquire about the "causes" of heterosexuality, because heterosexuality is assumed to be "normal" or "given." One asks instead about the cause of homosexuality. But many liberals concerned with human rights consider it reactionary to make inquiries into the causes of homosexuality because they feel it implies that homosexuality is a disease to be cured or a type of deviance to be prevented. This creates a situation in which it becomes impossible to subject either heterosexuality or homosexuality to investigation. The liberal view holds that homosexuals and heterosexuals do not differ much from each other and that sexual preference is a private matter. Discussion of the causes of homosexual preference, the liberal contends, simply serves to perpetu-

ate discrimination. Another facet of this position is that investigation of the causes of sexual preference, especially those guided by a psychoanalytic perspective, tend unfairly to blame parents for their children's "deviance" and thus cause misery not only among homosexuals themselves but among their parents. It is this sensibility that caused the National Institute of Mental Health Task Force on Homosexuality in 1972 to list questions of "etiology" as its lowest priority.

Although I sympathize with this general stance, I nevertheless believe that sexual preference is an important area to investigate both because of the scientific "need to know" and because such knowledge can inform a feminist analysis, since heterosexuality and homosexuality are connected with male dominance. I am primarily interested in parental relationships and sexual preference because they relate to my general project of showing the extent to which the nuclear family with father as head perpetuates male dominance. My concern is not to punish or change individuals' sexual orientation or to lay responsibility for either homosexuality or heterosexuality at the feet of parents but rather to consider the implications of the findings about parents for my hypothesis concerning the father's involvement and the mother's lack of involvement in the specifically sexual aspects of gender differentiation.

The largest, best-designed, and one of the least heterosexist investigations of how sexual preference develops in both genders was conducted in the early 1970s by Bell, Weinberg, and Hammersmith but was not published until 1981. The subjects consisted of 979 homosexual and 477 heterosexual men and women living in the San Francisco Bay Area. The homosexuals were recruited from diverse settings to insure their being more representative of gays in the general population than had been the case in previous studies.^[34] The only possible bias is that, because of the San Francisco location and the nature of the study itself, it is overweighted with activist, as opposed to closeted, homosexuals. If anything, this bias would probably work against finding support for any hypotheses concerning parental influences, because activist homosexuals have ordinarily been opposed to psychoanalytic speculations about parental involvements.

In fact the likelihood of disconfirming psychoanalytically based hypotheses concerning parental influences on sexual preference

was increased by the authors' policy concerning "knowledge." They asked all of their homosexual respondents whether they had read any books or articles or had attended lectures on the causes of homosexuality. They then tested each homosexual-heterosexual difference that was found to see if the difference could be accounted for by exposure to these theories. If the difference could be accounted for by exposure to "theories," the authors did not report the difference. When findings appeared that applied equally to homosexuals who had and had not been exposed to the literature on homosexuality, only the figures for the unexposed homosexuals were used. These were unusual and perhaps unprecedented precautions to rule out the impact of theories and previous research findings about the causes of homosexuality on the subjects' responses. But again, the effect of these precautions is to make associations with parental relationships all the more credible if they are found.

Much of the earlier research on sexual preference was based on percentage differences between homosexuals' and heterosexuals' answers to questions concerning parental influences, gender non-conformity, peer relations, first sexual experiences, and so forth. Bell, Weinberg, and Hammersmith went beyond this and used "path analysis" to develop a model that included fifteen variables associated with these and other factors arranged in a temporal sequence leading up to adult sexual preference. (In path analysis the temporal sequence is supplied by the investigator on the basis of logic. The sequence may be varied to some extent to see which way explains the most variance.) As a statistical technique, path analysis cannot tell one whether the temporal sequence is correct, but it can show whether the influence of a certain variable on adult sexual preference is direct or indirect. It can also show how much influence a particular variable would have if the influence of all the other variables were controlled. Path analysis provides a considerably more stringent test of influence than simply comparing homosexual-heterosexual differences on one variable at a time because the latter procedure cannot tell us whether the association is due to the action of another related variable or to the variable in question.

The results of the analysis indicate that "gender nonconformity in childhood," especially for boys, is the major influence on sexual preference, but even here the connection is far from overwhelm-

ing. Thus males who described themselves as being different from other boys as children, as not possessing stereotypical masculine traits or as not enjoying masculine games and activities, are more likely to become homosexuals. Gender nonconformity is somewhat more salient for males than for females, whereas family relationships are more salient for females than for males.^[35]

Joseph Harry, in a study largely focused on gay culture, also reports an association between gender nonconformity in childhood and adult homosexuality from a large survey of Chicago area gay and nongay males. In fact, a major thesis of Harry's book, Gay Children Grown Up, is that while gender-role preference is "neither necessary nor sufficient" to determine sexual orientation, it does make it "more likely than not" that the person who felt himself to be a gender nonconformist or "different" earlier will be gay later (pp. 12 and 15).

The finding of an empirical connection between gender role nonconformity and homosexuality in males may lead one to question the distinction I am trying to preserve between gender and sexual orientation. I do not claim the two are unrelated—certainly they have been thought to be related, especially in male thinking— only that they should be kept analytically or conceptually distinct. Whereas Bell, Weinberg, and Hammersmith tend to interpret this connection between gender nonconformity and homosexual preference as suggesting a biological basis for both, Harry suggests instead that a primitive gender identity is formed very early during the period of childhood amnesia, and therefore the presumably social interactional events that caused the identity (as opposed to biological causes) are inaccessible to consciousness and cannot be uncovered. Harrys findings suggest that male homosexuals are likely to be conscious of a preference for the female gender role between the ages of two and six. After this, many males consciously try to "defeminize" in order to conform more to male gender-typed standards. They do this, Harry says, under pressure from the male peer group and to avoid their fathers disapproval (p. 23).

Overall, the Bell, Weinberg, and Hammersmith study disconfirms far more than it confirms about background factors in homosexuality. The authors show that almost all the alleged causes of adult sexual orientation are either nonexistent or highly exaggerated. The two positive findings they do note, however, are the

aforementioned link between gender nonconformity and the development of homosexuality and that "poor relationships with fathers seemed more important than whatever relationships men and women may have had with their mothers."^[36] Thus while hypotheses concerning mothers tended to be disconfirmed, "poor father relations" were not. "The homosexual men's generally negative relationships with their fathers and the lesbians' experiences of their fathers as detached, hostile, rejecting, and frightening displayed a very modest but direct connection to their gender-role nonconformity and to sexual elements of their development as well" (pp. 189–90). These findings support my own conclusions concerning the greater salience of fathers with respect to sexual orientation, based on my canvass of earlier studies.^[37]

Gay Males

The most widely accepted psychoanalytic hypothesis guiding research about parental influences on male homosexuality has been that of a triangular system characterized by a close-binding (domineering-seductive) mother and a hostile or distant father. Within this triangular system, however, the mother relationship was generally considered to be the root cause. The importance of the mother was challenged in 1965 by Eva Bene, who, in a study of nonpatient heterosexuals and homosexuals using the Bene-Anthony Family Relations test, found that more than three times as many items showed significant differences in relation to the father than to the mother. According to Bene, "Far fewer homosexual than married men thought that their fathers had been cheerful, helpful, reliable, kind or understanding, while far more felt that their fathers had no time for them, had not loved them, and had made them feel unhappy. . . . Regarding the mother, the greatest difference between the two groups was that considerably more homosexual than married men thought that their mothers used to nag."^[38] Many other researchers using nonpatient samples have also attested to the greater importance of the father relationship for men.^[39]

Although the Bell, Weinberg, and Hammersmith study found that a larger percentage of homosexuals than heterosexuals are close to their mothers, this variable was the least important of those that were of enough significance to be entered into the path

model. Moreover, they found no evidence whatsoever for the "seductive mother" hypothesis, and only a small fraction of respondents said their mothers had acted like girlfriends or lovers with them.^[40] But their preliminary findings show that in a number of ways prehomosexual boys have less solidary relationships with their fathers than preheterosexual boys do. Rather than trying to isolate the various aspects of this negativity, the authors used a general measure—negative relationship with father—in the path analysis. These differences with respect to fathers contribute considerably more to the overall path analysis than do the differences with respect to mothers. Thus, the authors support the findings of Bene and others about the greater importance of the fathers. This negative relationship with father is somewhat positively correlated with homosexual arousal in childhood, childhood gender nonconformity, and feeling sexually different in childhood and is negatively correlated with identification with the father.^[41]

Because homosexual subjects reported having worse relationships with their fathers when they were boys than heterosexual subjects did, the authors are unable to establish what is cause and what is effect. The data are as amenable to the interpretation that poor father relationships dispose to homosexuality as to the view that the child's being different or feeling different in the first place makes it difficult for him to maintain a good relationship with his father (p. 190). For my purposes this is not as critical a question as it may seem. The influence is likely to go both ways. The key point for now is that the father-son relationship is more associated with sexual preference than is the mother-son relationship. The data are certainly consonant with the hypothesis that fathers are more concerned about their boys being "properly masculine" than mothers are and therefore respond more negatively to a son's gender nonconformity.

There is some evidence in Joseph Harry's work to support the possibility that gay males share more basic gender characteristics with straight males than with females. Many of Harry's male subjects who rejected "masculinity" were nevertheless aggressive and competitive (nonrelational?) in acting out their version of femaleness, which involved being glamorous and seductive. He called this response "actorization." Another group of gay males, more akin to transsexuals, were nurturant and domestic. Both groups had

rather stereotyped views of what women were like. It seems to me that there was still something very male and nonrelational about the femaleness of these gay males—perhaps especially among those who were actorized. The quasi-transsexuals, by contrast, had come nearer to adopting a maternal mother identification. Even with transsexuals, however, the maternal identification is not complete. Stoller notes that when one gets to know transsexuals outside of research questionnaires and evaluation interviews, one does not find mothering impulses in the femaleness of these transsexuals or the capacity to remain in extended relationships with males.^[42] These observations suggest that the more basic aspect of gender difference is women's mothering and greater relational capacities.

Lesbians

Many theorists, especially the psychoanalytically oriented ones, attribute lesbianism to a reversed triangular system consisting of a hostile or distant mother and a close-binding father.^[43] Other analysts, most notably Charlotte Wolff, argue that the quality of the mother relationship alone is central in the development of lesbianism. Eva Bene and others, however, again find evidence for the greater significance of the father relationship in differentiating heterosexual women and lesbians.^[44]

Bell, Weinberg, and Hammersmith's research suggests the latter findings. Although they report that lesbians have more negative relationships with mothers than heterosexuals do, they conclude that the mother-daughter relationship is not basic to the development of sexual preference among females. Although they warn that the importance of the father relationship must not be overrated either, variables connected with the father are more prominent than variables connected with the mother. Variables labeled "detached, hostile father," "weak father," "mother-dominated father," "father uninvolved in family decisions" all appear in the path analysis as being somewhat connected with adult lesbianism.^[45] (Lesbians did not rate their mothers as personally more domineering than heterosexuals rated theirs, but lesbians were more likely to describe their fathers as being dominated by their mothers. This tendency reflects an unusually nondominant father rather than an unusually dominating mother.) The effects of the four variables mentioned,

however, are all indirect and sometimes contradictory. For example, there is a negative correlation between "detached, hostile father" and "mother-dominated father," but both are associated with childhood gender nonconformity and through that with homosexual arousal in childhood or adolescence.

A larger proportion of lesbians described themselves as being "masculine" when they were growing up than heterosexual women did. Feeling masculine, however, had no connection for these women with feeling dominant. In the path analysis no connection existed between the degree to which respondents said they had been dominant (or submissive) as children and their adult sexual preference. This lack of connection suggests that lesbians do not seek dominance any more than heterosexual women, although lesbians did describe themselves as being masculine in other ways. I am suggesting that lesbians may share with heterosexual women a more egalitarian orientation.

The researchers found that the majority of female respondents, whether homosexual or heterosexual, identified with their mothers and, more important, that lesbians (to the surprise of the researchers) identified less with their fathers than heterosexual women did. More homosexual women said that while they were growing up, they did not want to be like their fathers and felt very little or not at all similar to their fathers (p. 133). Although identification with the father turns out to have no importance for the development of sexual orientation among females, it is an important difference to note because it supports other evidence that more-"feminine" women may be more (not less) "father-identified" than other women.^[46]

Whereas the parents' relationship to each other had little or no effect on sexual preference in males, "negative relationship between the parents" is a factor, albeit not a very strong one in female sexual orientation. Although a majority of both homosexuals and heterosexuals had described their parents' relationship "quite positively," more homosexual than heterosexual women thought their mothers had felt little or no affection for their fathers and vice versa.^[47] This finding would fit with Bene's finding that the statements about mothers that differentiated lesbians from nonlesbians are not about the mother's attitudes toward the child but about the mother's attitudes toward her husband. Charlotte Wolff, who speaks of the great importance of the mother in lesbianism, relates

the girl's lesbianism ultimately to the mothers attitude toward men. In her version the girl realizes that the mother values males more than females, so the girl becomes insecure about her own value.^[48] The findings that lesbians are more likely to report negative interaction between their parents suggests that they may be reacting against marriage and being a wife rather than rejecting or feeling rejected by their mothers qua mothers.

There is considerable anecdotal evidence among lesbians having to do with "coming out" to their parents that suggests mothers are less "devastated" than fathers. The stories relate that mothers, after a certain amount of reacting and overreacting, usually come around to the position that as long as the daughter is a good person, functioning effectively in the world, and happy, her sexual orientation does not really matter. Mothers often say, however, "But it will kill your father" or even "Don't tell your father."^[49] Mothers as mothers feel sexual orientation is not important. Mothers as wives, however, say, "You are going to disappoint your father because you are 'rejecting' men." Mothers qua mothers can understand "rejecting men."

The Social Meaning of Sexual Preference

In their summary, Bell, Weinberg, and Hammersmith comment that one of the major contributions of their study is "the lack of support it gives to many of the traditional notions about the causes and development of homosexuality."^[50] Indeed, they found nothing to be very important in accounting for homosexuality except gender nonconformity and homosexual ideation and experience, all of which are at least partially connected by definition and are therefore tautological. In the end, the authors suggest that perhaps biological factors are involved, but their study was not designed to provide any direct evidence concerning how such factors might operate. They presume that if sexual preference is a biological phenomenon, no one can be blamed. These conclusions tend to fit well with the liberal stance I outlined at the beginning of this section.

Reviews of this research in the popular press have emphasized the probable biological basis of sexual preference and have emphasized the degree to which parental factors were found not to be critical. In a sense this emphasis is all to the good. Of course, par-

ents do not cause homosexuality; to say they do simply reflects the nuclear family ideology (which psychoanalytic thinking has supported), which makes the family, and no one else, including the child, responsible for how children "turn out." Because the nuclear family has until very recently been central to a majority of children's upbringing, however, it stands to reason that parents would in fact be important mediators of societal expectations. Parents themselves are representatives of basic family roles in this society: mothers, fathers, wives, and husbands. It would be highly unlikely that parents and parental relations would have no connection whatsoever to their children's orientations given the salience of the nuclear family as "ideal" in this society and given the child's almost total material and emotional dependence on parents.

What neither Bell, Weinberg, and Hammersmith nor the popular press were able to deal with, and did not even attempt to deal with, were the findings concerning fathers. The authors argue that their findings disconfirm those psychoanalytic explanations of homosexuality that trace it back to unresolved oedipal feelings. But this is only half true. They note that "the connection between boys' relationships with their mothers and whether they become homosexual or heterosexual is hardly worth mentioning" and that they found "no evidence that prehomosexual girls are 'Oedipal victors'—having apparently usurped their mothers' place in their fathers' affections" (p. 184). This much may indeed be true, but they do not deal with the evidence that prehomosexual boys and girls have negative relationships with their fathers.

It is impossible from this kind of data to speculate in any detail about the causes of sexual preference, either biological or psychological, but the findings concerning both male and female homosexuals are compatible with Freud's version of the Oedipus complex, which makes the father the central character for both girls and boys. In effect, "feminine," that is, heterosexual and passive-submissive, girls (Freud's normal outcome) never really give up their fathers, and masculine, that is heterosexual and active-dominant, boys (Freud's normal outcome) internalize the father's rules. I will discuss the significance of Freud's ideas for the transmission of "patriarchy" in the next chapter. The differing perceptions of the father held by homosexuals and by heterosexuals certainly suggest that the system of male dominance is somehow involved.

Heterosexuals, both male and female, are accepting male dominance or are at least coming to terms with it, whereas homosexuals are refusing to play the male game, which in this society has come to mean male-dominated heterosexuality. The reasons or more proximate causes for this refusal may be varied and have by no means been sorted out. But many lesbian feminists and some gay male feminists have increasingly come to see being homosexual as a statement of resistance to the male dominance attaching to heterosexual institutions, including male-dominated heterosexual intercourse.^[51] The negativity toward fathers found among gay males may represent a rejection of the male peer groups insistence on both heterosexuality and doing male-typed things. Joseph Harry suggests that gay males' tendency to have been loners in childhood and adolescence was a defense against the pressures they knew they would get from their male peers to do male things and be heterosexual.^[52]

In 1963 I reported data that supported the hypothesis that "feminine" women are male-"identified," in terms of "felt identification," "assumed identification," or "solidarity" with the father.^[53] In other words, girls who understood, sympathized with, and were close to their fathers were likely to be "feminine" and, not surprisingly, well adjusted in our culture. They were not well adjusted because they were like their father, but they were well adjusted because they liked and understood their father. The reverse with men and their mothers did not hold. "Courtship progress" is a case in point. The sociologist Robert Winch showed long ago that whereas males who were attached to their mothers made slow progress in courtship, females who were most attached to their fathers were most advanced in courtship,^[54] and by that measure presumably well adjusted.

Nowadays, feminists sometimes refer to a lesbian as a "woman-identified woman," meaning a woman who likes and respects women. Would a gay male ever be referred to as a "man-identified man"? I think not. As far as I know, the phrase is never used and seems meaningless. Certainly, heterosexual males are not likely to call male homosexuals "male-identified." The unspoken understandings involved in this asymmetrical usage reveal that heterosexual women and heterosexual men are both seen as being identified with men in the sense of liking and respecting men. In a male-

dominant culture, the assumption is made that a real woman (as opposed to a lesbian) needs a man, while a real man may use a woman for sex but is essentially self-sufficient with his male peers. Real women "identify" with men, and real men "identify" with men.^[55] The data I have presented is consonant with this picture—a disturbing testimony to how gender difference has become defined in terms of male dominance.

Summary:
Mothers versus Fathers

All of the data presented in this chapter support, or at least are not incompatible with, the idea that the initial identification of children of both genders is with the mother and that both girls and boys develop their generically human qualities in connection with a maternal figure. The maternal aspect of women may be seen as generic and as symbolizing the common humanity of both males and females. Males to some extent are constrained to establish their own identity in partial conflict with this maternal principle and, as a countermeasure, to stress the "difference," "dominance," and a perception of women as sex objects for men. The father in the nuclear family is partly a representative of the male peer group and partly a protector of his wife and children from the male peer group. Either way, the father seems to be the focus of a second differentiation that occurs, a differentiation that turns the tables on girls and makes boys superior. Girls add another identification to their maternal identification and become "male-identified" in the sense of learning to please and play up to males, whereas boys abandon their mothers in order to identify as males and join their male peers. Fathers may indeed reinforce what we may think of as "femininity" in women, but that father-oriented "femininity" is an overlay that may not be in a woman's best interests because it is a stance that makes a woman dependent and childlike.

To the extent that women's attitudes as mothers rather than male peer group attitudes prevail in the society, it is likely that the society will move away from a focus on gender conformity, compulsory heterosexuality, and male-dominated heterosexual relationships. My emphasis on the mother, however, does not mean that I consider lesbianism or male homosexuality as solutions to male dominance. Certainly, heterosexuality as it is institutionalized in

marriage implies dominance for the husband and secondary status for the wife. Taken alone, this argument might imply that heterosexuality must go if male dominance is to go, and some feminists have taken this position. Although marriage today clearly does involve expectations of male dominance, the relations between women and men in marriage are at least potentially equalitarian, however, and they are already more equalitarian than the relations between women and men as defined by the male peer group. As we have seen, it is within male peer groups that women tend to be sex-objectified and that male dominance finds its most exaggerated expression. Male attitudes toward females in the context of the male peer group, whether homosexual or heterosexual in orientation, often embody male dominance to its greatest extent. The attitudes of males as "fathers" toward daughters may be less sexist than their attitudes as "one of the boys."

Thus, even though marriage still involves secondary status for the woman, marriage or heterosexual relationships in the context of commitment can work against male dominance by mitigating the views of women fostered in the male peer group. Since marriage constrains and limits women more than it does men, however, it may not be so salubrious for the individual woman who participates in it. In short, those of us committed to more equal gender relations may applaud some of the consequences of long-term heterosexual commitments for men, but until marriage itself becomes more genuinely a coalition of equals, heterosexuality for women seeking equality must remain problematic.