Chapter Three—
Defining Difference:
Psychological Perspectives

In Chapter 2, I made a distinction between women as mothers and women as wives. In this chapter I want to make a distinction between women's orientation toward interdependence and dependence. Too often in stereotypical thinking the orientations related to women's mothering, which include caring about others and seeing self in relation to others, are merged and fused with images of dependence and passivity. Confusing interdependence and dependence is akin to speaking about mother and wife in the same breath. In this chapter, I will review various attempts to conceptualize how women might differ from men in terms of personality orientations. I begin with Carol Gilligan's work and the further research it has stimulated, and then review other distinctions, going all the way back to Talcott Parsons's expressive-instrumental distinction.

Each of the conceptualizations I discuss can be viewed as attempts to recognize a vital "something else" besides the culturally dominant mode. There is something else besides instrumental action and agentic action to get things done; there is something else besides separate and bounded egos and protecting the individual rights of self and others. This "different" voice can be used as a critique of a culture in which male perspectives predominate. This "difference" attributed to women can also be turned against women, however, and used to discriminate against them.

Generally, though, feminists are coming to see the nondominant mode as more attractive and valuable and as deserving legitimacy.

At the same time, a clearer conception of just what this presumably more typically female mode may be is being developed. Gilligan's "different" moral voice, for example, is not new, but she gives this voice more cultural value and new depth and sophistication. Gilligan's book is a persuasive plea for the worth of this more typically female orientation, even though she offers little empirical evidence of the sort psychologists accept that it does indeed differentiate between men and women.

Beginning in the 1970s and using Parsons's instrumental-expressive distinction as a starting point, my colleagues and I began conducting research on expressiveness. Instrumental action is means-end action aimed at controlling the environment outside the immediate interactive system. Expressive action is directed toward interpersonal relationships and the socioemotional meanings of acts and objects. Because instrumental values are so prominent in our culture, expressive action has been downgraded and has been difficult to define in ways that do not sound pejorative.

In our research we begin to rehabilitate the concept of expressiveness by defining and measuring it in such a way that it cannot be confused (as it has often been) with emotionality, dependence, and incompetence. Our findings suggest that the most consistent difference between women and men across diverse samples occurs with regard to the expressive dimension and that instrumentality, at least as we measured it, does not constitute a unitary dimension nor does it consistently differentiate between women and men. I use these research findings to critique some of the reigning ideas of the 1970s in psychology, including the concept of androgyny, and some of the explanations for women's lack of achievement.

It is important to understand that an orientation toward expressiveness or relationality (I use these terms interchangeably) as such has little bearing on power. Power has been taken from women by the privatization of their more relational orientations by marriage and by the cultural emphasis on the male-dominated heterosexual couple. In the United States, women's expressiveness has been put in the service of individual men in the home. This association accounts for why some women are wary of claiming a more relational orientation for themselves: in this society, expressiveness or relationality has been associated with an individual woman "taking care of" a man and his progeny under his headship. Whereas a woman's

relational needs get defined as her "dependency," men may disguise their dependency needs because they are being met every day by women—wives, mistresses, secretaries, and daughters. In fact, women are financially dependent on men, but this dependence must not be confused with psychological dependency. Given a different power situation, women's maternal thinking and greater orientation toward interdependence could be the basis for a more humane society.

A Different Voice

Many of the issues concerning difference and feminists' stances toward difference discussed in the previous chapter resurfaced in response to Carol Gilligan's In a Different Voice. Gilligan, a developmental psychologist at the Harvard School of Education, contends that women tend to construct the moral domain differently than men. While men tend to think of morality in terms of individual rights and noninterference with the rights of others, women are more prone to think in terms of interdependence and the balancing of conflicting responsibilities in making moral decisions. Because masculine modes of approach to moral issues prevail in the public world of social power (including the construction of "tests" measuring moral development), they influence both genders, but women often find these approaches deficient. Gilligan argues that what may sometimes appear to be indecisiveness, lack of clarity, or dependence in women's moral reasoning is produced when women's different construction of moral issues comes into conflict with the controlling masculine view. Gilligan's aim is to articulate the moral voice associated with women so we can hear it as strong, viable, and legitimate in its own right.

Gilligan is careful to point out that she is not talking just about the cliché between male "justice" and female "mercy" or (worse yet) male "thinking" and female "feeling" (p. 69). Rather, she is describing two different ways of constructing the moral domain. Women are more likely to think in terms of equality, process, conflict resolution, not hurting, and caring for others. In masculine conceptions, these considerations may appear to be weak-minded or "unprincipled," but to women they may constitute an alternative perspective, not an unprincipled one but one that is difficult

to articulate in terms of principles or one that balances a complex variety of "principles." This perspective is not well captured by the questions and scoring procedures used to measure moral development. Even though adult females, especially educated ones, may give the "right" responses by pinpointing the "right" principles, the assumptions embodied in the measurement procedures remain flawed. Gilligan's point is that Kohlberg's scale and the culture's whole approach misses or sidelines another reality. The moral reality women see is by no means totally different, however; it is not "another world" but is very much a part of both men's and women's world, and it needs to become an integral part of our thinking about morality.

Gilligan lets us hear this different female voice by her own lucid analysis of selected quotations from individual girls and women in various contexts. She uses male and female children's responses to Kohlberg's moral dilemmas to argue that his "stages" of moral development are more often stages of male moral development because they do not adequately take into account the perspective more typical of females. She also draws on interviews with such diverse groups as college students from a class on moral and political choice, pregnant women considering the pros and cons of abortion, and matched male and female subjects in a study of life-cycle changes. Although her evidence is not of the sort to pass rigorous scientific muster, apparently many people, including feminists, felt intuitively familiar with what she was talking about and hailed her book as a major breakthrough.

Perhaps because her focus was on morality, or perhaps because the timing of her book coincided with a general trend toward an interest in difference, Gilligan's work has received more attention than similar efforts that preceded hers. Not all of this attention has been favorable, however, and negative assessments have become more prominent as those who see the dangers in stressing gender difference react to the positive responses Gilligan's work elicited in the popular feminist press.^[1] These criticisms run the gamut from those that claim the differences are not really there to those that claim the differences exist but they exist in response to oppression or they are the very differences that have kept women separate and stigmatized. For example, an interdisciplinary forum on the book, published in Signs, begins with a historians warning that Gilligan's

"different moral voice" sounds like and "reifies" the nineteenth-century doctrine of the separate spheres,^[2] an argument that is not helped when Gilligan implies that biology may have something to do with this difference. The two psychologists' comments that follow tend to deny any difference (except as stereotype) and complain that Gilligan does not provide any systematic evidence from existing tests that men and women actually do differ.^[3] The anthropologist, Carol Stack, concludes that Gilligan's typology is probably culture-bound and seems not to apply to African-American conceptions of morality. Stack suggests that there is an African-American model of moral development that women and men more nearly share and that sounds to me like Gilligan's more typically female (relational) approach.^[4]

I am oversimplifying these positions, but I believe they generally miss the point and do not address what Gilligan was trying to do, as I have described it above. In response to these criticisms, Gilligan notes that educated women's ability to display male moral reasoning (in a context where they presume it is expected) "has no bearing on the question of whether they would spontaneously choose to frame moral problems in this way." Her point is that the instruments are not measuring what needs to be measured.^[5] It is especially important to understand that "difference" does not mean a denial of the important degree to which men's and women's personalities and goals (and moral thinking) are similar. Overall, similarity and areas of difference need to be taken into account simultaneously; one does not preclude the other. Most differences are only tendencies.

Gilligan deals with the problem of similarity versus difference by recognizing that benefits have accrued to women from their partial assimilation to the masculine emphasis on rights. Women's increased emphasis on their own rights has allowed them to become less dependent on men, and, somewhat paradoxically, less dependent on typically male constructions of the moral domain. As women begin to see assertion of self as less dangerous, Gilligan tells us, they tend to move away from dependence toward a fuller expression of their belief in interdependence, and they also consider it moral to care for themselves as well as others. Thus the emphasis on rights may mitigate whatever tendencies women may have toward self-sacrifice. Put another way, the emphasis on rights

helps women to extricate themselves from the belief characteristic of an earlier stage in women's moral development to the effect that goodness means self-sacrifice, that caring for others must necessarily be at the expense of caring for self.^[6]

Gilligan's book has already inspired and will undoubtedly continue to inspire others to attend to the female voice in other fields. For example, the four authors of Women's Ways of Knowing apply a version of Gilligan's description of women's construction of the moral domain to women's approach to epistemology, or the question of how we know what we know. On the basis of 135 interviews with students and recent alumnae of colleges and "invisible colleges" (human service agencies supporting parents), these authors focus on a differentiation between connected knowing and separate knowing. The authors point out that the procedures used by connected knowers are less well understood than the more culturally legitimate procedures associated with separate knowing. Their task is to make connected knowing more legitimate and to describe it in such a way that it does not simply recapitulate and reaffirm the familiar objective-subjective, impersonal-personal type of dichotomies that have been used against women.^[7]

Gilligan's work and the more recent work it has inspired are the latest in a long line of efforts to delineate broad differences in orientation between the genders in a way that is not pejorative to women and that does not take these more typically female modes as deviation from the masculine standard. Although Gilligan's work concerns differences in typically male and typically female ways of thinking about morality, the contrasts she makes can be and have been compared to other qualitative typologies used by psychologists to delineate differences in masculine and feminine orientations. Examples of such typologies include David Gutmann's distinction between allocentric (masculine) and autocentric (feminine) ego styles, David Bakan's concepts of agency and communion, and Talcott Parsons's instrumental-expressive distinction, which has been in the literature somewhat longer.^[8] These typologies, especially Parsons's, were ignored or rejected in the 1970s by those feminist scholars concerned with similarity and inclusion.

Gilligan, in answering her critics, denies that she is creating a different typology and points out that she is simply criticizing Kohlberg's classification system because it was created using male

subjects and reflects a male point of view. Nevertheless, she has described in some detail what she considers to be an alternative to the male view; in this sense her distinction can be compared to the earlier typologies.

Bakan Revisited

In the early 1970s several feminist psychologists employed distinctions made by David Gutmann and David Bakan to analyze research findings or to design research that could elucidate difference in a way that did not turn out to be pejorative to women. David Gutmann directly attacked the masculine bias in the psychoanalytic concept of "ego strength," arguing that its definition was based on attributes of the male ego only and thereby caused women to appear to lack ego strength. Gutmann attempted to counteract this bias by defining two types of egos. He characterized the masculine ego as allocentric, an ego that tends to objectify others and to experience its own separateness from others. In contrast, the autocentric feminine ego is characterized by more permeable boundaries in which the distinctions between self and others, and between self and environment are less rigid and more merged. Psychologist Rae Carlson found some support for Gutmann's distinctions in an analysis of the words, approaches, and constructs used by college students in written exercises. She concluded, however, that although Gutmann's formulation is clearly the better elaborated in portraying the qualities of masculine and feminine ego functioning, Bakan's formulation using the concepts of "agency" and "communion" is capable of addressing a wider range of phenomena.^[9]

For Bakan, agency and communion are dynamic principles operating throughout nature. Agency refers to the organism acting as an individual in self-protection, self-assertion, and self-expansion. Communion refers to the organism as part of a larger whole and manifests itself in the sense of being at one with other organisms. "Agency manifests itself in isolation, alienation, and aloneness; communion in contact, openness, and union. Agency manifests itself in the urge to master; communion in contractual cooperation."^[10] Bakan argues that if societies or individuals are to be viable, "unmitigated agency" must be combined with "communion." Thus, Bakan's formulation, unlike Gutmann's, allows for both masculine and feminine qualities in one individual.

Both Carlson and another psychologist, Jeanne Block, have used Bakan's very broad behavioral distinction to characterize aspects of masculinity and femininity. Carlson reviewed two samples of one hundred abstracts each of research on sex differences from Eleanor Maccoby's Development of Sex Differences and finds agency and communion to be applicable to almost all of those studies that found significant gender differences.^[11] Carlson's own empirical work also supports the thesis that males express slightly more agentic themes than females in reporting significant emotional experiences. Moreover, she found that females combine agentic and communal orientations more often than males.^[12] This fact will become important to my later discussion. Jeanne Block used Bakan's agency and communion to describe gender role changes over the life cycle because it allowed one principle to be tempered with the other. Block predicts that, as their emotional maturity increases during adulthood, both men and women may eventually arrive at an integration of both orientations, although males will specialize in agency and females in communion.^[13]

Bakan's distinction is important in defining overall orientations, but it is also confusing. The concepts are so all-inclusive that their specific meaning is difficult to pin down. For example, the term agentic describes highly disciplined, self-oriented striving in the occupational world as well as the impulse-ridden self-assertion and self-extension of the infant. Such a remarkable equation may obfuscate more than clarify. If the referents of the concepts are carefully specified and not overgeneralized, however, they can be of use in producing a different and more complex picture of male and female differences and similarities.

Michael Finigan defined Bakan's concepts quite specifically and applied them in an observational study of the interaction in male-predominant and female-predominant groups, using groups of professionals in six social work agencies. In the interaction of these groups Finigan measures what Bakan calls "unmitigated agency" as "holding the floor for sixty seconds or more, interrupting, establishing the floor and reestablishing the floor." As one might guess, he found that males in both male-predominant and female-predominant groups were much more likely than females to engage in unmitigated agentic behaviors such as "floor holding."^[14]

Finigan found, however, that although males were more likely to engage in "unmitigated agency," both males and females en-

gaged much more equally in other agentic behaviors such as giving opinions and stating positions and in initiating ideas or solutions. Finigan also found that groups in which females predominated exhibited much more communal behavior—that is, they were more supportive of each other—than the groups in which males predominated. In male-predominant groups, although an individual might have an ally or so, an atmosphere of group supportiveness was absent. It is important to note that female professionals were supportive of each other. The giving of support was not something they did just for men, at least not in that setting. Finigan's work is important in showing that women do act agentically but in ways that are less likely to threaten the groups cohesion.

Expressiveness Revisited

Also in the early 1970s, I, along with several colleagues at the University of Oregon, attempted to use Parsons's more sociological and less global instrumental-expressive distinction as a basis for research on orientational similarities and differences between college men and women. Following Parsons, we defined expressiveness as a concern with the relations among the individuals within a social system, especially the attitudes and feelings of group members toward the self and each other. Instrumentality we defined as an orientation to goals outside the relational system itself. In Parsons's functionalist terms, instrumental action relates system to environment, and expressive action relates units within the system; both types of action are essential for the survival of any given social system. Both actions are goal-oriented, but the goal of instrumental action is external to the relational system and the goal of expressive action is internal to the system.

By the mid-1970s, Parsons's instrumental-expressive distinction had received much criticism from feminists. At the same time, however, the terms had been adopted by academic psychologists and were being used in "androgyny" research. Because the terms are still used, we decided to attempt to rehabilitate the concept of expressiveness by defining it carefully and by removing it from the context of Parsons's functionalism.^[15] We wanted to show that the pejorative connotations that had been attributed to the concept are not a part of its definition.

One major line of criticism of the terms instrumental and expressive had to do with Parsons's use of them in his functionalist description of the family. Taking the nuclear family as the relevant social system, Parsons maintains that the adult male role is specialized in the instrumental direction and the adult female role is specialized in the expressive direction. Parsons views this specialization as functional for marital solidarity because it eliminates competition between husband and wife. Many feminists at the time took this view to be a prime case of Parsons's attempt to justify or legitimate the middle-class mentality of the 1950s, when married women did not work outside the home or worked at minor jobs. In other words, the instrumental-expressive distinction was interpreted as a legitimation for the male provider role and the stay-at-home-wife role.

Actually, however, Parsons's argument has nothing to do with justifying housewifery, which he saw as a shrinking and unsatisfying role for middle-class women. His argument is that the male role is anchored in the predominantly instrumental occupational sphere, and the female role is anchored in the more expressive kinship sphere. Thus, if one takes Parsons's statement about role specialization in the nuclear family as description rather than prescription, the statement is an accurate general characterization of the division of labor that tends to prevail in husband-wife families. Husbands do tend to be the primary breadwinners (they certainly make more money), and wives do continue to be primarily responsible for "taking care of the relationship." Parsons's error is not in his description but in his failure to emphasize the differential power that accrues to the male provider role and the high valuation of instrumental action in this culture.^[16]

Our interest in the instrumental-expressive distinction, however, was not in using the terms to describe interaction in social systems but as general descriptors of gender-differentiated personality orientations. Here again, many feminists in the late 1960s and early 1970s rejected "expressiveness" because it had become associated with emotionality, incompetence, and dependence.^[17] My associates and I questioned whether these negative connotations formed an inherent part of the definition of expressiveness and tried explicitly to counteract this view of expressive orientations.

Although expressiveness does engage socioemotional skills, it is

misleading to view it as simply being emotional or emotionally labile. Expressiveness does not mean simply expressing emotion in an unpatterned way. Women, in this culture at least, are provided with patterned ways of expressing and negotiating socioemotional subtleties in interaction, whereas men are enjoined to be inexpressive or nonexpressive. Because of this inexpressiveness, men (when the inexpressive mask breaks down) are more likely to express raw emotion, spontaneous unpatterned emotion, than women. Women may resonate with, respond to, cope with, and even define emotion for others, but this is hardly the same as being emotional. Expressiveness then is an integrative skill, not unbridled weeping, so to speak. This view of expressiveness is consonant with Arlie Hochschild's description of women's "emotion work." Hochschild makes a clear distinction between being "easily affected by emotion" and the action of managing emotion.^[18]

Neither should expressiveness be confused with dependence and passivity. The seeking of interpersonal rewards is not the same thing as being dependent and lacking autonomy. Expressiveness does not mean giving in to other people, although people in an instrumental culture would tend to interpret expressiveness in this way. This instrumental bias leads us all to think in terms of who is getting the best of whom. Expressiveness denies this way of looking at the world and does not see people and interaction as instrumentalities but more as ends in themselves. Thus, we attempted to develop measures that could separate expressive orientations from an autonomy-dependence dimension.

Finally, to lay claim to expressiveness does not mean one cannot also be instrumental. The research I and my colleagues undertook does not assume that instrumental orientations and expressive orientations are at opposite ends of a continuum, so that if one is instrumental, one cannot be expressive, but rather that they may constitute two separate dimensions. Whether the two are related and how should be an empirical question.

On the basis of the theoretical considerations described above, we had judges select adjectives from the Adjective Check List often used by psychologists that seemed to them to represent the three dimensions with which we were concerned: instrumentality, expressiveness, and independence or autonomy.^[19] Only adjectives on which there was high agreement among the judges were used.

We presented these adjectives to male and female college students enrolled in introductory sociology courses at the University of Oregon and asked them to rate themselves on each adjective by checking "very true of me," "somewhat true of me," "somewhat untrue of me," and "very untrue of me." Multivariate analyses were applied to determine to what extent these students grouped items together into the three dimensions we had hypothesized.

The major finding from these analyses was that only the adjectives we had selected to represent the positive aspects of expressiveness seemed to form a unified group. Adjectives such as sympathetic, understanding, pleasant, considerate, good-natured, warm, and obliging clustered together as a distinguishable group.^[20] The theoretical dimension we had labeled instrumental broke up into two distinct subgroups represented by thorough, efficient, industrious, and planful on the one hand, and analytical, foresighted, and rational on the other. The autonomous dimension contained three separate groups: (1) a stern, forceful, aggressive, outgoing, and assertive cluster, (2) an independent, active cluster, and (3) a daring, adventurous cluster. Thus expressiveness, but not instrumentality, showed up as a distinct dimension.

The largest difference between men and women students shown by the scores on these dimensions was that men saw themselves as less positively expressive than women. The differences on the other dimensions were much smaller. This finding is consonant with the results of Bennett and Cohen's earlier study of self-attribution of traits: "the major difference in the self-concepts of the sexes is that women conceive of themselves as being richer in the positive qualities of social warmth and empathy."^[21] These authors go on, however, to say that women secondarily see themselves as more helpless, timid, and fearful than men and suggest that women's kindness may be developed out of fear of attack. That is, they offer the "psychology of the oppressed" explanation for women's greater expressiveness. Our findings based on internal correlations within sex groups, however, were that women tend to relate expressiveness to independence and that men do so much less clearly. Men strongly associated the words active and assertive with instrumental qualities and much less so with expressive qualities. These associations were not true for women, however. In other words, although these college women included positive expressiveness with positive instrumen-

tal qualities and autonomous qualities in their self-pictures, college men could not so easily include expressiveness with instrumental and autonomous qualities.^[22]

Ten years later, in 1982, we returned to introductory level sociology courses at the University of Oregon with the same list of adjectives and instructions we had used before. The list was also administered to high school juniors and seniors at a small rural school in Oregon, and a subset of the items was administered to a statewide representative sample of male and female nurses. Using a similar analysis, we again found that a strong positive expressive factor underlay the self-ratings of both females and males in these diverse groups.^[23] Indeed the expressiveness factor was even more clearly seen as a single factor by men in 1982 than it had been seen by them in 1972. Moreover, in every group in both years, women not only saw positive expressiveness as a unitary cluster but rated themselves significantly higher on these expressive qualities than men did.

The instrumental items continued to divide into an industrious cluster and an analytic cluster. College women in 1982 (but not in 1972) and women nurses rated themselves higher on industriousness than men. College men in 1972 and 1982 and the male nurses reported significantly higher levels of analytic characteristics (analytic, foresighted, rational). Not all males, however, see themselves as analytic, since among the high school students, women rated themselves as higher on analytic qualities than men. Moreover, the differences we found between adult men and women on the analytic items are much smaller than the difference on positive expressiveness.

With regard to what we call the autonomy dimensions, the factor structures were least consistent from one sample to the next, but two main factors appeared to be those of forcefulness (stern, forceful, aggressive, outgoing, assertive, independent, and active) and, secondly, adventurousness and daring. (The items "independent" and "active," which were separate in 1972, combined in 1982 with other items on the forceful dimension.) Although high school boys and male nurses, unlike college males, rated themselves as more adventurous than their female counterparts, there were no statistically significant gender differences in any of the groups studied on the large group of adjectives forming the forcefulness cluster.

Finding no gender differences in "forcefulness" contradicts the strong stereotype of forcefulness as "male."

The relative complexity and multidimensionality of both the instrumental and autonomy scales and the gender differences on them being small and related to life-cycle stages (high school boys see themselves as daring, and not analytic) strongly suggests that expressiveness is a more basic dimension of gender difference. At least as represented by our fairly diverse samples, males see themselves as having less of a socially desirable characteristic than females. This is quite different from the usual statement that females lack instrumental or autonomous qualities and thus by implication deserve their secondary status.

Our findings support a reconceptualization of gender difference in which it is males, not females, who claim less of certain desirable qualities. This reconceptualization is not a simple turning of the tables in which expressiveness becomes "good," and autonomy "bad." It is not what Alice Echols calls "the new feminism of yin and yang." Neither does it support Parsons's implication that instrumentalness and expressiveness are complements. Rather, our findings suggest a more complex view of difference in which females are able to combine expressive qualities with autonomous qualities in their self-pictures, and males deny expressiveness in theirs.^[24] Expressive qualities in women have tended to be defined as the opposite of those forceful-autonomous qualities that men use to explain or justify male dominance. Our research suggests that females see their expressiveness not as the weak opposite of forcefulness but as a desirable and positive set of characteristics that do not necessarily limit their autonomy. Men are more likely than women to see autonomous traits as contradicting expressive traits. This latter finding suggests further that men are more likely to downgrade expressiveness than women, because they are more likely than women to associate expressiveness with dependence.

The Trouble with Androgyny

During the 1970s, feminist psychologists developed the idea of measuring "androgyny" as opposed to difference. In the original masculinity-femininity scales, if one were more masculine, one had to be less feminine. This idea needed to be challenged. Although

the idea is not new that masculinity and femininity were multidimensional rather than unidimensional concepts and that they could not be represented by a single bipolar dimension with masculinity on one end and femininity on the other, feminist psychologists were the first to construct "androgyny" scales, which explicitly denied the bipolar assumption concerning masculinity and femininity.^[25] The feminist impulse behind these scales was to counteract the negative assessment of women that seemed to result from an emphasis on gender difference by suggesting that the ideal personality might combine both masculine and feminine traits in its self-image. The desirability of androgyny might be tested by demonstrating that people who saw themselves as having both masculine and feminine traits would be more likely to respond appropriately to a greater variety of situations than people who saw themselves as either decidedly masculine or feminine.

Bem's Sex-Role Inventory consists of a mixed list of twenty stereotypically masculine characteristics (such as "self-reliant" and "competitive"), twenty stereotypically feminine characteristics (such as "happy" and "sincere"), and twenty gender-neutral characteristics. Subjects are asked to indicate on a seven-point scale how well each characteristic describes themselves. Subjects who score high in both stereotypically masculine and stereotypically feminine traits are considered androgynous.^[26] The Personality Attributes Questionnaire is also based on stereotypes. On this test, subjects are asked to respond on a five-point rating scale ranging from strong denial of the characteristic to strong endorsement, or from one trait to its polar opposite; for example, not at all kind to very kind, very cold in relations with others to very warm in relations with others, and so forth.^[27] Both scales are based on the requirement that the items be considered socially valued behaviors.

Matters did not work out well for androgyny, however, for a variety of reasons. Both Bem's Sex Role Inventory (BSRI) and Spence, Stamp, and Helmreichs Personality Attributes Questionnaire (PAQ) were widely used for a few years to measure androgyny before they were overtaken by serious attacks on theoretical, methodological, and philosophical grounds.^[28] In my view, the basic problem with androgyny is simply this: Putting two wrongs together does not make a right; that is, combining sexist stereotypes of masculinity and femininity does not lead to a nonsexist solution.^[29] The problem with Bem's and with Spence and Helmreich's

measure of masculinity and femininity is that the adjectives they use do not accurately represent the concepts of instrumental and expressive, at least as I and my colleagues define them.

Although psychologists have adopted the terms instrumental and expressive, they have paid relatively little attention to their conceptual definition. Neither the BSRI nor the PAQ items were chosen with the terms in mind, but in summarizing their results, both Bem and Spence and Helmreich equate their masculine items with instrumentalness and their feminine items with expressiveness. Thus Bem considers words such as childlike, shy, and yielding, which were included on her scale, to be indicative of expressiveness, yet we would not associate these words with expressiveness by our or Parsons's definition. The PAQ includes "emotional" and "does not hide emotions" among its expressive items and includes both "independent" and "active" with its instrumental items. In our view none of these properly belong with expressiveness or instrumentalness. To some extent then, however well-intentioned, these conceptions of androgyny undercut the empowerment of women by reinforcing the stereotype of femininity (and expressiveness) as weakness and emotionality.

In fact, Bem's own early research showed that women who embraced the traits considered feminine-expressive on her scale to the exclusion of others were so passive that they were competent in neither masculine nor feminine spheres.^[30] Although it is true that "femininity" is stereotypically associated with passivity and dependence, expressiveness, technically defined as a relational orientation, does not have this implication and did not seem to have this implication for our women subjects and for the women subjects in other studies that isolated this expressive factor.^[31]

Another difficulty with the idea of androgyny as embodied in these scales is that it consists of equal parts masculine and feminine. This conceptualization precludes the possibility that expressive qualities are basic to the self-concepts of both genders and constitute the common human matrix from which men deviate in a negative direction. But this possibility is exactly what our research suggests: males see themselves as less expressive or relational than females, but females do not see themselves as less instrumental than males. Obviously, these findings challenge the perspective that tends to equate male with human and female as deviation.^[32]

Bem has abandoned the study of androgyny in favor of a devel-

opmental focus on how children form "gender schemas" that they use to encode and organize information. She proposes that we try to raise "gender-aschematic children", that is, children who are less gender-conscious and less gender-differentiated. She does not propose (wisely, I think) that we attempt to eliminate recognition of gender difference altogether, but rather suggests that we try to confine the "meaning" of gender to the genital differences themselves.^[33] Sensible as it sounds, it is difficult to limit gender to physiological differences, because these are themselves symbols. Physiological markers still leave the door wide open for masculinist interpretations—one has only to think what Freud was able to do with the fact that boys have penises! Physiological features are given social meanings with implications for social roles far beyond physiology. Even the physiological features one chooses to emphasize are telling. Bem tries to keep a strict parallelism (penis and testicles, clitoris and vagina). She ignores the uterus, undoubtedly because she feared it could be used against women. Simply warning children about sexist views as Bem advocates does not get at the male-dominant features of the nuclear family structure itself, where she assumes child-rearing should take place.

Interdependence versus Dependence

The confusing of psychological dependence with expressiveness partially explains why many feminists do not see expressiveness in a positive light. This confusion has also led some feminists into self-blame or a "women are their own worst enemy" position. The logic runs something like this: Women are expressive, which is erroneously assumed to mean dependent. This dependency is then seen as leading to any number of other defects such as blind conformity, passivity, inability to take control of one's life, and inability to achieve or to "get ahead." The next step is to conclude that women are their own worst enemy, which of course is what men have been saying, or suspecting, all along. Women are then exhorted to recognize these deficiencies, get over their dependency, take charge of their lives, quit complaining, and make something of themselves. Colette Dowling's The Cinderella Complex: Women's Hidden Fear of Independence, which appeared in 1981, is one example of this approach that became a bestseller. Dowling's main thesis is that

"personal psychological dependency—the deep wish to be taken care of by others—is the chief force holding women down today" (p. 21).

I do not deny that women have dependent motivations, but these are brought about by structural factors (not the least of which is women's economic dependency on men) other than those related to women's expressiveness. As our research and women's common sense suggests, expressiveness is not the same as dependency, and many women do not associate it with dependency. Miriam Greenspan, author of A New Approach to Women and Therapy, argues, counter to Dowling, that women must not deny their authentic need for "intimate connection" by confusing that need with helpless dependency and then blaming themselves for being dependent.^[34] As a therapist, she tries to show her women patients that their relational needs are legitimate.

The tendency to confuse expressiveness with dependency has also plagued studies of "achievement" in psychology. A widely reprinted article written a decade earlier by Judith Bardwick and Elizabeth Douvan provides an example. The article is entitled "Ambivalence: The Socialization of Women," and is itself a model of the authors' own ambivalence, created in part by the initial failure to separate dependency from expressiveness. They begin with a list of traits that they say are ascribed to women. This list includes psychological dependence, passivity, and fragility along with the more desirable traits of interpersonal orientation, nurturance, and supportiveness. They then argue that girls' problem is that they remain dependent on others for feelings of affirmation. When Bardwick and Douvan asked their students, both undergraduate and graduate, what they wanted, they answered, "When I love and am loved; when I contribute to the welfare of others; when I have established a good family life and have happy, normal children; when I know I have created a good, rewarding, stable relationship" (p. 231). While these statements do sound "traditional" in the sense that they imply children and marriage, they also express an active and responsible orientation to the general welfare of others. Bardwick and Douvan take a dim view of these statements, however, and conclude that "up until now very few women have succeeded in traditionally masculine roles, not only because of disparagement and prejudice, but largely because women have not

been fundamentally equipped and determined to succeed" (p. 233). What it takes to succeed in these roles, according to Bardwick and Douvan, is "objectivity rather than subjectivity, aggression rather than passivity, the motive to achieve rather than a fear of success, courage rather than conformity, and professional commitment, ambition, and drive" (p. 233).^[35]

Although Bardwick and Douvan criticize women for not meeting these standards, their point is that women are ambivalent about them. They point out that "the masculine [is] the yardstick against which everything is measured." They go on to say "since the sexes are different, women are defined as not-men and that means not good, inferior" (p. 234). The ambivalence women feel then is the product of women liking the way they are and at the same time, because both men and women esteem masculine qualities and achievements more, knowing that their own qualities are second rate. The authors then find it "disturbing to review the extent to which women perceive their responsibilities, goals, their very capacities as inferior to males" (p. 234). Although Bardwick and Douvan at this point try to upgrade women's contributions in traditional roles, it is a losing battle, and women end up sounding weak. The traits attributed to women become even more ridiculous when one thinks of women who are not white and upper middle class. Those women who have always worked and who also constantly take on responsibility for others on a reciprocal basis can hardly be called dependent and passive. To paraphrase Sojourner Truth, "aren't they women?"

I do not blame Bardwick and Douvan or other psychologists for their ambivalence concerning women in 1970. Many of us shared this ambivalence because we accepted the more typically masculine definition of what women were. Bardwick and Douvan's article made a positive contribution, however, in pointing out the ambivalence that accepting a male paradigm causes women to feel, and in recognizing the existence of a male paradigm in the first place. The task now is to extricate ourselves from that paradigm more clearly than was possible in 1970.

Increasingly, what seems to be happening is that women are able to attain and achieve in high-level jobs while retaining a sense of self as expressive. Lillian Rubin describes middle-aged middle-class women who, while sitting in their offices surrounded by work

and the symbols of work, still answered the question of "who and what they are" by describing themselves as "warm, sensitive, considerate and kind." No one mentioned work.^[36] In a sample of college-educated women born in the 1950s, Rosalind Gottfried documents that these women did not maintain that work took primacy in their lives over relationships. Although they did view work as important, these women were most concerned with their connections to other people, and these connections were more basic to their self-concepts. Moreover, these attitudes about connectedness were those of women who were the least passive, most feminist group— those most concerned with directing their own lives.^[37]

In 1976 Jean Baker Miller made the point clearly that women are being punished for making affiliations central to their lives and that women can be highly trained, high achievers, and still give great weight to affiliations. To simply dismiss women's need for affiliation as dependency is an error.^[38] Finally, if we look at school achievement, it is incorrect to assume that women perform less well than men. With regard to school performance rather than occupational performance, the "problem of female underachievement" ceases to exist. Women achieve beyond their potential as measured by IQ tests to a greater extent than men at every grade level through college.^[39]

Men and Dependency

Although the cultural stereotype has it that men are psychologically independent and women are dependent, from a structural standpoint this is hardly the case. The appearance of masculine autonomy outside the home is made possible by wives meeting men's dependency needs in the home, and other supportive females meeting their needs outside the home. Men's dependency needs are not readily apparent to outsiders because they are taken care of by women in the normal course of life. The key to how this operates rests on the males' dominance in relationships in which men are dependent on women.

Middle-class marriages are clearly structured in such a way that husbands' dependency needs are met by wives. Women as wives "take care of things" and continue to do so even when they have full-time jobs outside of the home.^[40] In fact, wives sometimes de-

scribe their husbands as "big babies"; the "bigger" in importance he is, the bigger the baby. Some men may be babies, but men do not present themselves as babies to their fellow males. Only in circumstances and roles in which they have a firm power base do men express dependency. Being the husband and male head of household provides such a power base.

Women may also play supportive roles to men at work. The clearest example here is the secretary—the office "wife." Often a boss will even admit to his dependency on a woman, but generally only in the context of his being in the superior position in the relationship. A boss might say of his secretary, "I do not know what I would do without Jenny here; I depend on her for everything." At the same time, in one way or another he will also make it clear that her ministrations to him ultimately testify to his superior judgment—he hired her, he pays her, she works for him under his direction, at his behest.^[41]

Whereas men's dependency needs are met by wives (and wife substitutes), wives' dependency needs often go unmet by husbands. Who soothes wives and binds up their psychological and physical wounds? Who takes care of things for wives? The poorer mental and physical health that characterizes married women compared to married men has been attributed to women's role as the mental and physical caretakers of husbands and children. Wives are the ones who cannot ever be dependent—to the detriment of their own physical and mental health.^[42] Yet women are called psychologically dependent. In reality, a sense of connectedness with others is different from and transcends both "taking care of" and "being taken care of." It involves a generalized stance that can be egalitarian and mutual. It can be shared by men and be a part of relationships between women and between women and men.

Women and Aggression

Aggression is often considered in common parlance to be the opposite of dependence, passivity, and failure to make it. This juxtaposition makes it abundantly clear that aggression is at the "good" end of the pole. In their comprehensive review of psychological studies on gender differences in children, Eleanor Maccoby and Carol Jacklin find girls to be less aggressive than boys. Although

Maccoby and Jacklin generally consider there to be fewer gender differences than many other psychologists, they argue that there is a clear-cut difference with respect to aggressiveness. They contend that males are more aggressive than females, both physically and verbally, directly and indirectly, and in a wide variety of settings. Does this mean women are doomed to passivity? Not in the least. Maccoby and Jacklin define aggression as action with the intent to hurt. It is not constructive and it is not the opposite of passivity.^[43]

It is true that girls have more anxiety about aggression than boys. Some have interpreted this to mean that girls have aggressive tendencies equal to those of boys but repress them out of fear of punishment or retaliation. Maccoby and Jacklin suggest, however, that if this were so, surely the aggression would come out in some attentuated form. What happens, though, is that boys act out aggressive impulses in play as well as in reality. In addition, boys are more likely than girls to aggress in the presence of weakness in another male. Girls do not respond to weakness in either boys or girls with aggression.

We know that women account for slightly more child abuse than men, but when one takes "opportunity" into account, that is, the time women spend with children compared to the time men spend, this tendency is reversed.^[44] A review of the experimental literature on aggression argues that women may act as aggressively as men under certain experimental conditions. These conditions involve a situation where the women do not empathize with the victim and in which they feel the aggression is justified. Over all situations, of all the traits discussed in the review, aggressiveness does appear to be most clearly a generalized trait that characterizes males far more than females.^[45]

It is important that readers be aware of their own biases about the value of aggressiveness. In a society governed by a masculine paradigm, aggression is likely to have desirable connotations. In the public mind, it is often associated with competitiveness, single-mindedness, or strong will. But why? Maccoby and Jacklin do not define aggression as a desirable characteristic and argue that it is as likely to interfere with constructive activity as it is to underlie it. Their definition of aggression as the intent to hurt emphasizes its antisocial nature. While our measure of expressiveness was only a tiny aspect of the whole picture, male aggressiveness may be re-

flected in the lesser amount of positive expressiveness males attributed to themselves in our study.

Like aggression, the term dominance tends to have a positive meaning in a society controlled by a masculine paradigm, which tends to define many situations in terms of "dominate or be dominated." A female paradigm, on the other hand, might deny the necessity for either dominance or submission. Resistance to domination would be desirable, but domination would not be. Male aggression could play into dominance-striving, but as Finigan's work suggests, dominance is not at all the same as constructive leadership.

In an article published after their book, Maccoby and Jacklin examine the question of male aggression cross-culturally and conclude that in other societies too, boys between three and six years old are more likely to initiate aggression than girls in the same age range. This initial difference may be either maintained or eliminated or even reversed by subsequent socialization experiences. Maccoby and Jacklin note that one type of socialization experience that has been found to moderate aggressiveness in boys is being assigned to child care. There are forms of aggression in which no gender differences are found, such as self-defense, quarreling over desirable resources, or even effecting a swift, silent kill. Maccoby and Jacklin suggest that future research on aggression will need to pay more attention to the particular interpersonal setting in which aggression takes place and also to differing types of aggression. Much aggression among boys occurs in male-male pairs in noisy contests over dominance. Girls may fight with each other but not in order to establish dominance.^[46]

When Maccoby and Jacklin speak of aggression, they clearly distinguish it from both activity and competition. Confusion over this issue has led Sarah Hrdy in her book The Woman That Never Evolved to argue against "maternal" feminists because she assumes that they envision women as passive and noncompetitive. Hrdy contends that studies of female primates show them to be highly competitive with each other and that "competition among females is one of the major determinants of primate social organization, and has contributed to the organisms women are today" (p. 189). In part Hrdy's stress on female competitiveness seems merely to be reiterating "the logic" of the sociobiological argument in its crudest form, namely, that whatever exists must have gotten there by having

won out in some sort of competitive struggle. In other words, a passive, noncompetitive woman could never have evolved. Hrdy envisions females' mothering as being the basis for their competition with other females. "Throughout millions of years of evolution, mammalian mothers have differed from one another in two important ways: in their capacity to produce and care for offspring and in their ability to enlist the support of males, or at least to forestall them from damaging their infants" (p. 189). In their efforts to secure resources for themselves and their offspring, mammalian mothers engaged in subtle competition with one another, which sometimes involved simply nonconfrontation and keeping out of the way. Hrdy is not saying females behaved as aggressively as males; her concern is to show that females had dominance hierarchies and were not politically insignificant or passive. This does not directly contradict the thesis that females are less likely than males to initiate aggression and make direct dominance attempts.

Indeed, the significance of Hrdy's descriptions of female primate behavior and speculations about early female human behavior is that only among humans do females become clearly subordinated to males. This subordination has to do with human cultural institutions. Hrdy speaks of various means to control women's sexuality, including clitoridectomy and castration of daughters and wives as well as those marital residence rules that separate women from their kin.

This picture of animal mothers that Hrdy draws has been systematically denied by the cultural image of women that defines them not as active mothers but as passive sideliners, cheering on their men. Niles Newton has argued that since the industrial revolution there has been a sharp conflict between what she calls a woman's "cultural femininity" and her "biological femininity." Newton says that although "cultural femininity . . . decrees that women should be passive," actually "the woman in her female biological role must be active, productive, and capable of concerted effort." Newton points out that not only is bearing and nursing a child physically and emotionally demanding, but the care of a child is not a matter of passively providing comfort; it involves active and sometimes aggressive interventions on behalf of the child in the physical and social environment. In sum, there is nothing passive about maternal emotions.^[47]

Finally, in addition to distinguishing expressiveness from emo-

tionality, dependence, ineptitude, and inability to aggress, it is also important to emphasize that the term must always be defined at a very general level. Only when expressiveness is thought of as a mode of personality organization or as a generalized stance rather than in terms of specific traits can it make sense to claim that expressiveness can differentiate women from men cross-culturally. Although we used descriptive adjectives defining expressiveness in our research, these are far from ideal measures because they are too narrowly confined to an empathetic dimension and also because they inevitably become intertwined with a very specific cultural and social structural context.

For example, these adjectives may be class or race bound. Consider, for example, the popularity of "interpersonal sensitivity." This term may mean nothing more than a style of interacting in which one listens to others and tries to assess their needs and agendas as a means of selling oneself or selling something else. It may simply be an expressive overlay to instrumental action—a style, not an orientation. "Interpersonal sensitivity" has become something that middle-class whites value highly, and it is a stance that the so-called new man is readily adopting.^[48] This is not what I want to convey by the idea that women are more relational, and I do not believe it is what women mean when they say they wish men could be more relational. Relationality is a stance that takes others into account not as "other" but as important in themselves. The forms that a relational orientation could take would vary greatly in terms of class and race or culture.

Some of the most interesting work now on gender similarity and difference focuses on interaction patterns rather than on personality orientations. Thus, for example, Marjorie Goodwin and Charles Goodwin have analyzed the "conversational procedures" used by preadolescent girls and boys in informal arguments. Their main finding was that both boys and girls engage in argumentation and generally in quite similar ways. They found, however, that girls argued in a more complex manner than boys. Girls used what the researchers call a "he-said-she-said" structure unknown among boys. This structure allows for extended and complex debate that nevertheless saves the faces of both accuser and accused by reference to third and fourth parties, thus avoiding direct confrontations of the "I win, you lose" variety. Such a method avoids direct aggression, while still getting the job done.^[49]

Summary

In this chapter I have presented empirical support for an assessment of women and men in terms of orientational similarities and differences that is less distorted by male bias. Women see themselves as more relational or interdependent or expressive than men see themselves. Instrumentality, however, does not hold up as a single dimension, and women and men do not differ consistently with respect to its components. Women's greater expressiveness is not to be confused with dependence or lack of autonomy. Men are more likely than women to make a mental connection between relationality and dependence. Moreover, structurally, men in fact have their dependency needs met. I used Maccoby and Jacklin's conclusions concerning greater aggressiveness in a wider range of contexts among males to suggest that this greater aggressiveness may play a part in our male subjects' reluctance to see themselves as being as positively expressive as our female subjects.

The reader may feel that much of the preceding discussion involves a very airy quibble over words and concepts—but words are never neutral. Words carry evaluative and power connotations, and definitions become important because we live in a male-dominant society. Feminist analyses or investigations must be highly self-conscious about terms. But one can only get so far with these large global assessments; the major intervening factor is power.

Expressiveness or relationality as an orientation does not determine power. Instrumentality, as we defined it, does not directly involve power either. Much of the purpose of my detailed discussion of the definition of expressiveness was to make clear that the terms that are obviously relevant to power (or lack of it), such as submissive, yielding, and dependent, do not properly belong in the expressive dimension. Being expressive does not render one powerless or powerful, but the structure of male-dominated marriages make women less powerful. Given a different power situation, women's somewhat greater expressiveness can also be the basis for human bonding.

In Chapter 5, I will discuss how women's mothering and women's greater relationality develop and how they might be connected. Before doing so, however, I will examine the hypothesis that gained popularity among some feminists in the 1970s, namely, that women's mothering contributes to male dominance. There is an element of

truth in this argument, but it is a partial truth at best. My emphasis will be on the positive motivation that being mothered produces. I will argue that the mother is the focus of the learning by both genders of common human expressive qualities, although males may later deny these qualities and substitute male dominance. This transformation is not the work of mothers, but of adolescent male peer groups and to some extent fathers.