R. W. Connell’s path-breaking conceptualizations of multiple masculinities (Connell, 1995, 2000) and hegemonic masculinity (Connell, 1987, 1995, 2000) have been taken up as central to gender theory and scholarship. (For an extensive overview of theory and research on masculinities, see Kimmel, Hearn, & Connell, 2005; for a summary of critiques of the concept hegemonic masculinity, see Connell & Messerschmidt, 2005). Although researchers have made widespread use of these concepts, femininity is still decidedly under-theorized (Connell & Messerschmidt, 2005; Martin, 1998; Pyke & Johnson, 2003). While there have been important attempts to theorize female masculinities (e.g., Halberstam, 1998; Messerschmidt, 2003) and hegemonic and subordinate femininities (e.g., Pyke & Johnson, 2003), a compelling and empirically useful conceptualization of hegemonic femininity and multiple, hierarchical femininities as central to male dominant gender relations has not yet been developed. In their most recent reformulation of the concept hegemonic masculinity, Connell and Messerschmidt (2005) call for more theory and research on femininities.

The concept of “emphasized femininity” focused on compliance to patriarchy, and this is still highly relevant in contemporary mass culture. Yet gender hierarchies are also impacted by new configurations of women’s identity and practice, especially among younger women-which are increasingly acknowledged by younger men. We consider that research on hegemonic masculinity now needs to give much closer attention to the practices of women and to the historical interplay of femininities and masculinities. (p. 848)

In this article, my goal is to recover the feminine other and place it in the center of a theory of gender hegemony. My use of the term “other” refers to the ways in which the feminine and femininity have been defined or displaced in work on masculinity. To recover the feminine other conceptually, I first outline Connell’s main contributions to our understanding of gender hegemony and multiple configurations of masculinity. I then discuss conceptual and empirical difficulties with applying Connell’s framework to femininity. Finally, building on the work of Connell and others, I offer an alternative conceptual framework for how gender hegemony operates through masculinities and femininities and that places men’s dominance over women at the center, allows for multiple configurations of femininity, and can be used for empirical research across groups and settings. In my discussion of the alternative model, I outline how it builds and improves upon, not only the original conceptual framework offered by Connell and developed in masculinity research, but also upon the recent reformulation by Connell and Messerschmidt (2005).

A brief description of Connell’s model

According to Connell (1995), gender can be defined as the ways in which the “reproductive arena”, which includes “bodily structures and processes of human reproduction”, organizes practice at all levels of social organization from identities, to symbolic rituals, to large-scale institutions (p. 71). As a central feature of gender relations, Connell defines masculinity as “... simultaneously a place in gender relations, the practices through which men and women engage that place in gender, and the effects of these practices on bodily experience, personality and culture” (p. 71). From this definition, we can summarize masculinity as having three components. First, it is social location that individuals, regardless of gender, can move into through practice. Second, it is a set of practices and characteristics understood to be “masculine”. Third, when these practices are embodied especially by men, but also by women, they have widespread cultural and social effects. There are individual effects-occupying the masculine position and performing it affects the way individuals experience their bodies, their sense of self, and how they project that self to others. While these are individual effects, it is important to point out that, for Connell, masculinity is not reducible to individual expression or experience. Masculinities and femininities can become “gender projects” in the lives of individuals, but they do not refer to features of or specific kinds of people. Instead of possessing or having masculinity, individuals move through and produce masculinity by engaging in masculine practices. In this way, masculinity is an identifiable set of practices that occur across space and over time and are taken up and enacted collectively by groups, communities, and societies. Through their recurring enactment over time and space, these practices structure the production and distribution of resources, the distribution of power in the form of authority, cathexis, by which Connell means the social arena of desire and sexuality, and symbolism or the production of meaning and values (Connell, 2000). In summary and to reiterate, masculinity is a social position, a set of practices, and the effects of the collective embodiment of those practices on individuals, relationships, institutional structures, and global relations of domination (Connell, 2000).

Importantly, and this is one of the major contributions offered by Connell, gender hegemony operates not just through the subordination of femininity to hegemonic masculinity, but also through the subordination and marginalization of other masculinities. Connell (1995) defines hegemonic masculinity as “the configuration of gender practice which embodies the currently accepted answer to the problem of the legitimacy of patriarchy, which guarantees (or is taken to guarantee) the dominant position of men and the subordination of women” (p. 77). Hegemonic masculinity, when embodied by at least some men over time and space, legitimates men’s domination over women as a group.

According to Connell, there are no femininities that are hegemonic (Connell, 1987). “All forms of femininity in this society are constructed in the context of the overall subordination of women to men. For this reason, there is no femininity that holds among women the position held by hegemonic masculinity among men” (187). Instead, there is what Connell refers to as emphasized femininity. Connell writes,

One form [of femininity] is defined around compliance with this subordination and is oriented to accommodating the interests and desires of men. I will call this ‘emphasized femininity’. Others are defined centrally by strategies of resistance or forms of non-compliance. Others again are defined by complex strategic combinations of compliance, resistance and co-operation. (pp. 184–185)

Here, Connell suggests that there are multiple femininities, but the focus is more on the relationships among masculinities and therefore Connell does not elaborate further.

Although emphasized femininity is central to men’s dominance over women, it is not the only mechanism for ensuring men’s domination over women. For Connell, the ascendancy of hegemonic masculinity over other subordinate and marginalized masculinities is equally important for gender hegemony. Complicit masculinities are “masculinities constructed in ways that realize the patriarchal dividend, without the tensions or risks of being the frontline troops of patriarchy...” (Connell, 1995: 79). To the extent that hegemonic masculinity ensures male dominance, all men benefit on some level even though most men don’t have to be “on the front lines” or embody hegemonic masculinity. This is an often overlooked but important contribution that Connell offers; some masculine practices and characteristics are hegemonic, but some are not. As Connell (1995) suggests, focusing on what most men do will not necessarily reveal how hegemonic masculinity is implicated in gender hegemony. We need theory that will allow us to distinguish masculine characteristics and men’s practices that perpetuate male dominance from those that do not, a point I return to later.

Perhaps most important for Connell is the subordination of gay men by heterosexual men. Gay men embody what Connell refers to as subordinate masculinities. When held up against hegemonic masculinity as the ideal, subordinate masculinities serve as the inferior “Other”. Connell writes,

Hegemony relates to cultural dominance in the society as a whole. Within that overall framework there are specific gender relations of dominance and subordination between groups of men. The most important case in contemporary European/American society is the dominance of heterosexual men and the subordination of homosexual men.... Oppression positions homosexual masculinities at the bottom of a gender hierarchy among men. (p. 78)

Connell suggests that subordinate masculinities are often conflated with femininity. However, as others have suggested, we are left with no conceptual apparatus with which to distinguish femininity from subordinate masculinities unless we reduce femininity to the practices of women and masculinity to those of men (Halberstam, 1998; Lorber, 1998; Martin, 1998), another point I return to later.

In Connell’s theory, subordination is one mechanism for the ascendancy of hegemonic masculinity, but it is not the only one; there are also marginalized masculinities. While hegemony, subordination, and complicity are aspects of the gender order, Connell offers marginalization to characterize the relationships among men that result as class and race intersect with gender. Marginalized masculinities are those of subordinated classes or racial/ethnic groups. This relationship is one of authorization and marginalization because hegemonic masculinity is conflated with whiteness and middle-class status, and it is conferred authority in a way marginalized masculinities are not. However, in Connell’s original conceptualization, it is difficult to distinguish different masculinities from different groups of men. We are once again left to the practices of particular groups of men rather than a configuration of practice some, but not all men embody. Further, because gender hegemony is so inextricably tied to heterosexual, middle-class, and white status, according to Connell, male dominance falls through the conceptual cracks when considering groups whose members are not white and middle class.

Applying the theoretical framework to femininity

Pyke and Johnson’s (2003) work on Korean and Vietnamese second-generation women is one effort to apply Connell’s framework to femininities to develop a definition for hegemonic femininity and subordinated femininities. Pyke and Johnson identify a relation of subordination and domination between white femininity and Asian femininity in both controlling images of Asian women and in the talk of Asian-American women. They suggest that, given this relation, it is useful to consider the characteristics attributed to white women as hegemonic femininity and those attributed to Asian women as a subordinated femininity.

... white women are constructed as monolithically self-confident, independent, assertive, and successful-characteristics of white hegemonic femininity. That these are the same ruling traits associated with hegemonic masculinity, albeit in a less exaggerated, feminine form, underscores the imitative structure of hegemonic femininity. That is, the supremacy of white femininity over Asian femininity mimics hegemonic masculinity. We are not arguing that hegemonic femininity and masculinity are equivalent structures. They are not. Whereas hegemonic masculinity is a superstructure of domination, hegemonic femininity is confined to power relations among women. However, the two structures are interrelated with hegemonic femininity constructed to serve hegemonic masculinity, from which it is granted legitimacy. (pp. 50–51)

While this study increases our understanding of how racialized gender performance is implicated in inequalities among women and demonstrates that there are ascendant femininities, I want to suggest that juxtaposing white and Asian femininities in terms of gender hegemony and subordination poses two significant problems. First, there is no way to identify the relationships between femininities operating within race and ethnicity. That is, if white femininity is hegemonic femininity and non-white femininities are subordinate, we have little conceptual room to identify multiple femininities within race and class groups, and more importantly, which raced and classed femininities serve the interests of male dominance and which do not.

Second, though Pyke and Johnson suggest that hegemonic femininity mimics hegemonic masculinity, there is no conceptual apparatus with which to identify how men benefit from the relationship between white femininity and Asian-American femininity. Although it is not difficult to understand how the construction of white women as “self-confident, independent, assertive, and successful” serve white men’s and women’s race and class interests, it is difficult to understand how these constructions serve men’s interests as men. I suggest that those valued characteristics (self-confident, independent, assertive, and successful) are not culturally inscribed as gender traits, but instead racial/ethnic traits and that the inequality between white women and Asian women is based on racial hegemony, not gender hegemony. Surely the relationship between white women and Asian women is an outcome of the intersection of gender and race as Pyke and Johnson demonstrate. However, their work does not offer better understanding of how the relationship between femininities is implicated in gender hegemony. We are still in need of a theoretical framework for multiple femininities that can account for the cultural hierarchy established between white women and Asian women as identified by Pyke and Johnson and can explain the role of femininities and masculinities in ensuring relations of domination that benefit men as a group.

Finally, I want to suggest that it is problematic to use Connell’s notion of subordination to distinguish between different femininities. As Connell (1987) suggests, within the context of a male dominant gender order, femininity is, by definition, a position of subordination in relation to masculinity. If we claim that racial and ethnic minority femininities are subordinate to white femininity, we obscure the subordination of white women in the gender order and we deny that racialized femininities might actually empower racial and ethnic minority women in a way that white femininities do not for white women (Hill Collins, 1990). For this reason, configurations of femininity that are not deemed normal, ideal, or desirable cannot be thought of as subordinate to an ideal femininity.

In summary, the goal here is to reclaim and re-work Connell’s theory of masculinities and gender hegemony in a way that 1) offers a conceptualization that does not reduce masculinities to the behavior of boys and men or femininity to the behavior of girls and women, 2) provides a definition of femininity that situates it, along with masculinity, in gender hegemony and allows for multiple configurations, and 3) is empirically useful for identifying how masculinity and femininity ensure men’s dominance over women as a group locally, regionally, and globally (Connell & Messerschmidt, 2005), and how they legitimate and perpetuate race, class, ethnic, and sexual inequality.

An alternative model

To use Connell’s insights to begin building an alternative theoretical model for masculinity, femininity, and their role in gender hegemony, femininity must be placed back into the theory without losing Connell’s invaluable conceptualization of hegemonic masculinity. Judith Butler’s theoretical framework for the heterosexual matrix provides a good place to start. According to Butler (1990), gender is the socially constructed binary that defines “men” and “women” as two distinct classes of people. The discursive construction of gender assumes that there are certain bodies, behaviors, personality traits, and desires that neatly match up to one or the other category. Both Connell and Butler agree that the categories “man” and “woman” include a whole repository of symbolic meanings. These symbolic meanings for gender difference establish the origins (e.g., biology, divine will, socialization), significance (e.g., defines subjectivity, is the foundation of society), and quality characteristics of each category (e.g., men are physically strong and authoritative/women are physically vulnerable and compliant).

My focus in this article is on the quality content of the categories “woman” and “man”. Embedded within the system of symbolic meanings that articulate and define gender positions and their relationship to each other are qualities members of each gender category should and are assumed to possess. I argue, in contrast to Connell and Messerschmidt (2005), it is in the idealized quality content of the categories “man” and “woman” that we find the hegemonic significance of masculinity and femininity.

Connell and others who theorize and research masculinities acknowledge that masculinity is always defined through its difference from femininity, however Butler places the relationship of difference more centrally in her conceptualization of gender. For Butler, heterosexual desire, as a defining feature for both women and men, is what binds the masculine and feminine in a binary, hierarchical relationship. In contemporary Western societies, heterosexual desire is defined as an erotic attachment to difference, and as such, it does the hegemonic work of fusing masculinity and femininity together as complementary opposites. Thus, it is assumed that men have a natural attraction to women because of their differences and women have a natural attraction to men. While there is far more to the content of masculinity and femininity than erotic desire, the construction of hetero-desire as the ontological essence of gender difference establishes the meaning of the relationship between masculinity and femininity. Regardless of one’s sex category, the possession of erotic desire for the feminine object is constructed as masculine and being the object of masculine desire is feminine. Heterosexual desire is defined as the basis of masculinity as others have argued (Anderson, 2002; Connell, 1987; Dowsett, 1993; Fejes, 2000; Garlick, 2003; Kimmel et al., 2005), but it is also, and importantly, the basis of the difference between and complementarity of femininity and masculinity.

Difference and complementarity alone, however, do not constitute hegemony. Hegemonic features of culture are those that serve the interests and ascendancy of ruling classes, legitimate their ascendancy and dominance, and encourage all to consent to and go along with social relations of ruling. Although heterosexual desire marks both difference and complementarity in Western societies, the cultural construction of embodied sexual relations, along with other features of masculinity and femininity, defines a naturalized masculine sexuality as physically dominant in relation to femininity. For example, despite women embracing and expressing sexual agency at different historical times and in different cultural settings, contemporary, Western constructions of heterosexual sex still reduce it to penetrating and being penetrated and that relation is consistently constructed as one of intrusion, “taking”, dominating (Segal, 1994).Footnote 1 Compulsory heterosexuality and hegemonic constructions of sexuality as natural or grounded in biology establish the “naturalness” of the complementary and hierarchical relationship between masculinity and femininity. Placed together in relationship to each other, these features of masculinity and femininity provide the hegemonic scaffolding for relationships between men and women as “naturally” and inevitably a relationship of dominance and submission.Footnote 2

We can take this focus of relationality and identify other characteristics that define the relationship between women and men as complementary and hierarchical. As identified in the vast empirical literature on masculinities, hegemonic masculinity can include physical strength, the ability to use interpersonal violence in the face of conflict, and authority. These characteristics guarantee men’s legitimate dominance over women only when they are symbolically paired with a complementary and inferior quality attached to femininity. To complement these characteristics in a way that subordinates femininity to masculinity, femininity includes physical vulnerability, an inability to use violence effectively, and compliance. Even if few women and men actually embody these characteristics in relation to each other, the symbolic relationship established through these hierarchical complementarities provides a rationale for social practice more generally. Thus, the significance of masculinity and femininity in gender hegemony is that they establish symbolic meanings for the relationship between women and men that provide the legitimating rationale for social relations ensuring the ascendancy and dominance of men.

In the on-going process of recurring patterns of social practice, the quality content of masculinity and femininity becomes not just the gender identities or gender displays of individuals, but also, and perhaps more importantly, a collective iteration in the form of culture, social structure, and social organization. The idealized features of masculinity and femininity as complementary and hierarchical provide a rationale for social relations at all levels of social organization from the self, to interaction, to institutional structures, to global relations of domination. As individuals, groups, and societies use masculinity and femininity as the rationale for what to do and how to do it, and collectively do so on a recurring basis in different institutional settings, not just gender difference, but also the implicit relationship between genders become a taken-for-granted feature of interpersonal relationships, culture, and social structure. That is, gender difference is institutionalized (Lorber, 2000; Martin, 2004) but, importantly, so is gender relationality.

For example, in her research in Vila Sao Joao, Brazil, Claudia Fonesca finds that gossip about wives’ infidelity is common despite strong norms against it (Fonesca, 2003). In the village, men are expected to provide economic support to their wives who, in return, promise sexual exclusivity. People in the village talk about cheating wives who violate these norms, but interestingly, it is the husbands of cheating wives who are the brunt of jokes. This loss of status for men “with horns” (the expression used by villagers for a cuckold) allows women to use the threat of infidelity as a mechanism of interpersonal power and, Fonesca argues, causes men to be protective and control women’s movement and employment. While Fonesca suggests that men’s control of women is an outcome of cheating wives, and this might be true in specific relationships, the prevalence of the gossip and joking, without an equal frequency of actual cheating or women being sanctioned, suggests that the gossip itself might offer a rationale for men controlling their wives. A focus on the relationship between masculinity and femininity as constructed through discourse rather than on the specific behaviors of husbands and wives shows how the symbolic construction of the man “with horns” serves a hegemonic function. The gossip and jokes are not simply stories about the behavior of individual men and women, but also about what is a good and desirable marital relationship between women and men. By constructing the cuckold as ridiculous, ineffective, and weak, the stories and jokes themselves expose and marginalize behavior by women and men that, if embodied, would threaten the hegemonic marital relationship between husbands and wives. Far from being a static set of characteristics embodied by even some women and men, the “man with horns” and the “wily woman”, as constructed through gossip and joking, legitimate and naturalize the economic and sexual exchange between wives and husbands and, thus, provide a legitimating rationale for men’s control of their wives, men’s sexual infidelity, and women’s economic dependence on men. The gossip and joking serve this hegemonic function but also offer space for women to use these meanings to negotiate interpersonal power dynamics in their relationships with their husbands and with other women.

It is through social practice that the hierarchical relationship between masculinity and femininity organizes material relations of social life. Practice, then, is not masculinity and femininity as Connell suggests; social practice, in all its forms, from embodied interaction to child raising, sexual activity, developing and executing policy, passing legislation, producing television programming, and invading countries, is the mechanism by which masculinities and femininities, as part of a vast network of gender meanings, come to organize social life. Masculinities and femininities provide a legitimating rationale not just for embodiment and behavior by individuals but also for how to coordinate, evaluate, and regulate social practices, and therein lies their hegemonic significance. Here we distinguish between what we are defining as contextually and culturally specific sets of meanings for what women and men are and should be (masculinity and femininity) and the mechanism (social practice) by which those meanings come to shape, influence, and transform social structure. In her critique of Connell’s book, Masculinities, Patricia Yancey Martin (1998) writes,

I can accept that man and woman are places in a system of gender relations and that masculinity is practice. But I have trouble understanding how masculinity is a place or an effect. When a man dresses “like a woman”, is he in a masculine or feminine place? How can we know? I think we have to know the substance of societal gender norms and/or ideologies to which people orient practice to ascertain whether it is (a form of) masculinity. Are we not otherwise forced to reduce masculinity to man and femininity to woman? (p. 473)

In the conceptualization offered here, masculinity and femininity are not a place, a practice, or a resulting structure. The “place” Connell refers to is the social locations “woman” and “man”, while the embodiment of masculine or feminine characteristics by individuals is gendered embodiment or display. Embodying and producing the relationship between masculinity and femininity in social interaction is “doing gender” (West & Zimmerman, 1987), and the extent to which a hierarchical and complementary relationship between masculinity and femininity is institutionalized is gender structure. This offers a conceptual and empirical path out of conflating the practices of men and women with masculinity and femininity and allows for those occupying the social location “woman” to engage in practices or embody characteristics that are defined as masculine and for “men” to embody features of femininity. Masculinity and femininity and their constructed relationship to each other are an available rationale for practice and a referent with which to interpret and judge, not just the gender displays and practices of individuals, but all social relations, policy, rules, and institutional practice and structure.

Connell and Messerschmidt (2005) argue that, although there is a symbolic component to masculinity, masculinity should not be reduced to a cultural norm. They write,

Discursive perspectives emphasize the symbolic dimension, whereas the concept of hegemonic masculinity was formulated within a multidimensional understanding of gender. Although any specification of hegemonic masculinity typically involves the formulation of cultural ideals, it should not be regarded only as a cultural norm. Gender relations also are constituted through non-discursive practices, including wage labor, violence, sexuality, domestic labor, and childcare, as well as through unreflective routinized actions. (p. 842)

By suggesting that a theoretical reduction of masculinity to the realm of the symbolic ignores the ways in which gender relations are constituted through non-discursive practices, Connell and Messerschmidt are conflating gender relations with masculinity. In the model presented here, masculinity and femininity are relegated to the realm of the symbolic, however, they are conceptualized as just one aspect of gender relations. If we do not collapse all gender relations into masculinity and femininity, there is no reason that conceptualizing masculinity and femininity as an available rationale for individual and social practice negates or ignores the non-discursive practices by which men as a group dominate women as a group. Masculinity and femininity, as a web of symbolic meanings, provide a rationale, or as Garlick (2003) suggests, a technology available for organizing social practice that, over time as recurring patterns of practice, become, produce, and legitimate male dominant interpersonal power relations, a gendered division of labor, an unequal distribution of resources and authority, global imperialism, and so on. Thus, masculinity and femininity are hegemonic precisely in the ideological work they do to legitimate and organize what men actually do to dominate women individually or as a group.

If, when, and how femininity and masculinity provide a rationale for practice are empirical questions and could be explored at all levels of social organization. However, instead of focusing either on masculinity or femininity or only on the practices of women and men, this model encourages an additional exploration of how a naturalized, complementary, and hierarchical relationship between masculinity and femininity is produced and deployed as a rationale or legitimating discourse for social practice, policy, or institutional structure that result in or ensure inequality and domination, not just along the lines of gender, but along the lines of race, class, sexuality, age, region, or nation.

Finally, Connell and Messerschmidt warn that reducing masculinity and femininity to the symbolic can lead to the same problems identified with theorizing femininity and masculinity as roles. Critics of gender role theory identify it as too static, as conflating behavior and norms, as ignoring variation among women and among men, and as not accounting for power (Connell & Messerschmidt, 2005). Although in this model masculinity and femininity are conceptualized as symbolic constructions, I am not suggesting, nor is it necessary to conclude that masculinity and femininity are static roles or a fixed set of behaviors that women and men adopt. Instead, the characteristics and practices defined as womanly and manly are constituted through the proliferation of a network of cross-cutting, sometimes contradicting discourses. The production, proliferation, and contestation of the quality content of masculinity and femininity are on-going, dynamic, social processes and include everyday practices such as gossip, story-telling, informal and formal sanctions, and wider-reaching practices and processes like policy development and implementation, legislation, social movements, media production and proliferation, global economic relations, and so on. As Foucault (1978) suggests, power operates through “a multiplicity of discursive elements that can come into play in various strategies” (p. 100). Masculinity and femininity are conceptualized here as produced, contested, and transformed through discursive processes, and therefore embedded within and productive of power relations. In this model, then, power dynamics are central, not only in the conceptual focus on the hierarchical relationship between masculinity and femininity rather than the specific characteristics idealized, but also in terms of the dynamics of the production, proliferation, and contestation of discourses articulating what men and women and their relationship to each other is and should be.

Gender hegemony and multiple masculinities and femininities

Now that we have established that it is the relationship articulated through the quality content of femininity and masculinity that is the central feature of gender hegemony, we can begin to think about multiple configurations of masculinity and femininity and their implications for gender hegemony. As Connell suggests, any conceptualization of hegemonic masculinity must be first defined in its difference from femininity. I would add, however, that any conceptualization of hegemonic masculinity must also be defined by the way in which it articulates a complementary and hierarchical relationship to femininity. Connell’s definition of hegemonic masculinity, with a few key changes (in italics) and the explicit addition of femininity, serves us quite well. Hegemonic masculinity is the qualities defined as manly that establish and legitimate a hierarchical and complementary relationship to femininity and that, by doing so, guarantee the dominant position of men and the subordination of women. Given the centrality of the relationship between masculinity and femininity in the new definition, we now have conceptual space for hegemonic femininity. Hegemonic femininity consists of the characteristics defined as womanly that establish and legitimate a hierarchical and complementary relationship to hegemonic masculinity and that, by doing so, guarantee the dominant position of men and the subordination of women.

Although the relationship between hegemonic masculinity and hegemonic femininity is one of ascendancy for the masculine and for men, there is, I argue, an ascendancy of hegemonic femininity over other femininities to serve the interests of the gender order and male domination. Connell (1987) writes, “Femininity organized as an adaptation to men’s power, and emphasizing compliance, nurturance, and empathy as womanly virtues, is not in much of a state to establish hegemony over other kinds of femininity” (188). This statement is only possible if femininity and masculinity are conceptualized in isolation from each other. A different picture emerges by placing the relationship between masculinity and femininity at the center of gender hegemony. If gender hegemony is produced through the relationship between femininity and masculinity, our efforts to identify multiple and hierarchical configurations of masculinities and femininities must also focus on this relationship. What emerges are gender qualities that cluster into configurations that are constructed, not so much in their difference from and inferiority to hegemonic masculinity as Connell suggests, but instead against the idealized relationship between masculinity and femininity, as exemplified by the cuckold and the wily woman in Fonesca’s work and identified above.

If hegemonic gender relations depend on the symbolic construction of desire for the feminine object, physical strength, and authority as the characteristics that differentiate men from women and define and legitimate their superiority and social dominance over women, then these characteristics must remain unavailable to women. To guarantee men’s exclusive access to these characteristics, other configurations of feminine characteristics must be defined as deviant and stigmatized. This is needed to define the ideals for femininity, but also to ensure swift and severe social sanction for women who take on or enact hegemonic masculinity. Practices and characteristics that are stigmatized and sanctioned if embodied by women include having sexual desire for other women, being promiscuous, “frigid”, or sexually inaccessible, and being aggressive. These are characteristics that, when embodied by women, constitute a refusal to complement hegemonic masculinity in a relation of subordination and therefore are threatening to male dominance. For this reason, they must be contained. They contradict or deviate from practices defined as feminine, threaten men’s exclusive possession of hegemonic masculine characteristics, and most importantly, constitute a refusal to embody the relationship between masculinity and femininity demanded by gender hegemony.

It is precisely because women often embody and practice these features of hegemonic masculinity, and because this challenges the hegemonic relationship between masculinity and femininity, that these characteristics, when embodied by women, are stigmatized and sanctioned. Hegemonic femininity is ascendant in relation to, what I suggest we call pariah femininities. I propose calling this set of characteristics pariah femininities instead of subordinate femininities because they are deemed, not so much inferior, as contaminating to the relationship between masculinity and femininity. The possession of any one of these characteristics is assumed to contaminate the individual, so by having the one characteristic, an individual becomes a kind of person – a lesbian, a “slut”, a shrew or “cock-teaser”, a bitch. Not only do the characteristics become master statuses for women who exhibit or enact them, these women are considered socially undesirable and contaminating to social life more generally. Examples of this can be found in Messerschmidt’s empirical work on adolescent violence and gender (Messerschmidt, 2003). In one of his case studies, Tina, a working-class white girl, successfully embodied femininity and was one of the popular preppy girls at school. After Tina began wearing tighter and more revealing clothing, one of her friends said she “dressed like a whore” (p. 89). When Tina physically assaulted her friend, she was expelled from the preppy friendship group. She was, however, praised by and recruited into the “badass group”. The “badass” girls were those who embodied a sexualized, heterosexual femininity and were also physically tough and aggressive. Not coincidently, the badass girls were lower in status than the preppies. Their status in the school hierarchy reflects not just the idealized relationship between masculinity and femininity (embodied by the preppy girls), but also between hegemonic femininity and pariah femininities (embodied by the badass group). The symbolic construction of girls’ sexual agency and ability and willingness to use physical violence as undesirable and deserving of sanction and social expulsion turns their potential challenge to male dominance into something contained and less threatening.

Although pariah femininities are actually the quality content of hegemonic masculinity enacted by women-desire for the feminine object (lesbian), authority (bitch), being physically violent (“badass” girl), taking charge and not being compliant (bitch, but also “cock-teaser” and slut), they are necessarily and compulsively constructed as feminine when enacted by women; they are not masculine.Footnote 3 When a woman is authoritative, she is not masculine; she is a bitch – both feminine and undesirable. The slut is decidedly feminine. Through popular media and heterosexual pornography, the lesbian is consistently constructed as the feminine object of masculine desire. Hegemonic masculinity must become something completely different when enacted by women for the characteristics to maintain their place squarely in masculinity and their only legitimate enactment solely in the hands of men. The symbolic construction of pariah femininities, then, is a central feature of gender hegemony and, as such, central to the very real, material sanctions exacted on women who embody them. This suggests that, in any empirical exploration of gender hegemony, one way to identify contextually specific features of hegemonic masculinity would be to identify locally defined pariah femininities – characteristics or practices that, when embodied by women in the setting, are simultaneously stigmatized and feminized.

Just as hegemonic masculinity must remain exclusively in the hands of men, hegemonic femininity must cohere with the gender category “woman”. When a man exhibits hegemonic, feminine characteristics – as in having desire to be the object of masculine desire, being physically weak, or being compliant – he becomes the target of stigma and social sanction, much like women who embody features of hegemonic masculinity. And, like pariah femininities, possession of one characteristic by a man is culturally defined as contaminating. Men having and acting on erotic desire for each other disrupts the assumed naturalized, complementary desire between men and women, and weak, ineffectual, and compliant men dislodge physical strength and authority from the social position “man”. And so we have the “fag,” the “pussy”, and the “wimp” – kinds of men who enact hegemonic femininity. And like women who embody hegemonic masculinity, men who exhibit hegemonic femininity are viewed as contaminating to social relations more generally.

We cannot, however, call these pariah masculinities. Men’s homosexual desire and being weak and ineffectual are not symbolically constructed as problematic masculine characteristics; they are constructed as decidedly feminine. Because femininity is always and already inferior and undesirable when compared to masculinity, it can sustain features of stigmatization and contamination. In contrast, masculinity must always remain superior; it must never be conflated with something undesirable. It is cultural insurance for male dominance that anybody who enacts or embodies hegemonic characteristics that do not align with their gender category is stigmatized as problematic and feminine. Masculinity maintains its position of superiority in relation to femininity and men maintain legitimate possession of those superior characteristics regardless of who is embodying femininity or masculinity. This means that there are no masculine characteristics that are stigmatized as contaminating or as subordinate. There are neither pariah masculinities nor subordinate masculinities. Thus, what were identified by Connell as subordinate masculinities, are, in this model, simply hegemonic femininity embodied or enacted by men. Halberstam (1998) and Messerschmidt (2003) identify specific forms of female masculinity by looking at how women embody masculinity. Building on Halberstam and Messerschmidt, I propose that there are specific forms of male femininity. However, they are not simply femininity embodied by men, as Halberstam’s and Messerschmidt’s work would suggest. I argue that we limit male femininities to the characteristics and practices that are culturally ascribed to women, do the cultural work of situating the feminine in a complementary, hierarchical relationship with the masculine, and are embodied by men. Because male femininities threaten the hegemonic relationship between masculinity and femininity, they are both feminizing and stigmatizing to the men who embody them. Just as we can identify hegemonic masculinity by looking at the practices that are most stigmatized and feminized when embodied by women, we can identify contextually specific hegemonic femininities by identifying locally defined male femininities – the specific practices and characteristics that are stigmatizing and feminizing when embodied by men. The benefits of not reducing homosexuality to a subordinate masculinity become clear when we look at the symbolic construction of masculinity and femininity within homosexual identity. Empirical research consistently demonstrates that gay men claim their status as “real men” by defining their embodiment of a gay identity in relation to in inferior feminine form – as a “straight gay” in relation to effeminate gay men (Connell, 1992; Smith, Kippax, & Chapple, 1998), as “bear” in relation to “twink” (Hennen, 2005), or as “top” in relation to “bottom” (Kippax & Smith, 2001; Lambevski, 1999; Underwood, 2003). Being a “straight gay”, a bear, and a top are masculine to the extent that they embody characteristics symbolically defined as masculine. Being effeminate, a twink, a bottom are male femininities in that they are symbolically constructed as men embodying femininity.

As defined here, there are features of femininity and masculinity that are not central to forming and legitimating a hierarchical relationship between men and women, and thus are neither particularly feminizing nor stigmatizing. For example, Matthew Gutmann, in his study of changing gender meanings and practices in Mexico City, found that working-class men often participate in childcare and do not lose status by doing so (Gutmann, 1996). As Gutmann explains, for lower-class families, economic changes have necessitated men’s participation in childcare, which, in turn, has led to ideological changes in the meanings of fatherhood and its centrality for defining manliness. In contrast, men in the higher classes still draw a distinct line between fathering and mothering where fathers provide economic support and mothers do all of the physical care and nurturing. At the same time, for all classes, caring for children and motherhood are still central features of femininity. Because there is little stigma attached to men’s fathering practices, nurturing children is not, in the context of working-class Mexico City, a male femininity. It is, however, a male femininity among the higher classes, illustrated by the consistency with which upper class men attached a feminizing stigma to men who “carry their babies”. Fathering, then, is not a feature of hegemonic masculinity in working-class Mexico City because the meanings for fathering as a component of masculinity do not establish a naturalized, hierarchical relationship between women and men. It is, however, still a feature of hegemonic masculinity among men in the upper classes.

Limiting hegemonic femininity and masculinity to only those characteristics and practices that articulate a complementary and hierarchical relationship between women and men offers conceptual and empirical space to identify idealized gender characteristics that do not perpetuate male dominance and therefore can be viewed as positive and valuable. For example, Lena Eskilsson’s (2003) research on a logging culture in the northern pine forest region of late nineteenth century Sweden suggests that, while there were characteristics considered masculine and feminine, valued masculine characteristics were not defined in hierarchical relation or in opposition to femininity. Masculinity consisted of a strong work ethic, skill, and maturity and was not juxtaposed to inferior and complementary characteristics valued in women. Not coincidentally from the perspective of the theoretical model provided here, Eskilsson found little evidence of male dominance in this culture despite a gender division of labor, gender segregation, and differences in the qualities valued as masculine and feminine.

Further, as I have argued elsewhere, a focus on the relationship between masculinity and femininity opens up the possibility for empirically identifying ways in which hegemonic femininity and masculinity, pariah femininities, and male femininities can be intentionally replaced with, what I call alternative femininities and masculinities (Schippers, 2002). In that study, I identified how members of a specific rock subculture reject hegemonic masculinity and femininity as defined within the specific cultural discourse of mainstream rock culture. Instead of valuing the characteristics and practices that symbolically situates the male rock musician in relation to and dominant over the female groupie, members publicly and consistently repudiate the practices of both musicians and groupies in mainstream rock through their music, their practices, and their talk. Through the proliferation of an alternative set of meanings for being a musician and for erotic attraction to musicians, men and women in the subculture establish alternative femininities and masculinities that provide a rationale for individual and social practice at the rock show. In this case, alternative femininities and masculinities are discursively valued traits and practices in women and men that do not articulate a complementary relation of dominance and subordination between women and men.

As the examples above suggest, what actual characteristics and practices are idealized as masculine and feminine is ultimately an empirical question and will vary by context, group, and society. The importance of context cannot be overstated here. Because of cultural, economic, political, and social variation across groups and societies, what specific features of masculinity and femininity ensure men’s dominance over women as a group will vary depending on context (Dellinger, 2004). While there might be overarching features of hegemonic masculinity and femininity that cross-cut contexts, groups, and perhaps societies, what those are is an empirical question. In fact, when looking at empirical explorations of masculinity in a Western context, we find that even heterosexual desire can sometimes take a back seat to other, more salient masculine characteristics within a particular context. For instance, Eric Anderson (2002) found that openly gay athletes were not stigmatized for their homosexuality as long as their athletic abilities provided, what he calls, masculinity insurance (p. 865), meaning the men were superior athletes and embodying this feature of masculinity overshadowed their gay identities. In the context of sport then, perhaps superior athletic ability is an equally central feature of hegemonic masculinity along with heterosexual desire (Messner, 1992). Interestingly, this masculinity insurance was effective in preventing the contaminating effects of homosexual desire only if the gay athletes did not publicly express any homosexual desire around their teammates. Further, when sexuality did emerge as salient, the gay athletes reported having to sometimes express desire for women sometimes, to prove to their teammates that they were just “normal guys”.

Finally, I suggest that we move away from defining variation in gendered practice across different races, classes, and settings as different masculinities and femininities, and instead understand this variation as hegemonic masculinity and femininity refracted through race and class difference. There is no reason to suggest that within the logic of gender difference, masculine and feminine qualities are not available to and required of women and men of color and to working and poor, white women and men. What appear to be different configurations of femininity and masculinity is instead group and cultural variation in the embodiment of hegemonic femininity and masculinity. Though the culturally specific forms of masculinity and femininity might vary, in their relationship to culturally specified characteristics of hegemonic masculinity within the setting or group, they reify hierarchical gender difference and legitimate male dominance. This opens space for empirically identifying hegemonic femininity in white, middle-class culture and non-white, non-middle-class culture. For instance, to build on the work of Pyke and Johnson (2003), we might ask what are the characteristics and practices valued in white, middle-class women that hetero-sexualize their relationship to white, middle-class men? What idealized features of femininity construct them as weak or ineffectual in relation to white, middle-class men? And what are the characteristics that do the same in Asian-American cultures and communities? We could also explore culturally and community specific forms of pariah femininities and male femininities.

But what of the relationship between white and Asian femininities identified by Pyke and Johnson? If gender identity is constructed through Western discourse as the ontological substance of human subjectivity, as Butler (1990) suggests, then excluding subordinate race and class groups from being “real” women and men provides a legitimating rationale for their social, political, and economic subordination. Thus, as is so clearly demonstrated by Pyke and Johnson, minority racial/ethnic groups and working and poor classes are constructed as undeserving “others” or as problematic because of their gender practices. As Pyke and Johnson show, stereotypes of or the actual gender practices of subordinate race and class groups or nations are often supportive of gender hegemony, but only as they intersect with race and class hegemony. However, race and class differences in gender performance or social organization, not embodied and institutionalized gender differences between women and men, provide the rationale for placing upper- and middle-class, white men and women higher in social status than others and rendering the gender practices of others as illegitimate.

As gender meanings cross-cut other systems of inequality, they are folded into and support race and class hegemony. It is through racist and class-based ideology that differences among women and among men in gender practice along the lines of race and class are constructed as differences in value or as moral or social problems. For example, Julie Bettie’s research on class and ethnic relations among high school girls interrogates differences in the girls’ embodiment, practices, desires, and goals (Bettie, 2003). She finds that differences in the girls’ race-class performances of femininity reflected and perpetuated structural inequalities along the lines of ethnicity and class. Importantly, Bettie found that peers and adults read all of the girls through the lens of heterosexuality across ethnicity and class. However, the “preps” or middle-class white girls’ performance of heterosexuality was interpreted as the only legitimate embodiment of “good girl” and “good student”. In contrast, Las Chicas, working-class Mexican-American girls, were interpreted by teachers and administrators as hyper-sexual and focused more on hetero-romance than school because of their performance of race-class femininity. It is significant to point out that Bettie found that Las Chicas’ race-class performance of femininity, as articulated by the girls themselves, was less about hetero-romance and more about resisting their class and ethnic location in school hierarchies. Las Chicas and the preps embody their gender identities and are intelligible as Las Chicas and preps by performing the characteristics, preferences, and desires idealized within their specific structural and cultural location. Despite these race and class differences, all the girls embody the social location girl and, as Bettie identifies, heterosexuality. Both groups of girls, Las Chicas and the preps, experience their embodiment of femininity as legitimate and desirable in relation to girls and boys within their ethnicity and class location. The privilege experienced by the preps is an outcome of the symbolic construction by teachers, administrators, and the preps of Las Chicas performance of femininity as inferior and outside the definition of good girl and good student. As Bettie astutely points out, though the derision expressed by teachers, administrators, and the preps is a discourse of gender, the hierarchies themselves are about race and class difference, not gender. Although Bettie does not focus on the relationships between boys and girls, one can imagine that the performance of femininity, across class, reflects and reproduces the heterosexual matrix in race-class specific forms. Only by analyzing hegemonic femininity in both groups of girls can we begin to identify how femininity is implicated in male dominance across race and class. This allows us to conceptualize and empirically analyze hegemonic gender relations between women and men of subordinated class and race groups while also recognizing how the cultural construction of those practices serve dominant class and race interests. By excluding members of some groups from being “real” or “good” women and men, white supremacy and class privilege are legitimated at the same time that the idealized quality content of masculinity and femininity is reinforced in both socially dominant groups and socially subordinate groups. Gender hegemony benefits from race and class hegemony when the gender practices of subordinate race and class groups are defined as problematic or deviant in order to reify and legitimate the ideal quality content for femininity and masculinity. However, the function of these hierarchies is not to reify and legitimate gender difference and gender hierarchy, though they rely on and ultimately support the gender order. The function of the stigmatization and material sanction is to establish hierarchies of value on the basis of race and class difference and legitimate local, regional, and global relations of race and class inequality.

Conclusion: implications for research

With the new model, any empirical exploration of masculinity and femininity and their role in gender hegemony must focus on relationality. The following questions can be explored in localized settings and in broader structures. 1) What characteristics or practices are understood as manly in the setting? 2) What characteristics or practices are womanly? 3) Of those practices and characteristics, which situate femininity as complementary and inferior to masculinity? Answering these questions, especially the third question, will empirically identify hegemonic masculinity and femininity, and importantly, the features of masculinity and femininity that are not hegemonic-that is, those characteristics and practices valued in women or in men that do not naturalize and endorse a hierarchical bond between the masculine and feminine and men and women. Other questions can also lead to better understanding of hegemonic femininity and hegemonic masculinity in particular settings. 1) What characteristics or practices of women are defined as feminine, contaminating, or disruptive? That is, what are the pariah femininities circulating? 2) What characteristics or practices of men are defined as feminine, contaminating, or disruptive? What are the male femininities? Answering these questions, however, is just the beginning of understanding gender inequality as it operates in local settings. The consequences of embodying these ideals and putting them into social practice in terms of the distribution of power, resources, and value are the true measures of gender inequality. Masculinity and femininity are configurations of meaning and not practice, but it is only by identifying how putting these ideals into practice results in unequal power relations and distribution of resources that we can truly know if they constitute hegemonic femininity and hegemonic masculinity. This suggests that simply asking people what the ideal characteristics are for women and for men and then deciding how the characteristics line up as complementary and hierarchical will not be enough. We would have to see which features of femininity and masculinity are put into practice, deployed as a rationale for practice, and institutionalized to establish and naturalize hierarchical and complementary social relationships between women and men and those who do not fit either category.Footnote 4 Our focus would not be identifying and describing the behavior of women and men, but the power relations and distribution of resources among women, men, and others and how masculinity and femininity as networks of meaning legitimate and ensure that structure. By researching masculinities and femininities as conceptualized here, we can better understand the operation of gender hegemony in local, regional, and global relations of inequality, and identify local, regional, and global ways to challenge gender hegemony.