My critique of Habermas’s account singles out three elements: the claimed neutrality of the observer perspective, the lack of differentiation in the notion of authority-governed complementarity, and the limits placed on the inter­pretation of moral development by the construction of competition and coop­eration as binary oppositions. A consideration of these concepts in light of the experiences of girls and women reveals their inability to accommodate sexual difference. Indeed, in relying on a reading of moral development which has dismissed women’s moral experience, Habermas’s account obscures the differ­ent developmental issues faced by women and men in each transitional phase. Thus, he can neither recognize his elision of the observer perspective into that of the generalized other, the organized expectations of a social group, nor can he suggest a more differentiated account of authority governed complemen­tarity which makes clear the masculinity of the conventional generalized other. Indeed, once sexual difference is introduced into the stages of moral development, the need for the category of connection as a supplement to competition and cooperation becomes clear if discourse ethics is to be seen as a reflective continuation of action types emerging at the preconventional and conventional stages. In brief, by attending to the importance of relationships of mutual recognition we can better understand how the hypothetical attitude

required of participants in argumentation demands the radical questioning of gendered and hierarchical roles if each is to recognize the other as a subject deserving equal respect.

Beginning with the perspective of the neutral observer, we see that, as he moves from the description of the objectivating perspective acquired by the child through engagement with external nature to the discussion of the role of the objectivating standpoint in the completion of the system of speaker perspectives, Habermas “quietly alters” the meaning of the observer perspec­tive. What was previously rooted in the first-person perspective of a person perceiving the world of physical events is moved out of the body of the child and into the body of a third person. But sexual difference disrupts the “neutrality” of the observer perspective at two points. First, in representing a member of the child’s social group, the body of the third-person is a sexed and gendered body. Second, as the object of the onlooker’s gaze, the child’s body itself is seen as sexed; the onlooker sees a “her” or a “him.” This is hardly surprising since learning social roles at the conventional level essentially involves learning sex roles. Indeed, although he ignores the differing reper­cussions for boys and girls, in an earlier essay Habermas at least acknowl­edges that one acquires a gendered role identity at this stage.37 In any case, the sex of the observer poses a question: is the child similarly or differently sexed than this third person?38

To be sure, sexual difference is not the only difference that disrupts the neutrality of the observer perspective. Racial difference in a racist society also comes into play in particularly powerful ways. As Patricia Williams writes, “blacks in a white society are conditioned from infancy to see in themselves only what others, who despise them, see.”39 While I am not focusing on the racial dimension of the observer, my hope is that the thematization of the exclusion of sexual difference will expose the false neutrality of the observer perspective to such an extent that the necessity of acknowledging the situat­edness of this perspective is revealed. The identification of the notion of sexual difference, in other words, works as a critical tool, alerting us to possibility of other sorts of exclusions enacted in the name of neutrality.

My claim is that under familial, social, and cultural conditions where men dominate, calling the third person “neutral” occludes its masculinity. To be sure, this perspective is not inherently masculine. As Habermas generally uses the term, the third-person or observer perspective refers to a particular sort of cognitive operation involving the ability to distance one’s self from a situation, generalize from the particular issues involved, and judge or assess them.40 However, in his account of the development of the stages of perspective taking, Habermas replaces the third-person observer perspective with the perspective of the generalized other, the organized responses and expectations of the social group. In fact, his account of the completion of world perspectives relies on this switch, since he is trying to show how the child generalizes from particu­lar interactions with parents to see them as part of a larger normative pattern. The perspective of the generalized other, then, provides a place for the third – person to stand, a particular set of shared norms and expectations which make assessment possible by furnishing the framework and standards for judgment. At the conventional level, the observer becomes engendered through its equa­tion with the generalized other—and thus represents a masculine perspective. The child is either like this third male person or different from him. Clearly, this affects the development of children growing up female differently from the way it affects children growing up male.

To be sure, the child’s acquisition of the “І-you” perspective at the precon – ventional level is also significantly affected by gender, namely, in the differing experiences of male and female children in their interaction with their fathers and mothers. Jessica Benjamin has looked at the denial of the mother’s subjec­tivity in traditional patriarchal family structures, analyzing its impact on the process of individuation for boys and girls. For both sexes, the father repre­sents the liberator, the way beyond dependency and identification with the mother into the outside world of agency and the recognition of one’s inde­pendence and individuality; corresponding to this idealization of the father is a devaluation of the mother.41 But, as the culmination of the pre-oedipal struggle to separate from the parents and grasp the sexual meaning of the difference between child and parent and mother and father, the oedipal phase presents the following limit: “identify only with the same-sex parent.”42 The boy repudiates the mother, denying her power and transferring it to the ideal­ized father and denying her subjectivity. The girl has to “overcome the primary identification with the mother and replace it with more generalized gender identifications that do not equate all femininity with the mother. If the girl tries to differentiate exclusively by repudiating the mother in favor of the father… she never really separates from the mother.”43 Well before they even reach the conventional stage children have a differentiated conception of themselves in relationship to each parent and conceive the father to have authority over the mother.

This leads to the second problem in Habermas’s account of moral develop­ment. The child is aware of two different types of authority-governed comple­mentarity: that of parent over child and that of male over female. The transition to the conventional level is thus a fundamentally different experience for girls than what it is for boys. Each child realizes that what might have a particular sort of interaction within her or his family is actually part of the generalized pattern of behavior of their social group. But, while the self-identity of boys is reinforced in a culture which values masculine roles, the girls’ sense of self is diminished. As Lyn Mikel Brown writes in her description of research on the moral development of girls from the second through the tenth grades:

This fifth grade girl has begun to see and question the consequences of a cultural norm that suggests, because of “like everything about names and stuff,” her mother is less than who she could be; who, according to the culture itself, “is not even alive.” These girls seem caught between their own relational experiences of female adults as powerful and trustworthy and a growing awareness that such experiences are not reflected in the wider society. Perhaps, then, things aren’t as they seem. What can it mean that those with whom they have felt the strongest bonds and with whom they have felt their strongest trust and value are made to “look like [they’re] not even there?”44

The work of Gilligan and Brown draws attention to the struggle between “self-silencing” and speaking in one’s own voice which occurs as girls take on feminine role definitions. The dilemmas Gilligan describes in In a Different Voice involve the difficulties women confront as they try to measure up to the societal expectations of the caring, maternal woman while nonetheless attempting to remain “responsive to themselves.” Similarly, Brown charts the preoccupation with being “nice” and “polite” which appears in girls’ narratives of their moral experience. While second grade girls understand being nice in terms of simple reciprocity-" if I’m nice to them, then they will be nice to me,” the fifth graders see being nice as a safe­guard “against oppression, mean treatment, or isolation, while being polite… should mitigate, even erase, the meanest and most hurtful behav­ior.”45 Moreover, by the time girls have reached the fifth grade, they have also begun to associate being nice with not mentioning things like unfairness or detachment: “they sometimes choose to silence themselves rather than to speak if by doing so they will risk exclusion or if speaking could be perceived as mean or thoughtless.”46 At seventh grade, girls see being nice as the key to social acceptability, learning that “to include their own real wishes and needs is to be called ‘selfish.’”47 Finally, facing the pressure of the feelings and thoughts they suppressed in order to be nice, girls in the tenth grade doubt their own voices, questioning the legitimacy of their ideas and express­ing self-doubt, confusion, and ambivalence.48

Because of the gendered inequality of our social and cultural institutions in which masculine roles are authorized and feminine roles are simultaneously devalued and glorified as instances of silent self-sacrifice, the internalization of the authority of the group which occurs at the conventional level reinforces the self-esteem of boys and diminishes that of girls. Moreover what they both learn is not that each role is in principle “interchangeable,” but that some roles are better, i. e., worth more, than others. In fact, once the gendered subtext of authority governed complementarity is brought into the picture, the more reflective form of reciprocity characterized by the formal symmetry of rights and duties at the conventional level appears fundamentally asymmetrical: the contents of social roles are distributed on the basis of men’s authority over women. To be sure, this does not mean that women themselves do not dominate; they, too, exercise authority over children, over other women, those disadvantaged by class and ethnicity, and, sometimes, over men of their own race and class. My point here, however, is that the hierarchy of sex roles instantiates a basic inequality at the conventional level that remains occluded when it is viewed primarily in terms of an increased reflexivity of reciprocal action perspectives.

A third problem with Habermas’s account of moral development becomes clear once we acknowledge the inability of the action categories of coopera­tion and competition accurately to conceptualize the loss of voice girls expe­rience at the transition to the conventional stage. In order to clarify this point, I want, first, to note the way in which Habermas’s taking over of Kohlberg’s framework leads to a distortion of the importance of relation­ships of mutual recognition in discourse ethics; second, to suggest an alter­native formulation which adds the concept of connection to the categories of competition and cooperation; and, third, to look at Gilligan’s and Brown’s findings in light of this more complex model.

First, Habermas wants to show how each stage is the reflective continua­tion of action perspectives acquired at the previous level. Yet, his action cate­gories at the preconventional level are limited to those of cooperation and competition. Accordingly, at the conventional level he is left with the oppo­sition between strategic and norm-guided interaction which he seeks to over­come at the post-conventional level. Argumentation is thus supposed to synthesize two orientations which Habermas usually seeks to separate—the success orientation constitutive of strategic action and the orientation toward understanding constitutive of communicative action.49 On the one hand, since argumentation is supposed to exclude attempts merely to influ­ence one’s opponent rather than convince her, it remains unclear how the split between strategic and norm-governed interaction is overcome at the post-conventional level. Presumably, Habermas means to suggest that norms are no longer taken as given but brought into question via reflection. Yet, with this move the meaning of “strategic” changes to refer to a purely cogni­tive competence; but such a cognitive understanding of “strategic” does not provide a way for overcoming the distinction between norm-governed and strategic action. On the other hand, even if it somehow were clear how this distinction were to be surmounted, it is difficult to see what Habermas wins with such a move. Indeed, he seems to lose more than he wins in that the notion of a “contest with arguments” filters out the element of intersubjec – tive recognition usually associated with ideal role-taking.

Second, because he begins with the binary opposition between coopera­tion and competition at the preconventional level, Habermas is forced to

bring in empathy “from behind.” Although he neglects empathy in his discussion of the stages of perspective taking, he sees sympathy for the fate of one’s neighbor as “a necessary emotional prerequisite for the cognitive operations expected of participants in discourse.”50 Further, he finds it “unlikely that one will be able to achieve this significant cognitive act [ideal role-taking] without that sensitive understanding which becomes real empa­thy and opens one’s eyes to the “difference,” that is the peculiarity and the inalienable otherness of a second person.”51 Finally, he claims that “without empathetic sensitivity by each person to everyone else, no solution deserving universal consent will result from the deliberation.”52 What is gained simply through a competition with arguments, then, may not be worth recognizing. Yet, for all its importance, Habermas fails to consider how the empathetic component of perspective taking emerges in the course of moral develop­ment. Empathy thus appears more as a some sort of motivational supple­ment than a constitutive component of intersubjective relationships.

To suggest how Habermas’s account might be filled out, I want to intro­duce the notion of an orientation toward connection. This orientation suggests itself because of the character of the relationship of mutual respon­siveness between the child and her primary reference persons. Rather than simply experiencing interactions with authority figures or friends as “rela­tions of exchange,” the child also experiences them as relations of attach­ment. In an article written with Grant Wiggins, Gilligan argues that “through the attachment or connection they create between them, child and parent come to know one another’s feelings and in this way discover how to comfort as well as how to hurt one another.”53 Benjamin also emphasizes the pleasure in being with the other which stems from the emotional attunement of child and parent: “Already at one year the infant can experience the wish to fulfill his own desire (say, to push the buttons on the stereo), and the wish to remain in accord with his parents’ will. Given such inevitable conflict, the desire to remain attuned can be converted into submission to the other’s will.”54 So not only is the child in a position of inequality and dependency, but she is also attached to her parents, wanting their approval and recogni­tion. In instances of conflict, the child’s desire for connection can be manip­ulated by parental authority. The parent may “make the child feel that the price of freedom is aloneness, or even, that freedom is not possible. Thus, if the child does not want to do without approval, she must give up her will.”55 The lack of symmetry in the authority governed relationship between the child and her parents, then, in this case in the child’s dependency on her connection with her parents, follows the same pattern in instances of conflict which Habermas describes. Yet, the issue involved is not merely avoiding punishment, but maintaining connection.

Things look somewhat different, however, when we turn to the desire for connection in symmetrical relationships. Gilligan stresses the way girls’ play manifests an emphasis on the relationship among those playing, keeping it intact as opposed to making the game itself the focal point of interaction.56 In these symmetrical relationships the possibility of deception in instances of conflict which Habermas discusses also appears, but this time as an element in the maintenance of connection. So here it is not only a way to achieve one’s particular goals, but also a device used to continue the relationship. Brown describes the negotiating strategies of second graders when they face conflicts with their friends, strategies which may include deception—as in the case of a little girl who puts on a disguise to avoid a friend whom she sees as want­ing to play with her too much—but which are centered around trying to come to an understanding in a way that avoids conflicts and includes everyone. While the fifth graders exhibit a more developed insight into the complexities of relationship and are able to understand conflicts from the perspectives of the various participants involved, their efforts to avoid conflict and maintain connection often lead them to associate “disagreement with fighting and loss of relationship” which pressures them “not to confront and work through differences.”57 Even more so, the attempts of the seventh graders to find inclu­sive solutions where there are no “winners or losers” reflect, on the one hand, a nuanced understanding of “public and private selves” and the variety of ways in which words and actions can hurt, but also a readiness to lie, remove themselves from the situation, or remain silent, on the other. By the tenth grade, although the girls again show an increase in social-cognitive complex­ity, they struggle to maintain connections with others to such an extent that they trust neither their perceptions nor their feelings and fear the potential costs of honesty to themselves and their relationships.58

What is striking in this account of the efforts girls go to in order to main­tain connection is the shift from the willingness to deceive another to the willingness to deceive or deny themselves. It is as if the honest recognition of their feelings and perceptions of their relationships would leave them detached and abandoned, disconnected and alone. But they are, of course, recognized to a certain extent. That is, they are recognized within the context of the structured expectations of women’s social roles. Brown’s research suggests that by the tenth grade, girls have internalized expectations of role recognition. Like Gilligan’s, her findings indicate that at the conventional level women learn that keeping a relationship intact requires conforming to the social expectation that women nourish relationships, even when it requires them to be suppress part of themselves.59 Habermas claims that at the conventional level complementary and symmetrical roles are synthesized at the cost of a polarization between strategic and normatively regulated interaction. While this may be the case when the categories of cooperation and competition frame the description, once connection is introduced additional costs come to light, namely the cost to one’s own sense of self when social norms can only be internalized through the denial of one’s own understandings and perceptions.

Deploying the research on girls’ development to reveal the gaps in Habermas’s discussion indicates that recognition and its place in relation­ships of mutual interconnection embody capacities essential to morality’s function in protecting vulnerable identities, capacities occluded when an orientation toward connection is left out of the description of moral devel­opment. Brown’s study shows that social cognitive complexity, although it may be a precondition for the perspective taking necessary at higher stages of moral development, is neither constitutive of nor sufficient for moral understanding. Along with Gilligan’s, her work draws attention to the importance of connection in the development of a moral sensibility. As the problems confronting the tenth-grade girls in Brown’s study and the women in Gilligan’s attest, there is a substantial difference between understanding a moral situation and being able to trust one’s understanding of the moral situ­ation—and this is not a difference between what I “ought” to do and what I “would” do in a situation (a confusion for which Habermas faults Gilligan),60 but a difference between having a perception and having the confidence to acknowledge that perception as one’s own. Indeed, the latter requires relationships of mutual recognition where one acquires a sense of confidence in one’s own voice.61 Although he associates this capacity more with the general category of ego development than with moral development per se, Habermas appeals to this sense of self-recognition and self-trust in an earlier essay. He argues that:

… ego-identity requires not only cognitive mastery of general levels of communication but also the ability to give one’s own needs their due in these communicative structures; as long as the ego is cut off from its inter­nal nature and disavows the dependency on needs that still await suitable interpretations, freedom, no matter how much it is guided by principles, remains in truth unfree in relation to the existing system of norms.62

As I have described it, the orientation toward connection can be inter­preted in terms of increasingly reflexive forms of recognition. Accordingly, at the preconventional level children recognize others particularly or natu­rally, that is in terms of the immediacy of face-to-face interactions.63 With the shift to the conventional level, recognition appears in terms of social roles. I have stressed that for women this stage is characterized by a split between norm conformity and self-trust. At the post-conventional level, then, the orientation toward connection requires overcoming role recogni­tion in favor of the mutual recognition of subjects deserving equal respect.

Mutual Recognition: Looking Beyond the Generalized Other Although it might be clear that post-conventional morality requires recipro­cal recognition, how it is possible to break out of the restrictions of role recog­nition which characterize the conventional stage, remains problematic, to say the least. I hope that by exploring the gaps between the observer and gener­alized other perspectives I might be able to suggest a plausible route. I begin by first looking at three ways in which we can find elements of indetermina­tion in position of the generalized other.

First, as I said above, the generalized other refers to the organized set of expectations of a social group. Mead offers the examples of the shared expec­tations institutionalized in the policeman or the state’s attorney.64 Using the example of the police, we could probably assume that the sort of expectations involved concern those of enforcing law and order or, perhaps, protecting property and securing the peace. Yet, from the perspective of poor urban blacks, say, these same expectations may take on a very particular meaning. We might understand law and order as a system which keeps us in our place, reinforcing our inequality. Protecting property might involve making sure that we don’t walk or drive into white neighborhoods. Securing the peace could evoke images of being beaten into submission—even after we have the strength to do nothing but submit. Thus different interpretations of the gener­alized other are available. The way the expectations organized in the gener­alized other are interpreted, the meanings they have for different members, is not fixed.

Second, the exact content of these expectations is not fixed. While we can assert with a high degree of confidence that the conventional expectations of sex roles involve the hierarchical domination by men and the concomitant subordination of women, further specificity is difficult. So although girls inter­nalize the awareness that some feminine roles are more worthy than others, their actual role choice remains indeterminate. For example, how does a girl interpret being a “good mother?” Is it the woman who stays at home, baking cookies—or is she a suffocating, overly attentive mother? Is it the Super-Mom of media hype—or is she a cold careerist barely able to manage a couple of hours of “quality” time? To be sure, these and other idealizations of Woman exert a pressure on girls and women to be the impossible. Nonetheless, these reflections point to the possibility of variability within the social role expec­tations which constitute the generalized other.

Third, there is not simply the generalized other, but a number of different generalized others. We internalize the expectations of more than one group. Indeed, the recognition and reinforcement we receive through one set of connections helps provide us with a standpoint for interpreting, and often combatting and rejecting, the distorted recognition or even lack of recognition we experience in other groups. For example, Janie Victoria Ward describes the role of the conventional generalized other in the process of identity formation in black girls. She writes:

As the black child sees herself as others see her, she knows that she is viewed in this society as a member of a devalued group. Transmitted daily to black children are messages that black people are undesirable, inade­quate, and inferior. Therefore, if she is black, she is undesirable, inade­quate, and inferior65

Yet, when the black child can call upon positive experiences from within the black community, when she feels reinforced and recognized in her connec­tion with black family members and friends, she can use this perspective to reject conventional expectations: “I am not what you believe black people to be, and I am black.”66 Clearly, learning sex roles can also be seen in such a light, bell hooks describes the way she countered “the right speech of woman­hood,” the speech which from the conventional perspective of the community provided a sort of background music which could be tuned out, claiming an authorship rooted in her valuation of speech among black women.67 The presence of generalized others, of groups with expectations and norms other than or beyond those of the conventional generalized other, provides us with a third perspective in the concept of the generalized other.

Thus interpretations of the generalized other are variable, the shared norms and expectations organized in the perspective of the generalized other are in some sense indeterminate, and the position of the generalized other can actually be broken down into a series of perspectives out of which generalized others are constructed. This openness means that we can never completely assume the perspective of the generalized other Instead, we take over an inter­pretation of it, an interpretation which arises out of our understanding of ourselves in the context of the relationships in which we are situated. Of course, becoming aware of the interpretations which we take over, and recog­nizing them as interpretations which can be questioned, requires a critical distance, that is, the perspective of a third person toward the perspective of the generalized other itself. In fact, it is this degree of reflection which makes possible the shift to the post-conventional level for this is the hypothetical atti­tude which we adopt when we question the norms and conventions of our social world. The gap between the perspectives of the generalized other and the third-person observer tells us how it is possible to break out of the confines of role recognition demanding that we take seriously the reciprocal recognition of each as a subject deserving of equal respect.

Awareness of sexual difference requires that we rethink the logic of moral development, attending to the importance of our interconnections if we are to recognize each other as subjects deserving of equal respect. Further, taking a hypothetical attitude toward the generalized other is an essential aspect of post-conventional morality. If we are to move beyond complementary forms of recognition to achieve the more reflective form of reciprocity, we have to continuously struggle against restrictive interpretations of the generalized other, to become aware of the differences among our expectations, and to assert the importance of the various groups of which we are apart. Achieving such recognition in practice remains dependent on our critical efforts and political engagement. We must fight against those cultural interpretations which limit our ability to see beyond conventional hierarchical and gendered notions of roles, rights, and duties. We must question given sets of expecta­tions, exposing the power differentials within them in order to create spaces for new expectations, for the expectations that we think should be shared. Finally, we must work to provide opportunities for the emergence of a vari­ety of groups and relationships founded on this recognition.