In this article, we attempt to understand the significance of these changes in the political landscape of gender. We are concerned with three issues. The first issue relates to the usage of notions of civil society from a feminist perspective. In particular, we are concerned with the ways in which definitions of political and civil society, as well as of democratization processes, include women as political agents. Secondly, we address changes within the women’s movement which enabled the Women’s National Coalition to achieve a significant presence in national politics during the 1990s. We address this issue particularly from the perspective of the emergence of a consciousness and discourse of gender within women’s organizations, although we recognize that other factors, such as the transitional process itself, played an important facilitating role. In this context, however, we are particularly concerned with understanding the unique character of the women’s movement in South Africa, as well as the relationship between the women’s movement, national liberation movements and civil society more generally. We deal with shifting relationships within ANC-aligned women’s organizations and the issue of an independent alliance of women’s organizations within civil society. Finally, the article addresses the question of whether the engagement of women’s organizations with the state, and the institutionalization of gender inside the state, has led to the demobilization of the movement. We argue that a simple demobilization thesis is not helpful for understanding the impact of the democratization process on the political opportunities available for women’s organizations.

Civil society has become a notoriously difficult concept to pin down in social science. Gordon White (1994; 376), highlighting the ambiguities of the term, points out that "though there is now a `paradigm’ of thought and a terrain of discussion…the term means different things to different people and often degenerates into a muddled political slogan". In South Africa, the term has acquired currency as a discourse which captured the range of organizations which were independent of both the state and the dominant national liberation movements (Glaser, 1997: 8). From this perspective, the notion of civil society functioned as political rationale in ways very similar to White’s (1994: 377) account of "an idealized counter-image, an embodiment of social virtue confronting political vice: the realm of freedom versus the realm of coercion, of participation versus hierarchy, pluralism versus conformity, spontaneity versus manipulation, purity versus corruption". As White argues, this political baggage may be serviceable from the point of view of activism, but obscures a range of considerations about relations of power within civil society, and of the precise abilities of civil society to achieve the democratic and developmental ideals which accompany its rhetoric.

Extensive discussions of the notion of civil society can be found in a range of sources (Keane, 1988, White, 1994, Glaser, 1997). For the purposes of this article, we consider civil society (both theory and practice) from a feminist perspective. We utilize the very broad definition of civil society offered by White (1994: 379): "an intermediate associational realm between state and family populated by organizations which are separate from the state, enjoy autonomy in relation to the state and are formed voluntarily by members of society to protect or extend their interests or values". This inclusive and descriptive notion allows for the diversity of organizations that may be found within civil society. We use the term `public sphere’ in the feminist sense, to include both the political realm as well as the realm of civil society.

Even at the broadest levels of description, feminists have critiqued the conception of civil society because it is predicated on the conceptual distinction between the public and the private spheres. This distinction is seen as having an exclusionary impact for women, both in terms of the ways in which the political and civil communities are defined as well as in the practical effects it has for the participation of women in the public sphere (Hirschmann, 1998). Both types of exclusion – the constitutive and the practical – obscure women’s agency as citizens, the former by limiting the definition of the political and the latter by failing to address systematic barriers to women’s participation.

The distinction between public and private sectors in society has also been criticized for its exclusionary impact in terms of participation in the public realm. This impact can be viewed at a number of levels. Firstly, feminists have argued that viewing the public sphere as the sphere of free equals, distinct from the inequalities of the private sphere, does not capture the reality that women’s unequal position within the family establishes the boundaries of their public participation.

As Gal (1997: 34) points out, despite claiming the equality of individuals in the political realm, idealized notions of civil society can obscure the economic and other social differences that constrain political participation. Civil society is often treated as a construction of individuality in opposition to the private space of the family. Far from being distinct spheres of human activity, however, Pateman has argued that the restrictions on women’s mobility and contractual powers, as well as ideological assumptions about ‘appropriate’ women’s roles, have a direct impact on their public capabilities. Although all citizens may have the right to participate, women ‘s private responsibilities and social stereotyping of women’s public roles which arise from these make it difficult for women to participate as equals in the public sphere. Indeed, David Held argues that reproductive rights are ‘the very basis of the possibility of effective participation in both civil society and the polity’ (quoted in Lister, 1998:18). Although women can be seen as being hindered in terms of their political participation in the same ways as other marginal social groupings (such as rural people, poor people, etc.), reproductive responsibilities and the division of labor in society act as particular gendered constraints. The impact is to narrow the scope for women’s participation in formal politics and in civil society.

A second level of understanding the ways in which women have been excluded is to examine explanations for democratization processes. Analyzing democratization debates from a gender perspective, Georgina Waylen (1994) has shown how the focus of what she terms ‘orthodox’ views of democratization omit women. At a constitutive level, this omission stems from a narrow view of politics as an elite-driven process. Analyses of democratization which focus purely on the high politics in effect value only male-dominated politics – women’s role as political agents is downplayed because their spheres of political activity are not seen as significant. This has implications for the way democracy itself is conceived, for it is narrowly limited to "an institutional arrangement to generate and legitimate leadership" (Waylen, 1994: 332). Issues of social and economic equality are divorced from that of political equality. The outcomes of democratization are excluded from analysis. As a consequence gender concerns beyond formal equality, such as issues about the redistribution of power, are marginalized.

Waylen argues that utilizing a wider notion of politics, one which encompasses not only institutions and processes among elites but also seeks to understand the relationship between popular mobilization and democratization processes, is necessary to capture the ways in which women have been political agents. While women’s presence in formal political organizations may be minimal, women’s politics has nevertheless often been robust and impacted on processes of democratization. In fact, as Waylen points out, the latitude allowed to women for ‘symbolic protests’ (such as silent demonstrations) by authoritarian regimes often kept alive struggles against tyranny when more direct forms of protest action were impossible.

Despite various constraints on women’s political activity, women have historically succeeded in forming organizations to take up their various causes. Women’s movements, defined as "those socio-political movements, composed primarily but not necessarily exclusively of female participants, that make claims on cultural and political systems on the basis of women’s historically ascribed gender roles" (Alvarez, 1990: 23) constantly challenge the boundary between the public and the private. By raising issues of "private" inequalities and oppressions such as responsibilities for childcare, reproductive freedom and protection from rape, for example, women’s movements redefine the landscape of politics. At the very least, the definition of the content of public discourse is re-shaped by the articulation of these concerns. For some feminists, the argument is taken further to suggest that the nature of politics is also altered by these new political concerns (Ruddick, 1989).

In South Africa, debates on the women’s movement have been somewhat peripheral to broader debates on civil society. The focus on the impact on democratization of movements in civil society has spawned a wide literature in South Africa (Friedman, 1991; Glaser, 1997; Humphries and Reitzes, 1995). What is absent from the literature is a consideration of women’s role in the process. Glaser (1997: 6) points out that two arenas have engaged analysts: the labor movement and the civic movement. Despite unusually high levels of women’s involvement in the civic movement in South Africa in the 1980s and 1990s, and a significant body of literature documenting women’s political activities, little attention is paid to the gender dimensions of these political activities and debates in mainstream civil society literature. And yet, as this article will argue, women’s organizations have been relatively vigorous in democratization struggles, and have addressed their demands at both the institutional design of the new democracy as well as at the outcomes of democratic processes. The demands of women’s organizations have deepened notions of democracy in the constitution as well as extended the scope of institutions in South Africa. Nevertheless, these contributions tend to be excluded from mainstream analyses of democratization in South Africa. The omission stems from both the constitutive as well as the practical limitations described above, and are reinforced by the weaknesses of the women’s movement in South Africa in sustaining a national impact.

This discrepancy between the country’s egalitarian norms and the inequalities between large groups in society has the potential to weaken the consolidation of democracy. White (1998: 28) identifies two dangers for new democracies which fail to address glaring inequalities: firstly, social discontent and political instability is likely to increase, and secondly, large sections of the population are excluded from access to the political process as a result of persistent poverty. The implication is that formal inclusion will not serve the democratic purposes of stability as long as it fails to be the basis for addressing socio-economic inequalities (including those of gender). This argument reinforces feminist demands that it is not only the formal institutions and procedures of democracy that must be addressed, but also the outcomes of democracy.

Although women’s organizations in South Africa have a long history (Walker, 1989; Wells, 1993), they have not always been nationally significant actors. Part of their marginality stems from the constitutive and practical exclusions discussed above. In addition, however, there are weaknesses all women’s organizations face: a dispersed constituency, the difficulties of identifying common interests in theace of cross-cutting racial, class and ethnic identities, and the intricate relationship between personalized discrimination and oppression on the basis of gender with systematic male bias.

In understanding why these weaknesses were overcome (at least temporarily) during the 1990s, we address two key factors identified in civil society debates - the development of a political consciousness of gender (Cohen and Arato, 1992), and the nature of political opportunities external to women’s organizations (Tarrow, 1994), particularly in terms of their relationships to both the state and the dominant liberation movements. Both these factors have had a significant impact on the ability of women’s organizations to position their claims within the larger demands of socio-political and economic transformation demanded by dominant sectors of civil society in the 1980s and 1990s. The emergence of women as a political constituency in South Africa in the 1990s can be traced to three interrelated developments: the changing spaces occupied by women within the African National Congress (ANC) and the United Democratic Front (UDF) during the 1980s, the opening of new political opportunities during the transitional process, and, allied with both of these, the maturation of a political consciousness of gender in the engagement with both the state and national liberation movements.

The relationship between the development of a consciousness or identity is central to the emergence of a social movement (such as the women’s movement) within civil society. Cohen and Arato (1992) distinguish between two models of explanation for this relationship, described as the stage model and the dual politics model. The stage model posits that social movements move from non-institutional mass protest action to institutionalized, routine interest groups in a linear process (Cohen and Arato, 1992: 556). In the first stage, demands are made that are non-negotiable and diffuse and usually involve mass protest action. During this stage, identity formation takes place and the aim is to form solidarity around this identity rather than with strategic gains and losses. In the second stage, routinization, inclusion and finally institutionalization take place. The organizational structure is changed from informal to formal. Successful inclusion and full institutionalization implies representation in "normal" politics, involving party competition.

Cohen and Arato (1992: 557) criticize this assumption of linearity for its inability to account for significant features of the feminist movement in the USA. They posit instead a notion of "dual politics", in which the discursive politics of identity and influence targeted at civil and political society exist simultaneously with the organized politics of inclusion and reform aimed at political and economic institutions (Cohen and Arato, 1992: 550). In this model, identity formation is inextricable from the development of movement objectives and strategies, each shaping the other.

Drawing on studies of the women’s movement in the USA, Cohen and Arato argue that the dual politics model is of greater explanatory value than the "stage or life cycle" model usually used to explain collective action. In the USA, such a dual logic existed in the women’s movement from the start. While routinization and institutionalization have taken place, these processes did not substitute collective action, grassroots mobilization or identity-oriented politics. The women’s movement was able to shift between mass action and political pressure on the state. They conclude (1992: 562) that the success of social movements on the level of civil society should be conceived not in terms of the achievement of certain substantive goals or the perpetuation of the movement, but rather in terms of the democratization of values, norms, and institutions that are rooted ultimately in a political culture.

Part of this political culture, in their view, must include the development of a specifically feminist consciousness (1992: 552). In our view, such a feminist consciousness cannot - should not - be defined a priori according to the abstract definitions of universal theory, but should be defined in the context of particular social formations and should have resonance in the historical experience and political culture of specific societies. Feminism is not a coherent ideology, nor is it uncontentious even among organizations committed to gender equality, as the South African debates reveal. Despite these limitations, we draw on Cohen and Arato’s (1992: 553) important insight that a successful movement requires a politics of identity aimed at norms, social relations and institutional arrangements and practices constructed in civil society, as well as mass action aimed at political society. A new collective identity for women that is simultaneously articulated at the grassroots level and accepted by the political elites is a necessity for success.

Cohen and Arato raise issues which lie at the heart of debates conducted within South African women’s organizations for decades. What is omitted from their discussion is the relationship between identities, interests and issues in gender politics. We deal with these issues in the South African context below. In South Africa, the women’s movement has contested the relative importance of gender vis-à-vis other social and political identities. As an ideology, feminism has been challenged not only by resistant men, but also by women themselves. The definitions of what constituted feminism, and concomitantly the women’s movement, have been fiercely debated.

The relationship between women’s equality and national liberation was central to this debate (Beall, Hassim and Todes, 1989). Despite the recognition of feminism as an umbrella term encompassing many variations, popular usages of the term tended to treat the ideology as coherent and distinct.

The context of a national struggle against apartheid framed many of these debates in a way that was often difficult to resolve. Some argued that a focus on national liberation weakened the possibilities for the emergence of an assertive women’s movement in South Africa (Beall, Hassim and Todes, 1989; Charman, De Swart and Simons, 1991; Horn, 1991). This constraint has been seen as both epistemological as well as strategic. At the epistemological level, the formulation of ‘triple oppression’, while capturing the inter-relationship between race, class and gender, treated these three identities as additive (Hassim, 1991:68) and led to an ‘unilluminating repetition of formulae’ (Walker, 1990: 3). Ironically, the specificity of gender was downplayed in activist approaches to women’s political roles. Analyzing contributions to the ANC’s Malibongwe conference in 1990, Charman et al (1991: 42-43) point out that there is an ‘androcentric legacy’ in political and popular discourse: When women are discussed it is very much in the masculine conception of women. Women are added on. Gender is not a fundamental category of analysis.

Others have argued that a reluctance to build a women’s movement separate from the national liberation movement stemmed from both the inseparable nature of race, class and gender oppression as well as from "a strategic choice made in the face of opposition from a seemingly invincible white nationalist party-state that was quick to exploit any sign of division in order to subjugate black people even further" (Kemp et al, 1995:142). Primo (1997:33) argues that there are positive aspects to women’s participation in national liberation movements that have been under-emphasized: The participation of women in the national liberation struggle aided the women’s struggle. For one, it raised the profile of women in the national liberation struggle, and placed some in leadership positions. Secondly, it created the space to raise issues of concern to women, to give public voice to many women’s private struggles.

Definitions of what would or did constitute a women’s movement in South Africa varied, and were overladen with prescriptive formulations of ‘progressiveness’ (Walker, 1990, Hassim, 1991). The dilemma of defining a women’s movement is not confined to South Africa; it is tied to the question of the linkage between interests and identities. Women’s interests cannot simply be read off from their economic or social position; women are divided by race, class, regional and other interests in ways which threaten to undermine the coherence of their gender identity. Addressing the tension between feminist theoretical perspectives on what constituted feminist politics and grassroots women’s action, Maxine Molyneux (1985) offered a conceptual distinction between ‘strategic gender interests’ and ‘practical gender needs’. Strategic interests can be defined as those claims which seek to transform social relations so as to promote the equality of men and women, while practical interests may be seen as those which arise from women’s gendered responsibilities within the family and community and which make no explicit claims to challenge power relations. While pursuing strategic gender interests assumes the existence of some form of feminist consciousness, practical gender needs can be articulated without such consciousness.

Although there has been some debate about the separability of these two concepts in actual struggles conducted by women’s organizations (Kabeer, 1993), the usefulness of Molyneux’s distinction was to open the space for a conceptualization of women’s movements which allows for this diversity of interests to be accommodated. A women’s movement can contain within it conservative elements that organize women from a particular social base but do not seek to question power relations within that base, let alone within society more generally. By contrast, in its broadest formulation feminism has a direct political dimension, being not only aware of women’s oppression, but prepared actively to confront patriarchal power in all its dimensions. From the point of view of making an impact on national political processes, we would argue that the development of political identity based on strategic gender interests is crucial to the success of the women’s movement in addressing gender inequalities through institutions.

The groundwork for the emergence of a strong political consciousness and advanced organizational forms among women was laid in the 1970s and 1980s, when community-based struggles drew in significant numbers of women. The central thrust of these struggles was a battle against the apartheid state (Meintjes, 1998), although in the process women’s organizations also defined a new space within civil society and new relationships with other organizations. The formation of the United Democratic Front (UDF) in 1984 added impetus to the organization of women, as it galvanized a range of forces at community level (Fester, 1997; Madlala-Routledge, 1997). Organizations such as the United Women’s Congress (UWCO) in the Western Cape, the Federation of Transvaal Women (Fedtraw) and the Natal Organization of Women (NOW) organized women around a core of issues which stemmed from local community concerns but made broader demands for the abolition of apartheid and for economic transformation. The membership of these organizations was mixed: veterans of the 1950s women’s struggles, young black women (both educated and working class) and a small number of white academic and professional women. Women’s organizations under the UDF umbrella in the 1980s developed strong branch formations as well as articulated linkages between women’s issues and national issues through marches and other forms of protest (Houston, 1998: 152). These in turn provided opportunities for leadership training which formed the bedrock of the Women’s National Coalition in the 1990s. In the process women’s issues increasingly were placed on the agenda of major progressive organizations, although not without continuing tensions about the status of gender in national struggles.

There are two positions taken in the literature which analyses UDF-aligned women’s organizations. One position is that women in these organizations were primarily focused on issues of immediate relevance to their gendered responsibilities (or in Molyneux’s terms, their practical gender needs) such as service delivery in the townships, rents and health, and that these became the basis for the emergence of a grassroots consciousness of gender (their strategic gender interests) (Beall et al, 1987; Patel, 1988, Jaffee, 1987). The formation of the UDF Women’s Congress was a mechanism for addressing the latter set of interests: "A concerted attempt to assert women’s leadership, bring women’s issues into the UDF in a more forceful way, and ensure that women’s struggle is an integral part of political struggle" (Jaffee, 1987: 74).

An alternative view, expressed most succinctly by Fester (1998), suggests that motherism, or women’s identities as mothers, was the basis of women’s mobilization. She points out that the re-launch of the Federation of South African Women (FSAW) and the United Women’s Congress (UWCO), both of which affiliated to the UDF, was accompanied by references to women’s roles as mothers and the impact of injustice in the country on women as mothers. Fester (1998: 6) argues that the overall political struggle for national liberation was an integral part of women’s lives as mothers caring for their children. She concludes that Motherist movements which defy oppression of all kinds, including patriarchal oppression, are more likely to inspire a feminism which will be effective in South Africa than would be the liberal feminism espoused by most first world feminists (Fester, 1998: 16).

We return to the issue of motherism as political identity below. In this context, we are concerned with the impact of the heightened involvement of women on the discourses and practices of major civil society organizations. The most significant impact was the declaration on May 2, 1990 by the ANC that the emancipation of women is not a by-product of a struggle for democracy, national liberation or socialism. It has to be addressed in its own right within our organization, the mass democratic movement and in society as a whole.

The significance of this declaration lies in the space it allowed women within the fold of the progressive movements to organize self-consciously on their own terms and in their own interests. It made possible the autonomous organization of women. This statement also provided the organizational latitude to the ANC Women’s League to open discussion in 1991 on the possibility of a broad-ranging, multiparty women’s coalition – a women’s movement.

The transitional process itself, during which political antagonists were engaging each other, facilitated this development as cross-party meetings were legitimated. Furthermore, the discourses of transition - the terms of the democratic society, the nature of new institutions, the values which would underpin the Constitution - were discourses in which the broader notions of social justice demanded by women could make an impact.

However, establishing such a movement of women was difficult. Firstly, as has been noted above, gender solidarity had historically been undervalued as a political resource and strategy. Secondly, feminism itself encompasses a diverse set of interests and identities, raising what post-structuralists have seen as the core dilemma in establishing and sustaining a women’s movement. Soper (1994:14-15) captures this dilemma well: feminism as theory has pulled the rug from under feminism as politics. For politics is essentially a group affair, based on the idea of making ‘common cause’, and feminism, like any other politics, has always implied a banding together, a movement based on the solidarity and sisterhood of women, who are linked by perhaps very little else than their sameness and common cause as women. If this sameness itself is challenged on the grounds that there is no ‘presence’ of womanhood, nothing that the term ‘woman’ immediately expresses and nothing instantiated concretely except particular women in particular situations, then the idea of a political community built around women - - the central aspiration of the early feminist movement - collapses.

An examination of the immediate past of the women’s movement in South Africa (the 1990s) would suggest that the possibility exists for the creation of a political community built around women, where that political community is prepared to accept that the terms of unity are rather narrow and constantly under negotiation. The experience of the Women’s National Coalition reveals that it is possible to build a movement while recognizing diversity, where solidarity rather than sameness is the basis of political action.

The formation of the Women’s National Coalition in 1991 was a turning point in the history of women’s organizations in South Africa. Cock (1997: 311) argues convincingly that the Women’s National Coalition was an organic women’s organization which brought together different constituencies from political parties to occupational and religious groups, service and special interest groups and community organizations, as well as the women’s sections of the different political parties (see also Kemp et al, 1995: 150). Initiated by the ANC Women’s League, the WNC soon comprised a range of women’s organizations with diverse interests, drawn together by the single issue of ensuring that women were included in the transitional and negotiations processes. The formation of the WNC also stemmed in part from the concern among some women’s organizations allied to the UDF that the unbanning of the ANC Women’s League would undermine the autonomy which had been achieved during the 1980s. From this perspective, a national organizational structure that retained some degree of independence from the national liberation movement was desirable.

The ANC Women’s League took an especially strong stance over women’s representation in constitutional negotiations, threatening to boycott the first non-racial elections with the slogan ‘no women, no vote’. Although it can be questioned whether the Women’s League had the organizational capacity and grassroots support to give effect to such a threat, this stance was instrumental in establishing the notion of women as a political constituency.

The central aim of the Coalition was to draft a comprehensive "Women’s Charter for Effective Equality", to be presented to negotiators for inclusion in the Bill of Rights. To keep this single-issue campaign together required intense debate about women’s needs and interests. The development of the Charter was a mechanism for resolving a range of competing concerns among women, but ultimately the document itself represented a collection of disparate demands, rather than a prioritization of particular women’s needs and interests. The affiliates in the Coalition were concerned not to superimpose a false universalism, but to ground the organization’s strategic direction in the diversity of South African women’s experiences and in addressing specific material conditions of disadvantage (Cock, 1997: 316).

The broader context of transition provided a framework within which an articulate women’s movement could make certain gains. The discourses of transition - the terms of the democratic society, the nature of new institutions, the values which would underpin the Constitution - were discourses in which the broader notions of social justice demanded by women could make an impact. This impact of the WNC on the constitutional debates as well as on the institutional framework for democracy provided immediate gains from collective action which sustained the organization in the short term.

Cock (1997: 317) argues that the through the Charter campaign, a collective identity began to emerge within the WNC. It brought together a hybrid of organizations that were very different from one another through a liberal feminist emphasis on rights and political representation. However, this identity appears to be rather muted. She points out that while the WNC represents an extremely wide cross-section of South Africa’s women, it does not have the common purpose that enables it to campaign effectively. On the other hand, the individual organizations that are able to mount campaigns tend to represent narrow groupings among women (Cock, 1997: 329).

It is unclear, therefore, whether the WNC was kept alive by the strategic spaces opened up by the transition process, or by the emergence of a common identity. Meintjes (1998: 79) argues that identity was built around a limited agenda of common interests - most notably the demand for equality- tempered by the recognition of diversity. An alternative view is that the consciousness that kept the WNC alive was rooted in motherhood and mothering and not the typically Western-inspired feminist consciousness (Fester, 1998). The Women’s National Coalition communications with UWO, UWCO, FSAW and the Women’s Alliance included the request to mobilize unorganized women through appealing, inter alia, to motherhood. Both Cock (1997: 320) and Fester (1997: 46) argue that "motherism" is a form of South African feminism that often brings women together and leads to mobilization around women’s issues. Meintjes (1988: 68) highlights the uses of motherist discourses within nationalist movements. This "motherism" is linked to an ethic of care which includes characteristics such as responsiveness, attentiveness, responsibility and nurturance (see Gilligan, 1982) and which causes women to perceive their lives as linked to that of children and others with whom they have close relationships. Yet, we need to ask whether "motherism" was used for the purposes of inclusion – to build solidarity among women, and to smooth over differences rather than forming a bond based on a common experience of motherhood (see Cock, 1997: 320).

Even though the WNC can be viewed as an organic women’s movement, the issue of collective identity remains a thorny one. Membership of the WNC was only open to organizations, not to individuals. The WNC was therefore already institutionalized through different organizations such as political parties, unions and service organizations even before it emerged as a coalition. This prior institutionalization is expressed in the close connection of the WNC with political parties, especially the ANC. This closeness was also expressed toward other political allies. A logistical problem arose, for example, when COSATU as a union federation could not join because its constitution makes it impossible to join another "federation". The problem was solved by allowing COSATU to be involved without affiliating (Kemp et al, 1995: 152).

These already constituted political groupings established limits on the ability of the WNC to consolidate a common feminist identity. Ethel Klein (1984: 3) argues that the rise of a political consciousness (such as a feminist consciousness) is a three stage process: (1) affiliation to a group, (2) recognition of group membership and (3) shared interests. If this is the case one should ask whether affiliation to a group and the recognition of group membership in the case of the WNC was with a broader women’s group or rather with the affiliated groups that belonged to the WNC. Although a feminist consciousness did not deeply root itself in the WNC, nevertheless a very strong sense of the relationship between gender issues and social justice did emerge, which came to be accepted formally by the political elites and was implemented in the drafting of the constitution.

However, when the single-issue charter campaign came to an end "motherhood" was not enough to sustain a common identity. Nor can the recognition of difference in itself sustain a movement. As Meintjes (1998: 81) points out, "Women’s capacity to work together depends upon mobilizing around unifying issues and, since the elections, these have not emerged in a way to engage the WNC". From this perspective, "demobilization" might have resulted from the fact that women went back to the institutionalized narrow interests of their different organizations, from which it is difficult to mobilize mass action, rather than from their incorporation into state processes. As Kemp et al (1995: 153) point out, after the Charter was adopted, the constant balancing of political forces and interests led many of the WNC’s key leaders, especially those within the major political parties, to advocate the Coalition’s dissolution.

Although Fester’s argument that motherism is a more fertile ground for the development of a women’s movement in South Africa than western feminism, is seductive, it does not take full account of tensions between conservatizing aspects of motherism and the project to transform gender relations. The problem with motherhood as an identity is that it locks women in the private sphere and reinforces their identities in relations to others as opposed to their autonomy as individuals. Thus, when the politics of feminism starts to challenge the boundaries of the private sphere resistance emerges even among women against issues of autonomy, bodily integrity and self-determination. A political strategy based on motherhood is not inherently progressive, no matter how strongly rooted it may be in women’s experience. As Gaitskell and Unterhalter have shown, the concept of motherhood is ‘very fluid and manipulable’ (Gaitskell and Unterhalter, 1989:75). It is not incidental that motherhood is central to nationalism in both its conservative and progressive variants. In comparing the uses of motherhood by the ANC and the IFP, Hassim (1993:20) argues against the reification of motherhood as a political concept. She makes the point that ‘both conservative and progressive forms of nationalism draw on women’s support in terms of their relationship to others rather than in their own terms. Women are equated with mothers, and families with women.

Freedom of choice, demand for control over one’s body and autonomy are more than a quest for equal rights in a bourgeois sense. Rather, they challenge the traditional identities of women and their gender roles and thus pose a serious challenge to the private sphere that cannot easily be accommodated by motherist politics. The resistance to this challenge in South Africa is expressed through the lack of support on the level of women’s public opinion for abortion. Abortion in South Africa was legalized through the efforts of law reformers and feminist reproductive rights groups, not by large scale women’s demands for autonomy and control over their bodies.

This is not to deny that motherism is an authentic, alternative South African feminist consciousness. It has played a central role in making space for the discussion of the relationship between nationalism and women’s rights in a context in which the more conventional western forms of feminism had little to offer South African women. This is not unique to South Africa – it is also the case in Eastern Europe, for example. Gal (1997: 42) points out the importance of distinguishing between structural issues of organizing in civil society and oppression that is discursively constructed. Civil society has ideological and discursive boundaries that determine what can count as "political" and these boundaries in turn shape what identities and activities can form part of the public in a given historical version of civil society (such as motherhood). In Eastern Europe women as a social category are depoliticized. Gal argues that what has not been discursively constructed is the political category of "woman" as an independent subject whose interests and issues can be publicly defined and debated.

However, the limits of discursive constructions have to be actively and continually confronted by social movements. While the discourse of women as "mothers" or "mothers of the nation" or "desexualized beings" allowed space for women’s political mobility in difficult circumstances (Primo, 1997), and even cleared the ground from which to speak of women as autonomous political agents (as in the ANC’s May 2 Document), its appropriateness as the dominant discourse needs to be under constant examination. This is particularly true in the context in which women have a constitutional right to equality, bodily integrity and autonomy. Are certain boundaries still discursively constructed through which new feminist identities cannot break?

Unlike Eastern Europe where the category "women" was depoliticized, in South Africa the category is infused with race. New feminist identities cannot be constructed as long as the women in South Africa do not deal sufficiently with the "politicisation of difference" that is a prerequisite for feminist identity politics (Watson, 1997: 22). The politicisation of difference on grounds of race and gender in South Africa is the subtext of identity politics that has led to the "politics of hostility". It has been the undertone in much of women’s organizing and mobilizing, even though it may not always be explicitly articulated. In order to develop feminist politics in South Africa—politics which includes issues of autonomy, self-determination and bodily integrity—we need to grapple with the shifting boundaries of the intersection of the identity markers of gender and race. Identities are constituted through the intersections between race, class, and gender which bear the mark of a particular cultural and historical context. At the same time, as De la Rey (1997:9) cautions, the problem of race should not be reduced to the general label of `difference’. She argues that any South African feminism that ignores the centrality of race will run the risk of making it invisible, and it will be a limited feminism. . . . We need to avoid the notion that all forms of difference are equivalent.

The WNC experience would suggest that the development of an identity for the women’s movement emerged from a pre-existing history of women’s organization, of intense debate about the meanings of gender in a society dominated by racial divisions, and the will of a politically committed and experienced core of women leaders. It was tied, crucially, to the successful identification of issues which cut across a wide set of differences and gave a common basis for solidarity. The formation of the WNC as an activist federation and the subsequent departure of women leaders into the arena of institutionalized politics have undermined leadership in the organization, but the mobilizing of women as a constituency has left a significant legacy. This legacy is both institutional, in the creation of the national machinery for women, as well as political, in that it has created a reference point of success for future movement battles.

To some extent, the successes of the Coalition have also opened up the discursive space for feminism to be debated in new ways. The demands for participation, representation and for changing gender power relations have become an acceptable part of the discourse of women’s organizations. The challenge for the women’s movement now appears to be centered more directly on the ways in which institutional gains can be used to address different women’s needs. In this process, as the Deputy Speaker, Baleka Kgositsile-Mbete (1998: 6) argued, the challenge is to create mechanisms for communication, support and accountability between the women’s movement in civil society and women inside the state.

New spaces have opened up in civil society to deal with issues of violence and poverty in ways which directly address male power in society. A different type of feminist consciousness may be developing that is transcending the politics of the private sphere and is taking on issues of women’s autonomy and independence. This politics transcends that of motherism, in that its reference points lie beyond women’s relationship to children (although these are often encompassed as well).

In order to understand why the "women’s movement" in South Africa has become relatively inactive since the 1994 election, it is necessary to examine the shifting relationships between the state and civil society. To what extent are the conventional arguments of social movement theorists about the link between institutionalization and demobilization useful to understand the trajectory of women’s politics in South Africa in the 1990s? In addressing this issue, we develop the argument made in preceding sections about the interconnections between constructions of the state, civil society and a feminist identity.

A key concern in social movement theory has been the impact of institutionalization on movements. By institutionalization is meant the creation of a repeatable process "in which all the relevant actors can resort to well-established and familiar routines". For political movements, institutionalization denotes the end of the sense of unlimited possibility (Meyer and Tarrow, 1997:21). Meyer and Tarrow identify three main components of institutionalization. Firstly, collective action becomes routinized. Secondly, a process of inclusion and marginalization occurs, in which those actors who adhere to the `rules of the game’ are granted access to key institutions while those who refuse are excluded. Thirdly, co-option occurs, in that movements alter their claims and their tactics into forms that are not disruptive to the political system.

From the point of view of women’s movements, these conventional markers of institutionalization are of limited benefit. They derive from an analysis of social movements characterized by high levels of overt protest action, perhaps even violent activism, and with a high potential to disrupt the political order. These have not been typical characteristics of women’s movements, which have particularly not had the power to radically disrupt the system. For these reasons, Katzenstein (1997:197) approaches the problem in a slightly different way, conceiving of institutionalization in spatial terms as the establishment of ‘organizational habitats’. ‘Such habitats are spaces where women advocates of equality can assemble, where discussion can occur, and where the organizing for institutional change can occur’. To what extent did institutionalization of gender, through the creation of the national machinery, result in the kind of institutionalization that Meyer and Tarrow highlight, and have the consequences been negative for sustaining the women’s movement? What alternative ‘habitats’ have the new institutional arrangements created, and can these habitats be useful spaces for women to articulate their interests?

An examination of the institutionalization process in South Africa reveals that at many levels, civil society appears to be demobilized. Despite the significant constitutional and political gains (not least of these the right to equality), it would seem that in the period since the first democratic election since 1994, women’s politics has been less effective and women’s organizations less visible than in the previous years. This does not mean that women’s issues were off the agenda; indeed, certain issues such as violence against women, and rape specifically, have enjoyed a higher profile than ever before. One of the reasons for this downward trajectory is that the departure of many highly skilled women organizers and leaders from civil society into government ‘has harmed existing networks and organizations of civil society, and placed increased pressure on the human, technical and financial resources of non-governmental organizations’ (Albertyn, 1995:11). Thus the shift to ‘engaging’ the state has at one level had the direct, albeit unintended and unavoidable, consequence of demobilizing women’s organizations, most notably the Women’s National Coalition. This pattern has been noted in other contexts by O’Donnell and Schmitter (1986), who argue that in the early phases of transition, social movements have considerable influence to push the transition further, but that in the latter stages of transitions political parties dominate in shaping political agendas.

The development of a range of institutions inside and outside government for the purpose of making and influencing social policy from a gendered perspective has shifted the terrain of politics more closely into and around the state, rather than in civil society. This range of institutions, referred to as ‘national machinery’, includes structures within the government, as well as independent statutory bodies such as the Commission for Gender Equality. The establishment of the national machinery has shifted gender discourses away from an engagement with politics in the broader sense of vigorous organizational articulation of, and contestation around, interests, to a preoccupation with internal state processes on policy-making, on the one hand, and with the establishment of structures, on the other. The danger of this is that, as Kathleen Jones argues, "what counts as political action (becomes) limited largely to formal interactions between citizens and the state" (Jones, 1990:799). This is consistent with a general shift in political focus as the consolidation of democracy in South Africa appears to revolve largely around state-centric, policy-related issues.

However, as the demand for inclusion is nominally met, and the demands for transformation sidelined, the limitations of inclusionary strategies are being confronted. An example of the limits of inclusion is the political battle which ensued between women’s organizations and the women-friendly Minister of Welfare and Population over the reform of the child benefit grant system. In this battle, women activists inside the state were ranged against women activists in civil society in contestation over the impact of policy outcomes on poor women. Despite the potential for a strong motherist politics to emerge during this battle between state and civil society, women’s organizations articulated their demands around a discourse of poverty and gender, rather than around the identity politics of motherhood. What was interrogated was the extent to which the state can in fact act as a medium for transformation, which parts of a state (which is evidently riddled with contradictory policies) are more likely to benefit women, and how to deal effectively with institutional resistance to the inclusion of women.

At the same time, the policy-making process itself has generated a new dynamic of civil society regeneration, and a changed political opportunity structure. The relatively open nature of policy formulation in democratic South Africa, in which the public has the opportunity to comment and challenge draft policy before it is put to Parliament, as well as the formalized processes of consultation, have provided a focus for organizational energy and demands. Unlike in closed political systems, social movements in South Africa do have the space to protest against unpopular policies. Thus, for example, the protest against the state’s child benefit grant galvanized women’s organizations; a ‘New Women’s Movement’ based in the Western Cape has emerged. Similarly, the prioritization of poverty as a concern by the Deputy President’s office was accompanied by "Speak Out on Poverty" hearings hosted by the Gender and Human Rights Commissions and the South African Non-Government Organization Coalition. Issues of violence against women have also been central to the public debate about gender, and have been the focus for a number of large-scale protests. New coalitions of women’s organizations have emerged which are based on common struggles around specific issues, such as the Reproductive Rights Alliance, and the Network Against Violence Against Women. These developments have revived the terrain of civil society by opening up spaces for alternative discourses, involving women’s autonomy, self-determination and bodily integrity.

This stratification reinforces the tension between engaging the state to extract gains from inclusion and the task of maintaining a critical distance so as to ensure that there is accountability to a political constituency. The issue has to be seen in historical context: we would argue that women’s organizations have much to gain from the process of engagement and constructive dialogue (a la Kgositsile-Mbete) in terms of shaping the form of policies and budgetary priorities. In the longer term, however, this dual politics of engaging/ criticizing can only be successful if the constituencies of women on whose behalf engagement takes place is broadened beyond the non-governmental organization sector. Accountability of elected women representatives to different constituencies of women is key in this regard. As Kemp et al have argued, ‘at their best, the dialogues women hold with each other disrupt authority and redefine all the issues’ (Kemp et al, 1995:159).

A feminist critique of civil society shows that it is premised on the divide between the public and the private sphere and that the important contributions of the women’s movement to the development of, and debate about, civil society in South Africa have been marginalized and neglected. In this article, we have attempted to reinstate issues of gender in civil society debates. In particular, we have foregrounded the dilemmas of defining identity within the women’s movement, particularly in a society deeply marked by class and race differences which appear to be more overt stimuli to political action than gender. We argue that the formation of the Women’s National Coalition was a significant turning point in history of women’s organizations, not only because of the gains it was able to make in negotiations to democracy but also because it established the possibility of women as a political constituency, able to act in defence of a set of interests, albeit narrowly defined. We argued that inclusionary strategies have the potential to open new political spaces for women’s struggles. The skilful positioning of the WNC during the negotiations period pushed the terrain of what was negotiated beyond the initial intentions of political parties. Furthermore, the achievement of formal equality in the context of a brutal apartheid past should not be under-estimated.

The article also posed the question of whether institutionalization of the women’s movement through the creation of a national gender machinery has contributed to the demobilization of women’s organizations within civil society. We argued that a simplistic demobilization argument does not suffice to explain the relative inactivity of the South African women’s movement presently, but that this "demobilization" should be viewed within the specific historical context of the development of civil society and other social movements in South Africa.

Using assumptions of the dual politics model of civil society we argued that identity is inextricably linked to the development of the strategies and objectives of the women’s movement. We also interrogated the development of a feminist consciousness and pointed out that it is linked to strategic gender issues. Feminist consciousness develops within the context of particular social formations. While we agreed that motherism is an authentic South African feminism, we argued that it may not be enough to sustain a political identity and solidarity among women. What is needed is a move to the politics of autonomy, bodily integrity and independence to transcend the politics of the private sphere.

Exactly what the impact of the institutionalization of women’s politics has been on civil society in South Africa needs empirical investigation. We leave this challenge open to ourselves and others who want to extend and deepen the existing body of literature on women’s citizenship and civil society.

1. At the AWEPA/GAP conference "Women at the Crossroads", this view was expressed frequently and without being challenged by both women Members of Parliament as well as by women active in civil society. For a written sample of these views, see Kgotsitsile (1998) and Watson (1998). Among commentators, this view has been expressed most succinctly by Madonsela (1996), Cock (1997) and Meintjies (1998).

2. This is not to deny that the notion of women’s interests needs further interrogation. This issue is discussed below.

3. These are very narrow measures, calculated on rigid criteria. The GEM, for example, looks at participation only in parliament. Nevertheless, they are used here to capture the enormous variation in South Africa’s scores on the two indices.

4. For a discussion of this point, see Kemp et al, 1995.

5. For example, women’s sewing groups.

6. See Hassim, 1991; Horn, 1991.

7. The journal Agenda carried a special issue devoted to "Women and Difference" in 1993.

8. For example, the demands that national liberation should take account of women’s oppression, the demand for women’s equality within the movement and in the constitution and the questioning of the customary authority of chiefs over women.

9. Dr. Ginwala is now speaker in the National Assembly.

10. Of course, the belief in the progressive value of a focus on the state was predominant among many actors in South Africa at this time.

11. This insight is derived from attendance at a range of forums, including most recently the WomensNet/Commission for Gender Equality Information Audit in October 1998.