Hostname: page-component-745bb68f8f-s22k5 Total loading time: 0 Render date: 2025-02-06T13:12:26.979Z Has data issue: false hasContentIssue false
  • English
  • Français

Quota Problems: Combating the Dangers of Essentialism

Published online by Cambridge University Press:  29 March 2006

Jane Mansbridge
Jane Mansbridge
Affiliation:
Harvard University
Rights & Permissions [Opens in a new window]

Extract

As I write, descriptive representation by gender improves substantive outcomes for women in every polity for which we have a measure. And as I write, significant representation by gender cannot be achieved in any existing polity without some form of quota. At this historical moment, therefore, quotas play an important democratic role in increasing gender equality. Yet because quotas potentially produce institutional rigidity and their need should decrease as structural and cultural conditions improve, it is best to institute them in their most flexible form. Because quotas also encourage essentialism, it is best to institute them in conjunction with major efforts to define in nonessentialist ways the reasons for their institution. Although quotas will inevitably increase essentialist beliefs, a conscious, concerted campaign could mitigate the most destructive effects of this tendency.Many thanks to Lisa Baldez for help in editing these thoughts. Comments welcome: jane_mansbridge@harvard.edu.

Type
CRITICAL PERSPECTIVES ON GENDER AND POLITICS
Information
Politics & Gender , Volume 1 , Issue 4 , December 2005 , pp. 622 - 638
Copyright
© 2005 The Women and Politics Research Section of the American Political Science Association

As I write, descriptive representation by gender improves substantive outcomes for women in every polity for which we have a measure. And as I write, significant representation by gender cannot be achieved in any existing polity without some form of quota. At this historical moment, therefore, quotas play an important democratic role in increasing gender equality. Yet because quotas potentially produce institutional rigidity and their need should decrease as structural and cultural conditions improve, it is best to institute them in their most flexible form. Because quotas also encourage essentialism, it is best to institute them in conjunction with major efforts to define in nonessentialist ways the reasons for their institution. Although quotas will inevitably increase essentialist beliefs, a conscious, concerted campaign could mitigate the most destructive effects of this tendency.

The general principle, for collectivities as well as individuals, is that the existence of major dangers should not prevent our taking steps that increase our autonomy and well-being, if at the same time we can take parallel steps that reduce those dangers. A process fraught with danger needs two tracks: one to move forward and one to establish safeguards against the danger. Stasis often proves a poor response to threat.

The case for quotas, which cannot fully be made here, rests on three separate arguments: 1) an argument that descriptive representation is substantively and symbolically important, even necessary, for the descriptively represented group and for the polity as a whole; 2) an argument that a group's lower than proportional representation in a representative assembly has been caused by some form of inappropriate discrimination against that group; and 3) an argument that quotas are the most effective way in practice to achieve descriptive representation. The first two claims are normative. In the case of women, in all countries of the world the evidence supporting both claims is strong. The third claim is prudential. It depends on several factors, including the likelihood of women achieving appropriate representation by other means and the institutional plausibility of achieving a satisfactory quota system in a given country in a given historical moment.

The case against quotas can be both institutional and cultural. Some specific institutional forms for producing quotas have highly problematic side effects. Here I focus not on the institutional problems but on the tendency of quotas to promote cultural beliefs in “essentialism”—the conviction that the individuals represented through quotas have some essential traits that help define them and that render them unable to be represented adequately by those without such traits. Essentialist beliefs reinforce stereotypes, trap the individuals in the group in the images traditionally held of the group, make it hard for those individuals to treat their identities flexibly and performatively, de-emphasize lines of division within groups to the advantage of dominant groups within the group, and harden lines of division between groups. The argument that men cannot represent women, for example, suggests that women cannot represent men. The argument that only women can represent women suggests that any woman can represent all women. The probability, amounting currently to close to certainty, that quotas will strengthen these essentialist beliefs provides a strong argument against them. It is not an argument that outweighs positive arguments in all specific historical situations. But if a polity institutes quotas, proponents of the quotas should strive to frame them in nonessentialist ways and mount an independent struggle against essentialist tendencies more broadly.

I. The Case for Quotas

Democratic theory does not mandate that every group, no matter how defined, should be represented in proportion to its numbers in the population. In 1969, when President Richard Nixon tried to appoint Harrold Carswell to the Supreme Court and opponents argued that Carswell was “mediocre” in his legal skills, Senator Roman Hruska (R-NE) rejoined, “Even if he is mediocre, there are a lot of mediocre judges and people and lawyers. They're entitled to a little representation, aren't they?” The comment provoked immediate ridicule because of the almost universal assumption that democratic norms allow, and the public wants, to choose representatives with greater than average skills.1

In a Google search for “Harrold Carswell” in June 2005, the first two hits referenced this line, suggesting that Carswell will live in history primarily as an example of mediocrity.

Along the same lines, the political theorist Roland Pennock proclaimed a decade later, “no one would argue that morons should be represented by morons” (Pennock 1979, 314, based on Griffiths and Wollheim 1960, 190). James Morone and Theodore Marmor also scoffed that “[c]ommon sense rebels against representing redheads or left-handers” (Morone and Marmor 1981, 437). More recently, Will Kymlicka concluded summarily, “the general idea of mirror [descriptive] representation is untenable” (1995, 139). Iris Marion Young concluded that “a relation of identity or similarity with constituents says nothing about what the representative does” (1997, 354).

In general, whenever interests conflict, the theory of aggregative (or “adversary”) democracy mandates the representation of those interests in proportion to the number of interest bearers in the population (Mansbridge 1980). In addition, whenever different perspectives might significantly improve a deliberation, the theory of deliberative democracy mandates the representation of those perspectives (Mansbridge 1999). These criteria answer what Anne Phillips (1995, 45) has called the “slippery slope” objection that no guiding principles can distinguish which groups have weaker and which stronger democratic claims for representation. Yet neither the democratic mandate for equal representation of interests in moments of conflict nor the mandate for representation of relevant perspectives in deliberation requires representation by actual members of the represented group. Members of a group need such descriptive representation only when their interests or perspectives cannot adequately be represented by individuals who are not themselves members of the group. The key question is when and why those interests and perspectives cannot adequately be represented by others.

In at least two circumstances, the interests and perspectives of members of a group cannot be adequately represented by others: 1) when representatives who are members of a group tend to respond to group-relevant issues with greater concern than nonmembers, and 2) when representatives who themselves are members of a group can communicate better among themselves, with other representatives and with constituents from that group. This capacity for better communication is accentuated a) when issues and even interests are uncrystallized, so that representatives who are members of a group have far more information about the relevant aspects of an issue than nonmembers; b) when a history of mistrust impedes communication and effective delegation between members of a subordinate group and representatives from a dominant group; and c) when the physical presence of representatives from the group induces other representatives to make greater efforts to understand that group's interests and give those interests more salience. In addition to promoting substantive representation in these contexts, descriptive representatives also play a powerful symbolic legitimating function by making the statement to the entire citizenry, including its other representatives, that members of that group are capable of ruling.2

This list of the circumstances in which descriptive representation enhances substantive representation expands on those referenced in Mansbridge 1999. For other positive symbolic effects of descriptive representation, see Mansbridge 1999.

These considerations apply equally to both powerful and marginalized groups, but powerful groups are far less likely in practice to experience proportional shortfalls in descriptive representation. When they occasionally have this experience, they also have a greater capacity to counterbalance in other ways the negative effects of not being descriptively represented.

Concern

We know that in the cases of gender and race, legislators who themselves are members of a group respond to issues affecting that group with greater concern than do nonmembers. Female legislators in almost every measured representative body, from the U.S. House of Representatives to the Indian panchayats, give more attention than do male legislators to issues, such as education and women's rights, that also typically concern female constituents more than they do men. Sue Thomas (1994), Michele Swers (2002), and others have shown that the differences between male and female legislators are most likely to appear on issues that absorb a great deal of the legislator's time and energy, such as the choice of committees to join and legislation to sponsor. By contrast, on simple roll-call votes, the representative's party has a greater effect than the representative's gender. Descriptive representatives from groups particularly affected by particular issues tend to care more about those issues, put their time in on them, and struggle to bring them to the legislative fore.

Barry Burden's (2005) recent research also demonstrates the importance of the personal experiences of the representatives. In the U.S. House of Representatives, he reveals, members who smoke are more likely than the nonsmokers to vote against, speak against, and sponsor bills against tobacco control measures. Members with school-age children are more likely than the others to be active on school choice issues. Members with children in public school are more likely to vote against voucher programs that help parents transfer their children out of public school. Members who themselves are Evangelical Protestants or Catholics are more likely to support faith-based initiatives and more likely to oppose stem cell research. The personal experiences of these representatives alert them to the needs of constituents like themselves and impel them to act on those needs, even when the numbers of individuals with those needs within their own districts do not make their pursuit of these issues electorally rational. Descriptive representation of this sort is particularly important on uncrystallized issues—that is, issues that have not found a salient place on the political agenda, so that political parties have not taken positions on those issues and politicians have not run for office on them (Mansbridge 1999).

Communication

In situations of relatively uncrystallized interests, marginalized groups in particular often need a critical legislative mass of descriptive representatives in order for the representatives to consult among themselves, as well as with their constituents, to try to understand what their descriptive constituents most need. In many such cases these representatives are acting as surrogate representatives for descriptive constituents outside their districts as well as within them (Mansbridge 1999).

Descriptive representatives have, moreover, several advantages in communicating with other legislators. In listening, they can respond flexibly, drawing not just from what they have heard from their constituents, but also from what they know from their own lives. In speaking, they can call up anecdotes from personal experience to describe how a piece of legislation may affect their group. They can speak vividly with facts and emotion drawn from their experience. They can speak with authenticity and be believed. They can also call on the relationships they have developed with other legislators for the empathy that the others might need to understand the descriptive representative's position. Even when the descriptive legislator is silent, his or her mere physical presence reminds the other legislators of the perspectives and interests of the group of which he or she is a descriptive member. In deliberative settings, many members of marginalized groups have had the experience of seeing members of dominant groups cut themselves short as they are about to say something demeaning to members of that group and substitute something more sensitive. This behavior is not simply a matter of surface conformity. The presence of someone who represents a group, particularly descriptively, physically reminds others of that group's interests, making it more likely that those claims and perspectives will be taken into account.3

The effects of descriptive presence may go even deeper. Laboratory experiments reveal that the mere presence of a Black experimenter reduced the level of the subjects' implicit (unconscious) negative associations with Blacks (Banaji, Bazerman, and Chugh 2003).

Descriptive representatives also have advantages in overcoming the mistrust that potentially impedes any interaction between representative and constituent. When the constituents are members of marginalized groups, with a justified history of mistrust of the dominant groups, the need to help overcome constituents' distrust of their representatives is even greater (Mansbridge 1999; Williams 1998). In the early years of the global violence-against-women movement, for example, women of the global “North” (broadly, the developed nations) primarily directed that movement. Women from the “South” (the less-developed nations) distrusted the leadership and failed to coordinate their efforts, with the result that the global movement had relatively little impact. After donors and other organizers realized that the women of the South needed funding to create their own autonomous organizations from which leadership could arise, descriptive representatives from the South began to lead and participate extensively in the global conferences. These representatives brought legitimacy, credibility, and open lines of communication with their “constituents” in their own countries, helping the movement become a model of successful transnational activism.4

Weldon (forthcoming) analyzes this dynamic, adding that the movement's commitment to a consensual process also gave the potentially marginalized groups of the South a protected voice. Note that “selectoral” (rather than electoral) representation in movements like this poses other legitimacy problems

Descriptive representation thus has positive effects on outcomes important to a group through at least the two mechanisms of investing more heavily through greater concern and communicating interests more thoroughly. The communication function is particularly important when the group's interests are uncrystallized, occur in settings of historical mistrust, or might be overlooked without the representative's physical presence.

The mechanisms of concern and communication apply whether or not a group is marginal in the larger society. White Americans, for example, seem just as alienated from Black representatives as Blacks from White representatives, perhaps more so. Claudine Gay shows that White constituents with Black representatives are less likely than similar constituents with White representatives to turn out to vote, to think that their representative would be helpful with a problem, to remember anything that the representative had done for the district, to rate the representative high on a feeling thermometer, or to approve of the way the representative had been handling his or her job (Gay 2001, 2002). Black constituents with White representatives when compared to Blacks with Black representatives reveal the same patterns, but those patterns, if anything, are less strong. Particularly in situations of historical communicative distrust, it seems that everyone wants descriptive representation, including and perhaps especially the members of dominant groups. In practice, however, the dominant groups usually get what they want.

Discrimination

The case for quotas rests not only on the substantive benefits of descriptive representation through concern and communication, but also, quite solidly, on discrimination. Regarding women, the evidence suggests strongly that both surface and structural forms of discrimination currently impede the proportional descriptive representation that one would otherwise expect. In surface discrimination, for example, members of the polity sometimes vote for a man rather than a woman to represent them, even when the woman and man are equally qualified.5

Surface discrimination has been declining in the established democracies. In 1937, only 33% of a representative sample of the U.S. population responded “yes” when asked, “If your party nominated a woman for President, would you vote for her if she were qualified for the job?”; by the late 1950s, 52% answered yes; by 1971, 66%; by 1982, 83%; by 1994, 89%; and by 1999, 98% (NORC General Social Survey and earlier surveys referenced therein). In addition, by now, laboratory ex-periments in the United States using male and female names for fictitious candidates tend to show no difference in support overall, at least among college students, although some differences in estimates of competence in specific policy areas persist along traditional lines (e.g., women are seen as better on nurturing issues, men on military ones). Matland and King 2002 summarize these studies, while reporting on their own finding that Republican voters are less likely to vote for a similarly described female candidate than for a man (probably because female Republicans tend in practice to be more liberal than male ones). In less-developed countries, surface discrimination still plays a major role, both in reported attitudes (Inglehart and Norris 2003) and in the experience of women trying to run for office.

In structural discrimination, women are expected to be the primary caregivers for children and the elderly, and are socialized not to see themselves as competitors in politics. They are therefore less likely to enter the competition for office (Lawless and Fox 2005). The case for past discrimination rests on an easily documented history in which women have been legally denied the vote, the capacity to serve on juries, and the like (see Williams 1998, 15–18).

II. The Case Against Quotas

Institutions

Whether or not quotas are the most effective mechanisms for delivering descriptive representation depends on a host of factors, including the degree to which the deck is already stacked institutionally against the descriptive group. We have known for a decade or more that women do better in proportional representation systems and perhaps also slightly better in multimember districts in general (Welch and Studlar 1990), in part because parties putting up lists of more than one candidate feel more compelled to put women on the list, and in part because multimember districts encourage policy-oriented candidates, as opposed to candidates who work specifically for benefits from their districts, and the middle-class women who run for office are more likely to have a policy orientation.6

By contrast, in the U.S., working-class and black candidates often do slightly better in single-member districts (Welch 1990), perhaps because such districts are more often homogeneously working class or black and the local communicative advantage of these candidates helps them to run on a platform of bringing home local benefits.

When list systems and multimember constituencies already exist, therefore, it is easier to insert women into the political process without imposing specific quotas.

The effectiveness of quotas also depends on the degree to which a particular culture at a particular moment in history can accept them as legitimate. In France, where a universalist political culture prevents the state from even collecting data on the ethnicity, race, or religion of individuals, women promoting parité had to resort to the philosophical ploy of arguing that women were not a “group” but, instead, half of the “universal” human race (Agacinski 2003; Gaspard 1994; Scott 1997). In the United States, the very word “quota” implies the negation of merit, individual worth, and fair competition, as well as the intervention of the state in individual freedom.7

An article advocating cumulative voting—a procedure sufficiently compatible with existing values in the United States as to have been in place in Illinois for many years—was, in the hands of opponents, enough for them successfully to label Lani Guinier a “quota queen” and cause President Bill Clinton to withdraw his recommendation of Guinier for Attorney General (Carter 1994).

In Africa, by contrast, where the idea of representing different territorial, ethnic, and linguistic groups descriptively has long been part of the accepted political culture, representing women descriptively through the allocation of denominated seats does not fall far from the existing norms.

The central institutional challenge is to devise a system of quotas with few unintended negative side effects. The opposition of the feminist journal Manushi to the proposed system of rotating seats reserved for women in India, for example, is grounded almost entirely in the unwanted institutional side effects of that system.8

Manushi, issues 96, 97, and especially 116. See also Baldez forthcoming.

Essentialism

The broader challenge, which cannot in practice fully be met, is to institute some form of quota system without encouraging essentialism. It is a commonplace these days to point out that we all have multiple identities, and that those identities are socially constructed, flexible, performative, and in many cases even, with some investment, discardable. The movements toward bisexuality, transgendered identities, multiraciality, and multiple nationality are all helping these forms of identity move down the road that religion in the United States has already traveled, from ascription to choice. In this more flexible world, quotas work on the side of rigidity.

The belief that there is some “essence” of womanness or femininity, blackness or negritude, that members of the group have and that is not accessible to an outsider, reinforces group stereotypes from both without and within. Social psychologists have shown us that human beings, perhaps innately, judge members of an in-group as more alike than they in fact are, and members of an out-group as more different from the in-group than they in fact are (Turner 1987). In order to process reality accurately, we need to combat these cognitive biases. Cultural conventions that reinforce group differences make this struggle harder. In the case of gender, beliefs in a gendered division of labor that make some tasks taboo for either males or females, jokes and traditions about female and male traits, divisions of the universe into the male and female principles, and even the linguistic division into male and female nouns all suggest some essence of maleness and femaleness. Implicit association tests reveal some of these associations, which by and large even the most politically sensitive continue to hold.9

For the “gratuitous gendering” of labor, parts of speech, and essences, see Mansbridge 1993. For implicit association tests, see Banaji 2001 and the website 〈https://implicit.harvard.edu/implicit〉, which allows one easily to test one's own unconscious associations.

Human cognitive capacities tend automatically toward some forms of essentializing. Thus, terms in the language tend to call to mind the dominant or majority variant of those terms. When we envision “table,” we envision something brown, thus marginalizing the green, white, and yellow tables of this world. When we envision “chair,” we envision something with four legs, and so forth (Bartlett 1990). Thus, “woman” does not automatically conjure up, in today's world, “woman with disabilities” or “lesbian.” Efforts to remember the heterogeneity within any human group must war against the way the brain itself generalizes.

Quotas can undermine some essentialist beliefs by, for example, bringing sufficient women of different kinds into political arenas to allow both men and women to experience them as capable leaders, thus undermining the belief that women qua women are incapable of rule. At the same time, they are likely to reinforce essentialist beliefs by suggesting that for essentialist reasons, only women can represent women (and therefore that women cannot represent men). Quotas that work by forcing members of a descriptive group to vote only for members of that group and not for members of other groups rigidify group lines in the most dramatic way.

It is particularly hard to avoid essentialist arguments when making the case that descriptive representation is substantively important for the descriptively represented group and for the polity as a whole. It is factually the case that in legislatures all over the world, women are more likely to take the lead on two kinds of “women's” issues: issues associated with women's traditional responsibility for the family, such as education and health, and women's rights issues.10

In the U.S. Congress, women are also more likely to take the lead on issues on which noticeable differences appear on average in public opinion polls between men and women. For this category, see Strauss 1998, with thanks to Patricia Connelly.

But explaining these differences, correctly, as the result of modal differences in experience leading to modal differences in interests and preferences skirts dangerously close to postulating essential differences in interests and perspectives.

The connection is not necessary. It is perfectly possible to talk about and act on the basis of experiences that many women have shared in one form or another without assuming that those experiences represent every woman's experience or that the forms in which a “shared” experience comes are the same for everyone. Nor need we collapse “experiences” in the plural into a singular “essence.”11

For alternate ways of conceiving of groups, see Young 1994 on “seriality,” Ferguson 2003 on family resemblances, and Zerilli 1998.

When we move from arguing that descriptive representation by gender will be good for women to the argument that such representation will be good for the polity, we come even closer to the essentialist danger. It is possible, for example, that on the margin and on average, women in many polities may be slightly more honest, caring, and cooperative than men. Laboratory and field experiments cannot usually detect significant differences between men and women on these and other stereotypically gendered traits. But interacting as a group may accentuate group differences, perhaps in part by activating group stereotypes, which women hold as well as men.12

In one experiment, although no significant differences appeared at an individual level, groups with a high proportion of women deliberating to a conclusion about how much to give the disadvantaged were more generous than groups with a high proportion of men (Mendelberg and Karpowitz forthcoming). See also Sunstein 2002 and others on group polarization.

In addition, relatively large differences between men and women do appear, on average, in some areas, such as physical aggression and attitudes toward war (Goldstein 2001; Maccoby 1998; Maccoby and Jacklin 1978). Building on the existing differences and influenced by the exaggeration of stereotypes to which the human brain tends, advocates often claim that the polity would benefit from more women in legislatures because of women's greater honesty, care, cooperation, or concern for peace. Some imply that these gender differences are innate.

The temptation to make such claims is great, both because some activists believe in the political significance of innate differences and because such claims have some weight with the general public. The campaign for suffrage in the United States, for example, did not acquire sufficient majorities to pass the required constitutional amendment until the proponents' argument had shifted from one based on women's rights to one based on women's contributions, as women, to the polity (Cott 1987; Kraditor 1965; Marilley 1996). Such claims slide easily from those based on different experiences, such as modally different experiences of parenting, of housekeeping, and generally of being “outside the beltway,” to those based on innate differences. Only a constant, explicit stress on experiences rather than innateness can mitigate this tendency. Optimally, the stress on experience should be coupled with some knowledge of the difficulty of finding any gender differences in the laboratory in most settings once most conditions are made the same (see, e.g., Hyde 1990).

Quotas tend to reinforce the existing human cognitive tendencies to see the members of the group as more similar than they are and more different from members of other groups. Membership in groups is already “something that you are ‘born’ into and that constitutes you as being who you are and is ascribed to you by others in a way that makes it involuntary from your point of view” (Gould 1996, 182; see also Williams 1998, 6). Based by their nature on group membership, quotas are also almost certain to reinforce the social and personal meanings of that membership, impeding from both outside and inside a flexible, performative relation with one's multiple identities.

The best way to fight the reification of essentialism that quotas inevitably suggest is to reinforce constantly the ways in which the great differences in existing systems of representation derive from historical and structural biases. For African Americans, for example, the near exclusion of black veterans in the South from business and home loans under the GI Bill by making applications local, a deliberate policy insisted upon for racial reasons by white southern senators, resulted in greatly lower rates of home ownership and, therefore, family wealth for blacks (Katznelson 2005). Yet the instability of politics as a career makes some family wealth a major asset in running for political office. For women, the currently near-universal assumption that women will take primary responsibility for child care makes an early political career difficult. Yet politicians who begin their careers late are less likely to reach the highest political office, and those who begin a political career after their children have gone to college are even less likely to advance far in the system. These deep structural biases, in addition to the unconscious assumptions that prejudice the voters, the party leaders who might ask someone to run for office, and the potential candidates themselves, work to make it unlikely that without some institutional intervention, African Americans and women will be represented in law-making bodies in proportion to their numbers in the population. Making arguments for quotas explicitly and only on the basis of these historical and structural biases helps, to some degree, to counter essentialist arguments.

We also know that the elected women who will represent “women” descriptively will not represent proportionally the full diversity of women in the polity. Iris Marion Young (1997) points out that whenever any one individual represents many, the process necessarily suppresses the differences among the many. Descriptive representation does not avoid this problem. The diversity within historically disadvantaged groups is no less than in any other groups; all forms of representation here too result in the suppression of difference. Thus, Young warns that any scheme of representation must rest on “acknowledging and affirming that there is a difference, a separation, between the representative and the constituents,” and she stresses the importance of authorization and accountability.13

Young uses this analysis to conclude that constituents should not seek descriptive representation, but rather focus on accountability and authorization (Young 2000, 129).

If biases in proportionality are corrected only for gender, the voting citizens in the polity will not be represented proportionally in their diversity (leaving aside the issues of representing noncitizens in the polity, nonresidents who are affected by the policies, and future generations). The highly educated, middle-class, dominant-ethnicity women who are most likely to be elected in systems engineered to represent only gender proportionally will represent “women's” issues primarily through their own experiences, which, although in some ways like, will also be unlike those of women who do not share their class or ethnic/racial background.14

See, e.g. re suffrage, Davis 1983 and Dovi 2002.

The writings and activism of women of color make it clear that specific problems and perspectives arise from specific intersections of social advantage and disadvantage, structure and culture (Collins 1990; Combahee River Collective [1977] 1983; Crenshaw 1991; Harris 1990). Yet separate quotas for women and for African Americans could result in the election of no female African Americans.15

In the United States in the 1980s and 1990s, for example, African-American men were proportionally represented descriptively in state legislatures. The under-representation of African Americans was concentrated on African-American women (Darcy, Hadley, and Kirksey 1993).

Some of these problems could be addressed through new democratic mechanisms. First, citizen assemblies drawn randomly from the population can serve both as deliberative forums and as checks on the inevitable misrepresentation in elected assemblies.16

For example, The British Columbia Citizens' Assembly (see 〈http://www.citizensassembly.bc.ca/public/extra/Whatis.xml〉 and Warren forthcoming). Such randomly selected assemblies are usually still not fully representative, because although the original selection is random and citizens are paid for their attendance, retired middle-class and upper-working-class people are more likely to accept the assignment than the working-age and poor. Moreover, the participants in citizen assemblies have no reporting and formal accountability relationship with constituents. What relationship there is works only through the feelings among the represented that the randomly selected representatives are “like” them, and the voluntary feelings among the representatives that they have an obligation to act both for the public good and for the good of any group they happen to represent descriptively or through conviction. These feelings on both sides are, however, neither trivial nor irrelevant to the relationship. Such feelings play a neglected but important role even in the elected representative/constituent interaction, where a representative's actual communication with all constituents is perforce relatively meager.

Second, political parties creating gender quotas through party lists could try consciously to represent the most relevant heterogeneities within gender.17

It is not clear, however, what incentives parties would have to diversify their lists in this way.

Third, systems could perhaps be organized to choose primarily descriptive representatives who have strong mutual relations with the most disadvantaged subgroups within their descriptive group (Dovi 2002). In current practice, only political self-organizing among marginalized groups can provide forums for the deliberation and articulation of interests, from which members at various intersections in those groups can enter the representative process. Given the relatively meager resources of many marginalized groups, that self-organizing must usually be funded from outside the group.18

See Weldon 2006. The best descriptive representatives will themselves contribute to organizing their own communities, but if they are concerned with reelection, they will probably not focus on the hardest segments to organize.

Although no practically imaginable reform will eliminate this problem, those involved in electoral reform should at least try to prevent the marginalization of subgroups from being worse under quotas than under nonquota systems.

In short, with the great costs in increasing essentialism that are likely to attend any move toward quotas, such a move should be combined with significant conscious efforts to point out that the need for quotas derives from biases in current systems produced by historical and structural discrimination, to anchor arguments for descriptive representation in differences in experience rather than innate characteristics, to find ways to represent institutionally the diversity within the group, to remind both members of the group and others of the great dissimilarities within the group and the great similarities with others outside the group, and to stress the fluidity of identity.

Quotas inherently rigidify identities. If adopted, they should be kept as flexible as possible, being instituted preferably at the party level rather than at the legislative level, by voluntary adoption rather than by legislation, and by legislation rather than by constitutional mandate. They should be used and portrayed as a practical and perhaps temporary response to centuries of discrimination, rather than as an eternally necessary recognition of essential differences. Even when quotas are the best practical route for achieving substantive gains, we should go forward with them conscious of the dangers and ringing them with safeguards. Subordinate groups have much to lose from practices that reinscribe difference.

References

Agacinski, Sylviane. 2003. “The Turning Point of Feminism: Against the Effacement of Women.” Trans. Mary Schwartz. In Beyond French Feminisms: Debates of Women, Politics, and Culture in France, 1981–2001, ed. Roger Célestin, Eliane DalMolin, and Isabelle de Courtivron. London: Palgrave, 17–22.
Baldez, Lisa. Forthcoming. “The Pros and Cons of Gender Quota Laws: What Happens When You Kick Men Out and Let Women In?” Politics & Gender 2 (1).
Banaji, Mahzarin R. 2001. “Implicit Attitudes Can be Measured.” In The Nature of Remembering, ed. H. L. Roediqer, I. N. Nairne, and A. M. Suprenant. Washington DC: American Psychological Association, 117–50.
Banaji, Mahzarin, Max H. Bazerman, and Dolly Chugh. 2003. “How (Un)ethical Are You?” Harvard Business Review 81 (12): 56–64.Google Scholar
Bartlett, Katharine T. 1990. “Feminist Legal Methods,” Harvard Law Review 103 (4): 829–88.Google Scholar
Burden, Barry C. 2005. Personal Roots of Representation. Unpublished book manuscript ().
Carter, Stephen L. 1994. “Forward.” In Lani Guinier, The Tyranny of the Majority: Fundamental Fairness in Representative Democracy. New York: Free Press.
Collins, Patricia Hill. 1990. Black Feminist Thought. London: Allen and Unwin.
Combahee River Collective. [1977] 1983. “The Combahee River Collective Statement.” In Home Girls: A Black Feminist Anthology, ed. Barbara Smith. New York: Women of Color/Kitchen Table Press, 272–82.
Cott, Nancy. 1987. The Grounding of Modern Feminism. New Haven, CT: Yale University Press.
Crenshaw, Kimberlé. 1991. “‘Mapping the Margins’: Intersectionality, Identity Politics and Violence Against Women.” Stanford Law Review 43 (6): 1241–99.Google Scholar
Darcy, Robert, Charles D. Hadley, and Jason F. Kirksey. 1993. “Election Systems and the Representation of Black Women in American State Legislatures.” Women & Politics 13 (2): 73–89.Google Scholar
Davis, Angela. 1983. Women, Race, and Class. New York: Vintage Books.
Dovi, Suzanne. 2002. “Preferable Descriptive Representatives: Will Just Every Woman, Black, and Latino Do?” American Political Science Review 96 (4): 729–43.Google Scholar
Ferguson, Michaele. 2003. “Sharing and Sharing Alike: Political Unity in Deeply Diverse Democracies.” Ph.D. diss. Harvard University.
Gaspard, Françoise. 1994. “De la parité: genèse d'un concept, naissance d'un mouvement.” Nouvelles Questions Féministes 15 (4): 42.Google Scholar
Gay, Claudine. 2001. “The Effect of Black Congressional Representation on Political Participation.” American Political Science Review 95 (3): 589–602.Google Scholar
Gay, Claudine. 2002. “Spirals of Trust? The Effect of Descriptive Representation on the Relationship Between Citizens and Their Government.” American Journal of Political Science 46 (6): 717–33.Google Scholar
Goldstein, Joshua S. 2001. War and Gender. Cambridge: Cambridge University Press.
Gould, Carol. 1996. “Diversity and Democracy: Representing Differences.” In Democracy and Difference: Contesting the Boundaries of the Political, ed. Seyla Benhabib. Princeton, NJ: Princeton University Press.
Griffiths, A. Phillips, and Richard Wollheim. 1960. “How Can One Person Represent Another?” Aristotelian Society, Supplement 34: 182–208.Google Scholar
Harris, Angela. 1990. “Race and Essentialism in Legal Theory.” Stanford Law Review 42 (3): 581–616.Google Scholar
Hyde, Janet Shibley. 1990. “Meta-Analysis and the Psychology of Gender Differences.” Signs 16 (1): 5–73.Google Scholar
Inglehart, Ronald, and Pippa Norris. 2003. Rising Tide: Gender Equality and Cultural Change Around the World. Cambridge: Cambridge University Press, 2003.
Katznelson, Ira. 2005. When Affirmative Action Was White. New York: W. W. Norton.
Kraditor, Aileen S. 1965. The Ideas of the Woman Suffrage Movement, 1890–1920. New York: Columbia University Press.
Kymlicka, Will. 1995. Multicultural Citizenship. Oxford: Oxford University Press.
Lawless, Jennifer L., and Richard L. Fox. 2005. It Takes a Candidate: Why Women Don't Run for Political Office. Cambridge: Cambridge University Press.
Maccoby, Eleanor E. 1998. The Two Sexes. Cambridge, MA: Harvard University Press.
Maccoby, Eleanor E., and Carol Nagy Jacklin. 1978. The Psychology of Sex Differences. Stanford, CA: Stanford University Press.
Mansbridge, Jane. 1980. Beyond Adversary Democracy. New York: Basic Books.
Mansbridge, Jane. 1993. “Feminism and Democratic Community.” In Democratic Community: NOMOS XXXV, ed. John W. Chapman and Ian Shapiro. New York: New York University Press, 339–95.
Mansbridge, Jane. 1999. “Should Blacks Represent Blacks and Women Represent Women? A Contingent ‘Yes’” Journal of Politics 61 (3): 628–57.Google Scholar
Marilley, Suzanne M. 1996. Woman Suffrage and the Origins of Liberal Feminism in the United States 1820–1920. Cambridge, MA: Harvard University Press.
Matland, Richard E., and David C. King. 2002. “Women as Candidates in Congressional Elections.” In Women Transforming Congress, ed. Cindy S. Rosenthal. Norman: University of Oklahoma Press, 119–45.
Mendelberg, Tali, and Christopher Karpowitz. Forthcoming. “How People Deliberate About Justice.” In Can the People Decide? Theory and Empirical Research on Democratic Deliberation, ed. Shawn W. Rosenberg.
Morone, James A., and Theodore R. Marmor. 1981. “Representing Consumer Institutions: The Case of American Health Planning.” Ethics 91 (3): 431–50.Google Scholar
National Opinion Research Center. General Social Surveys, 1972–1982: Cumulative Codebook. Chicago: National Opinion Research Center. http://webapp.icpsr.umich.edu/GSS/.
Pennock, J. Roland. 1979. Democratic Political Theory. Princeton, NJ: Princeton University Press.
Phillips, Anne. 1995. The Politics of Presence. Oxford: Oxford University Press.
Scott, Joan W. 1997. “‘La Querelle des Femmes’ in the Late Twentieth Century.” New Left Review 226: 3–19.Google Scholar
Strauss, Julie Etta. 1998. “Women in Congress: The Difference They Make.” Ph.D. diss. Northwestern University.
Sunstein, Cass R. 2002. “The Law of Group Polarization.” Journal of Political Philosophy 10 (2): 175–95.Google Scholar
Swers, Michele L. 2002. The Difference Women Make: The Policy Impact of Women in Congress. Chicago: University of Chicago Press.
Thomas, Sue. 1994. How Women Legislate. New York: Oxford University Press.
Turner, John C. 1987. Rediscovering the Social Group: A Self-Categorization Theory. Oxford: Blackwell.
Welch, Susan. 1990. “The Impact of At-Large Elections on the Representation of Blacks and Hispanics.” Journal of Politics 52 (4): 1050–76.Google Scholar
Welch, Susan, and Donley T. Studlar. 1990. “Multi-member Districts and the Representation of Women: Evidence from Britain and the United States.” Journal of Politics 52 (2): 391–412.Google Scholar
Weldon, S. Laurel. 2006. “Inclusion, Solidarity and Social Movements: The Global Movement Against Gender Violence.” Perspectives on Politics 4 (March): 55–74.Google Scholar
Williams, Melissa. 1998. Voice, Trust, and Memory: Marginalized Groups and the Failings of Liberal Representation. Princeton, NJ: Princeton University Press.
Young, Iris Marion. 1994. “Gender as Seriality: Thinking About Women as a Social Collective.” Signs 19 (3): 713–38.Google Scholar
Young, Iris Marion. 1997. “Deferring Group Representation.” In Nomos XXXIX: Ethnicity and Group Rights, ed. Ian Shapiro and Will Kymlicka. New York: New York University, 349–76.
Young, Iris Marion. 2000. Inclusion and Democracy. Oxford: Oxford University Press.
Zerilli, Linda. 1998. “Doing Without Knowing: Feminism's Politics of the Ordinary.” Political Theory 24 (4): 435–58.Google Scholar