Cambridge
UNIVERSITY PRESS
apsa
Preferable Descriptive Representatives: Will Just Any Woman, Black, or Latino Do? Author(s): Suzanne Dovi
Source: The American Political Science Review, Vol. 96, No. 4 (Dec, 2002), pp. 729-743 Published by: American Political Science Association Stable URL: http://www.Jstor.org/stable/3117507 Accessed: 27-12-2015 07:31 UTC
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/ info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
	(Si
m	4
mi	
American Political Science Association and Cambridge University Press are collaborating with JSTOR to digitize, preserve and extend access to The American Political Science Review.
STOR
http://www.jstor.org
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
American Political Science Review
Vol. 96, No. 4   December 2002
Preferable Descriptive Representatives: Will Just Any Woman, Black, or Latino Do?
SUZANNE DOVI University of Arizona
A body of theoretical literature has developed that explains why historically disadvantaged groups /\ should be represented by members of those groups. Such representatives are commonly referred \. to as descriptive representatives. This literature has also endorsed various institutional reforms aimed at increasing the number of descriptive representatives, e.g., party list quotas, racial districting, and proportional representation. However, this literature does not articulate criteria that should guide the selection of descriptive representatives to serve in these institutional positions. Indeed, some thinkers claim that such criteria cannot, or at least should not, be articulated. I argue that some descriptive representatives are preferable to others and that criteria for selecting preferable descriptive representatives can, and should, be articulated. Moreover, I recommend one such criterion: Preferable descriptive representatives possess strong mutual relationships with dispossessed subgroups of historically disadvantaged groups.
Democratic political institutions are often evaluated by the gender, ethnicity, and race of elected representatives (e.g., Guinier 1994; Paolino 1995). Implicit in these evaluations is the assumption that democratic political institutions that lack any representatives from historically disadvantaged groups are unjust. Moreover, these evaluations often assume that an increase in the number of representatives from historically disadvantaged groups can contribute to the substantive representation of those groups (e.g., Thomas 1991). For example, 1998 was declared the "Year of the Woman" in the United States because in that year the number of women in the House leaped from 28 to 48, and that in the Senate from two to six. This method of evaluating democratic institutions often assumes that the more women, Blacks, and Latinos, the better for democratic institutions.1
These assumptions justify the political practice of setting aside certain political and institutional positions for members of historically disadvantaged groups. These positions are specifically designed to increase the number of representatives from historically disadvantaged groups—that is, the number of what I call "descriptive representatives." Contemporary political theorists have directly and indirectly supported these assumptions by offering several explanations for why political representatives for a historically disadvantaged group should come from that group (e.g., Mansbridge 1999; Phillips 1995; Sapiro 1981). In their terminology, they have advanced reasons to think that "descriptive representation," "group representation," "the politics of presence," or "self-representation" is important.2
Suzanne Dovi is Assistant Professor, Department of Political Science, University of Arizona, 315 Social Sciences Tuscon AZ 85721 (sdovi@u.arizona.edu).
The author would like to thank Sigal Ben-Porath, Bernard Harcourt, Kristen Hessler, Jacob Levy, Melissa Williams and most of all, Houston Smit for their invaluable conversations, comments, and support. Thanks to Carrie Brennan, John Garcia, Cindy Holder, Patchen Markell, William Mishler, Jeff Spinner-Halev, John Schwarz, and Sue Thomas for reading and commenting on drafts of this paper. The author would, finally, like to thank the anonymous reviewers for their challenging and insightful comments.
1 For an opposing view, see Cameron, Epstein, and O'Halloran 1996.
2 I use these terms interchangeably in the rest of this paper.
Although the reasons they advance differ significantly, these theorists sound a common theme: To be fully democratic, a society that has denied full political membership to certain groups must be strongly committed to including those groups in its political life. Such a commitment, at least in many circumstances, requires society to take active steps to increase the number of descriptive representatives. On these grounds, these theorists endorse various institutional reforms such as party list quotas, caucuses, racial districting, and schemes for proportional representation. But these theorists have said remarkably little about the criteria that should guide democratic citizens in their choice of descriptive representatives.3 The emphasis of this literature so far has been on establishing the need for the presence of some descriptive representatives, not on investigating criteria for identifying preferable descriptive representatives.
Which members of historically disadvantaged groups are preferable representatives for those groups? My primary aim is to argue for the need for criteria that will help answer this question. I take the value of having descriptive representatives in public positions as a given.4 I advance existing discussions of group representation by explaining how democratic citizens should choose among various possible descriptive representatives. Toward this end, I propose one criterion for identifying preferable descriptive representatives: Preferable descriptive representatives have strong mutual relationships with dispossessed subgroups. I purposely set aside the question of whether descriptive representatives should be evaluated by the same or different criteria as other representatives in liberal democracies.51 focus on a different question: Are there any criteria for guiding
3 I use "democratic citizens" to refer to all citizens—that is, to both citizens who are members of historically disadvantaged groups and citizens who possess more privileged social locations.
4 The need for institutional reforms aimed at increasing the number of descriptive representatives for a certain historically disadvantaged group may be temporary. These reforms may be dropped, if and when the society has advanced to the point where a historically disadvantaged group is no longer politically marginalized.
5 I agree with Iris Marion Young's position (2000) that the difficulty in choosing descriptive representatives—what she calls the problem of one person representing the many—is a problem for all
729
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
Preferable Descriptive Representatives
December 2002
the appointment, nomination, or election of members of historically disadvantaged groups to positions that were created to increase the diversity of actors in the political arena and thereby the substantive representation of such groups? In other words, are there any principled reasons for preferring one descriptive representative to another? I offer my criterion to provide guidance for such decisions and to be more explicit about the political commitments that underlie a politics of presence. To put my position boldly, a commitment to a politics of presence would be more likely to support robust democratic relations if descriptive representatives were selected on the basis of their mutual relationships with dispossessed subgroups.
CONSTRUCTING A POLITICS OF PRESENCE: BUILDING AROUND A TENSION
Hannah Pitkin's classic work The Concept of Representation has set the terms of the debate over descriptive representation. For Pitkin, descriptive representation concerns what representatives "look like," rather than what they "do." For this reason, Pitkin (1967, 89) proclaimed that arguments for descriptive representation have "no room for representation as accountability." This line of reasoning—that a politics of presence is somehow incompatible with accountability—has defined the theoretical problem facing proponents of descriptive representation. Early attempts to articulate the need for members of historically disadvantaged groups to represent those groups focused on the conflicting interests of privileged and relatively less privileged groups. Most notably, Virginia Sapiro (1981) showed that trusting some groups to protect another group's interests, e.g., letting husbands take care of their wives' interests, was and continues to be foolhardy. The recurring betrayals of historically disadvantaged groups by relatively privileged groups partially explain why traditional mechanisms of accountability are insufficient. By emphasizing the conflicts between advantaged and disadvantaged citizens, Sapiro laid the theoretical groundwork for a politics of presence. She did so by standing Pitkin's point on its head: Democratic accountability sometimes requires descriptive representation.
Following Sapiro's lead, a rich theoretical literature has developed that defends the intuition that the chronic underrepresentation of women and ethnic minorities is a problem. Anne Phillips (1998,228) has, very helpfully, laid out four groups of arguments about why descriptive representation matters. The first of these, which Phillips calls "the role model argument," claims that members of historically disadvantaged groups benefit from seeing members of their group in positions
representation. The criteria for choosing descriptive representatives will overlap significantly with the criteria for choosing representatives more generally. However, descriptive representatives who have been appointed, nominated, or elected to positions aimed at increasing the substantive representation of historically disadvantaged groups have particular and unique obligations to those groups.
of power. Having a woman in office increases other women's self-esteem and their capacity to assume leadership roles. Second, Phillips maintains that descriptive representatives are needed to compensate for past and continued injustices toward certain groups. According to this second argument, past and present betrayals by privileged groups create a belief that trust can be given only to descriptive representatives. The presence of descriptive representatives can partially compensate for those betrayals. She refers to this second argument for presence as "the justice argument." The justice argument examines patterns of inequality to reveal the need for descriptive representation. Her third argument focuses on "overlooked interests." According to this argument, group representation allows historically excluded groups to get onto the political agenda their perspectives, issues, and interests that had been previously ignored. Deliberations about public policy will be improved by having a more diverse set of representatives. Finally, Phillips advances the "revitalized democracy" argument, which asserts that a commitment to diverse representation is necessary for increasing political participation and strengthening the legitimacy of democratic institutions.
Dismissing the role model argument as uninteresting and without bearing on democratic politics (Phillips 1995, 63),6 Phillips devotes much of her attention to the remaining three arguments. Her work reflects a tendency in the literature on group representation as a whole to stress the value of group representation for considerations of justice, for deliberation, and for re-vitalization of democratic institutions. For present purposes, I focus primarily on justice arguments—that is, arguments that unfair patterns of inequality indicate a need for an institutionalized voice. These arguments appeal to evidence ranging from formal political exclusions (e.g., the disenfranchisement of certain groups) to economic disparities (e.g., mean incomes falling below the mean incomes of other groups). Such evidence challenges the assumption that all groups in liberal democracies enjoy the political equality that democratic commitments demand. Appealing to this evidence, theorists of group representation assert that justice demands paying particular attention to those in liberal democracies who are worse off. In doing so, they invoke the spirit of John Rawls's difference principle (1971).7
Theorists of group representation unambiguously acknowledge that despite the importance of descriptive representation, some descriptive representatives fail to further, and can even undermine, the best interests of historically disadvantaged groups. For instance, Melissa Williams (1998, 6) states that "it would be absurd to claim that a representative, simply
6 In contrast, I maintain that the ability to inspire and to be an example of a political leader from a historically disadvantaged group could be crucial for mobilizing that group.
7 Rawls's difference principle asserts that social and economic inequalities are just to the extent that they are necessary to the institutional structure that is the greatest benefit to the least advantaged in the distribution. See Rawls 1971, 60,302.
730
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
American Political Science Review
Vol. 96, No. 4
because she is a woman, therefore represents the interests or perspectives of women generally, or that an African-American representative is automatically representative of all African Americans. The mere presence of members of marginalized groups in legislatures is not sufficient for the fair representation of citizens from those groups, even though it is often necessary:' Similarly, Phillips (1995,157) states that "if the presumption is that all women or all black people share the same preferences and goals, this is clearly—and dangerously—erroneous." In some circumstances, a politics of presence can be undesirable. Most theorists of group representation recognize that members of historically disadvantaged groups have diverse interests and beliefs and that a politics of presence by itself is insufficient for revitalizing democratic institutions.
A tension thus emerges in the literature on group representation. On the one hand, theorists of group representation have argued that certain patterns of inequalities justify having an institutionalized voice. Such arguments emphasize the shared obstacles facing certain members of particular groups. Highlighting how certain groups are unfairly excluded from political life substantiates their claims that certain policy remedies are desperately needed. On the other hand, these theorists increasingly acknowledge the diversity within historically marginalized groups. This diversity can seem to undermine the presumption that certain shared experiences of oppression justify giving some groups an institutionalized voice.
Generally speaking, the literature has responded to this tension by exploring its implications for the meaning of representation as a whole. In particular, it has emphasized that representation is a dynamic process that must negotiate seemingly contradictory demands. Here the literature makes some of its most insightful contributions to democratic theory. For instance, Williams shows how the contradictory demands placed on representatives lead to understanding representation as a kind of mediation. Williams (1998, 8) identifies three dimensions of political life that representatives must mediate: the dynamics of legislative decision-making, the nature of legislator-constituent relations, and the basis for aggregating citizens into representable constituencies. Williams's understanding of representation as mediation expands the traditional conception of representation, which focuses almost exclusively on the relationship between the representative and the represented.
Iris Marion Young also offers a revised understanding of representation in response to the problem posed by diversity within historically disadvantaged groups. Young warns that attempts to include more voices in the political arena can inadvertently suppress other voices. She (Young 1997, 350) illustrates this point using the example of a Latino representative who might inadvertently represent heterosexual Latinos at the expense of gay and lesbian Latinos. For Young (1997,351) the suppression of differences is a problem for all representation. Representatives of large districts or of small communities must negotiate the difficulty of one person representing many. Because such a
difficulty is endemic to all representation, the legitimacy of a representative is not primarily a function of his or her similarities to the represented. Representation should not be characterized by a "relationship of identity."
Instead, Young uses Jacques Derrida's concept of differance to recommend reconceptualizing representation as a differentiated relationship. Derrida's concept of differance is both extremely radical and subtle. The basic idea underlying the concept of differance is the rejection of polarities or dichotomies, such as that of cause-effect. For those who think that certain distinctions are straightforward, it will be tempting to reject the concept of differance out of hand. But Young (2000, 127) finds this concept useful for capturing the dynamic process of representation: for this concept, she suggests, encourages thinking of oppositions "in terms of the trace, a movement of temporalization that carries past and future with it." Young further explains that the concept of differance promotes a way of thinking about entities in their "plurality without requiring their collection into a common identity____Things are similar without being identical, and different without being contrary, depending on the point of reference and the moment in a process." Using this idea of differance, Young recommends understanding representation as a fluid relationship, instead of a "relationship of identity" between constituents and representatives.
Let me briefly explain how Young applies Derrida's concept of differance to the meaning of representation. The basic move is to argue that the differences between the represented and the representative need to be both acknowledged and affirmed. According to Young (2000, 127), "Conceptualizing representation in terms of differance means acknowledging and affirming that there is a difference, a separation, between the representative and the constituents." For Young, constituents should not look for representatives with their same identity; rather, they should look for traces of accountability and authorization. Representation should be understood as a dynamic process that moves between moments of authorization and moments of accountability (Young 2000, 129). For Young, the movement between these moments makes the process "democratic." A representative process is democratic to the degree that citizens authorize their representatives and then can hold them accountable. Assessments of representative processes will therefore partially depend on the past and future behavior of representatives. Young's description of the dynamic of representation emphasizes that citizens often cannot anticipate the issues that representatives will confront during their term in office. Democratic citizens should continuously suspend or "defer" their evaluations of representatives. Democratic citizens must assess representation dynamically, that is, assess the whole ongoing processes of authorization and accountability of representatives. Young would resist assessing a representative from any one point of reference.
Young is quite right that representation in general is a complex and dynamic process. However, something important is overlooked in the quick move from group
731
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
Preferable Descriptive Representatives
December 2002
representation to representation as a whole. Young's focus on the problems of all representation obscures the distinctive problems of representation facing historically disadvantaged groups. She loses sight of the fact that some differences between representatives and the represented are more politically relevant from the perspective of democratic theory. Some politically salient differences should not be affirmed, e.g., differences that result from unjust and systemic exclusion.8 As Young herself pointed out in her earlier work, some groups face structural obstacles. In Justice and the Politics of Difference, she described real representation as consisting of "the self-organization of groups, the group generation of policy proposals in a context where decisions makers were required to be responsive to their perspectives and a group veto regarding specific policies that affect a group directly" (Young 1990,184). In her later work, Young (1990, 372; 2000,144, fn 27) "defers" the question of institutional supports for group representation. Her emphasis on the problems common to all representation downplays how particular institutional supports are necessary for overcoming some structural obstacles. Her admirable concern about the ways in which such institutional supports can suppress differences among historically disadvantaged groups leads her to retract her earlier commitment to certain institutional reforms, e.g., group vetoes. Young does continue to support multimember legislative jurisdictions, caucuses, and party list quotas. Reserved legislative seats should be used as "a last resort and temporary option for representing otherwise excluded perspectives" (Young 2000, 150). Young's desire to avoid excluding certain opinions, interests, or perspectives of historically disadvantaged groups thus weakens her initial support for institutional reforms aimed at increasing the real representation of those groups.
The degree to which Young has changed her mind is of less immediate interest than how the desire to avoid suppressing differences among members of historically disadvantaged groups can prevent theorists from articulating and defending reasons for preferring certain descriptive representatives over others. Simply to affirm all differences, as Young eventually does, ignores how certain politically salient differences between different groups in society (specifically between those who have been oppressed and those who have not) can justify the need for an institutionalized voice. Consider Young's example of the Asian American, who, she claims, has the perspective of an African American. This example shows how much she has weakened the claim that a historically disadvantaged group should be represented by members of that group. If one extends one's understanding of group membership so far as to include certain Asian Americans as members of
8 Democratic citizens should remain vigilant about preventing unjust and systemic exclusions. One way that they can do this is by assessing their choice of descriptive representatives in light of such exclusions. Institutional reforms aimed at increasing the number of descriptive representatives can revitalize democratic institutions if citizens select descriptive representatives based on an understanding of whose interests, perspectives, and opinions are being systemically marginalized.
the African-American community,9 then existing proposals for group representation, e.g., party list quotas, become untenable, for it is questionable whether such a person could contribute to the self-organization of an African-American community (or should count toward a party list quota for African Americans.) If Asian Americans can possess an African-American perspective and thereby satisfy the requirements of being descriptive representatives for African Americans, so can whites. Young's revised understanding of representation could legitimate an all-male (or all white) legislature as adequately representing women (or people of color) provided that they shared similar interests, opinions, and perspectives. The central claim of the literature on group representation—that historically disadvantaged groups need representatives from those groups—is seriously diluted by Young's notion of representation as a differentiated relationship.
Young's description of the dynamic of representation and her explicit recognition of the diversity within historically disadvantaged groups are useful contributions to theoretical debates about the general meaning of representation. But they provide little guidance to those confronted with the task of evaluating a particular descriptive representative. Such evaluations could be improved by articulating some general criteria for preferring some descriptive representatives to others.
JUSTIFYING THE SILENCE ABOUT CRITERIA
Silence about the criteria for evaluating descriptive representatives is not accidental. Theorists of group representation have offered two kinds of arguments to justify their silence. I call these two arguments the autonomy argument and the contingency argument.
According to the autonomy argument, members of historically disadvantaged groups should decide for themselves who is a preferable descriptive representative. This argument assumes that autonomy is best equated with being left alone—at least in the case of a group's choice of its representatives. According to this line of thinking, respecting the autonomy of historically disadvantaged groups requires theorists to refrain from advancing criteria for evaluating descriptive representatives. This argument has two main versions. The first suggests that articulating criteria forjudging descriptive representatives attributes to historically disadvantaged groups a fixed identity. The second emphasizes the autonomy of the representative.
The first version of the autonomy argument asserts that any proposed criterion for evaluating descriptive representatives presupposes that a historically disadvantaged group has an essential nature. In other words, it presupposes that that such a group has a "fixed
9 I assume here that Asian Americans are not typically considered part of the African-American community. I recognize that biracial and multiracial identities might complicate this assumption. My argument does not assume an essential identity to these groups but is based on existing, historically contingent understandings of group membership.
732
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
American Political Science Review
Vol. 96, No. 4
essence given once and for all, and with traits that are homogeneously distributed among all the group members" (Gould 1996, 182). Such an assumption places undesirable constraints on the behavior of members of historically disadvantaged groups. According to Williams (1995, 6), "No defensible claim for group representation can rest on assertions of the essential identity of women or minorities; such assertions do violence to the empirical facts of diversity as well as to the agency of individuals to define the meaning of their social and biological traits." To explain why some members are less suitable descriptive representatives is to question the authenticity of those members' identity. Such explanations possess an implicit charge that "she isn't really a woman" or "he isn't really black." In this way, discussions about the criteria for selecting descriptive representatives are often interpreted as attacks on the "authenticity" of descriptive representatives. Not only are such discussions overly divisive (Jones 1993; Stasiulis 1993), but they prevent the group from determining its own boundaries. According to this first version of the autonomy argument, members of historically disadvantaged groups should define for themselves the meaning of their group identity, as well as choose their own descriptive representatives.
The second version of the autonomy argument emphasizes the autonomy of descriptive representatives. Phillips, in particular, argues that specifying the criteria for evaluating descriptive representatives would undermine the arguments for group representation. For Phillips, a politics of presence is justified because representatives have some political discretion about their political decisions. Because of this discretion, descriptive representatives should be present where political decisions are being made. To uphold standards of "strict accountability"—that is, to require descriptive representatives to act in certain ways, e.g., to endorse a particular policy— would undermine a primary reason for why certain historically disadvantaged groups need representatives from those groups: namely, that representatives are not mere puppets of their constituents but must exercise their own judgment. To suppose that there is a fixed set of criteria by which descriptive representatives should be judged is to fail to appreciate how the autonomy afforded to representatives justifies the politics of presence. According to this second formulation, it would be misguided to provide a laundry list of "good policies" that a female representative should support and to insist that preferable female representatives can vote only in ways consistent with that list. After all, male representatives could also vote according to a laundry list. The more one knows how a descriptive representative should act, the less it is necessary to have a descriptive representative. Such reasoning led Mansbridge (1999, 630) to conclude that descriptive representatives become less necessary when interests have crystallized.
The second kind of argument for remaining silent about the criteria for judging descriptive representatives is the contingency argument. According to this argument, it is impossible to articulate the criteria that should be used to evaluate descriptive representatives
because context matters. Some theorists of group representation, such as Williams (1998,17), stress that particular historical developments play "an important role in defining the groups whose moral claims are strongest." Others stress that the experiences of historically disadvantaged groups defy generalizations. A priori proposals for criteria will either be irrelevant or do more harm than good. The variations across groups prevent adopting any one set of criteria. Mansbridge provides a particularly illuminating discussion of the relationship between descriptive representation and contingency. She identifies four contingent conditions that could justify preferring descriptive representatives to nondescriptive representatives: "(1) adequate communication in contexts of mistrust, (2) innovative thinking in contexts of uncrystallized, not fully articulated, interests,... (3) creating a social meaning of 'ability to rule' for members of a group in historical contexts where the ability has been seriously questioned, and (4) increasing the polity's de facto legitimacy in contexts of past discrimination"(Mansbridge 1999, 628). For Mansbridge, descriptive representatives are needed when marginalized groups distrust relatively more privileged citizens and when marginalized groups possess political preferences that have not been fully formed. She emphasizes that descriptive representatives are necessary only under certain conditions—that is, when descriptive representatives perform certain functions in certain contexts. Mansbridge's discussion provides some important insights into evaluating when descriptive representation is necessary. Implicitly, her work offers some general criteria for evaluating descriptive representatives—that is, by their ability to satisfy these different functions. However, her emphasis on identifying the contingent conditions under which descriptive representation is preferable to nondescriptive representation makes the actual choice of descriptive representatives secondary, if not irrelevant.
In summary, theorists of descriptive representation refuse to specify any criteria for evaluating descriptive representatives for two good reasons. First, they see offering such criteria as violations of the autonomy of historically disadvantaged groups and/or their descriptive representatives, and second, they view such criteria as insensitive to contextual variation. For these reasons, theorists of descriptive representation avoid a tough question: Who is a preferable descriptive representative? Answering this question is not easy, because it requires privileging the interests, values, and perspectives of certain members of historically disadvantaged groups over those of other members. Answers to this question can therefore have the effect of downplaying, if not excluding, certain interests, values, and perspectives. For this reason, answers to the question of who is a preferable descriptive representative are more likely to be disputed than answers to the question, "Why have descriptive representatives?"
THE NEED FOR CRITERIA
Underlying both the autonomy and the contingency arguments is a legitimate concern about who gets to
733
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
Preferable Descriptive Representatives
December 2002
decide which criteria are best. The impulse to speak for others can be and often is paternalistic and imperialistic (Alcoff 1995). Standards for assessing political performance have often been used to disqualify historically disadvantaged groups from political participation. To articulate criteria for evaluating descriptive representatives runs the risk that those criteria can be used in unanticipated and possibly harmful ways. Some fear that articulating such criteria might also unduly influence members of historically disadvantaged groups.
Such concerns are understandable but ultimately un-persuasive. After all, to articulate such criteria is not necessarily to assume that all members of a historically disadvantaged group have some essential identity. In fact, the very real and politically relevant differences among members of historically marginalized groups point to the desperate need for a theoretical discussion of criteria. There is a difference between articulating particular policies that a descriptive representative must endorse to count as a legitimate descriptive representative and articulating general guidelines for identifying preferable descriptive representatives. For instance, theorists of group representation have agreed that institutional reforms are necessary because historically disadvantaged groups possess overlooked interests. To maintain that a descriptive representative should pay special attention to overlooked interests does not require that she possess a particular view about those interests.10 Descriptive representatives have autonomy to the extent that she can reasonably interpret those interests in a variety of ways. However, to say that descriptive representatives can legitimately interpret their group's interests in multiple ways is not to say that anything goes. Descriptive representatives who denounce their group affiliations or who deny that they have any particular obligation to their group would fail to achieve the ends for which descriptive representation was introduced (cf. Phillips's four arguments). Descriptive representatives who claim to represent only the common good might be desirable representatives for other reasons; however, they do not satisfy Phillips's "overlooked interests" argument.
One can articulate criteria for evaluating descriptive representatives without violating the autonomy either of historically disadvantaged groups or of their descriptive representatives. To pose criteria for judging descriptive representatives is not the same as imposing those criteria on members of historically disadvantaged groups. Obviously, to impose criteria on such groups,
10 Like representatives generally, good descriptive representatives should sometimes act as trustees and at other times act as delegates. The standards for good representation cannot be linked strictly to the policy preferences of the represented. Pitkin (1967, 166) expressed this point in the following way: "Neither 'follow their wishes' nor 'ignore their wishes' will do; the decision must depend on why they disagree... but the standard by which he [the representative] will be judged as a representative is whether he has promoted the objective interest of those he represents. Within the framework of his basic obligation there is room for a wide variety of alternatives." My criterion for assessing preferable descriptive representatives does not assume one particular understanding of the objective interests of historically marginalized groups.
or on democratic citizens more generally, is wrong-headed. It is crucial according to my view that members of historically disadvantaged groups retain the ability to choose to adopt any proposed criterion.11
I would remind those who fear that articulating criteria for evaluating descriptive representatives might exercise an undue influence on a historically disadvantaged group that silence holds similar risks. After all, members of historically disadvantaged groups are not always the ones who select descriptive representatives. Privileged citizens are frequently in charge of selecting political appointees and nominating candidates for public office. To assume that not articulating the criteria for selecting descriptive representatives for committees or party lists is the best way to protect the autonomy of historically disadvantaged groups ignores the power that privileged groups currently hold. As Audre Lorde (1984,41) aptly warned, "Your silence will not protect you." Silence about controversial subjects does not necessarily bring about desired outcomes.
Besides, citizens inevitably bring their general standards of representation to bear on their choice of representatives. Critically evaluating the standards for descriptive representatives can facilitate deliberations by democratic citizens, thereby "assisting, and not infringing on" the autonomy of historically disadvantaged groups.12 Articulating criteria for assessing descriptive representatives does not make the exercise of judgment unnecessary; rather, deliberations about these criteria can refine those judgments. Theorists can offer criteria for choosing among descriptive representatives and still maintain that members of historically disadvantaged groups must determine for themselves whether a specific criterion is appropriate at any particular moment. This leads to the second argument for remaining silent: the role of contingencies in evaluations of descriptive representatives.
Context undeniably does matter. Evaluations of descriptive representatives, like arguments for a politics of presence (Phillips 1995,46), depend on "historically specific analysis of the existing arrangements for representation." For instance, who is a preferable descriptive representative might depend on whose interests, opinions, and perspectives are currently being stigmatized and marginalized by existing political norms and institutional processes. "Because group identity is orchestrated and produced in part through political institutional processes" (Bickford 1999, 86), citizens should consider the unjust effects of those processes as relevant to assessments of preferable descriptive representatives. Which descriptive representatives are preferable might also depend on the reasons that descriptive representation is necessary, e.g., to increase the trust groups have in democratic institutions or to include overlooked interests on the policy agenda.
Espousing criteria is not the same as requiring that certain criteria be applied in all circumstances. Like
11 A description of the conditions necessary for promoting this ability is clearly beyond the scope of this paper.
12 I want to thank an anonymous reviewer for helping me formulate my position in this way.
734
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
American Political Science Review
Vol. 96, No. 4
most theorists of descriptive representation, I share the suspicion of a cookie-cutter approach to evaluating descriptive representatives. It would be foolhardy to propose a set of criteria that did not consider context or that did not require individuals to use their own judgment to determine whether the criteria apply to the particular case at hand. To recognize the importance of particularities does not preclude articulating criteria for evaluation. It requires having criteria that are sensitive to those particularities.
My final argument for articulating criteria for evaluating descriptive representatives is based on my understanding of a particular role that political theory can and should play. Increasingly, the need for political theory to inform contemporary political controversies has been recognized (Isaac 1998, chap. 7). Contemporary politics is plagued by controversies about the representation of historically disadvantaged groups. Consider the controversies over the leadership of African Americans, such as Louis Farrakhan. Angela Dillard (2001, 4) notes that conservatives among women (e.g., Phyllis Schlafly) and minorities (e.g., Thomas Sowell and Richard Rodriguez) "have been dismissed as traitors, as sell-outs, as self-loathing reactionaries who are little more than dupes of powerful white, male, heterosexual conservatives." Recently, Lee Freed, a female president of the Manitoba chapter of the First Nations Accountability Coalition in Canada, accused the male tribal leadership of corruption and nepotism. Such contemporary controversies over the leadership of historically disadvantaged groups confirm that historically disadvantaged groups can possess different understandings about who should be their representatives. They also confirm the suspicion that simply having descriptive representatives is not sufficient to meet the requirements of a democratic commitment to the concerns of historically disadvantaged groups. Descriptive representation can fail to revitalize democratic institutions. It can also undermine democratic institutions if the ruling elites of historically disadvantaged groups use their institutional positions to control those groups instead of mobilizing those groups or bringing their overlooked interests onto the policy agenda (e.g., Cohen 1999). For these reasons, it is important to clarify the criteria for judging descriptive representatives. Evaluations of democratic institutions need to go beyond merely quantitative considerations—that is, the number of descriptive representatives. Evaluations of democratic institutions need to consider the extent to which preferable descriptive representatives are present. The criteria for identifying preferable descriptive representatives need to identify principled reasons for preferring some descriptive representative to others that are in line with the arguments for group representation. By failing to discuss criteria for assessing descriptive representatives, this theoretical literature ignores certain persistent debates about descriptive representation in contemporary politics. It also disregards the possible dangers and disappointments of a politics of presence to democratic politics.
THE CRITERION FOR EVALUATING DESCRIPTIVE REPRESENTATIVES
My criterion for evaluating descriptive representatives is a general one: Democratic citizens should consider the degree to which a descriptive representative has mutual relationships with dispossessed subgroups as relevant to identifying preferable descriptive representatives. Preferable descriptive representatives will have strong mutual relationships with dispossessed subgroups. This criterion is composed of two aspects. First, preferable descriptive representatives should possess a particular kind of relationship (mutual), and second, they should have this kind of relationship with certain subgroups of historically disadvantaged groups (dispossessed). I explicate both aspects of my criterion below.
Mutual Relationships
The importance of relationships to group identity is not a new claim. David Truman (1951, 24) maintained that "interactions, or relationships, give the group its molding and guiding power." Other theorists of group representation discuss relationships between representatives and their constituents (e.g., Williams 1998, chap. 6; Young 2000, chap. 4). However, these theorists typically examine these relationships primarily in terms of whether the constituents "trust" their representatives. Young (2000, 128-30) evaluates the process of representation by the extent to which the relationship between representative and constituents "avoids separation" and "renews connection." She does not address how to evaluate particular individuals engaged in the process of representation.
What is distinctive about my criterion is its specification that representatives and members of historically disadvantaged groups must mutually recognize each other. Mutuality requires an interactive relationship between representatives and citizens. Mutual relationships require a historically disadvantaged group to recognize its descriptive representatives in a particular way as well as a descriptive representative to recognize that group in a particular way. Such reciprocal recognition is necessary for descriptive representatives and their groups to coordinate consciously chosen political activities. Descriptive representatives without mutual relationships could be "representative" in the sense that their behavior responds to the policy preferences of their group, but such responsiveness is not sufficient to make the form of representation democratic. Kings could be representatives of their subjects, in this sense, if they sufficiently polled the preferences of their subjects.
Democratic relations demand effective participation on the part of citizens (Dahl 1989, 109). Democratic representation requires that citizens can access and influence political institutions. Descriptive representatives with mutual relations would improve democratic representation by enabling historically disadvantaged groups to influence the political decision-making process. In doing so, historically disadvantaged groups
735
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
Preferable Descriptive Representatives
December 2002
act in concert with their descriptive representatives. Democratic relationships are therefore ones "in which both parties are active" (Plotke 1997, 29).
In proposing this criterion, I am advocating a new approach to assessing the performance of descriptive representatives. Political scientists often evaluate descriptive representatives' performance by focusing exclusively on notions of interests as identified by policy preferences. Assessing descriptive representatives solely by the way they cast their votes can lead to the conclusion that it does not matter who represents historically disadvantaged groups (Schwarz and Shaw 1976; Swain 1993). This approach ignores other reasons for having descriptive representatives, e.g., introducing overlooked interests or building trust in the political institutions. Cathy Cohen's analysis (1999) of AIDS activism in the African-American community revealed the general failure of black elites to recognize the particular needs of certain subgroups in the African-American communities, that is, black gays and lesbians and IV drug users. Cohen showed that although black leaders often maintained a "good" voting record in terms of AIDS policies, they failed to transform the nature of the political debate in ways that address the particular interests of specific subgroups in the African-American community, e.g., needle exchange programs.
Good descriptive representatives therefore cannot be identified simply by examining voting records. Just as presidential performance is partially judged by the president's selection of staff, descriptive representatives should be judged by who does and does not interact with them. Assessments of descriptive representatives need to consider whether these representatives reach out to (or distance themselves from) historically disadvantaged groups. Preferable descriptive representatives facilitate social networks. Formal as well as informal ties provide the channels through which democratic relationships could work and thereby the means to revitalize democratic institutions. I introduce mutual relationships into discussions of descriptive representation because these discussions need to reflect the fact that what determines policy is not only what political actors do but also whom they know.
It is important to emphasize a consideration implicit in my claim that preferable descriptive representatives possess mutual relationships: The commitment to democratic representation requires that democratic citizens should not be apathetic. Preferable descriptive representatives will inspire their group to act in concert with them. Although it is possible that a descriptive representative could adequately "represent" the concerns of the apathetic insofar as the representative takes positions that reflect the interests or preferences of apathetic citizens, the descriptive representative's actions would not be democratic to the extent that apathetic citizens do not care about the activities of that representative. My criterion prefers descriptive representatives who can and do mobilize a historically disadvantaged group, encouraging the active engagement of that group. Requiring preferable descriptive representatives to have mutual relations is very demanding and therefore likely to support robust democracies.
To possess mutual relations, descriptive representatives must recognize and be recognized by members of a historically disadvantaged group in two ways. First, they must recognize each other as belonging to a historically disadvantaged group, and second, they must recognize each other as having a common understanding of the proper aims of a descriptive representative of the group. To recognize each other mutually in these two ways is to possess a mutual relationship.
Preferable descriptive representatives are those who recognize and are recognized by members of their historically disadvantaged group as being "one of us." In particular, they have a reciprocated sense of having a fate linked with that of other members of their group.13 Michael Dawson (1994,77) defines the notion of linked fate in reference to African Americans as "the degree to which African Americans believe that their own self-interests are linked to the interests of the race." To possess a sense of linked fate is to experience "the group interests... as a useful proxy for self-interest." In other words, individuals who believe that their fates are linked to a group believe that "what happens to the group as a whole affects their own lives."
The notion of linked fate reflects the fact that a person's range of choices—that is, his or her perceived opportunities and goals— is both subjective and social. Group identities are partially formed by "the political processes through which concrete notions of collective interest are shaped,... who participates in those processes and who is advantaged and disadvantaged by them" (Reed 1999, 45-46). Recent political theorists have also endorsed the view that the social world— what they call culture—limits the range of choices available to individuals. For example, Joseph Raz and Avishai Margalit (1994,119) state that "familiarity with a culture determines the boundaries of the imaginable. Sharing in a culture, being part of it, determines the limits of the feasible." In this way, membership is both "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). Thus, even individuals who object to their identities being defined largely in terms of their group membership or who are critical of the ways in which group membership can constrain their choices can still have a sense of sharing their fate with a historically disadvantaged group.
To have a sense of linked fate with a historically disadvantaged group partially entails having a substantive conception of that group that is relevantly similar to those held by other members of that group. Group membership can be based on shared visible characteristics, e.g., color of skin, or on shared experiences (Mansbridge 1999). The substantive content of group membership can vary. Some individuals can belong to more than one group and therefore experience conflicting allegiances to different groups. Members can also possess conflicting views on their group's politics. I am not denying such differences.
13 For a discussion of how to measure this sense of linked fate using attitude and opinion surveys, see Dawson 1994, 77-80, 82-84.
736
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
American Political Science Review
Vol. 96, No. 4
But preferable descriptive representatives for a given group share an understanding of the group's boundaries with that group. Descriptive representatives who possess a narrower (and more exclusive) understanding of those boundaries are unable, or at least less likely, to satisfy the arguments that justify group representation. For this reason, an African-American descriptive representative who denies that gay and lesbian blacks are members of the group (or who excludes conservatives, IV drug users, Muslims, or other religious African Americans) would be less preferable than one who includes those members in his or her understanding of the group.14 After all, justifications for group representation tend to emphasize the extent to which descriptive representatives include overlooked interests, build trust, and foster deliberation. Descriptive representatives who overlook certain members of the group or who deem certain members "inauthentic" are less likely to fulfill these functions. Representatives who possess broader understandings of the group are more likely to overlap with the varied understandings of the represented and therefore satisfy the reasons for having an institutionalized voice.
Inclusive language also has its pitfalls (e.g., Cohen 1999; Minow 1990; Reed 1999,17). Generalized notions of a group can be so abstract and all-encompassing that they ignore significant differences among members of the group. Hence, U.S. suffragists who claimed to speak for women were justifiably criticized for speaking from an unreflective bias of being middle-class or educated or white (Davis 1983). Abstract notions of the group can also prevent elites from being held accountable to specific people. Speaking about the "underclass," "women," or "Latinos" in general terms can create an illusory unity among members of those groups that can be used to the detriment of vulnerable members of those groups (Reed 1999,5). For example, it is possible to diminish community support for policies aimed at helping vulnerable members of a community by portraying those policies as attacks on the community. Preferable descriptive representatives possess shared understandings of group membership that recognize salient differences of subgroups. I elaborate on my understanding of such salient differences in my discussion of dispossessed subgroups below.
To understand the importance of mutual recognition of belonging to the group for evaluating a descriptive representative, consider the following case. It is possible to imagine an African-American representative who grew up in a primarily white neighborhood, attended predominantly white private schools, has a white spouse, and has shown no demonstrable interest in the problems of other African Americans. In fact, such a representative could thrive politically by publicly distancing herself from the African-American community.15 The point of this example is not to
As can be seen, descriptive representatives with mutual relations do not necessarily possess "progressive" or "liberal" policy agendas. 15 My example bears a strong resemblance to Young's example of an Asian American who has an African-American perspective. However, my argument suggests that such an Asian American would be
question whether this woman is an "authentic" African American. I believe that she is.16 Rather, it is to question whether such a representative could satisfy sufficiently the reasons that theorists for a politics of presence gave for increasing the number of descriptive representatives. After all, such a descriptive representative lacks the relationships necessary to satisfy these reasons. She might individually face certain obstacles and experience forms of discrimination because of her identity; however, she lacks the relationships with African Americans that could enable her to achieve mutual recognition with them. The extent to which she disavows her relationships to African Americans indicates the extent to which she is less likely to possess mutual relationships with them. African-Americans would be more likely to distrust her. She would also be less likely to advance overlooked interests of the African-American community and to mobilize that community. Who perceives that representative as "belonging" to the group and whom a representative claims to act on the behalf of are important considerations for evaluating the qualifications of descriptive representatives.
Individuals in mutual relationships not only recognize each other as belonging to the same group, but also recognize that they share an understanding of the proper aims of their representatives. To have shared aims is to possess a similar vision for the future direction of politics—one whose goal is the improvement of the social, economic, and political status of particular historically disadvantaged groups. My understanding of aims has two components: policy preferences and values.17 A descriptive representative could disagree with members of a historically disadvantaged group about either component, yet still share aims.18 Some
a less preferable descriptive representative if those in the African-American community did not accept and identify him as a member of their community. Given the current divisions between racial groups, such a revised understanding of racial identity is possible, albeit unlikely. Note that Young's example is presented in a way that emphasizes its potential to be more inclusionary, while my argument suggests that relationships provide reasons for objecting to certain descriptive representatives for particular groups.
16 To articulate reasons for preferring some descriptive representatives to others is not the same as questioning the authenticity of a descriptive representative's membership. A full discussion of the relationship between preferability and legitimacy is clearly beyond the scope of this paper. Here I purposely limit my discussion to the desirability of particular descriptive representatives, not their legitimacy. I recognize that all members of a historically disadvantaged group are in some sense legitimate descriptive representatives of that group. In other words, Reverend Jesse Jackson, Marian Wright Edelman, Shelby Steele, and Allan Keyes are all legitimate descriptive representatives for African Americans; however, who is a preferable descriptive representative for African Americans depends on who possesses strong mutual relationships with dispossessed subgroups. Such subgroups can include conservative and/or poor subgroups.
17 For a discussion of the difference between measuring political values and policy preferences, see Stoker 2001 and Rasiniski 2001.
18 To explicate the idea of an aim, it is necessary to differentiate an aim from what Young (2000, 134) calls the "modes of representation"—that is, three aspects of one's identity that need to be represented. Those three aspects are interests (policy preferences), opinions (values, priorities, and principles), and perspectives (starting points of conversations). While Williams (1998,171) argues that interests and perspectives are more inextricably tied, Young stresses how these different aspects of a person's identity can conflict.
737
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
Preferable Descriptive Representatives
December 2002
members might experience a descriptive representative advocating certain public policies as a litmus test for shared aims with that representative, e.g., their position on abortion or affirmative action, while others see shared aims as resulting from a particular combination of policy preferences and values. Individuals can and will have different conceptions of what is necessary for having shared aims. Nevertheless, a descriptive representative who did not share either component with a historically disadvantaged group does not share aims with that group.
In this way, my criterion recognizes that people who share similar political values can justifiably disagree about the desirability of certain public policies. It also recognizes that individuals with different political values can agree about certain public policies. Consequently, I do not always want people who agree with my political values or with my policy preferences. I do want someone who shares my aims. The notion of shared aims recognizes the importance of the interaction between policy preferences and values for selecting preferable descriptive representatives. For this reason, shared aims must be measured in degrees: Descriptive representatives share aims with a historically disadvantaged group to greater or less degrees.
This notion of aims as a kind of direction for politics interjects into discussions of group representation my belief that the actions of descriptive representatives do matter. Pitkin was wrong to draw such a firm distinction between what a representative looks like and what a representative does. My criterion for evaluating descriptive representatives offers one way to follow Phillips's recommendation to integrate a politics of presence with a politics of ideas. Descriptive representatives are preferable to the degree that their actions are perceived by members of a historically disadvantaged group as improving their linked fate. My notion of aims is meant to capture the fact that members of historically disadvantaged groups, despite having different policy preferences and values, can still share a political vision aimed at relieving the plight of their communities. Thus, the actions of descriptive representatives are not irrelevant to who should be considered a preferable descriptive representative. Preferable descriptive representatives recognize themselves, and are recognized by members of a historically disadvantaged group, as sharing the aims of that group.
The importance of shared aims is most readily apparent when one lacks a representative who shares one's aims. One is less likely to accept differences of opinions with those who have different aims than with
For Young, the process of democratic representation relies on all three modes.
Young's analysis of the dynamic processes of representation, though, can divert attention from the proper standards for evaluating particular representatives. Individual representatives are less likely to satisfy all of these different modes of representation than are the processes of representation. I know of no representative who shares all of my interests, opinions, and perspectives. These modes of representation are too narrow to provide much guidance for identifying preferable descriptive representatives. For this reason, I argue that members of historically disadvantaged groups should seek descriptive representatives who share their aims.
those who share one's aims. My discussion of shared aims reflects evidence that African Americans tend to give their leaders the benefit of the doubt in the face of controversies, e.g., controversies over Louis Farrakhan. Historically disadvantaged groups are willing to "own" a representative with whom they disagree (Dawson 1994). The reason is that they possess a linked fate and shared aims with these leaders.
Both a sense of belonging to a group and shared aims are important for mutual relations, for individuals whose fates are linked can have different aims. For example, ultraorthodox Jewish women are forbidden from studying general subjects such as math and embrace their traditional role in the house. Moreover, many ultraorthodox women believe that women are incapable of making important decisions (such as voting). For some of these women, it is proper to arrange their social and political lives around the assumption that women are inferior. The presence of these ultra-orthodox Jewish women in Israel affects how Israeli women (including Jewish secular or even Palestinian women) are perceived. Nonpracticing Jewish women, Palestinian women, and ultraorthodox Jewish women thus share fates in Israel, even though these different subgroups have contradictory policy preferences and values. This example demonstrates that members of historically disadvantaged groups can share fates, even though their aims differ.
Preferable descriptive representatives are those who possess mutual relationships with their constituents. But, as has been shown, this relationship consists in descriptive representatives and their historically disadvantaged groups recognizing each other as having a linked fate and sharing aims. This mutual recognition thus provides some substantive guidance both for which descriptive representative is preferable and for what a descriptive representative should be doing. In these ways, mutual relationships between a descriptive representative and a historically disadvantaged group provide that group with a stake in politics: They can influence the political agenda through acting in concert with their descriptive representatives.
Dispossessed Subgroups
I now clarify the second aspect of my criterion—what I mean by a dispossessed subgroup. Dispossessed subgroups should not be understood as those groups that literally do not possess any private property, possessions, or resources. Rather, I use the term dispossessed in a narrower way to refer to groups that are unjustly excluded from and/or stigmatized by the political process and consequently lack the political and economic resources necessary for effective representation. Often dispossessed subgroups suffer oppression not only as members of their overarching group but also as members of the subgroup. They are therefore members of historically disadvantaged groups, yet they face further political obstacles—what Cathy Cohen (1999, 70) describes as secondary marginalization—that is, the ways in which members of marginalized groups construct and police group identity as to regulate behavior,
738
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
American Political Science Review
Vol. 96, No. 4
attitudes, and the public image of those groups. Perhaps it is in virtue of the combination of the forms of oppression that they lack the financial, time, and social resources necessary for political participation. Class, sexuality, drug use, geographic location, relationships to welfare, criminal records, and religion are all possible markers of dispossessed subgroups.
This second aspect of my criterion offers a way to return to the commitment found in the literature on group representation to those groups that have been and continue to be marginalized within the existing political system. A commitment to group representation entails a commitment to those whose interests have been overlooked, who have been and continue to be unjustly excluded from political participation, and whose presence could revitalize democratic institutions. Group representation therefore requires being vigilant about groups that lack a political voice. Preferable descriptive representatives would be those who seek out and establish mutual relationships with dispossessed subgroups.
To demonstrate the importance of mutual relationships with dispossessed subgroups, I focus on the ways in which class inequalities can constrain effective representation.19 Such inequalities can undermine democratic citizens' political resources. My discussion of dispossessed subgroups is by no means limited to the experiences of poor subgroups of historically disadvantaged groups. Other subgroups that lack the political and economic resources for effective representation would also count as dispossessed. I use poor subgroups to illustrate my understanding of dispossessed subgroups for two reasons. First, this example highlights the necessity of mutual relationships for improving the substantive representation of historically disadvantaged groups. Second, this example demonstrates the interactions among different forms of oppression.
Theorists of group representation often implicitly recognize the importance of class in their arguments. Almost all proponents of group representation (e.g., Phillips 1999, 151; Williams 1998, 15-18; Young 2000, 92-99) appeal to the economic structural inequalities that certain marginalized groups face, e.g., the rates of victimization, of poverty, of housing, and of job discrimination, to justify group representation. In this way, they recognize that economic inequalities are one indicator that a group deserves a political voice.
Nevertheless, just as in previous times women and ethnic groups were considered adequately represented by the presence of white male representatives, theorists of self-representation do not adequately acknowledge problems with poor subgroups of historically disadvantaged groups being represented by econom-
19 I explicitly reject an understanding of class that is based on categories and classification schemes; rather, I am concerned with how class relations are produced and maintained through political institutions. Phillips argues that the category of class is substantively different from conceptions of race and gender. For example, one loses one's class when one becomes an elected official. For an alternative understanding of class, see Acker (2000,197), who defines class as "social relations constructed through active practices, not as categories or classifications of people according to socioeconomic characteristics or occupational status."
ically more privileged members of their group. Some explicitly deny that class should be incorporated into political solutions for presence. For example, Phillips argues that the politics of presence should be treated as distinct from issues concerning class. Phillips offers several reasons for this distinction, e.g., the difficulty in defining class. However, Phillips (1995, 170-78) admits that these reasons for treating class separately are "insincere," stating that "when it comes down to it, the real reason for my silence on class is simply that it does not lend itself to the same kind of solutions."20 This admission implies that considerations of class cannot be adequately incorporated into the types of institutional reforms necessary for increasing the number of descriptive representatives. Interestingly, this admission contradicts her arguments for a politics of presence. Phillips is quite explicit that a politics of presence is not a guarantee for a robust democratic politics. A democratic politics must balance the commitment to presence with another commitment: what Phillips calls "a politics of ideas." For Phillips, the politics of ideas refers to the commitment to particular opinions, preferences, and beliefs. The politics of ideas would include one's position on class issues. In contrast, a politics of presence is a commitment to the intuition that it matters who expresses those opinions, preferences, and beliefs. For Phillips, democratic practices will nourish when democratic citizens start integrating these two commitments. Phillips's claim that class should not be incorporated into a politics of presence violates her own understanding of the conditions necessary for robust democratic relations.
Others minimize the significance of socioeconomic factors by choosing examples that focus almost exclusively on only one form of oppression. Often these examples explore the ways that groups are formally excluded from political participation. For instance, Williams focuses on the structural obstacles faced by U.S. women and African Americans in their efforts to gain full political standing. She cites economic inequalities as indicators that institutional reforms are necessary, yet her proposed institutional reforms are aimed exclusively at formal political exclusions.
Williams's emphasis on formal political exclusions reflects the tendency among proponents of group representation to notice the oppressive nature of socioeconomic status without incorporating this observation into their arguments for group representation or into their proposed institutional reforms. These proponents also have not incorporated the insight that one must understand the interactions among multiple forms of oppression (e.g., Collins 1990; Higginbotham 1992; hooks 2000).21 One cannot simply "add on" an analysis of class after advocating for increasing the
20 For her full discussion of class, see Phillips 1995, chap. 7.
21 The failure of the literature on descriptive representation to incorporate this insight is readily apparent in its proposed institutional reforms, e.g., party quotas. For instance, these reforms do not specify whether Asian Americans should count as descriptive representatives for African Americans and therefore should count toward an African-American party quota, as Young's example implies. Nor do they provide any way for determining whether African-American
739
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
Preferable Descriptive Representatives
December 2002
representation of women and people of color. Deborah King illustrates the inability to add on class by noting that education can increase the income potential among different groups disproportionately. King (1993, 223-24) claims that the economic benefits "of a post-secondary education, a college degree or higher, are greater for black females than for white females while among those with less than a college degree, black females earn less than white females." In this way, King reveals that focusing too much on only one form of oppression can mask the obstacles faced by certain segments of historically disadvantaged groups.
Williams's analysis of self-representation would have benefited from an example in which the dynamic of multiple forms of oppression was considered. For instance, Margaret Wilkerson and Jewell Gresham (1993, 297) have argued that the "feminization of poverty" cannot be understood as distinct from the "racializa-tion of poverty." Wilkerson and Gresham claim that the focus on the economic inequalities faced by women "negates the role played by racial barriers to black employment, particularly among males." Theorists of group representation tend to give examples in which the dynamics of race, class, and gender are prominent only a cursory treatment, if any treatment at all. They also tend to downplay how political norms and practices within the democratic institutions, e.g., recruitment practices, can marginalize certain subgroups. Consequently, their understandings of group representation ignore that inclusion in politics can promote instrumental political bargaining at the expense of transformative politics (Dryzek 1996; Reed 1999).
More specifically, theorists of group representation do not adequately address the particular barriers to effective representation experienced by poor subgroups of historically disadvantaged groups. For instance, given the practice of disenfranchising convicts, the high incarceration rates of poor blacks and Latinos cut off traditional avenues for seeking political representation:
Every state but three imposes some type of ban on voting by those convicted of serious crimes. Most states bar voting by felons while in prison, but restore the right to vote once the individual has served his sentence or completed parole. In 14 states, a felony conviction can mean a permanent ban on voting. (Braceras 2000)
The current practice of disenfranchising convicts cast doubts on whether certain subgroups of historically disadvantaged groups should rely primarily on electoral procedures for achieving substantive representation. Theorists of group representation need to address the obstacles that are produced through the interactions among different forms of oppression and that unjustly constrain certain subgroups.
Theorists who emphasize electoral reforms that increase the number of descriptive representatives also
women should "count" under the number of all women or have their own quota in the party lists. The failure to address these issues in this literature reveals the failure to incorporate an interactive understanding of race, class, and gender.
tend to ignore the kinds of resources necessary for poor subgroups to advance their political agendas. Traditional means for getting policy preferences onto the political agenda—studies, public relations campaigns, lobbying efforts—advantage citizens who are financially better off and resource-rich. Being able to stay informed about political issues, let alone to participate in politics, requires time and economic resources. Elected officials increasingly spend their time fund-raising. Citizens with economic resources can buy access, but those without economic resources tend to have relatively less access. Consequently, those with economic resources do not necessarily need as much of an institutionalized voice as those who lack those resources.
Typically, citizens who lack economic resources need to register their preferences through non-institutional and confrontational tactics. Frances Fox Piven and Richard Cloward (1979, 3) have argued that "protest tactics which defied political norms were not simply the recourse of troublemakers and fools. For the poor, they were the only recourse." According to Piven and Cloward, poor citizens need alternative tactics, such as protests, for effective representation. These tactics depend on numbers and relationships. Piven and Cloward's position reflects the common belief in the literature on social movements that more disruptive tactics are more likely to be successful (e.g., McAdam 1983; Tarrow 1994; Tilly, Tilly, and Tilly 1975). The social networks surrounding descriptive representatives are therefore relevant to their preferability. A descriptive representative who possesses mutual relationships to poor subgroups of historically disadvantaged groups is more likely to have the political resources necessary for advancing those subgroups' interests. Such descriptive representatives also hold the promise of expanding the boundaries of political participation.
One should not assume that class "perspectives" are necessarily better represented if ethnicity, race, and gender are better represented in legislatures. After all, research has documented the economic disparities within various racial groups (e.g., Dawson 1994, chap. 2; Hochschild 1995; Wilson 1980). Such disparities are increasing. Although they continue as a group to be economically and socially worse off than whites, African Americans are increasingly economically divided in ways that affect housing, jobs, death rates, and the likelihood of being a victim of crime. Such disparities among African Americans have led some to conclude that black identity will increasingly be tied to considerations of class as opposed to race—what Kilson (1983) called "status deracialization" and Wilson (1980) titled "the declining significance of race."
For Dawson (1994), the economic polarization of blacks does not necessarily lead to political polarization. Economic polarization, though, does affect the extent to which poor blacks participate politically. After examining falling rates of political participation among poor blacks, Hochschild (1995, 50) concluded that "the worst-off in general are losing political influence, and the worst-off blacks in particular are losing most." This observation is especially troublesome when considered in conjunction with Cohen's (1997)
740
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
American Political Science Review
Vol. 96, No. 4
claim that historically disadvantaged groups are policed internally. If Cohen is correct, then the choice of descriptive representatives is crucial for understanding why some dispossessed subgroups lack substantive representation. This choice is also crucial for the proposed institutional reforms, e.g., party list quotas and proportional representation, to be able to revitalize democratic practices.
Of course, low economic status is not the only reason that citizens do not participate in politics. Cohen (1999, 346) revealed that black gays and lesbians who were HIV-positive did not participate in AIDS protests from fear of being seen by other members of the black community. Such fears were particularly acute for members who relied on that community for support while sick. As Cohen's example demonstrates, other political norms and practices besides the formal exclusion of historically disadvantaged groups can exclude certain subgroups of historically disadvantaged groups. For this reason, preferable descriptive representatives can have mutual relationships with other types of dispossessed subgroups.
My criterion for evaluating descriptive representation should not be interpreted as arguing that the self-representation of women or of African Americans or other minorities is secondary to the representation of the poor. Such an argument would merely mimic the common claim that identity politics is divisive while class is more unifying (Gitlin 1995). Nor am I repeating claims that class is more politically salient than race (e.g., Loury 1987; Murray 1984; Sowell 1984). Instead, I maintain that who is a preferable descriptive representative depends on how different forms of oppression intersect, for example, how race can work in conjunction with class is relevant to determining who is a preferable descriptive representative. Democratic citizens need to evaluate descriptive representatives in ways that attend to how political institutions marginalize certain groups. Young was right that institutional reforms aimed at increasing the number of descriptive representatives can entrench certain interests, e.g., by privileging heterosexual Latinos at the expense of gay and lesbian Latinos. Moreover, evaluations of descriptive representatives are particularly messy when segments of a historically disadvantaged group reject a descriptive representative. For Young, the diversity within historically disadvantaged groups can be so great that schemes of group representation will necessarily result in the suppression of difference.
However, recognition of the diversity within historically disadvantaged groups does not change the fact that some groups are chronically underrepresented. In other words, it does not change the fact that some groups need institutional reforms to enhance their substantive representation. For the institutional reforms to work successfully, democratic citizens need to select descriptive representatives in ways that are sensitive to how institutional norms and practices unjustly marginalize dispossessed subgroups. My criterion offers one way to take into account the dynamic among different forms of oppression: Who is a preferable descriptive representative depends partially on
whose interests, opinions, and perspectives are being excluded. Recall that a descriptive representative's shared aims and sense of belonging to a group provide some substantive guidance for what that representative should be doing. In this way, my criterion depends on context. Those selecting descriptive representatives (for appointments, committees, or public office) need to attend to the mutual relationships that descriptive representatives possess with dispossessed subgroups. They should not assume that "just any woman will do" or that "just any black will do." Institutional reforms aimed at increasing the number of descriptive representatives are more likely to revitalize democratic institutions if citizens assess descriptive representatives using my criterion.
For this reason, I submit that when one has a choice between two descriptive representatives, one who has strong mutual relationships to dispossessed subgroups and another who does not, one should (ceteris paribus) prefer the former. I have so far avoided the question of what to do when choosing among descriptive representatives who possess mutual relationships to different dispossessed subgroups. Such moments do not have generalizable or easy answers. In such circumstances, citizens face tough choices that require exercising their own political judgment. To recognize that the dispossessed too can have diverse interests is to acknowledge that my criterion might not settle the question of who is a preferable descriptive representative. However, the refusal to examine the criteria being used for selecting descriptive representatives can reinforce the norms and practices that unjustly exclude dispossessed subgroups. Public deliberations about the proper criteria could therefore help refine those decisions and prevent such exclusions.
If historically disadvantaged groups do possess such deep divisions that they must consistently choose among interests, opinions, and perspectives of competing dispossessed subgroups, then those groups are less likely to be legitimately represented by only one representative. In other words, if certain groups possess intractable divisions, e.g., between liberal and conservative African Americans or between heterosexual and gay and lesbian Latinos, then such groups would need more than one descriptive representative. This observation affirms Young's conclusions (1999) about the need to pluralize group representation. The presence of multiple dispossessed subgroups indicates the need for more descriptive representation, not less. Unfortunately, there is often a limit to how many descriptive representatives a given group can have.22 Such limits require principled criteria for selecting descriptive representatives. As we have seen, the refusal to articulate any criteria for preferring some descriptive representatives to others has led some theorists, such as Young, to weaken their commitment to the position that historically disadvantaged groups should be represented by
22 This argument could lead to the proliferation of descriptive representatives. For a helpful discussion on how to identify historically disadvantaged groups, and thereby subgroups, that deserve group representation, see Williams 1998,15-18.
741
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
Preferable Descriptive Representatives
December 2002
members of their group. As I have also pointed out, this refusal could also prevent the proposed institutional reforms from revitalizing democratic institutions: Some descriptive representatives may perpetuate or even aggravate the marginalization of historically disadvantaged groups.
Proponents of group representation are likely to agree that it is important to attend to the relationships between descriptive representatives and dispossessed subgroups. In fact, my criterion arises from the same normative commitments that justify group representation. According to this logic, the extent to which a politics of presence can include those who have been systemically excluded from political life is also the extent to which a politics of presence can bolster democratic participation and the legitimacy of democratic institutions.
However, one needs to understand that democratic representation excludes as well as it includes. The act of excluding is not in itself objectionable on democratic grounds. After all, representative institutions require selecting some representatives at the expense of others. I introduce my criterion in an effort to provide some guidance for preferring certain descriptive representatives to others. I hope that others will expand on my criterion in ways that are sensitive to the reasons for supporting group representation and to data on the political marginalization of different groups. Introducing a criterion such as mine into existing discussions of group representation offers a principled way to balance a commitment to the diversity within historically disadvantaged groups with a commitment to group representation.
REFERENCES
Acker, Joan. 2000. "Revisiting Class: Thinking from Gender, Race, and Organizations." Social Politics (Summer): 192-213.
Alcoff, Linda. 1995. "The Problem of Speaking for Others." In Who Can Speak, ed. Judith Roof and Robyn Wiegman. Chicago: University of Illinois Press, 97-119.
Bickford, Susan. 1999. "Reconfiguring Pluralism: Identity and Institutions in the Inegalitarian Polity." American Journal of Political Science 43 (January): 86-108.
Braceras, Jennifer. 2000. "Massachusetts Shouldn't Let Convicts Vote." The Boston Globe (October 5): A27.
Cameron, Charles, David Epstein, and Sharyn O'Halloran. 1996. "Do Majority-Minority Districts Maximize Substantive Black Representation in Congress?" American Political Science Review 90 (December): 794-812.
Cohen, Cathy. 1997. "Straight Gay Politics: The Limits of an Ethnic Model of Inclusion." Nomos XXXIX: Ethnicity and Group Rights, ed. Ian Shapiro and Will Kymlicka. New York: New York University.
Cohen, Cathy. 1999. The Boundaries of Blackness: AIDS and the Breakdown of Black Politics. Chicago: University of Chicago.
Collins, Patricia. 1990. Black Feminist Thought: Knowledge, Consciousness, and the Politics of Empowerment. Boston: Unwin Hyman.
Dahl, Robert A. 1989. Democracy and Its Critics. New Haven, CT: Yale University.
Davis, Angela Y. 1983. Women, Race & Class. New York: Vintage Books.
Dawson, Michael C. 1994. Behind the Mule: Race and Class in African-American Politics. Princeton, NJ: Princeton University.
Dillard, Angela. 2001. Guess Who's Coming to Dinner Now? Multicultural Conservatism in America. New York: New York University.
Dryzek, John. 1996. "Political Inclusion and the Dynamics of Democratization." American Political Science Review 90 (September): 475-87.
Gitlin, Todd. 1995. Twilight of Common Dreams. New York: Metropolitan Books.
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, 171-86.
Guinier, Lani. 1994. The Tyranny of the Majority: Fundamental Fairness in Representative Democracy. New York: Free Press.
Higginbotham, Evelyn. 1992. "African American Women's History and the Metalanguage of Race." Signs 17: 251-74.
Hochschild, Jennifer. 1995. Facing Up to the American Dream: Race, Class, and the Soul of the Nation. Princeton, NJ: Princeton University.
Hooks, Bell. 2000. Feminist Theory: From Margin to Center.
Cambridge, MA: South End. Isaac, Jeff. 1998. Democracy in Dark Times. Ithaca, NJ: Cornell
University.
Jones, Kathleen. 1993. Compassionate Authority: Democracy and the Representation of Women. New York: Routledge.
Kilson, Martin. 1983. "The Black Bourgeoisie Revisited." Dissent (Winter): 85-96.
King, Deborah. 1993. "Multiple Jeopardy: The Context of a Black Feminist Ideology." In Feminist Frameworks. 3rd ed. Boston: McGraw-Hill, 220-36.
Lorde, Audre. 1984. Sister Outsider. Trumansburg, NY: Crossing Press.
Loury, Glenn. 1987. "Who Speaks for Black Americans?" Commentary 83 (January): 34-38.
Mansbridge, Jane. 1999. "Should Blacks Represent Blacks and Women Represent Women? A Contingent 'Yes.'" Journal of Politics 61 (August): 628-57.
McAdam, Doug. 1983. "Tactical Innovation and the Pace of Insurgency." American Sociological Review 48: 735-54.
Minow, Martha. 1990. Making All the Difference. Ithaca, NY: Cornell University.
Murray, Charles. 1984. Losing Ground: American Social Policy, 1950-1980. New York: Basic Books.
Paolino, Phillip. 1995. "Group-Salient Issues and Group Representation: Support for Women Candidates in the 1992 Senate Elections." American Journal of Political Science 39 (May): 294-313.
Phillips, Anne. 1995. Politics of Presence. New York: Clarendon.
Phillips, Anne. 1998. "Democracy and Representation: Or, Why Should It Matter Who Our Representatives Are?" In Feminism and Politics. Oxford: Oxford University, 224-40.
Pitkin, HannaFenichel. 1967. The Concept of Representation. Berkeley: University of California.
Piven, Francis Fox, and Richard Cloward. 1979. Poor People's Movements: Why They Succeed, How They Fail. New York: Vintage.
Plotke, David. 1997. "Representation is Democracy." Constellations 4 (November 1): 19-34.
Rasiniski, Kenneth. "Commentary: The Study of Values." In Citizens and Politics: Perspectives from Political Psychology, ed. James Kuklinski. New York: Cambridge University.
Rawls, John. 1971. Theory of Justice. Cambridge, MA: Belknap Press of Harvard University.
Raz, Joseph, and Avishai Margalit. 1994. "National Self-Determination." In Ethics in the Public Domain: Essays in the Morality of Law and Politics, ed. Joseph Raz. New York: Oxford University.
Reed, Aldoph. 1999. Stirrings of the Jug. Minneapolis: University of Minnesota.
Sapiro, Virginia. 1981. "When Are Interests Interesting?" American
Political Science Review 75 (September): 701-21. Schwarz, John E., and L. Earl Shaw. 1976. The United States
Congress in Comparative Perspective. New York: Holt, Rinehart
and Winston.
Sowell, Thomas. 1984. Civil Rights: Rhetoric or Reality"} New York: Quill.
742
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
American Political Science Review
Vol. 96, No. 4
Stasiulis, Daiva. 1993. "Authentic Voice: Anti-Racist Politics in Canadian Feminist Publishing and Literary Production." In Feminism and the Politics of Difference, ed. S. Gunew and Anne Yeatman. Sydney: Westview.
Stoker, Laura. 2001. "Political Value Judgments." In Citizens and Politics: Perspectives from Political Psychology, ed. James Kuklinski. New York: Cambridge University.
Swain, Carol M. 1993. Black Faces, Black Interests: The Representation of African Americans in Congress. Cambridge, MA: Harvard University.
Tarrow, Sidney. 1994. Power in Movement. Cambridge: Cambridge University.
Thomas, Sue. 1991. "The Impact of Women on State Legislative Policies." Journal of Politics 53 (November): 958-76.
Tilly, Charles, Louise Tilly, and Richard Tilly. 1975. The Rebellious Century, 1830-1930. Cambridge, MA: Harvard University.
Truman, David. 1951. The Governmental Process. New York: Knopf. Wilkerson, Margaret, and Jewell Gresham. 1993. "The Racial-
ization of Poverty" In Feminist Frameworks. 3rd ed. Boston:
McGraw-Hill.
Williams, Melissa. 1998. Voice, Trust, and Memory: Marginalized
Groups and the Failings of Liberal Representation. Princeton, NJ:
Princeton University. Wilson, William Julius. 1980. The Declining Significance of Race:
Blacks and Changing American Institutions. 2nd ed. Chicago:
University of Chicago. Young, Iris Marion. 1990. Justice and the Politics of Difference.
Princeton, NJ: Princeton University. 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. Young, Iris Marion. 2000. Inclusion and Democracy. Oxford: Oxford
University.
This content downloaded from 67.66.218.73 on Sun, 27 Dec 2015 07:31:10 UTC All use subject to JSTOR Terms and Conditions
743