Reading these essays gave me a thrill of excitement like the one I felt on hearing that the #MeToo movement had extended even into China. The ideas in this Critical Perspectives collection go much deeper into the nature of and reasons for descriptive representation than I could two decades ago. Anne Phillips (Reference Phillips1995) and Melissa Williams (Reference Williams1998), the two pioneers in this field who produced analyses far more thorough than mine, would, I think, agree with me on this. So would Iris Marion Young, whose early challenge in a talk over lunch inspired me to puzzle out my own take on the problem.
What I added to the conversation on descriptive representation was contingency. In my first book, Beyond Adversary Democracy (Reference Mansbridge1989), an empirical study of two participatory democracies, I realized that equal power as a democratic ideal had more or less value and urgency in different contexts. (There is less need to equalize power when interests do not conflict, when equal respect prevails, and when everyone can grow through taking responsibility in other ways). Later, I argued that whenever elected representatives and their constituents have relatively common interests (as I do with Elizabeth Warren, my representative in the Senate), accountability as giving an account and explaining is usually more appropriate and efficient than accountability as monitoring and sanctions (Mansbridge Reference Mansbridge, Bowens, Goodin and Schillemans2014). So, too, the need for descriptive representation depends on particular features of the context, including the importance in that context of uncrystallized interests, communicative distrust, a history of subordination, and low perceived governmental legitimacy. Who is a preferable descriptive representative will also vary according to context (Dovi Reference Dovi2002). Accordingly, the amount of energy and resources that movements and individuals should be willing to expend to increase descriptive representation should vary contingently by need.
These essays take the analysis to another level. Amanda Clayton, Diana Z. O’Brien, and Jennifer M. Piscopo, for example, make a crucial distinction between “exclusion by design,” or “bad faith” exclusion, and path-dependent exclusion. They rightly point out that my original analysis did not sufficiently stress contexts in which only power can counteract power. I would, however, not overstress my “optimism” about contemporary democratic politics. I do not expect dominant-group members to simply “learn” what they should do or to act in good faith without the influence of external incentives. In general, I am often more pessimistic than others in my cohort (e.g., alone among my friends, I expected that Donald Trump might well win the 2016 U.S. presidential election). I am also highly pessimistic about the capacity of the “pipeline measures in place” to recruit or promote more women in politics in the United States and elsewhere. A working-class woman of whatever race or ethnicity in the United States will find few venues eager to give her training in how to run for office. No law school I know has even considered providing scholarships for further training in public policy or running for office for its graduates who are members of groups that were historically legally excluded from the vote. The work of Jennifer Lawless and Richard Fox (Reference Lawless and Fox2005) and Shauna Shames (Reference Shames2017) provides little hope that women will want to run for office at the same rate as men anytime soon.
By contrast, at least outside the United States, I am far more sanguine about some kinds of quotas than I was in Reference Mansbridge1999, thanks to the work of many, including these authors and, most recently, Anna Catalano Weeks (Reference Weeks2022), who shows how, among OECD countries, legal gender quotas imposed on parties can come to be considered “natural,” how the very process of mobilizing for those quotas raises the salience of women’s issues positively, how quotas give women more power in the legislature and increase the chances of women in leadership, and how quotas have the greatest substantive effect on the political right and in relatively uncrystallized issues such as parental leave, which party leaders might otherwise avoid because those issues cut across the standard left-right party lines and therefore might divide the party. Legal quotas, of course, are easier to impose on parties in the proportional representation electoral systems characteristic of most OECD countries than in single-member plurality systems—especially nonparliamentary ones such as the United States, where the central party leadership has little power to dictate who should run in a particular district. It would be particularly hard politically, in both the United States and the European OECD countries, to go beyond gender to establish legal quotas for other disadvantaged groups. To benefit such groups and the appropriate intersectional proportions among such groups, more fluid methods may currently be more viable. Strong normative expectations, for example, seem to have produced a Democratic House Caucus that, as Clayton and coauthors point out, was only 57% white in 2020 when the percentage identifying as white in that year’s U.S. census was 64%. Yet such norms are likely to prove effective only when they align with electoral incentives, particularly among the activists. Whenever quotas are not politically practical, therefore, targeting training becomes all the more important.
Nadia Brown, Christopher J. Clark, Anna Mitchell Mahoney, and Michael Strawbridge focus on intersectionality. They use their analysis of 10,073 Twitter communications, backed by in-depth interviews with members of Congress, to demonstrate how important caucuses are to descriptive representation and how intersectionality both poses its own problems and generates its own opportunities within, between, and against those caucuses. Black women, as intersectional representatives, have opportunities to highlight their perspectives within the Congressional Black Caucus and the Congressional Women’s Caucus, but they have also needed to create their own spaces through the Congressional Caucus on Black Women and Girls and the Black Maternal Health Caucus. Such “enclave” deliberation (Mansbridge Reference Mansbridge1994) allows individuals who share some salient experiences to clarify their thinking together, come up with creative new ideas, and give one another mutual support in taking their ideas and demands out into potentially indifferent or hostile arenas. The intersection of race and gender falls into all four contexts that I singled out in 1999 as enhancing the value of descriptive representation, while the Black Maternal Health Caucus perfectly exemplifies how innovation on an uncrystallized issue depended on both the presence and the mutual communication of Black women representatives in Congress.
Magda Hinojosa and Kendall D. Funk emphasize the shared experiences of descriptive representatives and their constituents rather than only their visible characteristics (see also Allen Reference Allen2021 on shared experience). They usefully note that the individual self-presentation of representatives in different contexts can make the descriptive identity more or less salient. The contingent value of descriptive representation thus depends on the representative’s own actions. Key phrases, Twitter hashtags, stories on websites, hairstyles and dress, and symbolic actions can all heighten or dampen an implicit claim to be a descriptive representative for a particular group (see also Saward Reference Saward2014 on “shape-shifting”). Both the represented and other representatives can interpret such actions as signals that the representative will be particularly open to communication with descriptive constituents as well as acting in their interests. The value of such self-presentations varies by context, having fewer effects in uncrystallized and nonpoliticized issues, where simple shared experience might give a representative an insight on a committee or caucus without any self-awareness of where that insight came from. A representative who smokes, for example, may be more opposed to cigarette taxes and a representative whose family member has had cancer more predisposed to cancer research (Burden Reference Burden2007), thus descriptively representing smokers and those traumatized by cancer, without thinking of signaling such predispositions to the appropriate descriptive constituents.
Lara Greaves and Jennifer Curtin are absolutely right that when a treaty explicitly gives a specific right of representation to some individuals regardless of proportionality and that treaty does not abrogate human rights, in that context the treaty should prevail and members of the treaty-specified group should “always” represent that group. This general point becomes far stronger in the context of the systemic injustice of settler colonialism and the representation of Indigenous peoples. (We may leave aside as currently not realistically applicable the hypothetical contingency in which a majority of the treaty-specified group actively desires as a representative someone not originating from that group.) If a dominant group claimed that its members could be represented only by persons of its own “blood” because such representation was mandated by its unique cultural history, we would have strong normative concerns. But in the specific broader contingency that Greaves and Curtin describe, characterized by both a treaty and a larger context of deep historical injustice, the contingencies I singled out in 1999 are close to irrelevant.
Katherine Tate importantly stresses complexity. Yes, Black constituents feel better represented by Black representatives, but they do not become more trusting of the larger system. Yes, women constituents may give women representatives more favorable ratings, but they do not feel more efficacious in that larger system. Yes, Black and women representatives are more likely to pass legislation that responds to the concerns of Black and women constituents, but they often moderate these demands to get reelected. Yes, descriptive representatives can act as role models, but they cannot singlehandedly or in concert overcome embedded racism and sexism. I did not sufficiently stress these points, especially the last. The contingent value of descriptive representation can face serious structural limitations. Because of the many reasons and accumulated evidence that Tate adduces, complexity should be our future watchword.
Christina Xydias elaborates on this complexity, asking which shared experiences matter. She doubles down on contingency and context dependency, pointing out that the experiences that matter politically often depend on context and issue area. To understand whether and how a specific experience will affect a representative’s knowledge and attention, we need what practitioners of grounded political theory call a “recursive” mix of observing, theorizing, observing again, and theorizing again (Ackerly et al. Reference Ackerly, Cabrera, Forman, Johnson, Tenove and Wiener2021). All human beings have many facets, which become salient, or they may decide to make salient, in different contexts (see Young Reference Young1994 on “seriality”). As Lisa Disch (Reference Disch2021) has pointed out, descriptive representation not only reflects constituencies but also creates them through a complex interplay between representatives’ and constituents’ self-presentation and self-identification. Empirical and theoretical attention to these multiple facets and to representatives’ (and constituents’) self-presentation and self-identification is crucial, showing which intersectional identities become operative in which contexts.
Each of these essays, acknowledged too briefly here, reveals only the iceberg’s tip of all the authors’ in-depth work, which digs deeper and adds nuance to the short summary they present in this symposium. As a group, the essays brilliantly showcase our growing human understanding of both descriptive representation and the dynamic, creative aspects of representation in general.
In this rapidly expanding field, the four contingencies on which I focused in 1999 still seem relevant in most cases. I would like to pick up here only the two contexts of communicative distrust and uncrystallized interests.
In my view, normative theorists, empirical political scientists, and practitioners have failed to focus sufficiently on constituent-representative communication. In that communication, descriptive representation can be crucial. In the United States today, Black constituents are still more likely to contact Black representatives (Broockman Reference Broockman2014; Gay Reference Gay2002). Black representatives are also more likely to respond to Black citizens who ask for help, even when those citizens are outside the representative’s district (Broockman Reference Broockman2013, building on the concept of “surrogate” representation in Mansbridge Reference Mansbridge2003). Such representative-constituent communication is becoming easier by the year (incidentally making the categories of “delegate” and “trustee,” which were appropriate when representatives drove off in carriages to a relatively inaccessible capital, almost obsolete; Mansbridge Reference Mansbridge, Castiglione and Pollak2019). As Brown and colleagues and as Hinojosa and Funk point out, today representatives can reach out to constituents through email, Twitter, Facebook, and other social media, using a panoply of signals that draw attention to their descriptive representation. Representatives in the United States and Australia have also begun to access a new mechanism of “deliberative town halls,” using the internet to discuss issues with groups of randomly selected constituents (Neblo, Esterling, and Lazer Reference Neblo, Esterling and David2018). These changes make more possible an aspirational ideal of “recursive representation,” in which representatives would ideally have ongoing and mutually responsive contact with constituents (Mansbridge Reference Mansbridge, Castiglione and Pollak2019). Descriptive representation facilitates recursive representation.
Differences in class background between representatives and constituents present a growing problem in communicative distrust. Thomas Piketty (Reference Piketty and Goldhammer2020) shows how (among other causes) the decline of unions and industry in the United States, the United Kingdom, and France has resulted in the educated elite coming to dominate their left parties. Nicholas Carnes (Reference Carnes2013) documents the decline in many legislatures of members with a working-class background. Few members of democratic legislatures in the developed world today look at legislation “with the eyes of a working man” (Mill [1861] Reference Mill and Wollheim1975, ch. 3) or woman. This loss of descriptive representation may provide one reason for the rise of authoritarian populist leaders, who, although often not descriptive representatives, “speak what [some working-class people] think” (Mansbridge and Macedo Reference Mansbridge and Macedo2019, 72). Overcoming communicative distrust—whether by race, class, gender, or other form of marginalization—remains an important function of descriptive representation.
Uncrystallized interests also remain important in descriptive representation. Precisely because of the greater current descriptive representation of women, Blacks, and, intersectionally, Black women, political candidates now often take campaign positions on childcare, sexual harassment, racial bias in policing, and Black maternal health, issues previously uncrystallized. In addition, elected representatives often negotiate among themselves, formally or informally, on other issues that never reach the media or campaigns. A central lesson in good negotiation is to move beyond declared positions to underlying interests (Warren and Mansbridge et al. Reference Warren, Mansbridge, Bächtiger, Cameron, Chambers, Ferejohn, Jacobs, Knight, Naurin, Schwartzberg, Tamir, Thompson, Williams, Mansbridge and Martin2015). Each side declares that it wants X or Y, but these positions may be only means to achieving underlying interests that could perhaps be met by means less costly to the opposing party. Get to the interests, and you can possibly discover those less costly means. A good negotiating process can expose underlying interests that many representatives may not have made fully conscious even to themselves. Those interests are deeply uncrystallized. Descriptive representation helps representatives access within themselves the unarticulated interests of their constituents and weigh those interests in negotiation in ways closer to the ways their constituents would weigh them.
* * *
Let me conclude by turning to our own profession. A lunch conversation with Iris Marion Young at an American Political Science Association meeting prompted my 1999 article. We two women were talking and sharing experiences in a venue that earlier in our lives would not have welcomed us or brought us together. Today this symposium brings together 11 women and a male ally, three of them African Americans, one Mexican American, one Māori, and seven of European ancestry, writing within a deliberative community that one journal has created on politics and gender. These two instances suggest that deliberative enclaves and descriptive representation in the profession of political science can produce greater clarity of insight, intellectual as well as social support, and collective inspiration. The whole profession benefits.
So, too, the choice of topics to study often arises from a researcher’s personal experience. The study of representation probably has become so nuanced recently not only because time has passed and information accumulated, but even more because of changes in descriptive representation within the polity and within political science. Would we have become so conscious of intersectionality if Black women had not entered the profession in sufficient numbers to provide at least a small critical mass? Would we be so interested in descriptive representation if the numbers of minority groups in the profession had not grown? More subtly, would we be so interested in constituents, in relationships between constituents and representatives, and in the kinds of relationships among representatives that Brown and her colleagues are studying? At least in the last half century in the United States, women’s friendships have focused more than men’s on discussing relationships (cites in Mansbridge Reference Mansbridge, Chapman and Shapiro1993, n19). Women are also more likely to have participatory leadership styles (Eagly and Carli Reference Eagly and Carli2007). Such differences in experience and socialization help members of marginalized groups see more clearly and then open others’ eyes to previously ignored features of the political world. In short, the often subtle power that maintains the status quo and marginalizes some voices and interests also impoverishes the profession.
Because individuals’ own background experiences and socialization affect their academic interests, the academic world needs greater descriptive representation of existing human diversity, and particularly the forms of diversity that reflect systemic inequalities, so that academic knowledge and insight may focus more fairly on issues relevant to all members of the polity. Descriptive representation is not, of course, necessary for good writing and thinking. Several authors in this symposium, including myself, are not members of the groups about whom they write. But those not descriptively of a group learn from members of that group through the group members’ writing, the informal insights that arise in friendship, and the hard work of coalitional politics. Mutual communication is vital. Those descriptively in a marginalized group need their own deliberative enclaves to foster the goals of clarity, creativity, and mutual support, in academia as much as in legislatures. A neglected argument for diversity in admission to college, graduate school, and tenure is the key role of descriptive representation in the study of human experience and the need to create, as in this symposium, cutting-edge thinking on issues important to a range of humanity.
You have
Access
Reading these essays gave me a thrill of excitement like the one I felt on hearing that the #MeToo movement had extended even into China. The ideas in this Critical Perspectives collection go much deeper into the nature of and reasons for descriptive representation than I could two decades ago. Anne Phillips (Reference Phillips1995) and Melissa Williams (Reference Williams1998), the two pioneers in this field who produced analyses far more thorough than mine, would, I think, agree with me on this. So would Iris Marion Young, whose early challenge in a talk over lunch inspired me to puzzle out my own take on the problem.
What I added to the conversation on descriptive representation was contingency. In my first book, Beyond Adversary Democracy (Reference Mansbridge1989), an empirical study of two participatory democracies, I realized that equal power as a democratic ideal had more or less value and urgency in different contexts. (There is less need to equalize power when interests do not conflict, when equal respect prevails, and when everyone can grow through taking responsibility in other ways). Later, I argued that whenever elected representatives and their constituents have relatively common interests (as I do with Elizabeth Warren, my representative in the Senate), accountability as giving an account and explaining is usually more appropriate and efficient than accountability as monitoring and sanctions (Mansbridge Reference Mansbridge, Bowens, Goodin and Schillemans2014). So, too, the need for descriptive representation depends on particular features of the context, including the importance in that context of uncrystallized interests, communicative distrust, a history of subordination, and low perceived governmental legitimacy. Who is a preferable descriptive representative will also vary according to context (Dovi Reference Dovi2002). Accordingly, the amount of energy and resources that movements and individuals should be willing to expend to increase descriptive representation should vary contingently by need.
These essays take the analysis to another level. Amanda Clayton, Diana Z. O’Brien, and Jennifer M. Piscopo, for example, make a crucial distinction between “exclusion by design,” or “bad faith” exclusion, and path-dependent exclusion. They rightly point out that my original analysis did not sufficiently stress contexts in which only power can counteract power. I would, however, not overstress my “optimism” about contemporary democratic politics. I do not expect dominant-group members to simply “learn” what they should do or to act in good faith without the influence of external incentives. In general, I am often more pessimistic than others in my cohort (e.g., alone among my friends, I expected that Donald Trump might well win the 2016 U.S. presidential election). I am also highly pessimistic about the capacity of the “pipeline measures in place” to recruit or promote more women in politics in the United States and elsewhere. A working-class woman of whatever race or ethnicity in the United States will find few venues eager to give her training in how to run for office. No law school I know has even considered providing scholarships for further training in public policy or running for office for its graduates who are members of groups that were historically legally excluded from the vote. The work of Jennifer Lawless and Richard Fox (Reference Lawless and Fox2005) and Shauna Shames (Reference Shames2017) provides little hope that women will want to run for office at the same rate as men anytime soon.
By contrast, at least outside the United States, I am far more sanguine about some kinds of quotas than I was in Reference Mansbridge1999, thanks to the work of many, including these authors and, most recently, Anna Catalano Weeks (Reference Weeks2022), who shows how, among OECD countries, legal gender quotas imposed on parties can come to be considered “natural,” how the very process of mobilizing for those quotas raises the salience of women’s issues positively, how quotas give women more power in the legislature and increase the chances of women in leadership, and how quotas have the greatest substantive effect on the political right and in relatively uncrystallized issues such as parental leave, which party leaders might otherwise avoid because those issues cut across the standard left-right party lines and therefore might divide the party. Legal quotas, of course, are easier to impose on parties in the proportional representation electoral systems characteristic of most OECD countries than in single-member plurality systems—especially nonparliamentary ones such as the United States, where the central party leadership has little power to dictate who should run in a particular district. It would be particularly hard politically, in both the United States and the European OECD countries, to go beyond gender to establish legal quotas for other disadvantaged groups. To benefit such groups and the appropriate intersectional proportions among such groups, more fluid methods may currently be more viable. Strong normative expectations, for example, seem to have produced a Democratic House Caucus that, as Clayton and coauthors point out, was only 57% white in 2020 when the percentage identifying as white in that year’s U.S. census was 64%. Yet such norms are likely to prove effective only when they align with electoral incentives, particularly among the activists. Whenever quotas are not politically practical, therefore, targeting training becomes all the more important.
Nadia Brown, Christopher J. Clark, Anna Mitchell Mahoney, and Michael Strawbridge focus on intersectionality. They use their analysis of 10,073 Twitter communications, backed by in-depth interviews with members of Congress, to demonstrate how important caucuses are to descriptive representation and how intersectionality both poses its own problems and generates its own opportunities within, between, and against those caucuses. Black women, as intersectional representatives, have opportunities to highlight their perspectives within the Congressional Black Caucus and the Congressional Women’s Caucus, but they have also needed to create their own spaces through the Congressional Caucus on Black Women and Girls and the Black Maternal Health Caucus. Such “enclave” deliberation (Mansbridge Reference Mansbridge1994) allows individuals who share some salient experiences to clarify their thinking together, come up with creative new ideas, and give one another mutual support in taking their ideas and demands out into potentially indifferent or hostile arenas. The intersection of race and gender falls into all four contexts that I singled out in 1999 as enhancing the value of descriptive representation, while the Black Maternal Health Caucus perfectly exemplifies how innovation on an uncrystallized issue depended on both the presence and the mutual communication of Black women representatives in Congress.
Magda Hinojosa and Kendall D. Funk emphasize the shared experiences of descriptive representatives and their constituents rather than only their visible characteristics (see also Allen Reference Allen2021 on shared experience). They usefully note that the individual self-presentation of representatives in different contexts can make the descriptive identity more or less salient. The contingent value of descriptive representation thus depends on the representative’s own actions. Key phrases, Twitter hashtags, stories on websites, hairstyles and dress, and symbolic actions can all heighten or dampen an implicit claim to be a descriptive representative for a particular group (see also Saward Reference Saward2014 on “shape-shifting”). Both the represented and other representatives can interpret such actions as signals that the representative will be particularly open to communication with descriptive constituents as well as acting in their interests. The value of such self-presentations varies by context, having fewer effects in uncrystallized and nonpoliticized issues, where simple shared experience might give a representative an insight on a committee or caucus without any self-awareness of where that insight came from. A representative who smokes, for example, may be more opposed to cigarette taxes and a representative whose family member has had cancer more predisposed to cancer research (Burden Reference Burden2007), thus descriptively representing smokers and those traumatized by cancer, without thinking of signaling such predispositions to the appropriate descriptive constituents.
Lara Greaves and Jennifer Curtin are absolutely right that when a treaty explicitly gives a specific right of representation to some individuals regardless of proportionality and that treaty does not abrogate human rights, in that context the treaty should prevail and members of the treaty-specified group should “always” represent that group. This general point becomes far stronger in the context of the systemic injustice of settler colonialism and the representation of Indigenous peoples. (We may leave aside as currently not realistically applicable the hypothetical contingency in which a majority of the treaty-specified group actively desires as a representative someone not originating from that group.) If a dominant group claimed that its members could be represented only by persons of its own “blood” because such representation was mandated by its unique cultural history, we would have strong normative concerns. But in the specific broader contingency that Greaves and Curtin describe, characterized by both a treaty and a larger context of deep historical injustice, the contingencies I singled out in 1999 are close to irrelevant.
Katherine Tate importantly stresses complexity. Yes, Black constituents feel better represented by Black representatives, but they do not become more trusting of the larger system. Yes, women constituents may give women representatives more favorable ratings, but they do not feel more efficacious in that larger system. Yes, Black and women representatives are more likely to pass legislation that responds to the concerns of Black and women constituents, but they often moderate these demands to get reelected. Yes, descriptive representatives can act as role models, but they cannot singlehandedly or in concert overcome embedded racism and sexism. I did not sufficiently stress these points, especially the last. The contingent value of descriptive representation can face serious structural limitations. Because of the many reasons and accumulated evidence that Tate adduces, complexity should be our future watchword.
Christina Xydias elaborates on this complexity, asking which shared experiences matter. She doubles down on contingency and context dependency, pointing out that the experiences that matter politically often depend on context and issue area. To understand whether and how a specific experience will affect a representative’s knowledge and attention, we need what practitioners of grounded political theory call a “recursive” mix of observing, theorizing, observing again, and theorizing again (Ackerly et al. Reference Ackerly, Cabrera, Forman, Johnson, Tenove and Wiener2021). All human beings have many facets, which become salient, or they may decide to make salient, in different contexts (see Young Reference Young1994 on “seriality”). As Lisa Disch (Reference Disch2021) has pointed out, descriptive representation not only reflects constituencies but also creates them through a complex interplay between representatives’ and constituents’ self-presentation and self-identification. Empirical and theoretical attention to these multiple facets and to representatives’ (and constituents’) self-presentation and self-identification is crucial, showing which intersectional identities become operative in which contexts.
Each of these essays, acknowledged too briefly here, reveals only the iceberg’s tip of all the authors’ in-depth work, which digs deeper and adds nuance to the short summary they present in this symposium. As a group, the essays brilliantly showcase our growing human understanding of both descriptive representation and the dynamic, creative aspects of representation in general.
In this rapidly expanding field, the four contingencies on which I focused in 1999 still seem relevant in most cases. I would like to pick up here only the two contexts of communicative distrust and uncrystallized interests.
In my view, normative theorists, empirical political scientists, and practitioners have failed to focus sufficiently on constituent-representative communication. In that communication, descriptive representation can be crucial. In the United States today, Black constituents are still more likely to contact Black representatives (Broockman Reference Broockman2014; Gay Reference Gay2002). Black representatives are also more likely to respond to Black citizens who ask for help, even when those citizens are outside the representative’s district (Broockman Reference Broockman2013, building on the concept of “surrogate” representation in Mansbridge Reference Mansbridge2003). Such representative-constituent communication is becoming easier by the year (incidentally making the categories of “delegate” and “trustee,” which were appropriate when representatives drove off in carriages to a relatively inaccessible capital, almost obsolete; Mansbridge Reference Mansbridge, Castiglione and Pollak2019). As Brown and colleagues and as Hinojosa and Funk point out, today representatives can reach out to constituents through email, Twitter, Facebook, and other social media, using a panoply of signals that draw attention to their descriptive representation. Representatives in the United States and Australia have also begun to access a new mechanism of “deliberative town halls,” using the internet to discuss issues with groups of randomly selected constituents (Neblo, Esterling, and Lazer Reference Neblo, Esterling and David2018). These changes make more possible an aspirational ideal of “recursive representation,” in which representatives would ideally have ongoing and mutually responsive contact with constituents (Mansbridge Reference Mansbridge, Castiglione and Pollak2019). Descriptive representation facilitates recursive representation.
Differences in class background between representatives and constituents present a growing problem in communicative distrust. Thomas Piketty (Reference Piketty and Goldhammer2020) shows how (among other causes) the decline of unions and industry in the United States, the United Kingdom, and France has resulted in the educated elite coming to dominate their left parties. Nicholas Carnes (Reference Carnes2013) documents the decline in many legislatures of members with a working-class background. Few members of democratic legislatures in the developed world today look at legislation “with the eyes of a working man” (Mill [1861] Reference Mill and Wollheim1975, ch. 3) or woman. This loss of descriptive representation may provide one reason for the rise of authoritarian populist leaders, who, although often not descriptive representatives, “speak what [some working-class people] think” (Mansbridge and Macedo Reference Mansbridge and Macedo2019, 72). Overcoming communicative distrust—whether by race, class, gender, or other form of marginalization—remains an important function of descriptive representation.
Uncrystallized interests also remain important in descriptive representation. Precisely because of the greater current descriptive representation of women, Blacks, and, intersectionally, Black women, political candidates now often take campaign positions on childcare, sexual harassment, racial bias in policing, and Black maternal health, issues previously uncrystallized. In addition, elected representatives often negotiate among themselves, formally or informally, on other issues that never reach the media or campaigns. A central lesson in good negotiation is to move beyond declared positions to underlying interests (Warren and Mansbridge et al. Reference Warren, Mansbridge, Bächtiger, Cameron, Chambers, Ferejohn, Jacobs, Knight, Naurin, Schwartzberg, Tamir, Thompson, Williams, Mansbridge and Martin2015). Each side declares that it wants X or Y, but these positions may be only means to achieving underlying interests that could perhaps be met by means less costly to the opposing party. Get to the interests, and you can possibly discover those less costly means. A good negotiating process can expose underlying interests that many representatives may not have made fully conscious even to themselves. Those interests are deeply uncrystallized. Descriptive representation helps representatives access within themselves the unarticulated interests of their constituents and weigh those interests in negotiation in ways closer to the ways their constituents would weigh them.
* * *
Let me conclude by turning to our own profession. A lunch conversation with Iris Marion Young at an American Political Science Association meeting prompted my 1999 article. We two women were talking and sharing experiences in a venue that earlier in our lives would not have welcomed us or brought us together. Today this symposium brings together 11 women and a male ally, three of them African Americans, one Mexican American, one Māori, and seven of European ancestry, writing within a deliberative community that one journal has created on politics and gender. These two instances suggest that deliberative enclaves and descriptive representation in the profession of political science can produce greater clarity of insight, intellectual as well as social support, and collective inspiration. The whole profession benefits.
So, too, the choice of topics to study often arises from a researcher’s personal experience. The study of representation probably has become so nuanced recently not only because time has passed and information accumulated, but even more because of changes in descriptive representation within the polity and within political science. Would we have become so conscious of intersectionality if Black women had not entered the profession in sufficient numbers to provide at least a small critical mass? Would we be so interested in descriptive representation if the numbers of minority groups in the profession had not grown? More subtly, would we be so interested in constituents, in relationships between constituents and representatives, and in the kinds of relationships among representatives that Brown and her colleagues are studying? At least in the last half century in the United States, women’s friendships have focused more than men’s on discussing relationships (cites in Mansbridge Reference Mansbridge, Chapman and Shapiro1993, n19). Women are also more likely to have participatory leadership styles (Eagly and Carli Reference Eagly and Carli2007). Such differences in experience and socialization help members of marginalized groups see more clearly and then open others’ eyes to previously ignored features of the political world. In short, the often subtle power that maintains the status quo and marginalizes some voices and interests also impoverishes the profession.
Because individuals’ own background experiences and socialization affect their academic interests, the academic world needs greater descriptive representation of existing human diversity, and particularly the forms of diversity that reflect systemic inequalities, so that academic knowledge and insight may focus more fairly on issues relevant to all members of the polity. Descriptive representation is not, of course, necessary for good writing and thinking. Several authors in this symposium, including myself, are not members of the groups about whom they write. But those not descriptively of a group learn from members of that group through the group members’ writing, the informal insights that arise in friendship, and the hard work of coalitional politics. Mutual communication is vital. Those descriptively in a marginalized group need their own deliberative enclaves to foster the goals of clarity, creativity, and mutual support, in academia as much as in legislatures. A neglected argument for diversity in admission to college, graduate school, and tenure is the key role of descriptive representation in the study of human experience and the need to create, as in this symposium, cutting-edge thinking on issues important to a range of humanity.