In 1995, the fourth and most influential world conference on women delivered the Beijing Declaration, calling for “women's empowerment and their full participation on the basis of equality in all spheres of society, including participation in the decision-making process and access to power.” The idea that women needed not just the equal right to vote and participate in politics, but “full participation on the basis of equality,” had been gathering force for a number of years. Activists around the world had long challenged the underrepresentation of women in legislatures and decision-making assemblies. Political parties in the Scandinavian countries had been experimenting with voluntary gender quotas from the early 1980s. In the course of the 1990s, a number of countries, particularly in Latin America, made it a legal or constitutional requirement that something had to be done. Votes for women were the key demand at the end of the nineteenth century; parity of representation was the new demand by the end of the twentieth.

For those of us exploring the normative arguments that might underpin this political development, Hanna Pitkin's The Concept of Representation was both inspiration and foil (Reference Pitkin1967). It was inspiration because it forced us to think more carefully about the meanings we were attributing to “under-representation.” It was also, however, foil because it seemed so discouragingly critical of descriptive representation. It figured, therefore, as the position that had to be argued down.Footnote ¹

Rereading the book now, I see more clearly than I did then how thoroughly imbued the arguments are with antipaternalism and antielitism, and my earlier perception of it—as primarily a critique of descriptive representation—feels far from the truth. In making her distinction between descriptive and substantive representation, Pitkin stresses the importance of having representatives who use their discretion to judge and act, and she makes many telling points against writers who too exclusively focus on the composition of the legislature, or who look to descriptive “mirror” representation as the guarantor of better democracy. But she never loses sight of the key insight from the mandate school: the idea that representatives who persistently act against the declared wishes of their electorate cannot be said to engage in representation. Her views on this are grounded in the democratic assumption “that normally a man's wishes and what is good for him will coincide” (156) and that it is ultimately the electors who are the best judge. “We are individualists and democrats and relativists in our thinking, not quite content to tell a man what is in his interests without any regard to his wishes” (159). Even, then, in seeming to disparage the preoccupation with representatives who mirror our own views, Pitkin sets clear limits to what counts as representation. It is not, in her analysis, representation when people act as mere messengers; nor is it, on the other side, representation when they take it on themselves to make their decisions without reference to those who elect them. She carves a thoughtful middle way between our desire for creative leadership and our concern that representatives act for us.

But the middle way she offers us, wise as it is, is made easier because she does not address the concerns that have primarily shaped recent work on representation, the concerns about exclusion. It is notable, for example, that when Pitkin discusses descriptive representation, she elides debates about whether all shades of political opinion should be proportionately represented in a legislature with debates about whether to mirror a society's class or gender composition. She mostly concentrates on the former. Descriptive representation, in her account, is choosing “a representative who shares our values and commitments” (213), and it means attaching more weight to whether the legislature reflects the distribution of those values and commitments than to what it actually does. Much of her discussion then deals with classical questions in political science about whether legislatures formed under systems of proportional representation will be able to generate sufficiently stable government: whether, in other words, the focus on fair reflection will get in the way of the ability to act. Ultimately, she answers by saying that both matter and that there are limits on either side.

In more recent work on representation—perhaps most notably in the feminist work—we have tended to take the distinction between descriptive and substantive representation as a basis for thinking about the second kind of descriptiveness: not, that is, whether the political composition of the legislature should reflect the political composition of the electorate, but whether it should reflect the electorate in terms of gender, race, or class. And while Pitkin saw descriptive and substantive as different aspects of representation, much of the empirical agenda derived from these concepts has focused on assessing whether we need the first in order to achieve the second. The central question has been whether more proportionate representation along axes such as gender or race means better (substantive) representation of women's interests, or better (substantive) representation of minority concerns.

For Pitkin, there is clearly a connection between the two aspects. If a man's wishes and what is good for him normally coincide, she will presumably agree that a more diverse decision-making body, of the kind recommended in most of the gender literature, will normally provide fuller information about what people want and need. What we don't get, however, in The Concept of Representation is the sense of systemic exclusion of marginalized groups that has come to characterize more recent debate. To that extent, her project was very different.

As I read it, the empirical material on representation has not established enormously strong correlations between descriptive and substantive representation. When there are more women in a legislature, there tend to be more initiatives on employment equality or against domestic violence. Styles of doing politics change to some extent, and new issues appear on the political agenda. But on the less overtly gendered political issues, there is rarely a striking difference between the sexes, and on the most obviously gendered issues, there are pretty sharp areas of disagreement. There are gender differences, but it is not clear from the evidence so far that these are large enough to justify making descriptive representation a priority. Perhaps, then, we should follow what I take to be the Pitkin route: conclude that an assembly with no element of descriptive representation fails as representation, but leave it at that.

One riposte would be to say that descriptive representation has not achieved what was hoped for it because it is not yet descriptive enough. Certainly, few legislatures approximate the kind of equality signaled in the Beijing Declaration. For the world as a whole, women continue to constitute less than 20% of the members of legislative assemblies, and where the figures are higher, the improvements are mostly too recent to deliver clear effects. When, moreover, descriptiveness is limited to the axis of gender alone, it is unlikely to fulfill its promise, for women vary not just from men, but through their differential locations by race, sexuality, age, and class. In most instances, where measures have been introduced to achieve a more proportionate representation along characteristics of gender, race, or caste, people have voiced concerns about the relative advantages of those most likely to benefit. In India, where there has been a long history of quota initiatives to achieve the better representation of people from the most disadvantaged tribes and castes—initiatives dating back to the constituent assembly debates of the 1940s and revisited and extended at various key moments since thenFootnote ² —people talk of the problem of the “creamy layer.” They worry, that is, that an insufficiently differentiated program of affirmative action ends up benefitting those already privileged within each category. Similar concerns have been voiced in the United States in the context of measures—mainly redistricting rather than quotas—to increase the representation of African American and Latino politicians in state legislatures. It is, in addition, the standard complaint against gender quotas the world over that they provide opportunities for women who differ from the existing male politicians only in their gender but come from much the same socioeconomic and ethnocultural groups.Footnote ³

One way forward, then, would be to intensify and complicate the efforts at descriptiveness, to press not just for proportionate representation along one social axis—gender, class, race, caste—but for a proportionality that recognizes the cross-cutting intersections of the full range. Normatively, this strikes me as utterly compelling: I cannot see how one can identify the underrepresentation of women or dalits or African Americans as a problem, yet refuse on some principled ground to consider the underrepresentation of black women or female dalits or working class African Americans as a problem. But one might have practical reservations, for moving in this direction poses daunting challenges, and it is no easy matter to identify mechanisms that can adequately reflect the intersectionality of all the relevant forms of political exclusion. Faced with the difficulties of achieving complex proportionality, one might perhaps conclude (in ways that resonate with Pitkin's arguments) that putting all one's efforts into this will end up distracting us from key issues of what representatives actually do. One might also doubt that achieving even the most complex descriptive proportionality will really bring about the hoped-for changes in substantive representation. To be truly confident of this, we would have to believe the new representatives would be uniquely resistant to media distortions and robustly indifferent to the standard politician's defense that legislation can only move within limited parameters of change.

I confess that I lack this confidence, and yet I remain outraged by practices of representation that continue to exclude so much of the population. At this point, I suspect I am expressing something that goes against Hanna Pitkin's line of argument. I am saying that changing these patterns of representation matters even if it turns out to have minimal effect on what our representatives actually do. Descriptive representation is not just a tool for achieving better substantive representation. It is something that matters in and of itself.

Descriptive representation matters because of what it symbolizes to us in terms of citizenship and inclusion—what it conveys to us about who does and who does not count as a full member of society. This is not the symbolism Pitkin had in mind when she talked of symbolic representation, for she was referring to the symbolic representation of something like a nation or the unity of a people. The point, rather, is that changing patterns of representation matter because of the challenge this (re)presents to social and political hierarchies. We may additionally hope that more descriptive patterns of representation will change the direction of policy, but even if it has no such effect, it can still be an important objective.

From Pitkin's perspective, thinking of representation as a statement about citizenship might appear a misuse of the concept, might seem to forget the activity she stresses as so central to representation, might seem to turn it into a substitute for something else. I partially agree. I think that the issues we have addressed in recent years under the rubric of representation are not all, or not always, about that. They are often more about inclusion than representation, more about what it means to be recognized as a full member of one's society than how one can effect policy change. Think, for example, of Iris Marion Young's contributions to these debates, which move between notions of inclusion and notions of representation but with inclusion always as the central term.Footnote ⁴ If it is really inclusion that is at stake, then analysis of the relationship between descriptive and substantive representation and anxieties about how weak that correlation sometimes seems may be beside the point. What changes our lives are not just the policies introduced by our governments via some reasonably representative process (important as these are), but the social and political practices through which our society is organized, as reflected, among other things, in the people deemed worthy of representing us. Perhaps this is not representation per se, but it definitely matters.