One crucial element of Hanna Pitkin's (Reference Pitkin1967) definition of political representation has been relatively neglected in the voluminous literature her work has inspired. That element is what I will refer to as potentiality, the subjunctive idea that to be considered represented, citizens must feel that someone would defend their interests if those interests were threatened. Attention to potentiality provides a reason to value descriptive representation. Second, it illuminates the representation provided by nonelected leaders and social groups. Third, it clarifies the reciprocal links between participation and representation: persons who are participatory have better grounds to believe that their interests will be protected, and those who have such a belief participate more. Evidence to support this claimed relationship between participation and representation is presented for the U.S. case.

Potentiality appears toward the end of Pitkin's concluding, synthetic discussion of representation. Representation is a function provided by elites for the mass public, a function that requires the two “great moods” of form (institutions, rules) and substance (intentions, purpose). “Representing here means acting in the interest of the represented, in a manner responsive to them” (Pitkin Reference Pitkin1967, 209). Generally, however, one cannot match an individual's preferences with a certain legislator's preferences or actions. It is misguided to require such a match, given that, as Pitkin explains, representatives must have some freedom to act independently and no one can know with certainty the interests of another, or even at times of oneself. Representation resides in a system. Pitkin's ultimate definition hinges less on the specifics of the different “great moods” and still less upon dyadic matching, but rather requires potentiality. To be well represented, citizens must feel that someone would defend their interests were these threatened:

Representation rests not upon a simple matching of constituent with representative, but rather upon the representative's readiness to respond and the citizen's (reasonable) belief about what would happen if his or her interests were threatened or if he or she expressed a preference. To bolster these beliefs, representation also requires “machinery for the expression of the wishes of the represented” (Pitkin Reference Pitkin1967, 232) and “institutional arrangements for responsiveness to these wishes” (Pitkin Reference Pitkin1967, 233). Representation thus integrally requires potentiality, which I define as the individual's well-grounded, reasonable subjective sense that his or her interests would be defended were they expressed or were they at risk even without expression.

If representation rests upon potentiality, then descriptive representation may have a more crucial and less problematic role than Pitkin (Reference Pitkin1967) argues. A candidate's current detailed opinions may poorly predict future representation when representation means that individuals’ future interests will be protected, without present knowledge of what these might be. Those being represented are likely to feel more confidence in someone who is likely to share their needs and to believe that having shared needs will best be guaranteed by resemblance between the representative and themselves. That could be a representative malgré-lui (Miller and Stokes Reference Miller and Stokes1963) or a representative who shares an unattached fixed interest (see Pitkin's discussion of Burke Reference Pitkin1967, 168–89). In the contemporary United States, scholars often see gender, race, and ethnicity as signifying shared interests and thus commonly consider these as bases for descriptive representation. But people might disagree and consider other characteristics more important, such as religion or economic position. In fact, Pitkin (Reference Pitkin1967, 87) argues that this multiplicity of possible factors is one reason that descriptive representation is an inadequate basis for institutional design.

The desire for potential responsiveness, however, makes the average individual's turn to descriptive representation natural. Unlike the designer of an institution, the individual selects for herself or himself which characteristic or combination of characteristics seem most likely to ensure common future interests. The choice will vary with context as issues shift in salience. Individuals with the same objective characteristics may focus upon different ones for identifying a representative trusted to look out for their future interests. Further, as intersectionality suggests, individuals may require a representative to share multiple characteristics before having confidence in his or her potential responsiveness. Every combination of any identity is possible, and, as a result, descriptive representation appears to become even more muddled than Pitkin argues. It “appears to,” but it does not because individuals decide for themselves what matters, rather than institutional designers selecting the bases for representation. Admittedly, the multiple ways people can describe themselves indeed poses a challenge for empirical studies, but similar challenges have been addressed successfully. For example, researchers can ask people what characteristics they value or whether they feel there is someone who shares their interests. Thus, I argue that the persistent real-life attention to descriptive representation can be understood as a reasonable response by imperfectly informed citizens who seek representatives whom they have confidence will defend their interests should the occasion arise, a confidence they hold because they believe these representatives have similar interests to their own.

Representation and participation have multiple reciprocal relationships. A person who believes that his or her participation, were she or he moved to participate, would be paid attention (“external efficacy”) is someone who believes that there is potential for the representative to look out for her or his interests. At the same time, a person's belief that there is such a representative increases that person's level of political participation, even after taking account of the other factors known to increase activity. The first mechanism is direct. The belief in potentiality, that the system will respond, is essentially the same as a belief in one's efficacy, and efficacy has long been shown to increase political participation (e.g., Verba, Schlozman, and Brady Reference Verba, Schlozman and Brady1995, 346–47). The second mechanism operates via descriptive representation. If people consider themselves descriptively represented, then they see a representative who is “like me,” and this increases engagement, interest, and efficacy, especially for members of underrepresented groups who otherwise may look at politics as a “white man's game.”

Many studies support these claims. Participation rises when there are representatives—or candidates to be representatives—who share race, ethnicity, or gender with the potential participant. Latino turnout increases when there are Latino candidates or the likelihood of Latino candidates, with the result strongest for local elections (Barreto Reference Barreto2007; but see Leighley Reference Leighley2001). Other studies find that African American participation increases with local representation but not necessarily with representation in Congress (Bobo and Gilliam Reference Bobo and Gilliam1990; Leighley Reference Leighley2001; Tate Reference Tate2003; Gay Reference Gay2001 has mixed results). There is some evidence that women candidates and officials increase female turnout and interest in politics (Wolbrecht and Campbell Reference Wolbrecht and Campbell2007; Reingold and Harrell Reference Reingold and Harrell2010; Atkeson and Carillo Reference Atkeson and Carillo2007). State legislature composition allows a good test of the effect of descriptive representation on participation. Rocha et al. (Reference Rocha, Tolbert, Bowen and Clark2010) find that Latino and African American turnout increases when there is descriptive representation in the state legislature. Uhlaner and Scola (Reference Uhlaner and Scola2009) examine intersectional effects of gender and race or ethnicity in 2000 and 2004 and find turnout increases with more descriptive representatives in the state legislature. To be specific, non-Hispanic white women—and men—are more likely to vote where there are more white non-Hispanic women in office. Both African American men and women vote more where there are people of the same race in office, and the men turn out at especially high rates where these African American legislators are male. As argued above, we assume that this descriptive representation by gender, race, and ethnicity increases turnout in part by enhancing the sense of representation and thus increasing engagement and efficacy.

The representation literature naturally focuses on formal representation by elected officeholders, as these are the persons empowered to make binding decisions. But nonelected leaders also affect political outcomes and do so more effectively than the individual. Citizens may identify an interest group or other nonelected intermediary as the defender of their interests should the need arise. In fact, people can often more easily identify these potential defenders—defined as they are by interests and policies—than be confident of an elected official's views. Thus, a concept of representation that includes potentiality underscores the important role of nonelected elites and organizations in the system of representation. An intermediary may be any person or group whom the individual considers a potential defender of interests.

The attention to nonelected leaders suggests a third mechanism linking participation and representation. We know that recruitment—asking someone to act—increases participation (Verba, Schlozman, and Brady Reference Verba, Schlozman and Brady1995). As I have argued, individuals often connect to the larger political world via leaders who enhance a sense of duty and who, especially if recognized as looking out for the citizen's interests, can effectively recruit (Uhlaner Reference Uhlaner1989a; Reference Uhlaner1989b; Reference Uhlaner2002). A person who feels that there is a leader who will potentially defend his or her interests is more likely to respond to that leader's request for action than to one who lacks that linkage. In addition, persons who feel represented by someone are more likely to have a connection with that representative through which they can be recruited and receive benefits and sanctions.

Latinos are one group for which it is possible to explore the role of potentiality in the relationship between representation and participation. Numerous intermediaries have attempted to organize Latinos along ethnic and national origin lines. One item in the 1996 Latino National Political Survey (LNPS) taps subjective representation as conceptualized above; respondents are asked, “Is there any group or organization that you think looks out for your concerns, even if you are not a member?” About a quarter to a third of respondents (depending upon national origin and citizenship status) answer yes. The data show that respondents who feel that some group or organization looks out for their concerns are more participatory than those who do not share that feeling (Uhlaner Reference Uhlaner2002). This conclusion holds up for many activities, with some variation by national origin group (Mexican American, Cuban American, or Puerto Rican) and citizenship status. Of course, some of the same factors, notably organization membership, influence both participation and the sense of subjective representation. But even after taking account of a long list of other factors influencing political activity, subjective representation still significantly increases participation of both citizens and noncitizens. Notably, the sense of being represented dramatically increases particularized contacting, an activity that tends to be especially biased against members of minorities. These results provide an empirical basis for the theoretical claim that a sense of representation fosters political participation.

Pitkin's work still provides us with valuable guidance. I argue that the roles of subjectivity and especially of potentiality in Pitkin's idea of representation—the belief by the represented that their interests will be protected if they are threatened—have not received adequate attention. These ideas provide a means to reconcile the pervasive popular tendency to gravitate to descriptive representation with the strong scholarly and theoretical inclination toward substantive representation. Moreover, the linkage via intermediary groups can help ameliorate the tension Pitkin (Reference Pitkin2004) describes between representation and democracy by adding to local participatory democracy another avenue for enhancing participation. Representation can thereby increase democracy, not just constrict it.