THE LIMITS OF THE POLITICAL CONCEPTION

The discussion of this article is not directed to all the linguistic forms or Wittgensteinian family resemblances that might be captured by the term “representation,” it is rather on a specifically political version of the idea. A useful distinction that may be drawn here is between a concept of representation and a specifically political conception of the same idea. Ronald Dworkin relies on this distinction to suggest that we might think of arguments about the Constitution in terms of essentially contested conceptions, mutually inconsistent realizations of the same concept that involve the elements of complexity, normativity, and contestation (Dworkin Reference Dworkin1997, 7–12). Similarly, it is important to recognize that “political representation” names a specific conception (or set of conceptions) of the larger concept. That conception is contestable and may come in various forms (e.g., more or less formal modes) but they share a reference to a particular subset of the possible forms of the broader concept.

Political representation is ultimately about the aggregation and expression of preferences with respect to collective actions (Cohen Reference Cohen and Elster1998). How those processes occur is determined by the representative map, with all the limitations and exclusions that go along with privileging particular bases for constituency. Political representation is a practice that appeals to a set of legitimating principles to justify its limitations. A conception that is not sufficiently robust to generate legitimating principles for a set of representative limits is not a genuine political conception of representation.

This is not to suggest that the nature of the aggregation of preferences or collective decision-making process is unitary or unambiguous. Consider the logic behind the kind of mapping that takes place in quota-based or consociational systems. What is at work may be an attempt to create a subelectorate (assuming only women will vote for women), or to secure the expression of a worldview (women as representatives will think in a distinctly feminine way), to enhance legitimacy or to favor policy or deliberative outcomes (Phillips Reference Phillips1995; Sawer Reference Sawer2000; Schwindt-Bayer Reference Schwindt-Bayer and Mishler2005). Alternatively, descriptive representation may be taken to serve desirable goals in its effects on the constituency that is represented, by enhancing communication and trust, creating a social context for participation, or enhancing legitimacy (Mansbridge Reference Mansbridge1999; Sapiro Reference Sapiro and Phillips1998). Consociational models may be the best outcome for a deeply divided society. But in any case, the drawing of the map to include one set of categories excludes other, competing bases for identification. The features of the landscape that are included in the map define the aspects of individuals and communities that are given formal reflection in the system of political representation. The limitations of the map give it its meaning just as much as its positive content.

Exclusion, moreover, is only one form of limitation. The ability to move among identities as a political actor or to articulate competing interest claims may also be constrained by the logic of map-drawing or the strategic actions of political opponents, a phenomenon intrinsic to any formulation of political identity (Markell Reference Markell2003; Saward Reference Saward2014). Again, the acceptance of limits that exclude the articulation of a certain set of possible claims is the inescapably necessary condition of possibility for the articulation of others.

Even those elements that are included in the representative map are only imperfectly reflected. The loss of detail in the creation of a map creates an inevitable “gap” between even the most realistic depiction of an object and its experienced reality. This is yet another inescapable limit imposed by the practice of representation. Franklin Ankersmit describes this gap as the fundamental characteristic of “aesthetic” representation; “in a representative democracy all legitimate political power is essentially aesthetic.” (Ankersmit Reference Ankersmit2002, 118) The idea of such a gap, in fact, is the premise behind the institution of elections, based on the idea that elected representatives ought to be drawn from an elite subgroup of the electorate (Manin Reference Manin1997, 79). Even in the most strongly stated argument for what is sometimes called descriptive or “mirror” representation there is no presumption that the representatives and the constituents literally look alike; at most, what can be reasonably expected is that the presence of underrepresentated minorities in a legislative chamber will serve as a reminder of the fact of diversity in the larger society, not an accurate depiction of its scope (Goodin Reference Goodin2004). The same gap in representation is present in the reliance on a “representative sample,” as in experiments with deliberative polls and citizens’ juries (Fishkin and Luskin Reference Fishkin, Luskin and Saward2000). Reliance on the idea of a standard creates the danger of false essentialism (Smith 2000, 24).

While the discussion of an aesthetic gap is informative, the analogy to political representation is suggestive (or provocative) rather than direct. The purpose of artistic representation is to achieve an aesthetic effect, hence the gap between representation and subject is something to be overcome, illustrated, or manipulated. In political representation the purpose is not the aesthetic experience—not representation for its own sake or for what it illuminates—but the achievement of a set of political purposes.

The idea that the design of any system of representation involves the selection of limited aspects of the represented is not new. In Federalist 55 James Madison wrote, “if every Athenian had been a Socrates, the Athenian assembly would still have been a mob.” In Federalist 63 he wrote, “the “true distinction” between the new, American form of government and those that had gone before “lies in the total exclusion of the people, in their collective capacity, from any share” in the government (Madison et al. Reference Madison, Jay and Hamilton1788). These are not descriptions of an unfortunate arrangement necessitated by the existence of a large state, nor—as the comment about Socrates indicates—was Madison only referring to the idea of ensuring that elites would participate in governing. What Madison was describing was something else, a separation between the aspects of the citizens as represented constituents and the characteristics of the citizens tout court. As Nadia Urbinati argues, without the mediation between opinion and political action that representation provides democracy is “disfigured.” In Urbinati's analysis this disfigurement may be seen in the reduction of citizens to passive onlookers, the populist appeal to extreme and exclusionary positions, or the supplanting of rational deliberation by a politics of aesthetics (Urbinati Reference Urbinati2014). Each of these is a distorted, sham form of democracy precisely because it removes the moment of representation and the active effort to work within its limitations. Hannah Pitkin tells us that “representation” means “making present in some sense” something that is absent (Pitkin Reference Pitkin1967, 9); it is equally the case that representation means making absent in some sense what is actually present.

The gaps between representation and subject that appear in a representative map are the products of political artifice imposed on an unruly and sometimes uncooperative landscape. Where the drawing of the political map is undertaken self-consciously and deliberatively as the expression of agreed-upon political principle, the representative map becomes a profoundly democratic statement of constitutional self-government. By contrast, where the limitations of the representative map are taken for granted, the possibility of critique is severely limited. A focus on institutions and practices of political representation is the starting point for critical engagement with the legitimating principles of a democratic polity.

Moreover, making the limitations of representation explicit illuminates the choices involved in inclusion, the positive values that are ascribed to bases for identification or modes of political expression. Unlike the elements of the physical landscape, the elements represented in the political map determine for themselves what aspects of their entire selves will appear. The political representative map is the account a democratic polity gives of itself to itself.

Another term for this kind of reasoning is “constitutional.” Any constitutional precommitment establishes limitations on the discourse that is available for the expression of identities and interests in legitimate political debate (Schweber Reference Schweber2007). For example, the U.S. Constitution has been understood to exclude from political discourse certain appeals to religious purposes on the grounds that “political division along religious lines was one of the principal evils” the Constitution is designed to prevent.Footnote ³ These are not uniquely American concerns. In 1994 the Supreme Court of India upheld the act of the President of India dismissing elected governments in three provinces on grounds that the religious nature of the elected governments violated a constitutional commitment to secularism found, inter alia, in Article 51-A “Fundamental Duties,” a nonjusticiable provision declaring the duty of every citizen to “abide by the Constitution and respect its ideals and institutions.”Footnote ⁴ The European Court of Human Rights has recently declared that a system in which an elected body determines which churches will receive full state recognition violates the European Convention on Human Rights, on the grounds that “[a] situation in which religious communities are reduced to courting political parties for their votes is irreconcilable with the requirement of State neutrality in this field.”Footnote ⁵

The concerns of the U.S and Indian Supreme Courts and the European Courts of Human Rights echo the comment of Madison quoted earlier. None of the various judges in these cases were unaware that citizens have religious identities, and none was engaging in arguments about whether Rawlsian public reason permits the articulation of religious arguments as justifications for public policy. What all these jurists were saying was that in the operation of the representative process the religious identities of citizens ought not to be represented. The political map may exclude these features of the landscape because by an entirely artificial act of will the decision has been reached that it is not desirable to include them in the construction of the political landscape.

The positions of the U.S., Indian, and European courts are certainly contestable. There is a perfectly sensible argument that religious identity ought to be part of what is represented in the political process. More generally, arguments for adopting elements of descriptive representation can be understood as attempts to enhance the quality of the representative map. For multicultural theorists, the fact of limitation is a reason to question the legitimacy of constitutional exclusions per se (Benhabib Reference Benhabib1996; Tully 1995, 56; Reference YoungYoung 2000). These critiques are well taken, and they point to the necessity of exclusion as an element of any constitutional regime. The constitutional project is intrinsically limiting because it relies on representation, and a system of political representation reflects the limits of its map.

But representative limitations are not solely a function of constitutional reasoning. All forms of democratic government are based on the limitations of a representative map: the traditional sharp distinctions participatory, deliberative, and representative democracy are illusory. Imagine a utopian version of a town meeting, a system of collective decision-making that is perfectly participatory, deliberative, and consensual. In this imagine system individuals who will be affected by the outcome of a decision meet on equal terms, full information is universally shared, and deliberation continues until there is unanimity. Up to that point, it may be argued that this system reflects participation as opposed to representation. At a given moment consensus can be the basis for action, and normative evaluations of decision-making apply equally to representative and participatory democratic bodies (Rehfeld Reference Rehfeld2009, 215).

But without an institutional process of agreement leading to collective action that capacity is entirely evanescent, at the next moment the question of whether the perfect consensus persists has to be asked again, and then again at the moment after that. At each of these points there is the possibility of the withdrawal of consent. What happens then? If the answer is “then for that person there would no longer be any obligation to act in accordance with the consent of the community” then that decision was not representative, but neither was it politics (because it does not enable collective action), nor was it law (because it claims not obligation that it be respected). That is, a dimension of time is involved in any legitimate decision-making process that involves precommitments (Rubenfeld, Reference Rubenfeld2001; Weiner Reference Weiner2012). In order for the original decision to be legitimated as “law,” or the process of decision-making to have constituted something “political,” there must be a basis for asserting a binding commitment to accept a prior decision and the procedures on the basis of which reconsideration may occur.

The exclusionary effects of such commitments are the inescapable flip side of a process of inclusion; they are deliberative and procedural elements of the representative map. Roberts Rules of Order were not created as a weapon to silence voices that were otherwise being heard, they were created as a weapon against the tyranny of the loud that would prevent voices from being heard at all. As Michael Schudson puts it, “what makes conversation democratic is not free, equal, and spontaneous expression but equal access to the floor, equal participation in setting the ground rules for discussion, and a set of ground rules designed to encourage pertinent speaking, attentive listening, appropriate simplifications, and widely apportioned speaking rights” (Schudson Reference Schudson1997, 307). These restrictions on the practice of political will formation and expression define the polity through its deliberative practices.

The connection between deliberation, participation, and representation run through the possibility of self-representation. As it is conceived here, all political representation begins as self-representation because a representative deals with a limited version of a constituent, not with the complete person. The speaker who abides by rules of deliberative order accepts one set of limitations, and by participating in political discussions she accepts others as well based on accepted notions of what counts as relevant to political decision-making. Those norms may be challenged, to be sure, as in the slogan “the personal is political”Footnote ⁶ : the argument there is that the existing representative map is unduly limiting and should be altered. But that is not an argument against the existence of limitations, only about where they are drawn.

Consider again the analogy of a litigant who “represents herself” in court. It goes without saying that legal representation involves an exercise of narrative mapmaking, deciding what to leave in and what to leave out in order to create a version of events suitable to its purpose. When a litigant “represents herself” in a court, she is putting forward a certain limited narrative of herself—interests relevant to the case at hand, aspects of her worldview or character that serve her narrative—while leaving others out. The aesthetic gap appears in self-representation exactly as, and exactly because, it appears in the representation of others. Without the limitations of representation, including self-representation, every person is fully present, bringing all their attitudes, interests, preferences, momentary feelings, or long-held resentments to bear without mediation or translation into elements of a deliberately designed representative map. In Madison's words, they comprise a mob.

The role of representation in a democracy is thus nearly the opposite of what is sometimes portrayed. The inescapable element of limitation, the necessity of authorization, and the possibility of self-representation are at the core of self-government understood as the popular exercise of sovereign authority over law. In a democratic system of government, representation is what makes participation by political deliberation possible (Squires Reference Squires and Saward2000).

POSITIVIST THEORIES: THE LIMITS OF THE REPRESENTATIVE CLAIM

One implication of this argument is that a political conception of representation necessarily contains a prescriptive or normative element separate from the specific terms of a formal relationship, what Pitkin calls the duties of a representative “qua representative” (Pitkin Reference Pitkin1967, 31). Recently, some theorists have pushed against the assertion of normative definitions of representation by appealing to positivistic descriptions based on the assertion of a “representative claim.” Andrew Rehfeld draws on H. L. A. Hart's rules of recognition to propose a definition of representation that does away with normative content in the process of establishing a noncontested (or noncontestable) standard. In Rehfeld's version, “representation” occurs whenever the relevant audience believes that it does. “Representation thus depends only on an Audience recognizing that a particular claimant matches a person denoted by these rules, whatever these rules happen to be. As such, despite using the terms ‘qualified,’ ‘valid,’ and ‘appropriate,’ this account stands no matter what the content of these rules turns out to be” (Rehfeld Reference Rehfeld2009, 16).

Rehfeld's argument is motivated by his concern that no singular concept is sufficient to capture the range of possible forms of representative relationships and activities, defined in relation to “a range of normative ideals of authority, accountability, consent, interests, responsiveness, recognition, sovereignty, and policy correspondence, to name just a few.” Rehfeld argues that these different normative ideals refer to different and potentially incommensurate conceptions of political representation. “There is simply no reason to presume that these very different ideas must share some common covering theme, or that the creation of a concept to do so would be of much use to normative and empirical scholarship.” (Rehfeld Reference Rehfeld2011, 631). In part this is a debate about whether conceptual analyses of representation should “reflect fidelity to the world as it is, has been, or likely to be” as opposed to encompassing all the ways “the world might be.” (Rehfeld Reference Rehfeld2011, 632). Rehfeld's concern is that by focusing on recognized practices of representation in familiar contexts (democratic decision-making, legislation, elections) we are blinded to the forms that representation takes in less familiar contexts. In this argument, the focus is on the activity of representation rather than the claim to the status of “representative,” separating the “noun, verb, and adjective” form of the word. “So long as we use concepts that import democratic ideals. . .we risk confusing our treatment of the forms of representation with the conditions that make them democratic, legitimate, or just” (Rehfeld Reference Rehfeld2011, 635–6, 640).

Michael Saward similarly asserts that it is not the role of the political theorist to make “first order” claims about the legitimacy of representative claims, only to note the presence or absence of social indicia of legitimation as defined in the relevant cultural context (Saward Reference Saward2010, 114). Like Rehfeld, Saward advances a theory of representation in which the key act is the assertion of a representative claim (Saward Reference Saward2010, 38). Saward employs a complex set of terms to describe the claims-making process: “maker, subject, object, referent, audience.” A claimant identifies an actor as a “representative” of something or someone, to listeners whose acceptance of the claim is politically relevant. “Antiglobalization demonstrators (makers) set up themselves and their movements (subjects) as representatives of the oppressed and downtrodden (object) to the Western governments (intended audience)”; and “I (maker) claim that Bono (subject) symbolizes the needs of debt-ridden societies (object) to Western politicians (intended audience).” (Saward Reference Saward2010, 45, 60–1).

The examples of antiglobalization protestors and Bono are sympathetic. We can accept that the downtrodden and those threatened by globalization lack representation and, if asked, might be glad of the benefit of a visible spokesperson. But accepting that possibility is not the same thing as assuming its truth, and in other cases the claims are less deserving of sympathy. Benyamin Netanyahu claims to speak for American Jews in addressing Congress at the invitation of the Republican leadership, a claim that might be thought to have been accepted by its audience.Footnote ⁷ A definition that accepts Netanyahu's intervention between American Jews and their own elected representatives as a genuine instance of political representation lacks any critical force, regardless of whether—or perhaps particularly in the case where—those representatives are the “audience” ratifying the claim. And similar problems can arise when we examine cases such as lobbying organizations, “Astroturf” campaigns, or any other situation in which a representative claim is asserted without any relationship to a putative constituency. To take the most extreme possible case, it should not be a matter of discussion that the claim of slaveowners to be “represent” their slaves in a political context is meaningless, yet the infamous Three Fifths Compromise in the United States Constitution articulated just such a representative claim, which was formally accepted by its audience every time Congress seated a member from a slaveowning state.Footnote ⁸ That was a representative claim that was accepted by an audience, yet no definition of “political representation” can possibly include that claim as an example of a representative practice. But why not?

Further examination of Saward's arguments, in particular, point toward a more substantive understanding of representation as a relational claim. The account is constructivist, meaning that a representative claim may call a constituency into being rather than refer to an existing represented group (Saward Reference Saward2010, 78, 27; see Reference DischDisch 2012). Norms and cultural codes that define the terms of legitimation are not so much constraints as they are resources available for deployment and manipulation by a potential claimant (Saward Reference Saward2010, 75) The description is radically inclusivist: representative claims can be made by essentially anyone about essentially anything. (Saward Reference Saward2010, 46).

The idea that the assertion of a representative claim calls a constituency into being need not be understood in the sense of creation out of thin air. The idea may be rather one of mobilization or consciousness-raising, the effects of a representative claim as causing a potential constituency to become actually, thus coming into being as a constituency. Interpreting the idea of “calling into being” in this way, however, only emphasize the fact that to be an existing constituency is something different from having the potential to become an actual consistency, perhaps by being called into being. So long as the constituency is merely potential, the actor asserting the representative claim is declaring himself the representative of something that does not yet exist as the proper subject of political representation, as in the case of a post-Soviet nominee to the position of Poland's Ministry of Industry who declared “I represent subjects that do not yet exist” (quoted at Linz and Steppan Reference Linz and Steppan1996, 271). The nominee's claim to be a representative may be treated as metaphorical or aspirational, but it cannot be meaningfully said that he was a “representative” of a merely potential future constituency because there is no way to describe or contest the asserted relationship. A noncontestable representative claim is something like a nondefeasible proposition in science; it does not assert a relationship to any testable state of affairs. One may be a solitary revolutionary without followers, but to be a representative implies a relationship with a “represented.” A representative claim that does not assert an existing relationship is not a claim of representation at all, it is merely a declaration of hope that such a relationship will exist at some point in the future. If the potential constituency does not respond to the call to being the aspiration to that relationship will never be satisfied. This point is explored more deeply in the next section.”

Extending his argument, Saward proposes that a person asserting a representative claim creates resources for that claim's legitimation, calling into being not only the constituency but the very representative role itself. “Political actors do not simply occupy or exemplify (for example) types or forms which exist independently of their actions; types do not have a practical existence outside their enactment as roles by agents. Inherent to the act of claiming. . .is a constituting or reinforcing of the social availability of that role.” The representative is a “shapeshifter” with the ability to generate new versions of herself and of the meaning of representation itself. “[S]hape-shifting representation itself forges or enacts the particular manifestations or experiences of the (societal) systematic” (Saward Reference Saward2014, 725). Examples of “representatives” who can engage in such shape-shifting include “an election candidate, or a social movement figure, or a ‘shock-jock’ talk show host” who “positions him or herself as a subject with respect to constituents, supporters, or listeners” (Saward Reference Saward2014, 725–7). Constituencies are created, or “manipulated,” by the maker of a representative claim rather than vice versa. “[D]ifferent ‘containers’ of interests may be available for evocation by would-be representatives. . .religious groups, cultural or identity groups, social classes, even nonhuman interests and intergenerational or future generation interests.” The recognition of the multiplicity of the representative role is another argument against normative claims, which express a “perceived need to locate the one, correct definition” (Saward Reference Saward2014, 730, 724).

These are essentially positivistic accounts. The possibility that a putative representative might not be entitled to that status, or that a representative might act in a way that is outside the scope of her representation entirely is excluded. Yet at the same time, in the context of a democratic system Saward also identifies a standard that appears to involve a claim of objective legitimacy, the “acceptance of claims by appropriate constituency under reasonable conditions of judgment” (Saward Reference Saward2010, 144–5). “[D]emocratic legitimation of representation concerns ongoing acceptance of representative claims by specific appropriate constituencies. . .under certain conditions.” (Saward Reference Saward2014, 733). Legitimation occurs in “a specifically situated state of affairs whereby an instance of representation is regarded as democratically acceptable by, or is not rejected by, an appropriate constituency” (Saward Reference Saward2014, 733) But both “appropriate” and “reasonable” are normative terms that assert objective standards of legitimacy, not merely a positivistic account of legitimation. In Saward's account of representative claim there is thus a tension between a description of practices and a desire to articulate standards for the evaluation of the societal or cultural system in which those practices occur.

The reference to an appropriate constituency is in tension with some of Saward's most important examples. In the cases of antiglobalization protestors or Bono, it is difficult to describe an “intended” or any “actual” constituency. The desire to substitute “legitimation” for “legitimacy” does not escape the necessity of identifying the basis for the claim and the identification of the authorizing agents. Moreover, Saward's description of “acceptance” ambiguously includes silence (“is not rejected by”). How, again, do antiglobalization protestors, Bono, shock jocks, or social movement figures satisfy these criteria? What “ongoing acceptance” by what “actual plus intended” constituency is demonstrated to indicate the legitimacy of those representative claims?

Recognizing these problems, Christopher Lord and Johannes Pollak propose a variation on Saward's theory that they call “claims-making 2.” Using the European Union as an empirical test case, they observe that in conditions of highly complex systems of overlapping institutions of governance, the model of representation as “a more or less smooth transmission belt” in which voters’ preferences are transmitted to politicians through elections is inadequate (Lord and Pollak Reference Lord and Pollak2013, 518). They agree with Saward that a representative may call a constituency into being as well as respond to an existing aggregation of interests, that the relationship between representative and represented is mediated by a variety of intermediate institutions such as parties, and that modern democratic governance features informal as well as formal political institutions. The problem, as they note, is the lack of analytical and normative rigor in the analysis of these various phenomena. “How can we differentiate between legitimate and preposterous claims? Is it merely left to the constituency to decide about the legitimacy of claims?. . . How to determine whether they are claims to represent at all, contrary to mere claims?” (Lord and Pollak Reference Lord and Pollak2013, 521.) In response, they propose that the claim of representation carries implications that are “self-limiting and self-specifying.” The limiting element is the claim of a right to represent someone. “Instead of analyzing how the claims of some people are accepted, rejected, or ignored (claims-making 1) we should study what rights and obligations to represent others can be inferred from the manner in which claims of some people to represent others are accepted, rejected or ignored (claims-making 2)” (Lord and Pollak Reference Lord and Pollak2013, 522–3). The result is that a representative claim is always implicitly a claim to legitimacy. These claims are made, tested, and contested within an existing process, which may take place across a complex range of institutional settings but which nonetheless provides a coherent and testable set of legitimating norms.

In both Saward's model of shape-shifting representation and Lord and Pollak's discussion of rights claims the focus remains on those norms as they occur within a particular cultural and institutional setting. Yet both Saward's and Lord and Pollak's analyses acknowledge the necessity of normative standards that can be applied across contexts if a conception of political representation is to do the work that democratic theory requires of it. This emphasis is clear in Saward's description of “general conditions” such as “a plurality of sites, moments or opportunities for representative claim-making and reception,” “uncoerced equal access to subject-positional resources for claim-making,” “openness to different sorts of claims by different sorts of claimant[s],” and “the encouragement of contestation” (Saward Reference Saward2014, 734). These are not cultural- or context-specific norms of legitimation, they are democratic theoretic assertions of conditions of legitimacy that may serve as the basis for the very sort of “first order claims” that were earlier rejected. Rehfeld's focus on decision-making as the core activity of democratic representation implies a similar acceptance of normative standards in the definition, not merely the evaluation, of representative practices (Rehfeld Reference Rehfeld2011). As Lord and Pollak put it, “For all the benefits of using informal and decentered forms of representation, decisions need to pass at some point through representative processes that are formally structures to meet basic standards of liberal democracy such as public control, political equality, and public justification” (Lord and Pollak Reference Lord and Pollak2013, 526).

Yet if this is true, the basic conception of the claims-making approach is called into question. The willingness to distinguish between legitimate and illegitimate claims of representation raises the question of whether or under what circumstances a claim is not about representation at all. It is not necessary that political representation entail a single activity completely different in kind from activities engaged in by anyone else. Advocacy, deliberation, allocation of resources, and rule-making are all elements of the activity of political representation that fall under the umbrella of decision-making. But the idea that any time any one of these activities occurs political representation is taking place, regardless of context or relational claims, renders the concept meaningless. By the same token, an infinitely shapeshifting representative who claims to be all things to all people would not claim to “represent” anything at all. We need not accept static trustee-delegate binaries or formal institutional designations as the basis for discussion in order to recognize that a political conception of representation involves an invocation of roles that are defined by their limitations.

Moreover, the limitations of representation begin with participation by the represented; as noted earlier, all representation begins as self-representation. While normative standards may exist for the evaluation of representative practices, the definition of the concept remains positivistic and centered solely on the representative. The elements of claim and reception by an audience are all that is required (Saward Reference Saward2014, 732). Examples include corporate managers, participants in summit meetings, bureaucrats, lobbyists, and “assorted modes of network governance” (Saward Reference Saward2014, 732n). Earlier the examples of Bono claiming to represent the poor of Africa and Netanyahu claiming to represent world Jewry were mentioned; it is difficult to see how those cases are distinguished from participants in summit meetings (indeed, Bono and Netanyahu were acting pretty much exactly in that capacity at the moments in question). All of these qualify as political representation, in all cases the only authorization is provided by the audience (e.g., summit organizers or the U.S. Congress), none is defined by any particular relationship to an “appropriate constituency.” Rehfeld's discussion of the activity of political representation as decision-making similarly focuses “only on the representative's perspective,” to the exclusion of any role for the represented (Rehfeld Reference Rehfeld2011, 635). Lord and Pollak add the element of a normative claim of right as the basis for authorization, but they do not discuss the matter of who is entitled to evaluate or contest such claims.

Before going further there are two important potential critiques that should be addressed. It may reasonably be argued that the approaches of Rehfeld and Saward that I am describing as “positivistic” are essentially engaged in a different project from the one at work in this article. That is, the argument can be made that their theories are precisely positivistic (or perhaps “anthropological”) in the sense of describing the full range of observable practices rather than ascribing any particular normative evaluation to the claims of legitimacy that are involved. In Saward's case, one might argue that this is the distinction between the portion of his argument that focuses on the existence of a representative claim and the portion that focuses on an “appropriate constituency.”

Second, it might be argued that the focus on authorization by a constituency reveals that the project of this article is ultimately no less positivistic than those of Rehfeld and Saward, that the only difference is that I identify the existence of “representation” in the fact of acceptance by a constituency rather than the audience to whom a claim of representation is addressed.

A good analogy may be drawn here from theories of legal positivism, particularly the groundbreaking work of H. L. A. Hart referenced by Rehfeld. Hart proposed that to define “law” it was unhelpful to appeal to natural principles of justice or norms of fair practice because doing so inevitably involved the imposition of external, subjective norms to a question that is presented as an objective exercise in description and definition. Thus, Hart argued, the test for whether something is “law” has to be internal to the system in which the putative law appears, based on “rules of recognition.” Those rules, in turn, may or may not include normative elements such as “public good” or “justice” as elements of the defining characteristics of legitimate law (and claims to possess those characteristics may be contested) (Hart Reference Hart1997). As Jules Coleman helpfully summarizes the point, “[i]f the rule of recognition asserts that morality is a condition of legality, then morality is a condition of legality in that system,” but it nonetheless remains the case that “validity” is a question of “social facts - not moral argument” (Coleman Reference Coleman2001, 108, 107).

The relative merits of this approach for the comparative study of law has been a subject of debate that has spawned a large and intellectually rich literature (Leiter Reference Leiter2003; Raz Reference Raz and Alexander2001; Schweber Reference Schweber2007). Whatever the utility of a distinction between positivistic and normative approaches for law, however, the distinction dissolves in the political context of representation for three distinct reasons. As noted previously, the claim of representation entails a claim of legitimacy. It is for just this reason that politicians and activists attempt to push the boundaries of the concept as far as possible. It is possible to speak of better and worse representation just as it is possible) for anyone who is not a strict Hobbesian) to speak of good and bad law. But the point that has been emphasized throughout this discussion is that in the political context the identification of practice as “representation” at all contains a significant normative substance. If that assertion is taken to be true, then a purely analytical matter it will not do extend the term “representation” to describe any practice that is given that name. To put it another way, the argument of this article is that the definition of “political representation” is intrinsically bounded irrespective of the cultural system in which we observe it to occur. In this sense, then, the argument reflects a rejection of the totalizing relativism of the arguments I have described as “positivistic.”

The discussion of this point of criticism helps illuminate an important related question: in the sentence fragment “in which we observe it to occur,” who is “we”? The answer is “social scientists” (who are also readers of the American Political Science Review). That is, the argument is about the limitations of a conception of representation that are claimed to inhere in any useful formulation of its definition, not the limitations that result from the familiar practices or normative standards of one or another existing political culture.

At this point, the second potential critique becomes evident. If all that is at work in the discussion of the representative claim is a move from acceptance by an “audience” to authorization by a “constituency,” in what sense is this definition any more substantive than the formalistic recitation of claim and response that I have been calling into question?

The answer, I would argue, lies in the fact that in contrast to the approaches described as “positivistic” in this discussion, the argument presented in this article focuses equally on the claim of representation and what might be called the “claim of constituency.” In contrast to either the positivistic approach or the traditional approach, the argument of this article is that an analytically adequate definition of political representation requires consideration of both the relationship and the activity of representation, and a focus on both the representative and the represented. That is, I will argue that as a matter of analytical substance a claim that a relationship can be described as “representative” requires not only authorization by a constituency but also a valid “political constituency,” a conception that itself requires the same kind of explication that has been undertaken with respect to the political conception of “representation” and “representative.

What is missing in the discussion of representative claims, in other words, is any substantive consideration of constituency. The representative map is incompletely defined if it describes the limits of “representative” and “representation” but not “represented.” It cannot, after all, be the case that literally anything that can be named can be politically represented. The idea of a constituency “called into being” illustrates the point. The phrase does not literally mean that the individual constituents did not exist before the assertion of the representative claim, as in imaginary beings suddenly made real; there is no political representation of unicorns or the extragalactic poor. As Saward notes, a “flesh and blood” constituency is a prerequisite for political representation under any reading (Saward Reference Saward2010, 46, 53; compare Disch Reference Disch2012). What “calling a constituency into being” means is that prior to the assertion of the representative claim the individuals who are so affected did not yet form a “constituency,” a collective political actor with some defining characteristics having to do with the capacity for authorization. A corporate officer cannot exist in the absence of a corporation, a summit has to be organized, agencies with mandates and allocated resources are required before there can be bureaucrats. Yet in some instances, at least, the individuals who will be called into being as a constituency do not yet display the defining characteristics of that role. What are the limits of the concept of constituency, and what are the distinctive limitations that apply to a particular representative map?

THE CLAIM OF CONSTITUENCY: SHAREHOLDERS, STAKEHOLDERS, AND THE REQUIREMENT OF AUTHORIZATION

The accounts of representation based on claims-making focus almost exclusively on the activities and characteristics of the representative. To be sure, in Saward's analysis there is an intended “object”—the putative represented—but there is no discussion of their activities or characteristics. The same is true of Rehfeld's account of representation as activity and even Lord and Pollak's identification of the normative basis of a claim of representation focuses on an evaluation of the representative. The question, “Is it merely left to the constituency to decide the legitimacy of representative claims?” (Lord and Pollak Reference Lord and Pollak2013, 523) points to the need to consider when and under what conditions there is a “constituency” capable of legitimating a representative claim. The assertion of a representative claim, in fact, implies an equivalent assertion of a claim of constituency. Whether the appeal is to an existing and formally identified constituency, or seeks to call a constituency into being, the representative claim implies the additional claim that there is at least a potential category of represented who have the defining characteristics of a “constituency.”

What are those characteristics? We can begin, analytically, by recognizing elements of plausibility and relevance. The claim of constituency involves the identification of a plausible represented group. Here are two examples of implausible (or impossible) constituencies: unicorns, all persons who were wearing blue on April 19, 1986, and Halal hog butchers. The first is a category that cannot be aggregated, and the second is meaningless. In these cases the claim of representation is not merely “empirically empty” (Rehfeld Reference Rehfeld2011, 632–3), it is semantically null. Beyond being plausible, a constituency that is the subject of a claim of political representation is required to be relevant. “Persons who are fond of Beethoven” does not describe a category that is in any apparent way relevant to political representation, despite the fact that it is unquestionably a category that is capable of representation in some sense (for example, while lobbying a local classical radio station to alter its programming selections). The qualifier “apparent” is important here; a shared characteristic may be revealed to have unsuspected political relevance, and as was noted earlier the assertion of relevance is one of the ways in which existing definitions are challenged. It is possible that a fondness for Beethoven will be found to have a previously unrecognized political significance, and on that basis serve to call a constituency into being, but it remains the case that the identification of that point of political relevance is part of the process of asserting a claim of constituency. (It is also possible to advocate for changes to Sharia law that would permit the existence of Halal hog butchers, but advocacy is not the same thing as representation, a point to which I will return later.)

The substantive limitations of constituency in a given representative system derive from the significance of a specific political conception of the term. Rehfeld explores the idea of constituency within a particular political system (the United States) in his 2005 book, The Concept of Constituency (Rehfeld Reference Rehfeld2005). Challenging the unthinking assumption that geographically drawn electoral districts are the only suitable model for American democracy, he proposes random assignment to constituency groups at birth. Rehfeld uses this though experiment to explore the benefits of defining a fixed constituency as the basis of representation, including promotion of deliberation across difference to arrive at common understandings, prolonged engagement with co-constituents, fostering of shared values and expression of political will. Rehfeld's argument points to the observation that began this article, that the creation of any scheme of representation involves limitations and exclusions as the complementary function of privileging categories of identification. The claim to constitute a constituency is no less normative, nor less constitutionally significant, than the claim to act as a political representative. The question, then, is what warrants a claim of constituency capable and deserving of inclusion in the representative map? What are the limits of constituency?

To act as a constituency requires the same elements of constitutional self-constraint and self-representation that were earlier identified as essential conceptual elements of political representation. These are fundamentally deliberative capabilities; the conception of political representation collapses the sharp division between deliberative and direct democracy just as it does the division between representation and participation. As Urbinati notes, the deliberative and dialogic elements of representation preclude a purely passive electorate waiting for episodic moments of democratic action (Urbinati Reference Urbinati2008, 766–7). A constituency is required to engage in a variety of active roles: interlocutors, evaluators, bearers of opinions.

These conceptual arguments can be fleshed out in more concrete terms. A political constituency comprises a group of potential represented parties who are capable of (a) arriving at agreement that enables collective action or by disagreement impairing collective action; (b) entering into a normatively substantive relationship with representatives; (c) accepting self-imposed exclusions in the process of drawing the representative map; and (d) engaging in deliberative interactions with one another and authorizing representatives.

It is further the case, as Saward nicely points out, that a legitimate claim of constituency is one that permits contestation and a plurality of points of access. The representative claims of demonstrators, shock jocks, rock stars, and activists are not contestable in any meaningful sense precisely because they do not assert any special relationship to a putative constituency with the capacity to be politically represented. As a result, we are returned to the simple positivism of naming a social practice. But the discussion in this section began with the observation that political representation is a system of legitimized action. Equally important, Saward, like Urbinati, implicitly recognizes the point that a constituency is not a purely passive actor. To assert the claim that a group constitutes an actual or potential constituency is to assert the claim that the group has the capacity to perform that political role. The critical missing piece is authorization. Authorization, or a claim of authorization, by an “appropriate” constituency is simultaneously a foundational element of claims of both legitimation and legitimacy in any representative system. Without the element of authorization as an element of the description of a constituency, the inclusivist impulse once again becomes unhelpful. The examples of shareholder and stakeholder models of constituency demonstrate the point.

SHAREHOLDER MODELS

Saward suggests that corporate managers are the political representatives of their shareholdersFootnote ⁹ (Saward Reference Saward2014, 732n). This identification implicitly invokes a claim that the corporation, itself, is a representative body. It is not literally true that a corporation can speak, after all, any more than a church can worship or a bank can execute contracts. In all these cases the collective noun is used as a synecdoche to refer to the shareholders, members, or officers of the entity involved. The shareholder, in this model, engages in an act of self-representation in choosing to submit control over the use of her assets to the procedures of corporate decision-making while the officers of the corporation act as the shareholders’ representative. Thus we confront a not uncommon claim of nested representation: the corporation represents its shareholders, and based on that claim the corporation is entitled to political representation in its own right.

But while it is true that shareholders engage in voting that has some degree of influence over management and its policy decisions (depending on the relevant statutes and bylaws), the basis for that voting is entirely different from that of a democracy. In a democratic system the premise is of one vote per person. By contrast, shareholder voting in the standard corporate model is one share, one vote. Any such model of shareholder voting is problematic as an element of representative politics because its map is inconsistent with the majoritarian map of preference aggregation. To be sure, there are a wide variety of models of corporate governance (Lorsch Reference Lorsch2012; Stout Reference Stout2012). But the simple model of voting shares does not create a political constituency in any sense that is coherent with a democratic representative map.

The larger argument being made here is that in order for a corporate entity to satisfy the conditions for a political constituency, the representative relationship needs to be conceptually commensurate with the norms of representation reflected in the political representative map. An evaluation the representative qualities of a particular layer of representation (the corporation) requires an evaluation of the representative norms at work and testing them against the legitimating norms of the representative map. If the purpose of the map is to chart competing interests among sectors of the economy, identity groups, or clientilist networks, then the claim of “representation” may not be reached by the critique at all while other bases for challenge become available. What counts as a commensurate basis for representation depends on the representative map. This is not an attempt to impose some particular set of practices as a normative template, it is merely to observe that outright inconsistency between claims of constituency at one level (association) and another (participation in political decision-making) renders the representative claim incoherent. A system of decision-making may certainly be complex enough to encompass multiple systems of representation and multiple norms of legitimation—that is, multiple representative maps—but the differences among those systems need to be recognized and acknowledged rather than elided by the literal or metaphorical use of “shareholder.”

What is true of corporations, moreover, is true of associations generally. Many theorists look to voluntary association as alternatives and/or supplements to formal state structures of representation (Cohen and Rogers Reference Cohen and Rogers1992; Hirst Reference Hirst1997; Rossteutscher Reference Rossteutscher and Saward2000). These claims depend on a version of the shareholder model of representation insofar as members of these associations are thought to be capable of and entitled to representation through the association separately and in addition to their individual capacities, based on their share of “ownership” of the association. But if corporations are not properly a potential constituency because they are not themselves “representative” in an appropriate political sense, then other situations of similarly incommensurate forms of mapping should lead to the same conclusion. The implied critique of assertions of constituency might reach trade associations, religious organizations, and political parties. The point is that the possibility of such a normative critique depends on the proposition that genuine political representation requires some degree of consistency between representative maps operating at different levels of association and aggregation.

STAKEHOLDER MODELS

The stakeholder model is specifically intended to call constituencies into being in ways that force reconsideration of the limitations of representation. In this model, “collective outcomes are legitimate to the extent that all those subject to them have the right, capacity, and opportunity to participate in consequential deliberation about their content” (Dryzek and Niemeyer Reference Dryzek and Niemeyer2010, 14). A representative map that fails to provide such opportunity is thus presumptively illegitimate.

The problem lies in defining the phrase “subject to them.” Some theorists argue that the stakeholder model is only satisfied if all who are affected by decisions have a say in reaching them. (Saward Reference Saward2010, 145). By this logic, for example, the stakeholder model extends to “inhabitants of other countries experiencing negative ecological consequences as a result of productive or consumptive practices in the domestic political system; future generations. . .; and nonhuman species” (Mason Reference Mason1999, 74), as well as “trees, plants, and nonhuman nature” (Black Reference Black2003, 316). The argument for representation of nonhuman interests sometimes focuses explicitly on the idea of capacity, as in claims that some nonhuman animals are able to articulate interests directly (DeMello Reference DeMello2013, 6), but at other times the argument has been that “Nature” tout court can and should be represented in political dialogue (Sargisson Reference Sargisson2013, 639).

To deal first with the case of humans, it is trivially easy to identify an example of persons whose interests are affected by a decision yet have no legitimate claim to a say in the outcome: foreign invading troops. For that matter, no one seriously proposes that the people of China should have a say in the outcome of American elections despite the obvious potential consequences for their economic interests. For one thing, recognition that claims of representation occur in a specific of cultural context (Saward Reference Saward2010) cannot be squared with a model in which every person on earth is potentially part of the constituency of a local official. That does not mean that a consideration of stakeholder status is irrelevant to expanding our conceptions of political constituencies, but the stakeholder model cannot obviate the necessity of defining and accepting the limits that the claim of constituency implies.

The extension of the model to nonhumans raises additional conceptual difficulties. To speak of representing Nature tout court simply makes no sense as a definitional matter. What does it mean to represent quartz, or carbon rings, or the weather? Limiting the discussion to the biosphere does not help. Cancer presumably has an interest in seeing the greatest possible dissemination of carcinogens; algae have an interest in blooming in response to organic pollutants; the flu virus thrives best when the resistance of potential hosts is weakened. A global extinction event that radically diminishes biodiversity may create opportunities for new life forms to emerge or existing forms to expand their territory. True, this time around these are likely to be mosquitoes, beetles, and viruses rather than majestic land mammals, but who or what has a greater interest in preserving polar bears than in seeing the greatest possible expansion in the populations of Asian tiger mosquitoes?

The answer, of course, is humans, who also have readily articulable interests in a stable climate, the preservation of biodiversity, and agricultural food production. But there is a fundamental difference between speaking for Nature, as a representative, and speaking about Nature as a representative of interested humans advocating on behalf of natural objects. There are two crucial concepts at work here: authorization, and advocacy.

Earlier, I argued that all political representation begins as self-representation. Self-representation is the basis for the possibility of accepting constitutional constraints—substantive or procedural—in the formation of the representative map. The limits of representation, as I have argued throughout, define the possibility of political decision-making, which implies exclusion as the necessary complement to inclusionary structures. These observations, in turn, implies that a constituency cannot be the purely passive subject of its own generation any more than a constituency can be a purely passive recipient of rules and decisions if political representation is to be a meaningful concept. The capacity to act as a constituency through a process of self-representation, in turn, is what creates the conditions of possibility for authorization. Lord and Pollak are right to focus on the implicit claim of right involved in any representative claim. More specifically, a claim of representation is claim to have authorization to represent a constituency, and to constitute genuine political representation that constituency must be the source of that authorization, as in Saward's description of the audience for a representative claim as the “appropriate constituency.”

Rehfeld argues that authorization is not essential to political representation. “[T]he general claim that political representation necessarily involves being authorized by those you represent is just false.” For example, he says, representatives frequently represent citizens who did not vote. But here Rehfeld misses a key point. A constituent need not personally authorize each representative any more than a citizen expressly consents to every act of government. Rehfeld's own use of rules of recognition makes this point. A constituent need not personally authorize each representative or representative act to demonstrate acceptance of the claim to systemic legitimacy. Empirically as well as theoretically representation is recognized to exist based on the nature of the system of selection as a whole; studies find that voters will accept legislative outcomes in the absence of direct input if they believe that the process is a fair one (Dunn Reference Dunn2012; Marien Reference Marien2011).

Saward refers to “passive authorization” by “failing to object,” that is by silence (Saward Reference Saward2010:175n). But at a minimum, authorization based on silence should be accompanied by evidence that the constituency is aware of the representative claim and has both the capacity and opportunity to voice objections. In other words, the assertion of authorization by silence requires the prior establishment of a valid claim of constituency. As John Zumbrunnen puts it, there is a difference between “silence of contending voices” from the “collective silence of the demos” (Zumbrunnen Reference Zumbrunnen2008). Until a claim of constituency is established there can be no political representation because the test for legitimacy by authorization can neither be satisfied nor failed, it cannot be applied at all. And of course, authorization by silence is not the most extreme case: in a strictly positivistic theory, a systemically valid claim of political representation can exist even over the express objections of all members of the putative constituency.

It is important to note that in defining the conditions of possibility for political representation authorization is precedent to accountability. That is, while accountability may be necessary for normatively desirable forms of representation, authorization is required for representation in its political conception to exist at all. On the other hand, one basic mechanism of accountability is an inescapable: the possibility that authorization may be withdrawn or contested. A claim of representation that is not grounded in authorization is without any form of accountability. Moreover, such a claim cannot be effectively contested. Jane Mansbridge's category of “surrogate representation,” whereby Barney Frank, an openly homosexual member of Congress, “represents” homosexuals throughout the United States, is a case in point (Mansbridge Reference Mansbridge2003: 522; see also Mansbridge Reference Mansbridge2011: 276). Individual homosexuals or associations may publicly disclaim any connection to the representative, but that action does not reach his representative claim because that claim was never based on authorization by an identifiable constituency in the first place. Conversely, homosexuals may feel a connection to Congressman Frank, and a feeling of being “represented” in a certain way, but not the sense that satisfies the work we require of a political conception of representation. Congressman Frank himself distinguishes between his constituents and his “surrogate constituents.” The first group grants him authorization to function as a political representative, including advocacy for measures beneficial to gays and lesbians (Mansbridge Reference Mansbridge2003, 523).

A claim of constituency without authorization can neither be contested nor disproved. In an extreme case, such a claim could be made with respect to a potential constituency that would actively reject the representative claim if given the opportunity. Authorization may be systemic rather than specific, but a claim of constituency that is forced upon an unwilling represented is a sham.

On the other hand, the absence of authorization by a constituency does not in any way preclude advocacy. Advocacy, unlike representation, is not a relational claim and therefore does not imply any authorization by constituency. I can talk about future generations even though I cannot speak for them because I cannot act with their authorization. When I express their interests I am describing current people's conceptions of the future, not any future that has or necessarily every will come into being. In the Netherlands, the Party for the Animals does not represent animals, it represents people who care about the interests of animals.Footnote ¹⁰ The decision of the Spanish Parliament to extend some rights to simians was undertaken by representatives of human voters (across several parties) concerned to protect the well-being of a certain class of exceptionally capable animals.Footnote ¹¹ That did not make those Senators representatives of animals any more than advocating for a reduction in carbon emissions makes me a “representative” of the current climate system. That the constituency authorizes the representative to advocate for a third party—simians or future generations—does nothing to alter the relational claim of representation that exists between representative and constituency.

A claim to engage in advocacy is also uncontestable. One may question whether I am advocating effectively, or even whether I am advocating in good faith, but the fact that I am engaged in the activity of advocacy and therefore am enacting the role of an advocate is self-ratifying in precisely the way that the representative claim and claim of constituency are not. The possibility of contestation inheres in the implicit assertion of normative claims—a “right” to represent, in Lord and Pollak's terms—which is what creates the basis for legitimation.

Recognizing these distinctions between advocacy and representation causes us to reconsider the idea of a “pure trustee” or surrogate versions of representation. These are claims of representation that lack any relationship between representative and constituency capable of sustaining authorization, let alone the contestation or withdrawal of authorization. Similar questions are raised with respect to the idea of “calling a constituency into being.” To call for the creation of a presently nonexistence or noncapable constituency is not representation, it is a special case of advocacy for representation, addressed to the potential represented and seeking to persuade them to enter into a relationship that will permit the possibility of authorization. In this process the authorizing constituents engage in self-representation. That is, they accept the limitations imposed by participation in a formal or informal political process. Authorization legitimizes the actions of a representative with respect to a limited, politically relevant aspect of her constituents, not their entirety. And authorization legitimizes a limited, politically relevant subset of the possible behaviors in which the representative may engage, not each and everything she chooses to do while occupying her position. What aspects are defined as politically relevant, in what manner they are represented, and through what activities that representation is carried out are all contestable questions; the answers to those questions define the representative map. But in the absence of conditions of capacity and authorization, both representative and constituent claims can neither be contested nor tested.