In his recent survey of the legitimacy claims of unelected representatives, Saward (Reference Saward2008) opens with a quote from Bono, and so shall weFootnote 1:
I represent a lot of people [in Africa] who have no voice at all. . . . They haven't asked me to represent them.
Nobody elected Bono, he is not formally accountable to anybody, and most of the people he claims to represent have no idea who he is or what he proposes. Nevertheless, his representation claim makes some sense. It makes most sense not in terms of representing African people, nor in terms of representing a place called Africa, but rather in terms of representing a discourse of Africa. “Africa” as constructed in this discourse may bear some relation to people and places, but more important is that it constructs them in a particular kind of way: as victims of an unjust world and the caprice of nature, lacking much in the way of agency themselves, with claims on the conscience of the wealthy.Footnote 2 These claims stop at a better deal within the existing world system, falling short of structural transformation. A cynic might also see a place for celebrity and conspicuous charity in the discourse. This discourse is transnational, may be only weakly present in Africa itself, and is generally only one among several or many discourses that particular individuals who engage it (be it at G8 meetings or live8 concerts) subscribe to.
Discursive representation does, then, already happen, although as our invocation of Bono suggests, it is not necessarily done without controversy. Bono himself might insist he is representing real people, not a discourse. His critics might accept that he is indeed representing a discourse, but not one that actually benefits people in Africa. In this article we make the case for representing discourses as an integral aspect of democracy, especially deliberative or discursive democracy. We link discursive representation to theories of deliberative democracy that emphasize the engagement of discourses in existing institutions of government and the broader public sphere, and those that ponder the design of deliberative institutions as part of the architecture of government. Discursive representation is one way to redeem the promise of deliberative democracy when the deliberative participation of all affected by a collective decision is infeasible. We show how to organize representation in a Chamber of Discourses and how to evaluate representation practices in the more informal interplay of discourses in the public sphere. We draw contrasts with more conventional notions about representing individuals and groups, and identify discursive counterparts to concepts of authorization and accountability that figure in most accounts of representation. Representation is conventionally defined as “substantive acting for others,” in Pitkin's (Reference Pitkin1967) terms; “others” may be captured in terms of the discourses to which they subscribe. Whether discourses are represented by particular persons is an open question. We show not only how to designate discursive representatives, but also demonstrate less tangible ways in which discourses can find representation. We do not claim that representation of discourses is always preferable to that of individuals, just that it is different, sometimes feasible when the representation of persons is not so feasible (especially in transnational settings lacking a well-defined demos), and, on some criteria and in some settings, may do better.Footnote 3
A discourse can be understood as a set of categories and concepts embodying specific assumptions, judgments, contentions, dispositions, and capabilities. It enables the mind to process sensory inputs into coherent accounts, which can then be shared in intersubjectively meaningful fashion. At a basic level, any political discourse will normally feature an ontology of entities recognized as existing or relevant. Among these entities, some (e.g., individuals, social classes, groups, or states) will be ascribed agency, the capacity to act, while in competing discourses the same entities will be denied agency (e.g., liberal individualists deny the agency of classes). For those entities recognized as agents, some motives will be recognized, others denied. So, for example, administrative discourses recognize the agency of managers motivated by public interest values, whereas market liberal discourses ascribe to administrators only rational egoism. Any discourse will also contain an account of the relationships taken to prevail between agents and others. So, economistic discourses see competition as natural, whereas feminist discourses would see the possibility of cooperation while recognizing pervasive patriarchy. Finally, discourses rely on metaphors and other rhetorical devices. So, a “spaceship earth” metaphor is central to some environmental discourses, whereas horror stories about “welfare queens” and the like are central to individualistic conservative discourses on social policy (Alker and Sylvan Reference Alker and Sylvan1994; Dryzek and Berejikian Reference Dryzek and Berejikian1993).
Discourses enable as well as constrain thought, speech, and action.Footnote 4 Any discourse embodies some conception of common sense and acceptable knowledge; it may embody power by recognizing some interests as valid while repressing others. However, discourses are not just a surface manifestation of interests because discourses help constitute identities and their associated interests. The relevant array of discourses depends on the issue at hand (although some discourses can apply to a number of different issues) and can evolve with time. For example, when it comes to economic issues, relevant discourses might include market liberalism, antiglobalization, social democracy, and sustainable development. When it comes to international security, pervasive discourses might include realism, counterterror, Islamic radicalism, and neoconservatism. Discourses do not constitute the entirety of nonindividual political phenomena that may demand representation. In particular, discourses should not be confused with groups defined by ascriptive characteristics such as race, class, or gender; coalitions of actors who may favor a policy for different reasons; interests, which although they may be constituted by discourses, can also exist independent of discourse; interest groups, which have a tangible organization that discourses lack; or opinions on particular issues, which may be embedded in particular discourses, but need not be.
WHY REPRESENT DISCOURSES
Given that other modes of representation already exist, why might discursive representation be attractive? We begin our argument through reference to the rationality of systematically involving multiple discourses in collective decision. We then turn to an ontological justification of the priority of discourses, grounded in the discursive psychology of a world of fractured individual commitments. This account enables an ethical argument that even the individual autonomy prized by liberals can be promoted by representing the multiple discourses each individual inhabits. We then show that discursive representation is especially appropriate when a well-bounded demos is hard to locate, and helps realize the promise of contemporary theories of deliberative democracy.
Rationality
In a long tradition encompassing, among others, J.S. Mill, John Dewey, Karl Popper (Reference Popper1966), and Charles Lindblom (Reference Lindblom1965), democracy is seen as more rational in the production of collective outcomes than its alternatives. It provides opportunities for policy proposals to be criticized from a variety of directions, both before and after their implementation, thus providing the ideal setting for systematic trial and error in policy making. Democracy is, in Mill's terms, a “Congress of Opinions.” The key consideration here is that all the vantage points for criticizing policy get represented — not that these vantage points get represented in proportion to the number of people who subscribe to them. When it comes to representing arguments, proportionality may actually be undesirable because it can pave the way to groupthink and the silencing of uncomfortable voices from the margins or across divides. Sunstein's (Reference Sunstein2000) deployment of social psychological findings on group polarization show that if members of a group (e.g., a jury) start with an inclination in one direction, deliberation will have the effect of moving the average position in the group toward an extreme version of that inclination. If a substantial majority of the population lean in one direction, proportionality in their representation in the forum may produce this movement to an extreme. Thus, it is important from the point of view of responsiveness to the initial distribution of positions, let alone collective rationality, to have countervailing discourses well represented in the forum at the outset to check this polarizing effect.
For policy-making rationality, then, all relevant discourses should get represented, regardless of how many people subscribe to each. Rationality may even benefit from the presence of a vantage point to which nobody subscribes; such was presumably the rationale for the use of a “Devil's Advocate” when evaluating cases for sainthood in the Catholic Church (which is, of course, not a paragon of democracy in any other sense).Footnote 5
Now, it is one thing to ask that for the sake of rationality all vantage points, perspectives, or viewpoints get represented more or less equally in a forum, but quite another to ask that all relevant discourses get represented. Our justification here is that discourses have a solidity that perspectives do not. Furthermore, discourses can be measured and described (we explore methods later), whereas perspectives can be more elusive. In contemplating the representation of perspectives, Young (Reference Young2000, 143–44) solves this elusiveness problem by assuming that “to the extent that persons are positioned similarly in those [social] structures, then they have similar perspectives,” such that analysis of social structure can guide the selection of representatives. In contrast, we are open to discourses having a force independent of, and possibly prior to, social structure. Discursive representation is a conceptually simpler matter than the complex representations of perspectives, interests, opinions, and groups that are the ingredients of what Young calls “communicative democracy.”
An Ontological Justification
What are the key entities that populate the political world and merit representation? In the liberal tradition, the answer would be “individuals.” However, as Castiglione and Warren (Reference Castiglione and Warren2006, 13) point out,
. . .from the perspective of those who are represented, what is represented are not persons as such, but some of the interests, identities, and values that persons have or hold. Representative relationships select for specific aspects of persons, by framing wants, desires, discontents, values and judgments in ways that they become publicly visible, articulated in language and symbols, and thus politically salient.
Thus, the whole person cannot be represented (see also Young Reference Young2000, 133). Which “aspects of persons” merit representation, and what happens when they point in different directions? This question has received a number of analytical treatments. Sagoff (Reference Sagoff1988) distinguished between the “consumer” and “citizen” preferences of individuals; so (to use one of his examples), the same individuals who would as consumers love to use a ski resort will as citizens oppose its construction in a wilderness area. Sagoff resolves the problem by asserting the superiority of politics and citizen preferences over markets and consumer preferences; economists wedded to contingent valuation would disagree, as would market liberals. Goodin (Reference Goodin, Elster and Hylland1986) speaks of “laundering preferences” before they are ready to be put into collective choice processes. When it comes to elections, Brennan and Lomasky (Reference Brennan and Lomasky1993) argue that the very fact that any one person's vote is almost always inconsequential releases voters' “expressive preferences” as opposed to their material self-interest in deciding whom to vote for. Thus, in choosing whether to emphasize voting systems or markets, we also choose the relative weight of individuals' expressive preferences and their material interests. Expressive preferences might involve ethics or identity politics.
We prefer a less analytical and more empirical treatment of what Elster (Reference Elster1986) calls the “multiple self.” Speaking in terms of preferences is unduly restrictive, for aspects of the multiple self may not be reducible to preferences (and the instrumental form of rationality it implies). This question can be illuminated by discursive psychology (Edwards and Potter Reference Edwards and Potter1992). Discursive psychology takes seriously the Wittgensteinian notion of language games as the framework in which cognition is possible. The mind itself lies at the intersection of such games: “I inhabit many different discourses each of which has its own cluster of significations” (Harré and Gillett Reference Harré and Gillett1994, 25). Subjectivity is, then, multifaceted: “most of us will fashion a complex subjectivity from participation in many different discourses” (25). This is not a matter of an autonomous self picking and choosing across discourses because the multifaceted self is constituted by discourses; we cannot think outside discourses because they also enter and help constitute the mind. However, the very fact that each individual engages multiple discourses provides some freedom for maneuver, such that “fluid positionings instead of fixed roles” are possible (36), which is crucial when it comes to the possibilities for the reflection that is central to deliberative and democratic interaction. Thus, persons are not simply bundles of discourses; autonomous individuals can reflect across the discourses they engage, even as they can never fully escape their constraints.
The individual selves prized by liberals can, then, be quite fractured by the discourses that the individual engages. Group representation is no less problematic in this light. Group representation is normally tied to descriptive representation, where “blacks represent blacks and women represent women,” as Mansbridge (Reference Mansbridge1999) puts it. There may be more than one discourse relevant to black interests or women's interests, which a unitary framing of that group's interests will not capture. Some of its advocates recognize the need to “pluralize group representation,” but then face indeterminacy in how far to go in representing different subgroups (Dovi Reference Dovi2002, 741). This indeterminacy can be ameliorated (although not eliminated) if we can show how the range of relevant discourses can be described.
Ethics
The liberal argument for the representation of individuals has an ethical as well as an ontological aspect, on the grounds that individuals are capable of self-government, and the repositories of moral worth. There are nonliberal arguments in which groups, social classes, and communities have similar moral standing, but what about discourses? There is actually no need to give discourses any moral standing that is not reducible to that of the individuals who subscribe to them. Yet, there is still a moral (as opposed to ontological) argument for discursive representation. Once we accept the insight from discursive psychology that any individual may engage multiple discourses, it is important that all these discourses get represented. Otherwise, the individual in his or her entirety is not represented. Discursive representation may, then do a morally superior because more comprehensive job of representing persons than do theories that treat individuals as unproblematic wholes. Liberals might reply that each individual should manage the demands of competing discourses him- or herself prior to seeking representation because an autonomous person is one who chooses not just among options, but also among reasons for that choice (Watson Reference Watson1975). Yet, demanding this management prior to representation may paradoxically disrespect individual autonomy, if it requires the individual to repress some aspect of his- or herself. For example, a government employee may choose to vote for party X because he or she fears that party Y, whose platform he or she otherwise prefers on moral grounds, will undertake budget cuts that endanger his or her job. Their moral preferences are repressed in their voting choice. Discursive representation would ensure both aspects of the self of this government employee get represented in subsequent deliberations.
The Decline of the Demos
Rationality, ontology, and ethics can justify discursive representation in any time or place. We now introduce some developments in contemporary politics that reinforce the case. Democratic theory has traditionally been tied closely to the idea of a well-bounded demos: no demos, no democracy. Correspondingly, in Pitkin's (Reference Pitkin1967) classic statement about representation, the definition of the people is logically prior to contemplation of their representation. Representative democracy in this light requires a precisely bounded citizenry, normally defined by membership of a political unit organized on a territorial basis, which then elects representatives. However, today's world is increasingly unlike this. Authority increasingly escapes the sovereign state, to be located in, or diffused throughout, the global system. Sometimes authority is transferred to an international governmental organization such as the World Trade Organization (W T O). When a tangible organization such as the WTO exists, it is possible to imagine global elections to its board, but impossible to institute them in any feasible future. The most that can be hoped for is the representation of states, which entails representation of peoples at one very considerable remove (and, of course, not all states are internally democratic). Currently, the WTO runs according to a single discourse, that of market-oriented neoliberalism. A more democratic WTO would be responsive to a broader range of discourses, such as the counterdiscourses constructed by antiglobalization activists.
Political authority is also increasingly diffused into informal networks made up of governmental and nongovernmental actors, be they businesses, professional associations, unions, nongovernmental organizations (NGOs), social movements, or individual activists (Rhodes Reference Rhodes1997). Networked governance is almost impossible to render accountable in standard democratic terms because there is often no unique demos associated with a network. This is especially true when networks cross national boundaries. If networks cannot be held formally accountable to any well-defined demos, we have to look in other directions to render them accountable. One way of doing this is to try to ensure that a network is not dominated by a single discourse whose terms are accepted uncritically by all involved actors in a way that marginalizes other discourses that could claim relevance. For example, the international networks of finance and capital described by Castells (Reference Castells1997) have generally been dominated by economistic discourses to the exclusion of social justice discourses. International environmental networks have often been dominated by a moderate discourse of sustainable development that by the lights of more radical green discourses is too easily accommodated to economic growth rather than effective environmental conservation.
If the demos is in decline, then Ankersmit's (Reference Ankersmit2002) contention that the process of representation itself constitutes any “people” gains in plausibility. This kind of indeterminacy can be embraced by discursive representation, under which different discourses can constitute the relevant people in different ways. So, for example, in a cosmopolitan discourse, “the people” is global; in a nationalist discourse, it is always more particular.
Discourses in Theories of Deliberative Democracy
Deliberative democracy ought to be less wedded to conventional notions of representing persons than is the aggregative kind of democracy to which it is often contrasted because it puts talk and communication at the center of democracy (Chambers Reference Chambers2003, 308). From the viewpoint of the discursive self in deliberative democracy, it may then be more important for the quality of deliberation that all relevant discourses get represented, rather than that all individuals get represented. As Mansbridge (Reference Mansbridge2003, 524) points out, in deliberative democracy there is no requirement that perspectives get “presented by a number of legislators proportional to the numbers of citizens who hold those perspectives.” Weaver et al. (Reference Weaver, Garcia, Schwarz and Miller2007) show experimentally that the “weight” of a message in the forum depends more on the frequency with which it is repeated than on the number of people who present it, a finding that further undermines any argument for proportionality in representation in communicative settings.
The account of deliberative democracy presented by Dryzek (Reference Dryzek2000) highlights the generation and engagement of discourses in the public sphere. Public opinion is then defined as the provisional outcome of the contestation of discourses as transmitted to the state or other public authority. This feature fits nicely with discursive psychology because the reflective agents who populate a deliberative democracy can be seen as negotiating the field of discourses in which they necessarily participate, with more or less competence. This conceptualization of deliberative democracy is compatible with, but more precisely connected to discourses, and so their representation, than the formulations of Habermas (Reference Habermas1996) concerning diffuse “subjectless communication” that produces public opinion and of Benhabib (Reference Benhabib and Benhabib1996, 74) concerning an “anonymous public conversation” in “interlocking and overlapping networks and associations of deliberation, contestation, and argumentation.” Benhabib, Dryzek, and Habermas all assign the public sphere a central place in the architecture of deliberative democracy. Spaces in the public sphere have proliferated along with new communications media. Democratic legitimacy is generated in the extent to which collective decisions are consistent with the constellation of discourses existing within the public sphere, in the degree to which this balance is itself under the decentralized control of reflective, competent, and informed actors (Dryzek Reference Dryzek2001). This is not the place to debate the pros and cons of these approaches; suffice it to say that there are versions of deliberative democracy for which discourses and their contestation or engagement are a central feature, begging the question of how they might be represented. There are also versions of deliberative democracy that have moved beyond the idea that effective deliberation requires a demos with a well-specified set of actors united by bonds of social solidarity. Indeed, it is in such settings that key features of deliberative democracy are now being tested (Scheuerman Reference Scheuerman, Dryzek, Honig and Phillips2006).
In Joshua Cohen's classic statement, deliberative legitimacy is to be found in all those subject to a decision participating in deliberation about its content (Cohen Reference Cohen, Hamlin and Pettit1989). However, given the impossibility of organizing participation by all those affected, another solution needs to be found in order that communication from those not in the deliberative forum itself be somehow represented inside the forum when the deliberation of all yields to the deliberation of some (Manin Reference Manin1987). Critics of deliberative democracy (Shapiro Reference Shapiro and Macedo1999; Walzer Reference Walzer and Macedo1999) have pointed out that effective face-to-face deliberation can only ever involve a handful of people. Thus, seeking democratic legitimacy via participation in deliberation by all those affected looks futile. Conventional electoral representation to reduce the number of deliberators is one solution, but elections themselves are not necessarily deliberative affairs (and proposals to make them more deliberative rarely involve anything like the deliberation of all). Furthermore, the number of legislators in a general purpose legislature is still generally too large for them all to deliberate together.
Discursive representation offers a solution to this key problem of scale that confronts deliberative democracy. The number of discourses that need representing on any issue is generally much smaller than the number of representatives in general-purpose legislatures, so it ought to be possible to constitute a small issue-specific deliberating group that contains representatives of all relevant discourses. We now ask how such small groups might be constituted formally in order to combine effective deliberation with discursive representation. Then we turn to more informal ways of securing discursive representation that resonate with accounts of deliberative democracy emphasizing engagement of discourses in a broad public sphere.
FORMALLY CONSTITUTING THE CHAMBER OF DISCOURSES
It is possible to imagine a Chamber of Discourses corresponding to more familiar assemblies based on the representation of individuals. Existing parliamentary chambers do of course feature discourses, but only unsystematically, as a by-product of electoral representation.
We have already argued that there is no need for proportionality in discursive representation. Epistemic justifications for deliberation (Estlund Reference Estlund, Bohman and Rehg1997) also suggest that the composition of the deliberative forum need not mirror that of the population at large. However, it is important to ensure that each relevant discourse gets articulate representation, and we should be wary of the “lottery of talent” introducing inequalities across discursive representatives. Having multiple representatives for each discourse ought to ameliorate at least chance factors. It may also be true that the nature of a discourse is associated with the capacity of its adherents to articulate its content. Here, deliberative democrats would stress the need for forum design to bring out the “communicative competence” of representatives. Experience with deliberative “minipublics” shows that ordinary citizens can become capable deliberators.
In thinking about the Chamber of Discourses, we must allow that in deliberation individuals reflect on the discourses they engage and can change their minds. Mansbridge (Reference Mansbridge2003, 524) suggests that when “deliberative mechanisms work well” they should select against “the least informed political positions in the polity.” It is entirely possible that particular discourses initially identified for representation in the forum will not survive deliberation unscathed, but that may not be so bad if the transformation renders the constellation of discourses more publicly defensible. Niemeyer (Reference Niemeyer2004) demonstrates this process empirically. On an environmental issue deliberated in a citizen's jury, he shows that a discourse that tried to assuage anxieties on both sides of the issue was transformed for its adherents toward a more clearly preservationist discourse. The possibility that discourses get transformed once represented does mean that discursive representation is inconsistent with a “delegate” model of representation.
We should also recognize that discourses can be transformed, or even constituted, by the very fact of their representation. Representing a previously marginalized discourse may mean that a particular category of people gets constituted as agents within the discourse. For example, the fact that the discourse of environmental justice became heard in policy-making processes in the United States in the late 1980s validated the agency in environmental affairs of low-income ethnic minority victims of pollution. The discourse of Africa associated with Bono perhaps exists mainly in the fact of its representation at high-profile international events. However, this last feature is by no means unique to discourses. As Ankersmit (Reference Ankersmit2002, 115) puts it, perhaps overstating the point, “without political representation we are without a conception of what reality—the represented—is like; without it, political reality has neither face nor contours. Without representation there is no represented.”
To constitute formally a Chamber of Discourses would require to begin a way of identifying and describing the array of relevant discourses on an issue. We would then need a way to designate representatives of each discourse (or of positions in the array of discourses). Members of the Chamber of Discourses could not be elected because then they would represent constituencies of individuals. Another option would be through random selection—as advocated, for example, in Leib's (Reference Leib2004) proposal for a fourth “popular” branch of government in the United States. The problem with random selection is that large numbers are needed to guard against the possibility that a relevant discourse might be missed. However, the larger the number of representatives, the harder it becomes for them to deliberate together. This is why large-scale processes such as deliberative opinion polls and citizens' assemblies subdivide their participants into smaller deliberative groups of no more than 20 or so each. Thus, we need a procedure better than random chance to ensure that all discourses are effectively represented in each group.
A more economical alternative would involve constituting a deliberative minipublic of around 15 to 20 citizens, the kind of number now used extensively in institutions such as citizens' juries, consensus conferences, and planning cells. The standard procedure is to begin with an initial random sample of citizens, and then target individuals with particular social characteristics—age, education, place of residence, income, ethnicity, and so forth. This is essentially a “politics of presence” kind of approach to representation (Phillips Reference Phillips1995). However, discursive representation involves (in Phillips' terms) a “politics of ideas.” There is no guarantee or even strong likelihood that people with different social characteristics will in fact represent different discourses, or that a reasonably full range of social characteristics will guarantee a reasonably full range of discourses is present in the forum. Discursive representation can improve the deliberative capacities of institutional designs featuring random selection by ensuring that a comprehensive range of discourses is present. Fortunately, there are methods available to both (a) map the constellation of discourses relevant to an issue and (b) determine which individuals best represent each discourse. We now describe some methods. These methods illustrate what is possible. Our basic argument for discursive representation does not depend on commitment to any or all of them.
SYSTEMATIC SELECTION OF DISCURSIVE REPRESENTATIVES
Davies, Sherlock, and Rauschmayer (Reference Davies, Sherlock and Rauschmayer2005) show how Q methodology can be used to recruit individuals who best represent particular arguments to deliberative mini-publics. Q methodology involves measuring an individual's subjective orientation to an issue area in terms of his or her ranking of a set of 35 to 60 statements about the issue in a “Q sort.” These statements can be keyed to the five features of discourses listed previously: ontology, agency, motives of agents, relationships, and metaphors (although this is not done by Davies, Sherlock, and Rauschmayer). For example, in a study of discourses surrounding local sustainability issues in the United Kingdom, Barry and Proops (Reference Barry and Proops1999, 342) deploy 36 statements. One of their statements about ontology is “LETS [local employment and trading systems] is a new type of economy in which sustainability is a key aspect.” A statement referring to agency is “We all have to take responsibility for environmental problems.” A statement on motives is “People are taking a short-term view: they're not thinking about the long term.” One about relationships is “You can't look at one part of the planet, because all the parts interact.”Footnote 6
The ranking process is itself reflective, so consistent with the notion that discourses can be transformed and winnowed in the process of their representation. Individuals from the subject population are asked to order the statements into a manageable number of categories from “most agree” to “most disagree.” The subject population could be several hundred individuals selected at random (Q methodologists are happy working with much smaller numbers of subjects, but the link we are trying to make here to representation means that a larger number might be required to help us find particularly good representatives of each discourse). The Q sorts so produced can then be factor analyzed; factor analysis is essentially a summary procedure that produces a manageable number of (in this case) discourses. We can compute a loading (correlation coefficient) between each discourse and each individual. Those individuals loading highest on a particular discourse will make particularly good discursive representatives—at least in the sense that they are characteristic of the discourse in question, although, of course, they can vary in how articulate they are when it comes to deliberation itself. However, there may be circumstances in which it is desirable to select more complex individuals who load on more than one discourse (as we see later).
To take an example, consider the study of political discourses in Russia in the late 1990s reported in Dryzek and Holmes (Reference Dryzek and Holmes2002, 92–113). This study identified three discourses. The first, chastened democracy, remained committed to democracy despite current political disasters. The second, reactionary antiliberalism, regretted the demise of the Soviet Union and opposed the postcommunist status quo. The third, authoritarian development, disapproves of both the Soviet Union and the postcommunist present, seeking a better economic future under a disciplined autocracy. Among the Russians interviewed, the person with the highest loading on chastened democracy (70, where 100 would indicate perfect agreement) is a public relations manager who describes herself as a liberal. The person with the highest loading (59) on reactionary antiliberalism is a teacher who describes himself as a Russian nationalist. The two people with the equal highest loading (60) on authoritarian development are a student who describes herself as a nonpartisan atheist and a construction worker who claims not to care about politics. Assuming they are articulate, these individuals would on the face of it make particularly good representatives for any forum in which representation of the discourse in question is required—within Russia, or even internationally. (At meetings of international economic organizations, it could be instructive to have somebody representing authoritarian development.
Most Q studies seek only to map discourses present. Davies, Sherlock, and Rauschmayer (Reference Davies, Sherlock and Rauschmayer2005) use Q to select participants for deliberative forums. We can designate those participants as representatives of discourses. In most theories of representation, those represented somehow authorize the representation. The method we have described seems to substitute social science for political process, with the risk of empowering an unaccountable social scientific elite. We address procedures for countering this hazard in the “Authorization and Accountability” section, but one check might be to expand the range of methods used, to which we now turn.
OTHER METHODS FOR SELECTING DISCURSIVE REPRESENTATIVES
Although we have discussed Q methodology as a particularly systematic way to identify relevant discourses and choose discursive representatives, there are other ways. When it comes to discourse identification, there exist in many issue areas enumerations of relevant discourses based on historical analysis. So, for example, for U.S. environmental politics, Brulle (Reference Brulle2000) enumerates seven discourses on the environmental side: wildlife management, conservation, preservation, reform environmentalism, deep ecology, environmental justice, and ecofeminism, along with an antienvironmental discourse of “manifest destiny.” In criminal justice policy, at least four discourses can be enumerated. One stresses the psychopathology of criminals, a second treats crime as a matter of rational choice, a third emphasizes the social causes of crime, and a fourth the social dislocation of individual offenders. Each discourse comes with a range of treatments: respectively, retribution, deterrence, social policy, and restorative justice. In the criminal justice policy area, one could imagine constituting a chamber with representatives from these different discourses. In these environmental and criminal justice examples, it is not difficult to identify individual activists, publicists, or politicians associated with each discourse, who could serve as discursive representatives.
Q is an interpretive methodology that happens to be quantitative, but other interpretive methods for discourse analysis are qualitative or ethnographic. Both in-depth interviews with individuals and focus groups could be used to map relevant discourses in an issue area. Hochschild (Reference Hochschild1981) analyzes 28 in-depth interviews of rich and poor Americans in order to map different beliefs about distributive justice and the sorts of distributive rules that should be applied to different policy areas. Despite considerable ambiguity and inconsistency among her subjects, Hochschild's analysis could be mined for discourses and their representatives. Notably, Hochschild finds six kinds of distributive rule applied by her subjects, although their application is issue-area specific. However, for example, when it comes to policy for financing schools, it would be possible to identify using her analysis an individual who subscribes to a discourse of need, one that stresses performance, and so forth. An ethnographic study that began with the intent of identifying discourses and their representatives would enable a much sharper focus.
Opinion surveys could also inform the identification of relevant discourses, although their lack of interpretive depth may mean that they have to be supplemented by other sorts of analysis. So, for example, Kempton, Boster, and Hartley (Reference Kempton, Boster and Hartley1995) combine surveys and semistructured interviews. They find a vernacular environmental discourse that appears to be shared by most ordinary people in the United States (including categories of people they targeted for explicit antienvironmental sentiments), although for some individuals in-depth interviews reveal that it is overridden by discourses that stress either employment and social justice or cynicism about the way environmental values get deployed (215). Discursive representation here would mean identifying individuals who prioritized the latter two discourses, as well as those who did not.
There are then a number of methods that could be deployed to select discursive representatives. Different methods might yield different representations, just as different electoral systems produce different configurations of political parties. Triangulation across different methods might increase our confidence in the validity of any particular representation, although it would be of little help should representations differ. However, in the latter case, there would be no problem in using different methods to pick different discursive representatives. One method might simply pick up on a discourse that another method missed. For example, opinion surveys would miss subjugated or marginal discourses that were not preconceived by the survey designer; it might take in-depth interviews or Q methodology to reveal these. We should also allow that particular discourses may only crystallize in the process of selection of their representatives. In-depth interviews might well have such an effect, especially if they have the salutary effects that psychotherapists claim.
Among alternative methods for the selection of discursive representatives, Q methodology or in-depth interviews should be used when the content and configuration of relevant discourses is weakly understood. In-depth interviews should be used to tease out discourses that have yet to crystallize fully in the understandings of any actor. Opinion surveys can be used when the content of relevant discourses is well understood and/or financial constraints suggest a low-cost method. Historical methods are appropriate when conducting interviews is impossible, too expensive, or the population from which one might select discursive representatives is highly dispersed (as in transnational affairs).
DIFFERENT SORTS OF DISCURSIVE REPRESENTATIVES
Choosing as representatives for participation in deliberation only those individuals who are strongly identified with particular discourses is not necessarily the most defensible procedure. Discursive psychology suggests that the typical individual actually has access to more than one discourse. In this light, choosing individuals who identify strongly with a single discourse might look a bit like selecting for extremism. One solution here might be to constitute two deliberating subchambers, one made up of individuals initially identifying strongly with single discourses, the other made up of individuals identified with two or more discourses. The first group might then be best at opening up the relevant range of issues, whereas the second might be better at reaching reflective judgment across discourses. Alternatively, we might decide what we actually want the deliberating group to do, and select for extremism and moderation accordingly. If the deliberating forum is akin to a jury delivering a verdict (say, a health care committee deciding whether an expensive lifesaving treatment is warranted in a particular case), we might want to select for moderation across discourses. If we want the forum to generate ideas (e.g., on a novel policy problem), we might want to select for extremism in discursive representation.Footnote 7 However, in light of the possibility of deliberation-induced change in individuals' commitments to particular discourses, and even the content of discourses, these suggestions remain speculative. Designing empirical studies to test the effects of different forum compositions along these lines would actually be quite straightforward.
It might even be useful to have a Chamber of Extremism and a Chamber of Moderation sitting in parallel. This would be analogous to the way lower and upper houses currently operate in bicameral parliaments, with the upper house expected to be a moderate house of review controlling the partisan excesses of the lower house. In practice, lower houses are themselves vulnerable to excessive moderation as parties converge on the median voter in elections, so an explicit Chamber of Extremism might actually improve the quality of debate by sharpening differences.
Another possible institutional design might involve a Chamber of Moderation adjudicating the presentations made to it by individuals strongly associated with particular discourses. Such a design would resemble the way mini-publics such as citizens' juries and consensus conferences already operate, although citizen-adjudicators in these forums are currently selected on the basis of their lack of any prior partisanship, rather than sympathy with multiple discourses.
In thinking about discursive representation, it is important to stress that discourses are not necessarily reducible to the opinions of a well-defined set of subscribers. Discursive psychology accepts, and Q methodological studies typically confirm, that any given individual may subscribe partially to several different, perhaps competing, discourses, each of which resonates with a particular aspect of the “self.” For this individual, different situations may then invoke different discourses. Discursive representation then involves representing discourses, not selves, even when we need to identify individuals to articulate the discourse in question. It is even possible that a particular discourse may find no complete resonance with any individual, although partial resonance with many, attracting minor aspects of a number of “divided selves.” How exactly might the representation of any such discourse be organized? One solution might be to find the individual or set of individuals loading most highly on this discourse, even if they load more highly on another discourse. The likelihood of any such fugitive discourse on any issue is an empirical question. However, such a discourse could conceivably represent a new understanding currently at the margins of public opinion, with the potential to become more significant in the future. It might, of course, also represent an understanding on the way out, or one that is destined to remain marginal. However, from the point of view of problem-solving rationality discussed previously, marginal discourses may still be important. Representation of marginal discourses is especially important from the point of view of democratic equality to the degree dominant discourses embody privilege and power.
DECISION AND POWER
How should decisions be reached in any formal Chamber of Discourses? A theory of representation is not a full theory of democracy, so one can imagine a variety of decision mechanisms, including voting. Consensus may be a plausible rule if the chamber is composed of a small number of individuals, each of whom can be associated with more than one discourse (so featuring moderation as defined previously), although undesirable conformity pressures may accompany small size. Consensus is less plausible as numbers increase, or to the degree each participant is strongly associated with a particular discourse, although even here we should not assume that discourses are necessarily incommensurable.Footnote 8 “Working agreements” may still be possible in which participants agree on a course of action for different reasons, but understand as morally legitimate the reasons of others (Eriksen Reference Eriksen2006). Even if they cannot agree on major issues, participants might still practice the “economy of moral disagreement” advocated by Gutmann and Thompson (Reference Gutmann and Thompson1996), what we could style an economy of discourse disagreement, searching for aspects of issues representatives can agree on. There is no justification for giving discursive representatives veto power over decisions that affect their discourse, of the sort that Young (Reference Young1990) believes should be possessed by representatives of oppressed groups.
How much power should any Chamber of Discourses possess in relation to other sorts of representative institutions, such as legislatures? Again, a theory of representation is not a complete theory of democracy, and so the idea of discursive representation cannot itself adjudicate across any competing representative claims. Discursive representation might complement the work of familiar institutions, rather than replace them. A formal Chamber of Discourses could take its place in existing institutional architecture in a variety of ways. Mini-publics deployed so far have generally been issue specific, authorized by legislatures or political executives, constituted for one occasion, and then dissolved immediately afterward. This is the normal procedure when it comes to consensus conferences, planning cells, citizens' juries, and the citizens' assemblies used to frame referendum questions on constitutional reform in British Columbia, Ontario, and the Netherlands. If (in contrast to the Canadian Citizens' Assemblies, which had a specified role in decision making) the Chamber of Discourses is advisory, then it needs an audience, which may be found in the broader public sphere, as well as in the legislature. Taking the idea of a Chamber of Discourses very literally would suggest that it could begin with several hundred citizens serving a term as members of the upper house of a bicameral legislature (a proposal of this sort was made by the Demos think tank in the context of debates about reform of the House of Lords in the UK). Subsets of the house could then be chosen along the lines we have specified to deliberate particular issues. Alternatively, these citizens could constitute Leib's (Reference Leib2004) proposed fourth “popular” branch of government, reviewing policy proposals generated in executive or legislative branches, or generating proposals for review by the other three branches. In nonstate and transnational contexts, it is easier to imagine granting more substantial and perhaps even final authority to a Chamber of Discourses, if other sorts of representative institutions are not available. Within more familiar governmental contexts, legislative mandates for public consultation and participation present opportunities for experimentation, especially in cases where established forms of consultation are recognized as ineffective. Liberal democratic governments are occasionally willing to experiment, as for example in UK Prime Minister Gordon Brown's 2007 “big idea” for citizens' juries on major policy issues, plus a Citizens' Summit to deliberate basic national values.
AUTHORIZATION AND ACCOUNTABILITY
Theories of representation from Pitkin (Reference Pitkin1967) to Young (Reference Young1990, 128–33) require not only the selection of representatives, but also their authorization by and accountability to those represented. Issues concerning authorization and accountability become pressing to the degree a Chamber of Discourses has an explicit share in decision-making authority. Authorization is, on the face of it, problematic in the methods we have described for the selection of discursive representatives, which would involve social science rather than political process. Such use of social science is already practiced when it comes to the constitution of familiar mini-publics such as deliberative polls, citizens' juries, consensus conferences, and citizens' assemblies. Random selection itself is a social scientific technique that often makes little sense to those not versed in social science. The use of social characteristics to narrow down an initial random sample into a smaller deliberating group is again soaked in social science theories about what individual characteristics matter, as well as assumed links between social characteristics and points of view. Furthermore, when it comes to the engineering of electoral systems, social scientific theories inform the selection of alternative systems (Reilly Reference Reilly2001) (although as the British Columbia Citizens' Assembly shows, such theories can be made intelligible to lay citizens). Yet, the authorization problem remains.
There are several ways to ameliorate this problem. To begin, the social science itself can be done as democratically as possible. In the case of Q methodology, this principle would entail using only statements that appear in ordinary political language, not ones contrived by the analyst. The initial set of statements should be as comprehensive as possible in capturing the variety of things that could be said about the issue at hand, so as to enable capture of the extant variety of discourses. For all methods, data should be analyzed in ways that minimize the observer's discretion in interpreting results. Multiple methods can be used to ensure no discourse is missed. Once results are produced, they can be presented in plain language for validation by citizen participants. In Q methodology, it is easy to summarize an identified discourse in narrative form (Dryzek and Holmes Reference Dryzek and Holmes2002), and those designated as representatives of a discourse can be asked if the narrative really does describe them.
Once we have identified a set of individuals loading highly on a discourse, we could ask them to select a representative. This would require informing this set about both the content of the discourse and the way it was delineated. Furthermore, it may be possible to involve citizens themselves in doing the social science. Social scientists could still be technical consultants, but defer to citizens when it comes to judgments about (say) the items to be included in a Q sort or survey, or the interpretation of in-depth interview transcripts. Whatever use is made of social science, it is important to make it transparent to nonexperts involved in the forum in question. Analysts could then be accountable before hearings, just as executive officials can be called before hearings of nonexpert representatives in an elected legislature. However, those conducting the hearings could not be discursive representatives because they would have been selected by the procedure they are assessing. Instead, they could be drawn from the larger pool from which discursive representatives are selected. As a final check, validation of the configuration of discursive representatives could be sought from actors in the broader public sphere by inviting their comments on forum composition.
Accountability cannot in discursive representation be induced by the representative's fear of sanction because there is no subsequent election at which the representative might be punished. Discursive accountability must be understood instead in communicative fashion. To be accountable to the discourse (or discourses) they represent, representatives must continue to communicate in terms that make sense within that discourse (or discourses), even as they encounter different others in the Chamber of Discourses, and even as they reflect and change their minds in such encounters. If, in the limiting case, representatives seem to be abandoning their discourse and adopting another (as happened in the environmental citizens' jury we discussed previously), then discursive accountability requires that any shift make sense in the terms established by the original discourse.Footnote 9 This requirement is not necessarily met when, for example, social justice advocates get drawn into the language of stability, security, and efficiency; when environmentalists abandon the language of intrinsic value in nature, and start speaking in terms of how preserving nature has economic benefits; or when advocates of an ethical foreign policy slip into the language of realism. Discursive accountability can be facilitated by publicity, such that representatives are always mindful of how what they say will be received in the terms of the discourse(s) that validate their representation. Discursive representatives do not have to be “delegates” of discourses, unable to reflect and change their minds. However, if they do change their minds, they must justify the change in terms set by the discourse(s) they represent.
A MORE INFORMAL CHAMBER OF DISCOURSES
We noted at the outset that discourses currently get represented in mostly informal fashion (e.g., by high-profile activists such as Bono). Contemporary democratic theory can welcome this kind of activity, especially in conceptualizations of deliberative democracy that emphasize the engagement of discourses in a broad public sphere (see the “Discourses in Theories of Deliberative Democracy” section). Discourses are generated within and populate the public sphere, and so a more informal Chamber of Discourses could be grounded in this public sphere. Historically, new discourses have been brought onto the democratic agenda from oppositional public spheres, outside the formal institutions of the state. Think, for example, of how environmentalism and feminism arrived in the 1960s. In some cases, these discourses were brought very quickly into governing processes—environmentalism in the United States in 1970, and feminism in Scandinavian countries around the same time. (However, the result in these countries was rapid attenuation of any radical critique associated with the discourse.) This informal chamber could coexist with the formal chamber we have described, and they could be linked as elements in what Hendriks (Reference Hendriks2006, 499–502) calls an “integrated deliberative system.” Within that system, representatives in the informal chamber could present discourses for validation in the formal chamber. These informal representatives could also exercise critical oversight over the constellation of discourses identified for the formal chamber (as indicated in our previous discussion of discursive accountability).
If we think of a Chamber of Discourses in these informal terms, then it would seem at first sight that all that needs to be done is to leave it alone. In Habermas' terms, the public sphere is a “wild” zone that can be protected by, for example, a standard range of liberal rights to free belief, expression, assembly, and association. Beyond that, critics might need to expose and counter agents of distortion in the public sphere, such as the influence exercised by large media corporations, lack of material resources meaning that some sorts of voices do not get heard, hegemonic discourses that serve the interests of the powerful, and so forth.
In this light, discourses get represented by the normal array of actors present in the public sphere. However, the idea of discursive representation enables and provides criteria for reevaluation of some standard normative treatments of civil society. Putnam (Reference Putnam2000) disparages “checkbook” groups such as the Sierra Club that demand nothing more than money from their members and that have little in the way of internal participation of the sort that might help build social capital in the larger society in which the Sierra Club operates. In light of discursive representation, Putnam's criticism misses the point. Checkbook groups may build discursive capital (in the sense of facilitating the articulation of discourses), if not social capital. The Sierra Club exists to represent a particular discourse of environmental preservation, and contributors to the Sierra Club express solidarity with that discourse. Discursive accountability can be sought by these leaders continuing to communicate in terms that make sense within the discourse of preservation (even as they engage other discourses). If leaders could not justify their actions in these terms, contributors can back other groups instead.
Discourses engaging in the broad public sphere get represented to more authoritative political structures (e.g., states) through a variety of mechanisms. Public opinion defined in the engagement of discourses can reach the state or other public authority, and so find representation. In his “two-track” model of deliberation in the public sphere influencing deliberation in the legislature, Habermas (Reference Habermas1996) (very conventionally) eventually stresses elections. In this light, we might evaluate electoral systems by how well they represent discourses. For example, preferential voting as practiced in Australia almost guarantees a two-party system in parliament. However, minor discourses get represented even when nobody in parliament is formally associated with them because the two major parties need to cultivate minor parties in order to receive their voters' second preferences, so preferential voting may at least be better than first-past-the post plurality voting in representing discourses. However, elections are not the only transmission mechanisms. Others include the use of rhetoric by activists, influence on the terms of political discussion that can change the understandings of government actors, and arguments that are heard by public officials (Dryzek Reference Dryzek2000). Conceptualizing such transmission mechanisms as forms of discursive representation drives home the need to subject them to critical scrutiny. Rhetoric, in particular, is often treated with suspicion by democratic theorists, on the grounds of its capacity for emotional manipulation and coercion (e.g., Chambers Reference Chambers1996, 151). However, rhetoric may be vital in representing a discourse to those in positions of political authority not initially subscribing to it. The solution here would be to hold rhetoric to standards such as noncoerciveness and the need to connect particular interests to general principles. The latter could, for example, curb the racist or ethnic nationalist rhetoric of demagogues.
In addition, all forms of transmission need to be held to the discursive accountability standard introduced previously. People claiming to represent a discourse or discourses should always communicate in terms that make sense within the discourse or discourses in question, even when they contemplate shifting in relation to the constellation of discourses they subsequently encounter. This standard is probably met more easily to the degree representatives keep their distance from explicit participation in collective decision making in, for example, corporatist arrangements.
In the case of networked governance, discursive accountability could be facilitated by specifying that a network does not require as the price of entry that participants commit to the hegemonic discourse of the network and renounce other relevant discourses. This kind of accountability would be hard to secure in transnational financial networks, which currently exclude discourses of sustainability and social justice.
Informal discursive representation may currently be found directed toward familiar and conventional authority structures (e.g., states and international organizations). However, this informal representation could also mesh with any formal Chamber of Discourses. In this context, public sphere activism could provide a check on the degree to which the formal chamber features a comprehensive and accurate set of the relevant discourses, and promote discursive accountability by calling changes of language in the formal chamber to account.
TRANSNATIONAL DISCURSIVE REPRESENTATION
Representing discourses in transnational political action is actually more straightforward than representing persons (especially in the absence of elections). Indeed, it is already happening. In recent years, even economistic global institutions such as the World Bank and (begrudgingly) the International Monetary Fund (IMF) have begun a program of outreach to global civil society, meaning accountability no longer runs strictly to states. Who elects the NGOs? Nobody. Is there an identifiable constituency or category of people with which each NGO is associated and to which it is accountable? Not usually. International relations scholars have started to think about accountability (Grant and Keohane Reference Grant and Keohane2005), albeit mainly in terms of how sanctions can be levied on advocates, rather than discursive accountability as we have characterized it. However, NGOs pushing for human rights, fair trade, sustainable development, demilitarization, transparency, and so forth, may best be thought of as representatives of particular discourses in international politics. Is the world any more democratic for their activities? Clearly, yes, the international governmental institutions they target now have to justify their activities in light of a variety of discourses, whereas previously they either felt no need to justify at all, or did so in narrowly economistic and administrative terms. Thus, the idea of discursive representation provides democratic validation for the activities of NGOs and other transnational activists.
This kind of transnational discursive representation is currently informal in character, but more formal representation can be imagined. Thompson (Reference Thompson1999) suggests that cross-border policy impacts can be brought into democratic accountability by the device of a “tribune for non-citizens.” Such a tribune could not easily be elected —the appropriate electorate would be dispersed and extraordinarily hard to organize. However, for particular policy issues, it would be possible to identify relevant extranational or transnational discourses, and identify a good representative for them. For example, there exists a very well-defined transnational discourse of sustainable development. Perhaps global sustainability tribunes could be identified to represent this discourse in particular national governments. The problem, of course, is that those representatives would be least welcome where they were needed most. One can imagine them being welcomed by countries that are exemplary international citizens (Sweden), but resisted by countries that are poor international citizens, those that subscribe to hard-line notions of sovereignty, superpowers, and rogue states.
For most states, transnational discursive representation will probably have to be informal in any foreseeable future, constituted mostly by NGOs and networks of political activists in transnational public spheres exerting pressure. It is easier to envisage more formal Chambers of Discourses established in association with international organizations. Organizations such as the WTO, IMF, and World Bank have (as we have noted) accepted the need to legitimate their activities beyond the states that are their members, funders, or clients. Constituting formal Chambers of Discourses would be one very public way of discharging this obligation.
Conclusion
Once the basic idea of discursive representation is accepted, choices need to be made on several dimensions. Should discursive representation be formal, informal, or an integrated combination of both? Discursive representation could be formalized, especially in connection with growing enthusiasm for the constitution of mini-publics to deliberate complex and controversial policy issues, and as a way for governments to meet mandated requirements for public consultation. If discursive representation is formal, what method should we use to select representatives? How much authority should any Chamber of Discourses possess in relation to other representative institutions? There is no universal answer to any of these questions, although we have provided guidance about how each might be answered in particular contexts.
We have argued that discursive representation already occurs, although it is not always recognized as such. Whether formal, informal, or an integrated mix of both, discursive representation can help render policy making more rational, respect individual autonomy by more fully representing diverse aspects of the self, assist in realizing the promise of deliberative democracy, and make democratic theory more applicable to a world where the consequences of decisions are felt across national boundaries.
Comments
No Comments have been published for this article.