Representation roles

A core feature of contemporary analysis of political representation is the construction and use of typologies of representation. Often conceived as roles, not least those of trustee and delegate, they are also and variously conceived as views (Pitkin Reference Pitkin1967), forms (Mansbridge Reference Mansbridge2003), ideal-types (Rehfeld Reference Rehfeld2009), varieties (Pettit Reference Pettit and Shapiro2009), or conceptions (Disch Reference Disch2011). The different terms reflect subtly different analytical starting points and perspectives. Representation “types” and “forms” reflect isolable, persisting, and distinct features of the political world; “ideal-types” are more abstract generalizations of such isolable, persisting, and distinct features; “roles” reflect isolable, persisting, and distinct practices in the political world (ways of doing representation); and (d) “views” and “conceptions” of representation are more explicitly situated (historical or contemporary) perspectives on what representation is or what it is for.Footnote ¹

The argument for identifying and analyzing the importance of shape-shifting representation is built upon the performative and constructivist definition of political representation as the contingent product of “representative claims.”Footnote ² According to this perspective, representation exists primarily by virtue of its being done—practiced, performed, claimed. Representative roles and relations gain a presence in our politics because myriad actors make claims to speak for others (and for themselves). Representation is a performative product in two linked senses: it is performed in the theatrical sense (i.e., it is both done and shown to be done (Schechner Reference Schechner2002)) and in the speech-act sense (it is a speech or other act which establishes, or contributes to establishing, a state of affairs) (Austin Reference Austin, Urmson and Sbisa1975; Butler Reference Butler1997). A performative account initially emphasizes roles over types or forms because role emphasizes the crucial place of practice, or acts, in constituting representation. 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—implicitly or explicitlyFootnote ³ —to represent a constituency is a constituting or reinforcing of the social availability of that role.

The representative claim framework emphasizes the situated or contextual dynamics of producing relations of representation, rather than a wholesale break with a stress on typology construction (e.g., in the work of Mansbridge and Rehfeld). The emphasis on roles and practices highlights three features of representative politics which provide crucial underpinnings for the concept of the shape-shifting representative. First, it stresses representation's variability: it is a protean phenomenon that can be formal and informal, electoral and nonelectoral, national and transnational, potentially happening in multiple spaces and possessing many guises. Second, it stresses representation's contingency and dynamism: there's a lot “going on” in representation, a constant process of making, receiving, accepting, or rejecting representative claims. And third, this approach highlights representation's aesthetic and cultural character: would-be representatives need to “make representations” of their constituents (in the sense of artistic portrayals or depictions, such as candidates for office constantly using phrases such as “hard-working families,” “strivers,” or “battlers”) to try to get the latter to recognize themselves in the claims being made (Saward Reference Saward2010).Footnote ⁴

Representation, on this account, is produced through the performance of roles. Consider a wide though not exhaustive snapshot of accounts of major representative roles (or types or forms that are best reinterpreted as roles) in the literature:

1. 1. trustees and delegates (and politicos)

2. 2. functional roles played in governmental systemsFootnote ⁵

3. 3. promissory, surrogate, gyroscopic (Mansbridge Reference Mansbridge2003)

4. 4. descriptive and substantive representation (e.g., Celis et al. Reference Celis2008)

5. 5. a politics of ideas and a politics of presence (Phillips Reference Phillips1995)

6. 6. liberal and republican models of representation (Bellamy and Castiglione Reference Bellamy and Castiglione2013)

7. 7. conceptions of roles of the “good representative” (Dovi Reference Dovi2008)

8. 8. formal or positional governmental roles (prime minister, member of parliament, etc.)

9. 9. principals and agents

10. 10. likeness and distinction (Chabal and Daloz Reference Chabal and Daloz2006)

11. 11. indicative and responsive (with the latter category divided into “directed” and “interpretive”) (Pettit Reference Pettit and Shapiro2009)

12. 12. modes of “informal” representation: e.g., stakeholder (Macdonald Reference Macdonald2008), advocate, champion.

An extensive literature discusses derivations and applications of these examples, often spinning off Pitkin's (Reference Pitkin1967) extensive analysis. They vary greatly in their founding assumptions, political motivations, cultural and geographical reach, institutional focus, and empirical applications. But note that several categories in this rich list are primarily performative, ways of doing or carrying out representation, whether in specific, perhaps policy-to-policy terms (such as trustee, delegate, or directed-responsive), or in more general orientational terms (such as liberal, republican, gyroscopic, substantive, or likeness). Others are more distinctly nominal (functional roles, such as senator or mayor), relational (principal and agent, ideas and presence, for example), or characteristic (such as descriptive).

However, despite those multidimensional variations each entry on the list can best be conceived as a representative role that can in principle be combined (in a number of ways, as we shall see) with a range of other entries. Arguably, even relational or characteristic categories of representation are parasitic on a performative or role-oriented conception of (for example) acting as a principal or a descriptive representative. A representative claimant plays the role of delegate, champion, descriptive representative, good or moral representative, and so on. In the practice of a would-be representative such roles can be mixed and matched outside and across their original theoretical or political points of derivation. For example: a representative may claim a descriptive likeness to the majority of members of his constituency (“I grew up in this town, and I know you can all recognize me as one of your own”), to be a trustee of constituency interests (“You can rely on me to do what's right for the town, even when that's not easy”), and claim the mantle of a good representative on the basis that he makes all of his decisions in a public and visible way (“With me, what you see is what you’ll get”). In principle, a wide range of plausible examples can be gleaned by working the above list.

From role to subject position

These points add up to a reframing of representative roles as factors of mobility rather than occupancy, reflexive positioning rather than comparatively fixed positions. If we were to insist on describing shape-shifting representation as a “role,” it would be a meta-role, a role of roles deploying shifting shades and aspects of a range of representative roles. The shape-shifting representative is not just one more role character alongside (e.g.) the trustee, the delegate, the surrogate, or the gyroscopic representative. The “role” of the shape-shifting representative is highly distinct; in theory and in practice it disrupts, conjoins, de- and re-attaches other, more familiar, roles. The shape-shifting representative is—they are—the linking mechanism, the creator of representative personas, the forger of hybrid roles of speaking and acting for, moving in and among a range of familiar roles.

However, importantly, these “roles” function most clearly as resources for representative claim-making. Indeed, it is at this point that the notion of a “role” becomes seriously strained, inadequate in its implied fixed or static status next to the need for concepts which can capture the more dynamic—and, I argue, for that reason more realistic—mobility and positioning work (betwixt and between “roles”) that is central to claims of representation. Despite its superiority to “types” or “forms,” the concept of role ultimately fails to capture the full importance of the very attribute which placed it above those alternatives; it can capture active practice of representation within the bounds of particular roles, but not for example when the performance of representative claims breaks those bounds, as a regular if not routine political phenomenon. Its more or less neutral, descriptive overtones deflect analytical attention from the very constitution of those roles through representative claim-making, and why this or that role is salient or prominent in a time and place. It cannot capture the wider subtleties of representative practice that are crucial to a more rounded understanding.

For these reasons, I propose displacing the concept of role in favor of positioning. Social psychologists made this productive move several years ago, arguing that “the concept of positioning can be used as a dynamic alternative to the more static concept of role” (Harré and van Langenhove Reference Harré and van Langenhove1991, 393). According to Henriksen (Reference Henriksen, Moghadden, Harré and Lee2008, 41), positioning theory's emphasis on “interactivity, movement and fluidity” has major advantages over “role-theory as a static tool for understanding social interaction.” Just as the claim-making conception stresses representation's dynamism, so roles become (in principle) malleable resources for would-be representatives who position themselves to exploit those resources. In this light, for example, an election candidate, or a social movement figure, or a “shock-jock” talk show host, positions him or herself as a subject with respect to constituents, supporters, or listeners; in other words, they adopt subject positions. Subject positions are intersubjective, culturally and discursively constituted stances that are (differentially) available for adoption by actors. For example, the subject position of descriptive or sociological likeness, and another of trusteeship, is available to potential Western representative claimants (at least) as a social resource. If Chabal and Daloz (Reference Chabal and Daloz2006) are right, for instance, a claim such as “I can speak for you because I am like you an ordinary person, doing the things you do and concerned with the things that concern you” expresses a local cultural resource within which a Swedish politician may fruitfully position herself.

Subject-positioning occurs due to, and against a background of, a complex array of available resources in specific cultural-political landscapes (such as national, religious or linguistic communities). It fosters analysis of both stability and dynamic change, depending on how actors or subjects (are able to) deploy available cultural resources. Unlike with the concept of role, neither stable subject-positioning nor dynamic multi- or re-positioning is downgraded. Further, it introduces into work on representation an interactive dynamism: claims by representatives position themselves and their audience, and claims by the represented position both them and the representative. Representatives do not so much have or occupy roles as “pause at” or “move through” available subject-positional resources, which in turn they play a part in creating or reshaping. It is the relational and changeable array of such resources which defines the dynamic playground of political representation, including the situated capacities of would-be representatives to shape-shift, i.e., to reposition themselves among the array. Consider for example Fenno's view of “trustee” and “delegate” not as representative roles but as resources for congressmen to use to justify their actions to constituents. Notions of trusteeship and delegacy are deployed in representatives’ “presentational and explanatory activity” (2003, 168):

Patterns

Actual or potential movement through an array of culturally sensitive subject positions—shape-shifting—is critical to grasping the empirical and theoretical dynamics of representation. Crucially, this shape-shifting creates patterns of representation. It is not so much the separate types of representation that can provide purchase in explaining representative practices as the patterns of the worldly combination and recombination of subject positions. The most productive way in which to capture shape-shifting representation is through a novel shift in emphasis from role to contingent positioning and patterning. Together these concepts give us the practice and the outcomes of representation, where “role” runs together practice and outcome.

Consider a hypothetical example: a leader of a national trade union who claims to represent the substantive interests of his country on the basis of his union role speaking for significant numbers of workers. As that claim is challenged, he may shift to claiming representation by virtue of his likeness to ordinary people, not least by telling his story of rising through the union ranks from the shop floor. The decline of the trade union base prompts him to gain party support and to stand and win a parliamentary seat. Here, he adopts an outsider stance as champion of particular groups of low-wage and vulnerable workers, moving away from the “likeness” claims. Being a promise-keeping, unwavering good representative of the constituencies he cultivates becomes the core claim around which he functions. We have here a sequential pattern of subject positions: the actor offers himself as a representative by virtue of (a) substantive policy positions, then (b) on the basis of likeness or similarity to constituents, then (c) in terms of the champion of particular interests, and finally (d) in terms of his moral consistency. Deploying concepts of positioning and patterning by would-be representatives presses us to take such dynamic trajectories seriously. Consider further the former president of Brazil, Luiz Inácio Lula da Silva, or more commonly Lula. From local union official to national union official, to co-organizer of the Workers Party to congressman, to multiple times presidential candidate to two-term president, Lula is noted for shifts from radicalism to reformism, and in styles of self-presentation. Of course, shape-shifting is not just sequential; as we shall see, time and space are more complexly deployed in the politics of representation. Such patterns and patterning of representation are crucial to our efforts to gain robust theoretical and empirical understanding of political representation.

“As fortune and circumstance dictate”: Constraint and enablement

Are we dealing primarily with shapes or shifters, constraining structures, or enabling agency? Some actors may experience little choice as to which representative positions they adopt or in which they are placed—“delinquent youth,” perhaps. Others may have much more choice as to whether they appear as a “champion” of certain interests, for example. Regarding subject positions as relationally defined opens up important questions of power and choice for would-be representatives, whether they occupy stable positions or shapes or move among them more actively. To explore patterns of shape-shifting representation, we need to understand two key dimensions: (a) structural options for shape-shifting, and (b) the degree to which would-be representatives activate the agentic opportunities available to them. The variations in shaping of representative roles which these two dimensions produce are set out in Figure 1.

FIGURE 1.

Shapes and shifters

Figure 1 captures key lines of variation of constraint and enablement. With respect to structural options for shape-shifting, aspects of socio-economic and political context will frame the extent and type of subject positions that representative claimants might be able to shape themselves for. For example, political parties in the U.S. remain more fluid ideologically than their European counterparts (despite the recent rise of more overtly partisan congressional politics). This fact can foster greater opportunities to shape-shift among U.S. party politicians. Further, a pluralistic and multicultural society may offer greater repositioning options to greater numbers of representative claimants. On the other hand, actors may experience different degrees of desire (or, good reasons) to choose whether to shape-shift, i.e., the extent to which they activate opportunities for shape-shifting. So, for example, a context where an actor has little wish or need to shape-shift and there are few structural options for shape-shifting may lead to a “shape-accordance” pattern of representation.

A context where there are more structural options but strategic reasons for an agent not to shape-shift among representative roles may lead to a “shape-retained” pattern of representation (e.g., a strong and secure party leader who revels in her reputation for spotless consistency). The shape-retained pattern bears a close relation to Mansbridge's (Reference Mansbridge2009) “selection model,”, while emphasizing the degree of choice exercised by the would-be representative in how her representing is to be characterized and performed. This point underscores the utility of the shape-shifting approach; a representative may act in accordance with constituent preferences, whether by “selection” or “sanction” (Mansbridge Reference Mansbridge2009), but using a model that allows for consideration of the degree of choice he or she may have had in so doing adds important theoretical and empirical nuance. Positioning as a representative of sort A is one thing, but understanding how and why it is not B (or not-now B, or not-here B) can help us to begin to deepen our grasp via situated and relational analysis.

An agent who would wish to shape his or her representative claims to differing or changing positions, but faces fewer enabling structural resources, may give rise to a “shape-confined” pattern of representation. Similarly, a shape-confined pattern may arise where competitors or opponents claim that an actor stands for this or that goal or group, and the actor herself is unable to combat the claim effectively. Consider a challenge faced by President Obama five months before the U.S. presidential election of 2012:

On this account, Obama recognized the dangers of shape-confinement—struggling to overcome strong characterizations of himself, not least from his opponents—and seeking the political space to shape-shift. The “shape-shifting” pattern in its strictest sense may arise where structural options for, and agentic choices and strategies favoring, shape-shifting among representative positions are both greater. To the question “what matters most, shapes or shifters, structure or agency?” the answer is: it depends. Attention to specific contexts and strategies is crucial.

The positions mapped in Figure 1 reinforce the idea that subject positions such as delegate, champion, and so on are adopted or traversed by actors, freely or reluctantly. This is one part of what it means to highlight positioning and patterns rather than roles in analyzing political representation. It is not simply that representatives may play a particular role; it is that certain contextual constraints and strategic opportunities and choices may lead to that “role” being played in the context of positioning and patterning behavior. Sometimes the shape-shifting is done by representatives out of strength, at others out of weakness; sometimes it is effectively imposed on representatives by others. The category of role tends to obscure agent choices and capacities—are they, for example, shape-confined or shape-retaining? The answer to that question is crucial to grasping the nature and force of their representative claims. There is only one box labelled shape shifter (in Figure 1), but each of the other three categories starts from the assumption that shape-shifting is normal, a core ingredient in the dynamics of politics across contexts.Footnote ⁸

Dimensions and patterns of shape-shifting representation

Although just one of the four poles in Figure 1 is denoted “shape-shifter” strictly speaking, they all share a common conceptual space. They bleed into each other, and develop their texture only by virtue of the ever-presence of the others (i.e., relational). They all acknowledge the crucial place of (potential or actual) shape-shifting at the heart of our ideas and practices of representation. But what sort of shape-shifting? The two key dimensions to shape-shifting are the temporal and the spatial, key aspects in examining political representation which, despite for example the classic work of Fenno (Reference Fenno2003) and Eulau et al. (Reference Eulau, Whalke, Buchanan and Ferguson1959), play little role in contemporary theories. I will comment on each in turn.

Among theorists of representation there is an underlying, and largely unspoken, assumption that representatives play one distinct representative role at a time. Likewise, there is a common linear assumption about time in representative politics—that any given moment in a passage of time is in principle equal in political significance and intensity to any other. Where issues of temporality are considered at all, they tend to be large-scale epochal ones, for example Pitkin's (Reference Pitkin1967) view of the period of the “fascist theory of representation.” A focus on shape-shifting representation brings these assumptions into question. On one level, we have the importance of representatives being able to “mark moments,” to heighten or intensify the significance of particular times or junctures. We are familiar with the importance of election days and nights, for example, as key components of political “timescapes” (Goetz and Meyer-Sahling Reference Goetz and Meyer-Sahling2009). But more significant are the ways in which shape-shifting representative practice upsets common assumptions of temporal linearity and singularity. The shape-shifting representative may, for example:

1. 1. offer representative claims on the basis of being one thing at one time, and other thing at another time—e.g., a delegate at one time and a trustee at another;

2. 2. offer multiple representative claims at the same time, or in overlapping periods of time, to different (or even the same) audiences; or

3. 3. offer claims to be, at one or at overlapping periods of time, one sort of representative (e.g., a descriptively representative delegate of a particular group) while in fact acting as a different sort (e.g., pursue a view of the common good while not subject to sanction).

Similar issues arise with respect to space and the closely linked notion of identity. Common assumptions among empirical and conceptual analysis of representation include the view that the electoral district and the nation-state (and little else) are the fixed and given “containers” of the primary issues, interests, and identities that may call for or require political representation.Footnote ⁹ Closely allied with this is the common view that the identity of the constituency or constituents is given (especially with regard to electoral representation).

Sometimes these assumptions will be accurate, but they do not go far enough. The significance of space(s) and place(s) and identities associated with them is a question of subjective and intersubjective perception, not an objectively given property of a system. Even a stable nation-state structure requires constant evocation through speeches, ceremonies, rituals, symbols, and so on (Rai Reference Rai2010). And different “containers” of interests may be available for evocation by would-be representatives, many of them noncontiguous in their distribution: religious groups, cultural or identity groups, social classes, even nonhuman interests and intergenerational or future generation interests. A politically salient sense of what spaces and identities there are and how and why they matter (or need representing) is manipulable, with multiple possibilities in practice.

So we can combine an opened-up sense of both time and space to get a sense of the key dimensions of shape-shifting representation. A representative claimant may claim to be playing representation role A in one place (or to and for one group), and role B in another, at the same time. He may play roles C and D to a common or overlapping potential audience at the same time, perhaps by using dog-whistle tactics—“telling one group of voters one thing, while allowing or encouraging another group to believe another” (Goodin and Saward Reference Goodin and Saward2005). He may claim the reality of a little thought-of space in order to (try to) render it politically significant to his advantage—witness the Northern League in Italy, and in particular its leader Umberto Bossi, and his invocation of “Padania” as a would-be separate state from the rest of Italy (Giordano Reference Giordano1999).

Taking these two key dimensions, Figure 2 sets out four basic patterns of shape-shifting representation—I shall refer to these four as “level 1 patterns.” Recall that shape-shifting refers to processes of positioning and repositioning the persona and the nature of his or her claims to audiences or constituencies. Where (in more conventional terms) roles reinforce roles, the argument here is that positioning creates and recreates patterns of representative practice. Each of the patterns A–D in Figure 2 are simplified snapshots of what in fact will be points on a spectrum, where for example the spectrum on the temporal dimension would cover varying degrees of (and shifts among) concurrency and non-concurrency of shape-shifting.

FIGURE 2.

Patterns of shape-shifting representation

The example of the trade union leader as a shape-shifting representer outlined above offered a sequential shape-shifting. This is an instance of pattern D, where the actor shifts from claiming national-level representation to more specific spaces and groups (thus multiple spaces) through non-concurrent shaping of his persona (i.e., moving from one persona to another in a sequence). The example of President Lula of Brazil may exemplify Pattern C: sequential positions on policy and ideological orientation were adopted by Lula largely within (and for citizens within) one political space (the nation-state of Brazil).

The movement from roles to shapes and patterns of shape-shifting sets in train a continuing process of conceptual elaboration. If, as I have argued, shape-shifting representation ought to be a primary unit of analysis of political representation, then we need to pinpoint also the succeeding steps in the argument. Let us identify, then, three consequent key levels of patterning of shape-shifting representation.

Level 1 patterning—pattern A or B or C or D. The representative claimant positions herself, or is positioned, with respect to a number of representative “roles” (resources), and in so doing retains a particular (though complex) pattern of concurrency and spatial reference.

Level 2 patterning—A to B, or D to B, etc. The claimant moves from one pattern to another. For example, she moves from offering a range of representative personas in one space to acting likewise over a number of spaces or would-be constituencies (a move that may be advantageous because, in principle, a claimant might be judged unreliable or inconsistent if addressing the same group of people with different representative persona). Consider for example a member of parliament who is in favor of developing green energy technologies nationally, but in whose own constituency a major campaign opposed to the creation of a local wind farm has become popular. When he campaigns in his constituency, he softens his message, saying the technology might not be right for us, here. When campaigning elsewhere, he positions himself differently. This can be seen as a shift from pattern A to pattern B: from consistent positioning in all political spaces to differential positioning in different political spaces.

Level 3 patterning—A to B to C, or B to C to D, and so on. Here, the level of complexity of the patterning increases further.

In principle, of course, there will be a great many possible levels of patterning. Which patterns are discovered empirically will depend a good deal on political cultures of representation and the political positions that are being pursued or contested by representatives. For empirical analysis, this reconceptualization of political representation—from the prioritization of roles to the prioritization of patterns—can be expected to give rise to a number of distinct research questions. Detailed research can be expected to reveal an array of specific patterns, and their detection should tell us a good deal about the dynamics of political representation across geographical and political-cultural contexts. One might object that would-be representatives merely perform ad hoc or opportunistic shape-shifting. But, in my approach, shape-retaining for example may be just as “ad hoc” or “opportunistic” as shape-shifting proper. Further, we are not searching here for moral patterns—terms such as “opportunistic” unduly prejudge representation and overlook important structural reasons for, and patterns of, shape-shifting. Such objections beg the question: why take this opportunity to shape-shift, or indeed shape-retain, in this context, where another actor may not (be able to)? What are the patterns of constraint, enablement, and shape-shifting?

A number of questions may guide the generation of hypotheses about such patterns, not least for example:

1. 1. Are identifiable patterns of shape-shifting representation correlated with different political actors (heads of state, members of parliament, interest group leaders, protest group leaders, and so on)?

2. 2. Do specific countries or political cultures foster or in some way require distinctive adherence to certain patterns of representative claim-making?

3. 3. Is there a relationship between shape-shifting or shape-retaining from a position of strength (or weakness) and a specific pattern (or level of pattern) of shape-shifting?

4. 4. On a reasonable measure of political visibility, what pattern(s) in what contexts foster greater or lesser degrees of political transparency?

Let us take the final question by way of illustration. Surely, in cases of multiple, hidden, and (on the face of it) inconsistent claims by shape-shifting representatives, the latter will be “found out”? Political scientists have long offered evidence that American voters lack information about their political choices in the polling booth. They rely on cues from candidates, parties, and other actors which offer shortcuts (Popkin Reference Popkin1994). But even this view may rest on questionable assumptions about the ready availability of sufficient knowledge, or sources of knowledge, in addition to the problem that not enough people avail themselves of it. Amid contemporary tendencies to “narrowcasting” media, people get their information from limited sources, even when many further sources are feasibly available to them. Trends in access to and accessibility of knowledge of politics may underpin capacities for shape-shifting representation. With imperfect and differentiated distributions of perceptibility come opportunities to manipulate perceptions of political time, space, and identity. Capacities and opportunities for shape-shifting representation (and for example for shape-retaining representation) may increase where dense and divided modes of knowledge mediation form particular regimes of political visibility. Footnote ¹⁰

Systemic representation

There remains a crucial issue with respect to space: where representation is understood to happen within a polity. Is it, for example, an instruction or practice relevant only to formal elective office? Recent innovations in thinking about political representation have, following one thread in Pitkin's work (Reference Pitkin1967), shifted their analytic and normative focus away from strictly dyadic relationships between represented and representative, and towards systemic views. Two factors underpin this move. First, in terms of defining representation, it is recognized that a range of actors, for example unelected as well as elected ones, may succeed in making effective representative claims (Montanaro Reference Montanaro2012; Rehfeld 2006; Saward Reference Saward2010, 82–110). Second, in terms of normative democratic criteria, it is the overall systemic quality of representation (however that quality itself is assessed) which counts, however much dyadic instances of representation still matter, normatively. However, we need to question this new conventional wisdom and further unpack this distinction. It acts as cover for further important distinctions concerning the “systemic,” and may obscure what is at stake in the linkages and mutual dependencies between dyadic and systemic representation. Consider Figure 3.

FIGURE 3.

Dimensions of systemic representation

The systemic-governmental view holds that representation happens in the executive and legislature of a state, and perhaps in governmental and quasi-governmental international bodies. The systemic-societal view holds that representation occurs across society, including in governmental bodies but also for example in interest groups, social movements, and businesses. In terms of democratic legitimation, one can look to specified qualities of representative dyads within (box A), or within and beyond (box B), the governmental system. Without entering into a full defence here, I hold that boxes C and D in Figure 3 are the relevant ones to consider where the analysis of the quality of political representation is at issue.

Writers who advocate a “systemic” view of representation, notably Pitkin, Urbinati, and Mansbridge, especially as far as normative judgements are concerned, tend to assume that systemic = systemic-governmental. The key reasons behind my preferred systemic-societal interpretation include the argument that so-called “informal” representation—often located outside the state, often nonelective or informally or partially elective though still based on claim-making and reception—is political representation and is potentially democratic representation.Footnote ¹¹ Representation is claimed or enacted by a wide array of local, national, and international groups and individuals, elected or chosen or not-elected and rejected. Even democratic representation need not to be understood as confined to a set of statal institutions, but rather understood more broadly (and indeed more complexly) as a set of practices more or less present in a wide array of diffuse locations, including transnational contexts. On this basis, analysts should work with a systemic-societal basis of representation's domain, along with a dyadic+systemic view of representation's democratic quality. The key point is that the shape-shifting representative has in principle a wide and variegated societal playground (I return to the more strictly normative aspect of the argument below). Indeed, this framing rightly suggests that (shape-shifting) representation occurs throughout social, including personal and group, lives as well as in more conventionally political contexts. As Lloyd writes, “Humans are representing animals. . . Human life [. . .] is largely a cycle of making and interpreting representations” (cited in Slezak Reference Slezak2002). I acknowledge this fact and focus on the more clearly political aspects of shape-shifting representation.

A further crucial issue concerning the relationship between the dyadic and the societal-systemic arises in the specific context of shape-shifting representation. The “system” in this potentially broad systemic account is not a consistent, singular, or clearly bounded set of spaces, actors, or institutions. Nor do “dyadic” examples of representation exist in isolation from each other. Rather, shape-shifting representation itself forges or enacts the particular manifestations or experiences of the (societal) systemic. An instance of shape-shifting representation will characteristically involve a pattern of representative dyads. These patterns, in turn, enact the systemic. Urbinati (Reference Urbinati, Alonso, Keane and Merkel2011, 46) has written of the “. . .intricate network of interdependency between representatives and the represented.” Unstable patterns of such networked interdependency result from (among other things) the activities and claims of shape-shifting representers. The latter, for example, may succeed in claiming to speak for several overlapping constituencies with respect to different sets of interests. That makes for a pattern of dyads that becomes in effect a patterned system of representative claim-making and reception. In this respect, the very distinction between the dyadic and the systemic views of representation begins to blur, or even to break down. The numbers, character, and timing of representative claims and their reception do not map at all neatly either onto a series of discrete dyads or a single clearly bounded conception of a system of representation. Shape-shifting representation challenges fundamentally such neat distinctions, just as it challenges unduly tight separations between types or roles of representation.

Patterns of shape-shifting representation shape in turn the character and boundaries of the larger representative system. Their claims to (patterns of) roles may suffer varied fates: fading, transforming, lingering, or being rendered effective or ineffective. Whichever fate they meet, their representative claims rest not on something that they are, but rather on what they do, and what constructions of the systemic societal domain, persistent or ephemeral, their actions weave.