A tale of two demoi

Introduction

It was not so long ago that one could begin an article on the democratic boundary problem by noting how little attention the topic has received in democratic theory literature. One cannot make that claim today. In brief, the democratic boundary problem concerns the question of what the principles of democratic theory have to say about the make-up and boundaries of the demos itself. The central idea of democracy is that the exercise of political power must be legitimated by ‘the people’. But how is it to be decided who makes up ‘the people’? Is it possible to address the issue of who is to be included and who excluded from the demos in a way that can itself be described as ‘democratic’?

Until recently, the answer to this question was ‘no’: since democracy already presupposes a demos, democracy cannot choose the demos. Even if the boundaries of existing democracies were not determined by anything resembling democratic means but usually through war, conquest, colonialism, or some other form of coercion, there is little if anything that democratic theory has to say about them. ¹ In the last two decades, however, democratic theorists have come to find this answer increasingly dissatisfying. In the present age of mass immigration, human rights, international organizations and new social movements; in an age when decisions made in one country increasingly affect the lives of citizens elsewhere; and in an age when the power of the state and state-centered democracy appears to be shrinking in the face of globalization: it is becoming starkly clear that we need to rethink the justification of boundaries and their relation to democratic communities. As a result, the literature on what is variously called ‘the boundary problem’, ² ‘the problem of inclusion’, ³ ‘constituting the demos’, ⁴ ‘the legitimacy of the people’, ⁵ or ‘the question of the “who”’, ⁶ has grown massively. If the question of how to constitute the demos were once ignored by democratic theory, today it is an inescapable part of it.

The purpose of this article is not to solve the boundary problem but to clarify the issues that come up when engaging it. In my view, the debate over the boundary problem has been hindered by the fact that contributors frequently work with different understandings of democracy, of what it means to make something democratically legitimate and of what it means to participate in a demos. ‘Democracy’, for example, can refer to a decision-making method, a form of governance, or a way of organizing society. We can use the idea of ‘democratic legitimation’ in reference to the process of establishing something in a democratic way, but we can also use it to refer to the process of democratizing something already established. Even the concept of a ‘demos’ harbors a certain ambiguity. As I will show, the demos can be theorized from both a third-person, observer’s perspective and a second-person, participant’s perspective. The former is the perspective from which the demos appears as a whole entity, a thing, a concrete and discrete totality; it is the perspective from which the demos becomes something that can be represented by a state. But a demos is also an association of ongoing relationships among thinking, speaking and acting participants striving to organize a common life together. From this second-person point of view, the demos is not intrinsically tied to the perspective of the state; rather, it is constituted through the numerous person-to-person relationships that participants negotiate among each other.

Different understandings of these concepts lead contributors to different conclusions about how the boundary problem can be feasibly addressed. In what follows, we will work through these conceptual issues, and in doing so we will see if there are not new insights to be gleaned that can make the boundary problem a productive one instead of a paralysing one. Ultimately, I will show that the dual structure of the concept of the demos and its proper formulation will be decisive here. While the concept of the demos requires both the second-person and third-person perspectives, democratic theory has long privileged the third-person observer’s perspective over the second-person participant’s perspective in a way that obscures the differences between the logic of democracy and the logic of the state. This distorts how we think about boundaries, the role they play in relation to the demos and how they may be subjected to democratic criteria. Drawing on literature from deliberative democratic theory, I will show that the principle of democratic legitimation that governs the demos ultimately resides in the second-person perspective, the perspective of the participant in politics engaged with fellow participants in politics. If we reconstruct the demos from this second-person perspective up to the third-person perspective, we get a rather different conception of the demos that provides more fertile ground for thinking about how demotic boundaries may be subjected to standards of democratic legitimation.

I Three (or four) meanings of ‘democratic’

Joseph Schumpeter, giving an early formulation of the boundary problem, described democracy as simply a method for making decisions. ⁷ Under this democratic method model, there are no democratic grounds on which to contest the make-up of any particular demos, as democracy, being a mere method, speaks only to the procedures through which decisions are made within that demos. Thus, if a state were to exclude a certain race, religion, or social class from participating, this would be perfectly consistent with democracy, and we would need some other, non-democratic grounds on which to protest. It is this way of thinking about democracy as a decision-making method that forms the basis of the oft-cited boundary paradox ⁸ in contemporary democratic theory. ⁹ According to the boundary paradox thesis, any attempt to legitimate the boundaries of a demos on a democratic basis would lead to an endless regress: in order to decide democratically who is to be included in the demos and who excluded from it, there would already have to be some kind of ‘demos’ in place authorized to make that decision; this then raises the question of how to democratically constitute the demos that chooses the demos; and so on. Versions of this argument have been used to show that the question of boundaries is simply beyond the scope of democratic theory; alternatively, it has been used to argue that the only democratically legitimate demos is one that includes everyone ¹⁰ or is at least ‘unbounded in principle’. ¹¹

Yet most political theorists do not think of democracy as a mere decision-making method. A more common way to think of democracy is as a form of governance whereby ‘those subject to law as its addressees can at the same time understand themselves as authors of law’. ¹² Above all, the democratic governance model seeks to institutionalize a set of values protecting those subject to a government’s political power against tyrannical abuse by requiring that those subject to power should have an effective voice in how that power is constituted and deployed. Democracy’s function is, on the one hand, to secure the freedom of those subject to governance from domination and, on the other hand, to ensure that the decisions made by government cohere with the interests and values of the governed. ¹³ This approach informs a number of proposed solutions to the boundary problem: while it may not be possible to determine the demos by means of democratic method, this does not prohibit us from deducing who should be included on the basis of the values democracy is meant to embody. Thus the principle of ‘all affected interests’ states that the demos should encompass all those whose ‘interests’ are (‘actually’ or ‘possibly’, depending on the account) affected by a governmental decision, ¹⁴ while the principle of ‘all subjected persons’ specifies participation in terms of those persons subjected to laws or coercive power of a given institution. ¹⁵ If democracy is about the participation rights of the governed, the all-affected and all-subjected principles are tools for discerning who ‘the governed’ are.

But there is a third way of thinking about democracy. Decades before Schumpeter offered his ‘minimalist’ view of what democracy signifies, John Dewey insisted that democracy is much more than a form of government, but is rather a way of organizing society. ¹⁶ Evident in both the political liberalism of John Rawls and Jürgen Habermas’ discourse theory of democracy, this model of democratic society has become mainstream in contemporary political thought. ¹⁷ Under this model, the ideal of democratic self-determination has not to do just with the negative values of checking power, but it is equally concerned with the positive values of developing and maintaining a form of life that participants share with one another, and which finds anchorage in a common system of laws and robust institutions. It invokes not just the vocabulary of power, rule and domination but of solidarity, recognition and obligation. Thus the hierarchical relation between those exercising power and those over whom power is exercised must be grounded in the reciprocal relation of free and equal citizens who share a sense of mutual responsibility to one another and who understand themselves as co-participants in a common collective project. This more ‘maximal’ understanding of democracy thus presupposes a deeper set of social bonds among members of a demos than the democratic governance model. Hence defenders of the nation-state status quo frequently argue that the boundaries of the demos cannot simply be removed, expanded, or otherwise redrawn without disturbing the sense of solidarity that makes democracy viable, and it is therefore best that we accept boundaries as historically given. ¹⁸

David Miller casts democratic governance and democratic society as two fundamentally different models of democracy that adhere to different sets of democratic values. On his reading, any attempt to adjust demotic boundaries must involve a ‘trade-off’ between the negative values of checking power (democratic governance) and the positive values of sustaining the necessary sense of mutual responsibility (democratic society). Because the form of reasoning that descends from each value-set produces a different conclusion about where (or where not) to mark the boundaries of the demos, we can espouse one set only at the expense of the other. ¹⁹ However, the vision of democratic society Miller and other liberal nationalists assume holds only so long as the solidaristic conditions necessary to sustain these positive values of democracy can be said to obtain independently of the political-legal guarantees that secure the negative values. Sarah Song, in arguing that demotic boundaries should remain as defined by existing states, identifies solidarity as an ‘instrumental’ condition for democratic society and thus antecedent to realization of democracy itself. ²⁰ But as others have shown, this and other versions of what Arash Abizadeh calls the ‘prepolitical ground thesis’ of the demos are highly problematic. ²¹ At best, they obscure the multiple ways in which relations of solidarity are always themselves permeated by political controversy, internal diversity and division, rival interpretations and periodic efforts at redefinition and renewal. ²² At worst, they can give licence to paternalistic and extra-democratic attempts to enforce a particular idea of the ‘true’ demos over and above the views of existing and potential members. ²³

We should be cautious, however. Proponents of the democratic governance model sometimes use this fact that solidarity relations are themselves politically contested as an excuse to disregard them altogether. Thus Abizadeh asserts bluntly that ‘there is no collectivity prior to institutional articulation’ and, consequently, the only real consideration when considering the boundaries of the demos should be the relation of ruler and ruled. ²⁴ But showing that solidarities are contested is not the same as showing that they are irrelevant. Even if relations of power are, quite justifiably, the ultimate consideration when considering the proper scope of the demos, ²⁵ as a democratic matter we cannot simply override the self-understanding of democratic participants by appealing to superordinate principles of right inclusion. We can agree with Abizadeh’s point that liberal democracy need not require a ‘cultural’ or ‘prepolitical’ nation and still heed Margaret Canovan’s warning against taking for granted the resources necessary for democratic cohesion. ²⁶ For the boundary problem to be addressed democratically, both current and potential members must be able to see themselves as part of a democratic way of life they can claim as their own.

This brings us to one more way to characterize the democratic society model, one by which these positive and negative values of democracy emerge together. This is the conception, for example, envisioned by the deliberative theory of democracy. Habermas posits his conception of deliberative politics in contrast to the classical liberal and republican conceptions: whereas the liberal conception interprets democratic politics as the aggregation of individual private citizens defending their various rights and interests against the imposition of public power, the republican conception casts the democratic community as a singular ‘macrosubject’ whose citizenry combines into a unified general will. The deliberative conception, in contrast, takes the subject of democratic politics to be neither the aggregation of individual rights-holders nor the united general will but the process through which citizens navigate diverse identities, exchange claims, test arguments and assess decisions. ²⁷ On this account, the ‘common good’ provided by democracy is not the sum of the ‘subjective’ goods of individual participants, nor is it an ‘objective’ good that exists apart from and above its participants; instead, it is an ‘intersubjective’ good that accrues in the way citizens engaged in political, moral and ethical discourses accept responsibility both for building a common institutional order of civic life and for mutually guaranteeing each other full rights and entitlements as co-equal citizens. ²⁸ The role of democratic institutions in society – at the level of basic constitutional guarantees, participatory democratic procedures and fair processes of adjudication – is to secure the conditions under which citizens can exercise the communicative freedom necessary for them to see themselves as the authors of the law as well as its subjects.

Taking this conception of a deliberative democratic society as our starting point allows us to bring the contours of the boundary problem into sharper focus. Instead of asking whether it is the positive values or the negative values of democracy that should take priority in determining the proper scope of the demos, we can frame the democratic boundary problem as a discrepancy that arises between, on one hand, the level at which citizens (and would-be citizens) intersubjectively deliberate the conditions under which they coexist in a democratic society and, on the other hand, the level at which these conditions are institutionalized in positive law. It is in the discrepancy between these two levels that the concept of the demos itself opens up into a dual structure. The premise of the deliberative democratic model is that democratic legitimation arises both initially and ultimately from the level at which participants in a demos engage each other as coequal citizens, which is constituted by the second-person relationships that they take up with one another. But the institutional mechanisms for securing the communicative freedom that makes effective participation possible require a third-person perspective that can claim an administrative view over the demos as a whole. In addition to the boundary-related issues of distinguishing who is and who is not a member, it is only from this view that the diverse voices circulating throughout the demos can be brought together and channeled into binding decisions. My argument is that conceptualizing the democratic boundary problem requires us appropriately to conceptualize these two levels of the demos and how they relate to each other. ²⁹

II Two meanings of ‘democratic legitimation’

Even if we settle on a deliberative conception of democratic legitimacy, there still remains an ambiguity in what we mean by democratic legitimation when it comes to the boundary problem. The idea of ‘democratic legitimation’ can be used in reference to the process of establishing something in a democratic way, but it could also mean democratizing something already established. Some treatments of the boundary problem seek principles for establishing the boundaries of the demos outright, as if we were beginning from the ‘state of nature’ of social contract theory. Robert Goodin argues, for example, that ‘unless we know what the right way to constitute the demos is in the first place, we have no way to know whether any given empirical tendencies within democratic politics will … lead them to (re)constitute the demos in precisely the right way’. ³⁰

It should not be surprising that establishment-focused approaches to the boundary problem tend to end up favoring global demos as the proper scope of the demos. One argument stems from the boundary paradox: given that it is impossible to decide democratically who is included in or excluded from the demos without including all those who might be excluded as a result of that decision, the demos tasked with democratically justifying boundaries can only be global, since there is no democratic justification at the outset to exclude anyone. ³¹ The only other way to determine the proper scope of the demos at the outset is to introduce some superordinate principle like ‘interlinked interests’ or ‘all affected interests’. ³² But such principles do not really address the problem of establishing boundaries through democratic means; instead, they seek work-arounds that would make the problem of democratic process superfluous. ³³ Democratic legitimacy requires that members be able to recognize the demos collectively as ‘their’ demos. Even if the correct solution in theory is the establishment of a global demos, in practice it is difficult to imagine how a global demos could be simply imposed upon existing demoi and still be considered democratic. ³⁴

In any event, drawing a mere logical conclusion about what the demos ‘ought’ to look like runs into the additional problem of how the ideal demos is to measure against the actually existing demoi and their non-ideal boundaries. Albena Azmanova points out the tendency of ideal theories to fall into a ‘judgment paradox’, whereby the more rigorously a theory is constructed according to precise principles of justice, the less relevant it becomes to actual political situations. ³⁵ Indeed, the more idealist visions of the proper demos diverge from the global order of demoi we currently have, the more fanciful, utopian and impractical such visions are bound to appear. Arguments for a global demos, even if correct in theory, are a long way off from the way democratic politics is organized today. Moreover, claiming the demos ‘ought’ to be global gives us little guidance by itself on how to achieve that goal, nor does it tell us much about how to evaluate the legitimacy of the boundaries of existing democratic polities.

A better approach might be to frame the boundary problem as a question of democratizing boundaries – that is, of beginning with demoi as they exist and seeing if their policies toward boundaries can be reimagined on a more democratic basis. This can have the advantage of putting the boundary problem more in touch with actual political controversies that have to do with boundaries, such as the status of immigrants and refugees, accounting for historical injustices, or the effects of globalization. In contrast to Goodin, the question here is not to discover what the correct result of a democratic decision on boundaries would look like; rather, it is to find ways to mediate controversies over boundaries in ways consistent with the ideal of democratic process.

Favoring a democratization focus over an establishment focus does not rule out the possibility that we would need to establish institutions that do not yet exist to adjudicate certain issues. There are after all a number of questions related to boundaries that simply cannot be addressed democratically except on a supranational if not global basis. But there are two points to be kept in mind. First, while there are some boundary issues (e.g. secession disputes, certain effects of climate change) that demand the establishment of and deference to higher-level if not global democratic institutions, there are other issues (e.g. immigration) in which existing nation-states can at least make some headway in allowing contestation on a more if not completely democratic basis. Second, even if the institutionalization of higher-level demoi does prove necessary, the movement from state-centered to supranational demoi must still be able to draw on democratic processes of legitimation that originate from within already existing demoi. To the extent that democracy must be conceived as more than a decision-making method or a form of governance, but also a way of organizing society, the path to establishing demoi in a democratic way can run only through the further democratization of demoi that are already established.

III The dual structure of the demos

We can approach the problem of democratizing the demos by recognizing that the concept of the demos has a dual structure, consisting of third-person and second-person levels. In brief, the terms ‘third-person’ and ‘second-person’ follow the conventions of grammatical reference and its system of pronouns. We say we are referring to someone or something in the third-person when we can grammatically replace that referent with a pronoun such as ‘she’, ‘he’, ‘it’, or ‘they’. The referent is that which is spoken of in a given speech act. In the second-person, in contrast, the referent of a speech act is also the addressee: she is not merely spoken of but spoken to. Second-person referents can be replaced with some version of the pronoun ‘you’ and sometimes ‘we’. ³⁶ When we take up a third-person point of view on a referent, we are referring to it from the external standpoint of an outsider looking in; the referent is not a participant in its own reference. When we take up a second-person point of view, in contrast, we are looking for an internal point of view; that to which we refer in the second person is, by virtue of its being addressed, being brought as a participant into the conversation in which it is referenced.

Implicitly, democratic theory has long recognized that the demos contains both third-person and second-person aspects, but it generally privileges the third-person point of view on the demos over the secondperson point of view. This is because, until recently, democratic theory has always assumed the paradigm of the sovereign state. As John Ruggie (1993: 159) has argued, this paradigm has its own particular point of view from which ‘political space [comes] to be defined as it appear[s] from a single fixed viewpoint’. ³⁷ This is a particular version of the third-person point of view that Ruggie traces to the perfection of ‘single-point perspective’ in Renaissance aesthetics. Its conceit lay not in its aspiration to representational objectivity and precision per se but in the assumption that said objectivity and precision can be encapsulated within a particular point of view: ‘a single point of view, the point of view of a single subjectivity, from which all other subjectivities were plotted in diminishing size and depth toward the vanishing point’. ³⁸

Ruggie places the single-point perspective at the core of the ‘social episteme’ of modern sovereignty. It is the perspective that entitles the sovereign observer to bring its society under its gaze as a unified whole: coherent, transparent and – to use James Scott’s term – ‘legible’. ³⁹ This way of ‘seeing like a state’ had practical as well as epistemic implications: it served as a template not just for ‘observing’ society but for making society ‘observable’, for the active reorganization of political space around a common center. It is implemented, if not through outright coercion, in systems of documentation, identification numbers, standardized languages and education curricula, communication and transportation infrastructures, and the like, which organize a polity into a single continuous space within its boundaries. ⁴⁰ In Torpey’s terminology, the single-point perspective is a means not just of observing a society but of ‘embracing’ a society. ⁴¹

It makes sense that democratic theory would emulate this single-point, sovereigntist perspective. Since the 17th century, the sovereign state has provided both the venue in which modern democracy developed and the problem with which it was concerned. The very term ‘popular sovereignty’ indicates that the problem to which it is a solution is not the sovereign perspective as such but merely who should occupy it. Thought-experiments about a ‘state of nature’, ‘social contract’, or ‘original position’ were not designed for considering the proper boundaries of society – they took it for granted that they already had the relevant demos squarely in view. This is not to say that they ignored the second-person aspect of the demos, but these second-person relations were taken to hold good only within boundaries presupposed by other means. Since the 17th century, the demos of democratic theory has always been the state’s demos.

This is not a problem so long as it is assumed that demoi derive their boundaries from those of existing states. But if it is the boundaries of the demos that are themselves the topic under investigation, this association can no longer be taken as given. Instead, the concept of the demos must have a validity prior to the existing state, ⁴² since whether or not it fits the state is precisely what remains to be answered. If we accept that the demos can have a conceptual coherence independently of the state, if it forms itself ‘from the bottom up’ as a community which only then submits itself to the authority of a state, then it is no longer obvious that it is from the single-point perspective of the state that the demos should be defined.

A further advantage of distinguishing between third-person and second-person perspectives is that it brings into focus the distinction between claim-contents and claim-making. Philosophers of language commonly distinguish two levels of linguistic understanding that undergird practical rationality: the semantic level, which concerns the contents of statements; and the pragmatic level, which concerns the way statements are used in communicative practice. ⁴³ Any third-person statement about something can be made only in the context of a speaker addressing to a hearer a second-person claim about that thing. This is not peculiar to the representation of demoi; it is a feature of the way we represent all things through language. However, it becomes an issue with particular practical implications when we are talking about social entities and groupings, for it is here that a gap begins to open between the perspective of a participant engaged with other participants and that of an observer making a judgement about the identity of the whole. The danger here is a familiar one in interpretive social science: it is the risk of reifying the social entity by means of an over-hasty objectification of what may in fact be a diffuse, heterogeneous and contested collection of perspectives, practices and relationships.

Yet there are reasons to think this issue is pressing for the concept of the demos in a rather particular way. For starters, a demos is more than a diffuse collection of agents, relationships and practices; it also possesses a centralized organizational form that integrates it through formal rules, procedures and membership criteria. An important feature of a demos is that it is codified in law, and it is this centralizing feature that allows us to refer to it objectively in a way that is less problematic than when we refer to other types of entities, such as language groups, cultures, or even nations.

At the same time, what makes a demos a demos is that it aspires to be organized not just through law but through democratically legitimated law. ⁴⁴ Because democratic legitimation requires participants to be able to see themselves not just as subjects of law but also as authors of law, the demos cannot be reduced to its mode of legal institutionalization in a way that excludes the second-person voices of the speaking, hearing and acting agents it comprises. The dialectical tension between third- and second-person perspectives – as Habermas puts it, between ‘facticity and validity’ – is incorporated into the very concept of the demos in a way that remains latent or at least informal in other contexts. For here it is not simply that third-person claims about the demos are ‘pragmatically’ dependent on second-person acts of claiming; ⁴⁵ they are made so dependent by virtue of the idea of democracy as a form of self-determination by a citizenry of politically autonomous agents.

To overlook the significance of this distinction is to overlook the significance of the deliberative turn in democratic theory, for which the second-person level of claim-making and reason-giving is central to the idea of democratic legitimacy as such. This can be demonstrated in reference to the ‘boundary paradox’, the thesis that the boundaries of a democracy cannot be legitimated democratically because any attempt to do so would produce an infinite regress. This account of the boundary paradox appears an intractable theoretical problem only to the extent that we define ‘democracy’ as a mere decision-making method in which the sole means participants have for effective political communication is voting. This takes us back to a Schumpeterian ‘aggregative’ model where democratic politics is little more than the weighing and tallying of exogenously given claim-contents of the individual citizens’ preferences. ⁴⁶ In addition, because this model of democracy as a method of aggregating preferences elides the way individual preferences are generated through the second-person engagements of claim-making and reason-giving that demotic participants have with each other, it forecloses attributing any content to the demos itself that is endogenous to the polity’s democratic character, such as solidarity or mutual responsibility. If the only relevant relationship citizens have to their democracy is their individual relation to the state via the medium of the ballot, then any broader sense of affiliation to each other as co-participants in the demos can be attributed to them only exogenously to the political process. The aggregative model presupposed by the boundary paradox abets the assumption that democratic solidarity requires some kind of ‘prepolitical’ ground.

Moreover, while voting and free elections are essential to democracy – along with the right to stand for office, they remain without doubt the most express formal embodiment of democratic legitimation – they neither exhaust the requirements for democratic legitimacy, nor do they guarantee it. Activities such as free and open expression, public debate, civil society organizations, civil disobedience, mass demonstration, protests and strikes are also essential to democratic legitimation, and sometimes they are even more effective than the simple casting of a ballot. ⁴⁷ Conversely, the mere availability of universal suffrage and free elections is no guarantee that the laws and policies adopted will actually reflect the beliefs and wills of those casting the votes. ⁴⁸

Opening up our conception of democratic politics to these additional means of democratic expression – which rely more directly on second-person relations of association than more formal procedures – will not allow us fully to escape the boundary paradox, but it does allow room for greater nuance and maneuverability. This is why, in approaching the democratic boundary problem, it would make more sense to follow the deliberative formula for democratic legitimation forwarded by Habermas and Rainer Forst: that those subject to law also be able to understand themselves as authors of law. At first glance this looks like a retreat into a weaker standard than that of formal suffrage, since it need not in principle mean inclusion in formal procedures. But in another sense it is actually a stronger standard, since in the end it assumes a more active affinity between popular will and enacted law than formal procedure can often guarantee on its own.

IV Beyond the sovereigntist perspective

Most accounts of demotic boundaries present the concept of the demos in third-person terms: ‘The demos is X’, ‘The demos ought to be X’. This is not a problem in and of itself: it is impractical if not impossible to avoid referring to the demos in this way, and the concept of the demos, being dual, ultimately requires both the second-person and third-person levels. The question is how we relate these two levels to each other.

The problem with many attempts to address the boundary problem is their tendency to tacitly adopt the conceptual assumptions of the sovereign state, which presupposes the sovereigntist perspective of an observer ‘looking down’ from a single fixed point. This traps the theoretical purview within a particular kind of third-person stance divorced from the second-person perspectives of demotic participants, and it restricts our understanding of how demotic boundaries function and the theoretical alternatives available to us. If, on the other hand, we were to begin from the second-person perspective and ‘work up’ to the third-person perspective, these assumptions change substantially.

The chief problem with taking this sovereigntist perspective for granted with regard to the boundary problem is that, as Ruggie notes, it utilizes a very particular conception of the political space in which demoi are situated. The sovereigntist perspective envisions political space as absolute space: the demos is all that is within the ‘frame’ of the sovereign point of view; the boundaries of the frame represent the external limits of the demos, which hermetically contains all that transpires within those limits. This view cultivates a presumption of demotic isolation – a presumption that a given body politic can be taken and treated from the outset as a discrete, self-contained unit unto itself, disconnected and unaffected by anything that might be happening outside of it. Of course, few truly believe that demoi subsist as isolated units. Democratic-governance approaches like the all-affected-interests principle, for example, aim precisely to address the fact that people in nominally ‘separate’ societies can greatly influence each other’s lives in ways that might require democratic redress. But the principle still retains the general idea of a demos circumscribed from a central point, where the relevant relation is that between a hypothetical locus of decision-making and the particular persons being ‘affected’ by the decisions issuing therefrom. What relationship the included persons have to one another is defined by their common relation to this central locus; their relation to everyone else is defined negatively, by the latter’s absence of such a relation – they are ‘out of frame’.

The difficulty with this top-down conception of bounded space is that it excludes the contexts in which contests over boundaries are made. This problem is in line with Sofia Näsström’s critique that arguments like the all-affected-interests principle are better suited for criticizing existing boundaries than setting new ones. ⁴⁹ A person outside of a given boundary can claim to be affected by decisions made within that boundary and thereby entitled to democratic inclusion, but such acts of claim-making can be incorporated into our model of boundary politics only if we first presuppose a second-person level of exchange between insiders and outsiders in which these claims can be exchanged and evaluated. Liberal nationalists who appeal to the ‘pre-existence’ of a certain form of collective identity face a similar problem. By judging the existence and contours of democratic solidarity from a top-down sovereigntist perspective, they disregard the second-person contexts that make collective identities and solidarities possible in the first place.

But if the sovereigntist perspective conceives political space in absolute terms, the second-person perspective constitutes political space relationally, through the interactions of thinking, speaking and acting agents ‘on the ground’ who engage one another as ‘I’, ‘you’, or ‘we’. Agents may have third-person views of their own on what the space they inhabit does or should look like, but these views remain exposed at the pragmatic level to second-person challenges and demands for justification. ⁵⁰ Political space is not merely observed; it must be constituted through the way actors navigate and negotiate their relations to one another. ⁵¹ Boundaries likewise take on a different character at this level. Absent a sovereign observer to decide who is ‘in frame’ or ‘out of frame’, boundaries can be produced only through the interactive relations that agents take up in second-person perspective: boundaries acquire salience only in virtue of the reciprocal relationships of claim-making and reason-giving that already exist between actors on either side of the boundary in question. Boundaries, in other words, are already dependent on trans-boundary processes, and their constitution and maintenance must be understood in the context of networks of second-person relations that extend beyond them. Conceptualizing a bounded community from this perspective forces us ‘to examine how it was formed as a community out of the interconnected space that always already existed’. ⁵²

This gives us an insight into how we might think about the democratic legitimation of boundaries from a second-person point of view. The very act of taking up a second-person relationship in the space of communicative rationality already presupposes a certain level of mutual recognition: each must be in a position to recognize the other as an accountable agent capable of accepting and giving reasons. ⁵³ As participants in interaction approach a condition whereby each is willing to give and respond to reasons ‘without reservation’, ⁵⁴ the recognition each gives to the other’s capacity to be rationally accountable momentarily relativizes any boundary that is otherwise structuring the relationship. This does not necessarily eliminate the boundary, but it does leave the third-person ideas that participants may have about the boundary exposed at the second-person level to new interpretations and reasons, which in turn prompt a second-person response.

Reconstructing the demos from this second-person perspective thus allows us to transcend the sovereigntist paradigm in two ways. First, it allows us to situate the demos within the broader network of interactive relations in which actors draw boundaries and make claims about them. Despite the way they have been portrayed in modern political thought since Hobbes, demoi do not exist as insular, self-contained units. Instead of taking boundaries as the limits of meaningful interaction in a demos, beginning from the second-person level forces us to understand boundaries as products of interaction. Second, in addition to making us rethink how demoi and their boundaries are constituted in space, the second-person perspective also requires us to rethink how demoi reproduce themselves in time. In contrast to the ‘static’ top-down view of a discretely bounded demos in absolute space, the demos, situated in a network of multiple second-person relations that extend both within and across its nominal boundaries, must continually reproduce itself through various exchanges of claim-making and reason-giving, making it permanently subject to discursive renegotiation and revision. It might even be better to think in terms not of a singularly defined demos but of an ongoing discursive process of demosgenesis.

This is consistent with the deliberative-democratic formula that legitimacy arises out of ‘the process of discursively, argumentatively, and deliberatively reaching a generally justified political decision that is always only provisional and revisable’. ⁵⁵ We can provisionally stipulate that a boundary established between two persons is ‘legitimate’ if the actors on either side of it are able to understand themselves as ‘authors’ of the boundary as well as its subject. This will require further elaboration, to be sure, and we still need to relate this principle to the third-person level of the demos – the level at which the demos is institutionalized in democratically legitimate law. While the second-person level is indeed primary to the constitution of the demos, our picture of the demos cannot be complete without an additional level at which second-person acts of claim-making are converted into claim-contents.

A look at Arash Abizadeh’s ‘unbounded demos thesis’ shows why we need this additional, third-person level. Like here, Abizadeh insists that the boundaries of the demos must be justifiable from an ‘internal, second-person’ perspective; he further notes the inherently ‘sovereign’-like point of view from which democratic theory has been developed since the early contract theorists. ⁵⁶ Referring to the boundary paradox, Abizadeh claims the question of whether a community can determine its boundaries democratically recedes once we recognize that the very logic of democratic governance defies boundedness. Using a version of the all-subjected-persons principle, he argues that, not only do states engage in exercises of coercive power that subject people outside their extant boundaries, the boundaries states use to exclude people from citizenship are themselves exercises of political coercion that stand in need of democratic legitimation. The scope of persons entitled to participation in the demos by virtue of subjection to power thus always extends beyond the polity’s formal boundaries by definition. In an earlier article, he claimed this demanded the establishment of a global demos; ⁵⁷ more recently, he claims that the demos, as a regulative idea of democratic theory, evades concrete definition. In his view:

… the collective subject of democratic politics is in principle unbounded, or, more precisely, bounded only by the capacity of communicative decision-making practices to track the outward reach of political power … Hence the demos is both everywhere and nowhere. It is everywhere: in principle unbounded. It is nowhere: a regulative ideal that no actual, politically articulated collectivity can ever fully succeed in instantiating. ⁵⁸

V Institutionalizing the demos

Even if we accept that claims about the demos must originate at the second-person level of claim-making and reason-giving, participants cannot glean a conception of the demos about which they are making claims and giving reasons unless they can also refer to the demos of which they are members, from which they are excluded, or whose boundaries they are affirming, contesting, or otherwise negotiating. For a demos to actually function as a demos, actors must be able to attribute to it contents that can be broadly recognized as constituting features around which it is possible to envision the organization of a collective form of life, to which particular actors engaged in negotiation over the demos can see themselves as possibly belonging or not belonging. It is through the third-person perspective on the demos that actors can take up a reflexive relation to the demos they are negotiating through their second-person acts of claim-making and reason-giving.

Yet while the ability of actors to negotiate third-person accounts of the demos and its boundaries provides a necessary condition for maintaining a demos, the demos cannot fully serve as a venue for the democratic organization of a shared form of life unless it is institutionalized in democratically legitimate law. The role of institutions is to regulate and coordinate the various narratives about the demos into a coherent, general, third-person account that can be broadly recognized as legitimate by those engaged with the demos. Positive law, official policies and legitimate coercion function as tools for managing the diversity of possible narratives about the demos and staving off the centrifugal effects of too much divergence in actors’ ideas about the character of a political community. But even here, the third-person perspective on the demos institutionalized in law cannot be severed from the second-person perspectives of the participants themselves.

In Habermas’ discourse theory of democracy, political power unequivocally arises from below. The legitimacy of power stems ultimately from the position actors take as those who authorize the coercive power to which they are made subject, and this requires them to possess both the freedom and the authority necessary to see themselves as the authors as well as the addressees of the law. In Habermas’ words, ‘Legitimate law is compatible only with a mode of legal coercion that does not destroy the rational motives for obeying the law: it must remain possible for everyone to obey legal norms on the basis of insight’. ⁶¹ This is why the legitimacy of law depends in the last instance on the mobilization of what he calls communicative freedom, which exists ‘only between actors who, adopting a performative attitude, want to reach an understanding with one another about something and expect one another to take positions on reciprocally raised validity claims’. ⁶² Communicative freedom allows actors taking up a second-person relation to one another to understand the reasons for which they act in society as their reasons and the norms according to which they act as their norms. It is not just the right to speak; it is the right to be heard; it is the right to be acknowledged as a rationally accountable actor and to have one’s claims acknowledged with reasons and counter-reasons. ⁶³

As such, the effective use of communicative freedom is most secure when it is anchored in the legal guarantees of formal citizenship: equal liberties, due process, free expression, free assembly, rights to petition, voting and so forth. This puts communicative freedom and legitimate law in a virtuously circular relation to each other: legitimate law secures the conditions under which participants can most effectively mobilize their communicative freedom, and participants mobilize their communicative freedom to shape the content of legitimate law. It also makes the balance to be maintained between the communicative generation and the institutionalized enforcement of legal norms a highly precarious one. Robert Cover and Frank Michelman analysed this relationship with regard to the role of what they call ‘jurisgenesis’ in American practices of judicial review. ⁶⁴ The key point of both concepts is that even formal procedures of democratic law-making do not suffice to guarantee legitimacy. For the contents of law to remain legitimate, they must be able to continue to resonate over time with the various interpretations circulating in society – that is, at the level of second-person exchanges of claim-making and reason-giving. So long as the law claims democratic legitimacy, the state has no dictatorial authority to decree for its citizenry what the law is or means, nor does it have privileged access to its ‘original’ or ‘true’ meaning; rather, its function is to coordinate and manage various interpretations that arise from within the society itself. This does not rule out the state’s occasional need to override the interpretations of at least some part of the community, favoring one interpretation at the expense of another in order to preserve the integrity of the law as a coordinating mechanism; just as often, however, it requires new interpretations of law to bring it into accord with evolved understandings. If the coercive power of the state oversteps its coordinating and managing function to take an ‘authoritarian’ position on what the law is or means, it becomes ‘jurispathic’: it breaks the circle between communicative freedom and law. ⁶⁵

Seyla Benhabib has recently used a similar concept of ‘democratic iterations’ to analyse changing ideas about citizenship and the legitimation of international law. ⁶⁶ But for her the issue concerns not potential conflicts between the communicative and institutionalized levels of the demos but an ongoing mediation between moral universality and ethical-political particularity. ⁶⁷ This leads to the criticism that Benhabib’s account is a little too triumphal, downplaying the risks of real conflicts and reversals while envisioning a forward march toward cosmopolitanism marred by only light friction. ⁶⁸ Taking the dual structure of the demos developed in the preceding pages, we can give the above reading of democratic legitimacy a more schematic reading. If the democratic constitution of the demos originates in the second-person acts of claim-making in which actors engage to generate the third-person claim-contents of the demos, then it must exhibit the same dependence on communicative freedom: the democratic legitimation of a demos and its boundaries ultimately depends on the ability of participants to understand the demos as their demos and for those subjected to boundaries to see them as their boundaries. The role of laws that govern the third-person conception of the demos is to coordinate among and manage the diverse and ever-evolving second-person interpretations of demotic identity that emerge from within and across its nominal boundaries. Neither the state nor international law can accord to itself any privileged standpoint on what constitutes ‘a people’. When they attempt to enforce a sovereigntist point of view on the demos that stands over and above the understandings of the participants themselves, they risk suppressing the very communicative freedom that makes democratic legitimacy and solidarity possible in the first place: they become, as it were, ‘demospathic’. Instead, the role of institutionalized law must be to maximize the effective communicative freedom necessary for participants to negotiate boundaries and demotic identities that they can claim as their own.

VI Legitimating and contesting the demos

A crucial implication of the above is that the make-up of a given demos can never be static. The ability of a democracy to maintain its demos in a way consistent with deliberative-democratic principles depends on its ability to cultivate a virtuous circle of ongoing ‘demosgenesis’ that takes place between members and non-members alike. Putting the above arguments together, we can reformulate the criterion for legitimating boundaries in a way consistent with the idea of democratic legitimacy as follows:

The communicative freedom necessary to legitimate a boundary presupposes the communicative freedom necessary to effectively challenge the boundary, and vice versa.

At the very least, we can demand of existing demoi adherence to the minimum guarantees necessary to set in motion the virtuous circle between effective exercises of communicative freedom and legitimate law. ⁶⁹ This can be interpreted here as the minimum rights necessary for immigrants and refugees, global civil society groups and world policy organizations to pursue a fair hearing of claims within various demoi (including claims based on all-affected or all-subjected principles), as well as the reasonable expectation for nominal outsiders to make valid claims for greater inclusion. The formal right to vote is a powerful, legally secured means of effectively exercising one’s communicative freedom in deliberations over a demotic order in which one can see oneself as an author as well as a subject. But it is not the only means. Since the democratic value of formal suffrage lies precisely in its ability to maximize and protect the communicative freedom of those who exercise it, voting represents but one end of a broad and varied spectrum of possible avenues for exercising communicative freedom, with greater or lesser efficacy. This allows us to think about how demotic boundaries may be viewed as more or less democratically legitimate, depending on the extent, security and effectiveness with which insiders and outsiders are able to exercise the communicative freedom in negotiations over them. In the case of immigration, for example, it should not be beyond the means of social scientists and NGOs to tabulate, rate and monitor, across democratic countries, the capacities of documented and undocumented migrants to enjoy legal and civic protections, publicly organize, protest, strike, vote in local elections, have some form of political representation, and so on up the scale of political participation. Though hardly a solution by itself, such a cross-country rating (perhaps in the style of Freedom House’s ‘Freedom in the World’ reports) could provide an argumentative resource for groups pushing for greater rights in the public sphere. Measuring the ‘effectiveness’ of claims by persons not living in or trying to enter a given country, but who are claiming participation rights based on affectedness or subjection, would be admittedly trickier, but one could still imagine possibilities.

Specifying the legitimacy of demotic boundaries in terms of the effective freedom to contest them helps us ‘push back’ on the boundary paradox, but it does not overcome it altogether. The boundary paradox can still approach situations of intractability when the third-person level of the demos divorces itself from some or all persons making second-person claims. One such situation is when formal democratic procedures operate in such a way that a minority group within the established demos is routinely marginalized to the point where members of it are no longer able to understand the laws to which they are subject as their laws or the demos in which they are included as their demos. These are instances in which a minority group might reasonably make claims for demotic autonomy or even secession. ⁷⁰ While there may still be sufficient goodwill among the parties within the existing democracy to arrange amicable terms of separation, if the democratic process has broken down completely, the only way to settle a secession conflict on democratic grounds is to reintroduce the circle of communicative freedom and legitimate law at a higher-level institution. Nancy Fraser, for example, advocates the establishment of global democratic institutions charged specifically with the task of adjudicating boundary disputes. ⁷¹ But there is no reason – at least not with regard to local boundary disputes – that such an institution has to be global in every case. An adequately democratized EU, for example, whose democratic authority were accepted by Catalans, Basques and Spaniards alike (or Scots, Irish and Britons), would likely be sufficient to adjudicate their various disputes – though the make-up and boundaries of the European ‘demos’ are subject to the same standards of legitimacy as lower-level demoi.

Higher-level institutions might also sometimes be necessary for more ‘complex’ forms of demotic autonomy – for example, of minority groups that span existing demotic boundaries or inhabit well-defined territories. Understanding the constitution of demoi in terms of second-person relations can also allow us to consider the demotic viability of groups that do not today fit the traditional nation-state mold, such as diasporas (the Hmong), identity groups divided among several states (Kurds), demoi that also share identitarian bonds with persons in other states (Hungarians), or groups that live among each other within a common territory (Bosnia and Herzegovina). From the perspective of democratic adjudication, identity claims such as ‘ethnic belonging’ or ‘common culture’ can count as valid reasons among others for formalizing in law certain kinds of associative ties, but they cannot stand alone as a basis for exclusionary boundaries that override other possible claims. Like any set of third-person claim-contents about the constitution of a demos, narratives of ethnicity and culture are contested and subject to revision in broad networks of second-person claim-making, with regard both to their content and to their perceived importance to demotic identity. But this does not rule out their validity as possible resources for empowerment and autonomy.

This of course points to the challenge of democratizing international institutions to a sufficient extent that they can legitimately take on such adjudicatory burdens, which points in turn to the task of preparing the ground for generating demoi beyond the bounded territorial state. The first step is to recognize that the question of the democratic legitimacy of the demos can never be fully separated from the question of democratic legitimacy in general. If the boundaries of a democracy cannot be perceived as legitimate by both insiders and outsiders, the legitimacy of the democracy itself is likely to suffer as the pressures of transnational flows continue to increase.

Conclusion

In this article, I have argued that democratic theorists dealing with the boundary problem rely on a truncated conception of the demos – one that implicitly relies on the sovereigntist, single-point perspective of the modern state. In doing so, they privilege the third-person observer’s view of the demos over the second-person relationships of participants who give demotic relationships meaning and validity in the first place. It is this level, the level at which actors make use of their communicative freedom to inform, replenish and periodically revise third-person definitions of the demos, that must be prioritized when we consider problems associated with making boundaries susceptible to standards of democratic legitimacy. Yet it is obscured by models of democracy and definitions of democratic legitimacy that presuppose the single-point perspective of the sovereign state.

Understanding the demos as having a dual structure in the way presented here may not be as amenable to the kind of decisive ‘solutions’ preferred by ideal theory, such as the ‘all-affected principle’ or the ‘global-demos thesis’, but it offers greater focus on the politics of demotic meaning and validity as encountered by those who must contend with boundaries in today’s globalizing world. Because the reproduction of the demos depends on the ongoing exchanges of second-person claim-making and reason-giving within and across boundaries, the only way for the state to ‘freeze’ the contours of demotic identity is by exerting an increasingly paternalistic position over the make-up of the demos and the capacities of insiders as well as outsiders to make claims. As is plainly shown in the legacy of modern nationalism, notably in times of war and economic crisis, the hardening of boundaries and the hardening of liberties for citizens as well as foreigners go historically hand-in-hand. To defend stasis in the composition of demotic boundaries, one must be prepared to defend not just the forms of coercion, colonization and domination that first created them but those forms that will be required to maintain it in the future. ⁷² The dual model of the demos outlined here should be both useful and necessary for social scientists and theorists alike to distinguish the sense of community demanded by the democratic state as a state from the senses of solidarity citizens build for themselves as a people, and to identify the subtle and overt ways in which the former asserts itself over the latter. The lifeblood of a democracy is dependent on the ability of insiders and outsiders continually and with maximal communicative freedom to reproduce, renegotiate and reinvent their understandings of democratic communities both within and across their nominal boundaries. The democratization of the demos must be understood as an integral part of the democratic project at large – within, between, as well as across demoi.