THE ETHNOS BEHIND THE NATION

The problem of political legitimacy appears in both empirical and normative guises. We may, from the third-person perspective of the social scientist, ask Weber's question: Which causal factors explain why people in a particular society (act as if they) believe that those with political power have the right to exercise it? Or we may, from the second-person perspective of the political philosopher, ask Rousseau's question: What justificatory reasons can be given for the belief that political power is exercised by right? The first asks what legitimates political power in a particular society; the second asks what legitimizes political power as such, by reference to a rationally defensible principle of legitimacy.Footnote ²

Any configuration of ideas can be approached from these two perspectives. On the one hand, from an external, third-person perspective, we may treat cultural–nationalist doctrine as an ideology, i.e., a configuration of ideas, principles, and interpretations that as a matter of empirical fact inform people's attitudes and are embedded in social institutions and practices (Freeden Reference Freeden1998). From this same external perspective, we may also treat cultural nationalism as a type of sociopolitical phenomenon or movement corresponding to such an ideology. On the other hand, from an internal, second-person perspective, we may engage with (the same) cultural–nationalist doctrine as a sociopolitical theory, critically testing and evaluating the rational coherence and justifiability of the empirical and normative claims it makes (Miller Reference Miller1995; Tamir Reference Tamir1993).

It is the distinctive feature of cultural–nationalist doctrines that they suppose that the exercise of political power is legitimate only to the extent that it is an expression of, or conforms to, the prepolitical culture of a nation. Prepolitical here means prior, both causally and constitutively, to the exercise of political power. The nation is supposed to be prepolitical in the causal sense insofar as its defining feature—a distinct common culture—is not the mere effect of “political imposition,” but arises “more or less spontaneously” from the “authentic” beliefs, values, choices, and/or relationships of its members.Footnote ³ And it is prepolitical in the constitutive sense insofar as co-nationals share some distinct set of properties that do not amount either to being subject to (the power of) the same political institutions, or to its having been decided, by authoritative political procedures, that they share a national culture. (Thus the distinct nature of the nation's culture must in principle be specifiable independent of political procedures, so that any politically enforced judgments about what comprises the nation's culture are meant to discover or track independently specifiable cultural properties rather than constitute them.Footnote ⁴)

Cultural–nationalist doctrine therefore combines the prepolitical ground thesis—the thesis that legitimizing political power requires some prepolitical ground—with the claim that the cultural nation furnishes the required ground. The cultural–nationalist demand for a prepolitical ground of legitimacy in the first, causal sense may seem entirely sensible: It is a demand that draws natural inspiration from the thought that the exercise of power cannot be self-legitimizing, that might does not by itself make right. And the demand for a prepolitical ground in the second, constitutive sense draws inspiration from the thought that legitimately subjecting a group of persons to political power requires that those persons already be, independent of the fact of their common political subjection, bound together as a collectivity. Hence the cultural–nationalist idea is that prior to the political apparatus of the state stands a nation and its culture, and politics is legitimate only insofar as it achieves a proper “fit” with this prepolitical realm. Moreover, the nation is supposed not only to provide the prepolitical basis for legitimacy, but also to compose the collective self over whom political power is exercised. The cultural nation thus answers both the legitimacy and the boundary questions: Political power is legitimate in virtue of reflecting the nation's culture, and the nation's members are simply those who share a common culture. This congruence is why cultural nationalism is a modern theory promising freedom in the face of political power. Political institutions reflecting a culture alien to the nation, by contrast, spell its subjugation to an alien power.

This is precisely the view advanced in one of the foundational texts of cultural nationalism: Johann Gottlieb Fichte's Reden an die deutsche Nation, a series of Addresses delivered in Berlin in the winter of 1807–08 under French occupation. Fichte's stated goal was to rouse the German nation from its slumber to assert its freedom and throw off the Napoleonic yoke. Fichte defined the nation in terms of a common language, which he depicted as the state's prepolitical ground of legitimacy: “The first, original, and truly natural boundaries of states are without doubt their inner boundaries. Those who speak the same language are already, prior to any human art, by mere nature linked together by a multitude of invisible bonds; they understand each other . . . they belong together, and are naturally one, an inseparable whole” (Fichte [1808] Reference Fichte, Lauth, Fuchs and Gliwitzky2005, 267). The nation is prepolitical (indeed, for Fichte, it is natural) because the shared language that is its defining feature and that binds a Volk together arises spontaneously from its shared cultural life and experiences—and not as the mere effect or determination of political power. And it is the nation defined in terms of a shared language and culture that provides the prepolitical ground for political legitimacy and the legitimate boundaries of the state.

Because a ground, as I have been using that term, is some (nonnormative) factual aspect of the world, and because of course (nonnormative) facts do not by themselves legitimize anything, to show how a fact can serve as a ground for legitimizing political power requires showing that it serves some further normative ideals or values. Fichte himself (Reference Fichte, Lauth, Fuchs and Gliwitzky2005, 191–93) was clear about this: The fact of shared nationality is supposed to legitimize political power because the linguistic–cultural nation makes possible, and is the locus for, freedom in its highest form; and the nation is the locus for freedom in part precisely because it is (supposedly) prepolitical.

If the cultural nation is what legitimizes the exercise of political power over the nation's members, and if having a distinct culture (or language) is what tells us where the nation's boundaries are, then we must ask: (1) Which individuals qualify as members by virtue of sharing a distinct culture (or language)? And (2) What criteria of qualification determine membership, i.e., what does it mean to share a distinct culture (or language) in the first place? The trouble with these questions is that the boundaries of culture are notoriously fuzzy. They face a problem of closure: the problem that neither specifying the putative boundaries of cultural groups (which individuals?) nor specifying those of cultures themselves (what criteria?) admit of determinate answers.

Of course it is sometimes thought that nationalists have a determinate answer ready at hand, in a subjective criterion of membership: Perhaps individuals share a distinct culture, and so are members of the same nation, just insofar as they believe that they do. In fact nationalists frequently do claim that co-membership of the nation is partly a matter of identifying with each other and having a sentiment of belonging together. But subjective identification is insufficient for the cultural nationalist: the judgment that we are co-nationals is supposed to track (and be justified by) the set of cultural characteristics that we uniquely share.Footnote ⁵ What the cultural nationalist is after is an account of culture that identifies a set of properties uniquely shared by co-nationals and that does not collapse into merely being recognized, as a matter of social fact, as sharing a culture. What is needed is criteria available from within an internal, second-person perspective, and not merely from the external, third-person one.

Most definitions “insist that cultures embrace the ideas as well as the institutions and practices of a society” (Fleischacker Reference Fleischacker1992, 160). Thus to distinguish one cultural group from another, to say that these individuals share a culture whereas those individuals share a different culture, it is presumably necessary to identify the set of ideas, institutions, and practices in which these, but not those, individuals are invested. The trouble is that no two individuals will ever be invested in all the same ideas, institutions, or practices. Between any two individuals there will be some similarities and some differences—indeed, an infinite number of similarities and differences—and in practice these will be cross-cutting. Thus in one respect individuals A and B may be similar, because both speak German, but different from C and D, who speak French; whereas in another respect A and C may be similar, because they both are Catholic, but different from C and D, who are Protestant. If culture were individuated according to linguistic practice, then A and B would seem to share one culture, C and D another; but if it were individuated according to religious beliefs or practice, then A and C would seem to share a culture, B and D another. Given the ubiquity of these cross-cutting variations, appealing in general to shared ideas, institutions, and practices will invariably fail to pick out distinct cultural groups: In practice there will always be both internal variation in cultural features among the members of the putative group and external overlap with putative nonmembers (Patten Reference Patten2011).

We therefore presumably need to know very specifically which subset of ideas, institutions, or practices are decisive for differentiating cultures and identifying their members, and why this more restrictive subset should be decisive. One problem here is that the individuals uniquely picked out by the most plausible subset candidates do not correspond to the groups that social actors themselves subjectively pick out. Furthermore, not only is there no non-question-begging justification for choosing one specific subset of features as decisive over all others, but the seemingly most plausible candidates are themselves subject to the problem of closure: not just specifying cultural groups (which individuals?), but specifying the boundaries of culture as such (what criteria?) is problematic.

Consider, for example, language. Like Fichte, many cultural nationalists have singled it out as the decisive criterion for shared culture. The initial problem here is that sharing a language does not seem sufficientFootnote ⁶ for sharing a culture: that the Guyanese and Irish share a common language does not thereby demonstrate a common Guyano–Irish culture. The dilemma is that if a criterion such as language still seems to include too many individuals within a putative cultural group (because of internal variation), then adding further restrictive criteria to raise the bar for membership will invariably end up excluding too many individuals.Footnote ⁷ A typical justification for individuating cultures by language, moreover, appeals to the supposed fact that language is the necessary medium for all other cultural features (beliefs, institutions, and practices). Yet however central language may be, sharing one does not prevent variation in other features, and it is not clear why such variation should be irrelevant for individuating cultures. The selection of language (or any other feature) as the essential criterion of shared culture should be seen for what it is: itself a sociopolitical artifact, not some intrinsic feature of cultures (Shore Reference Shore2000, 23).

The even deeper problem is that the boundaries between supposedly distinct languages are themselves fuzzy. This is obviously true of related languages, such as the Romance languages, which historically have simply bled into each other village to village. But the problem of closure afflicts even attempts to specify the boundaries of a language in relation to apparently quite distinct ones, say, German in relation to French. This is in part because the circumstances under which the demarcation of cultures or languages is politically relevant are those in which the supposedly distinct cultural entities are in contact. Languages in contact affect each other, not just via loan words, but also via loan coinages, that is, words that are “domestic” in construction, but that “would never have been coined, or would never have acquired their modern meaning, were it not for the influence of a foreign model” (Martyn Reference Martyn1997, 305). An example is the German word Geist, which came to mean “spirit” only as a translation of the Latin spiritus. So Iranians who insist that their culture is Persian (and thus Indo-European) and not Arab (and thus Semitic) will likely omit mentioning that over a third of Persian vocabulary consists of Arabic loan words (and that Iran's majority religion is a Semitic one); Spaniards who define themselves against their “Moorish” neighbors will likely do so by helping themselves to Arabic words as well.

A typical cultural–nationalist response to the problem of foreign cultural interference is to see the interference not as calling into question the possibility of delineating unequivocal cultural or linguistic boundaries per se, but as a sociopolitical problem to be dealt with through a defense of cultural or linguistic purity. We see this clearly in Fichte's Reden, where he excoriated the use of foreign (Latinate) loan words. The nostalgia for purity is also already found in Rousseau's Considérations sur le gouvernement de Pologne. If this purism is a form of xenophobia, it is not yet an ethnic xenophobia, but it still supposes that the notion of linguistic purity makes sense. As David Martyn has noted, this is a rather dubious supposition:

A parallel argument about purity evidently applies to any of the other customary candidate criteria for distinguishing culture, such as religion or history.Footnote ⁸ And if purity is chimerical in this way, then the problem of closure goes all the way down: It afflicts not just the attempt to specify which individuals meet the relevant criteria, but the very attempt to specify the criteria themselves. Hence the elusiveness of any satisfactory answer to the problem of closure: because there are an infinite number of features on which different individuals can be similar and different in cross-cutting ways, because the selection of some subset of features as decisive for individuating cultural groups begs the question of why other features are irrelevant, and because any candidate set for what essentially differentiates cultures is itself subject to the problem of closure, the problem of closure remains intractable. Any attempt to specify once and for all the members of a distinct cultural group, and the set of features that constitute its boundaries, faces the insurmountable problem that further difference can always be found within, and similarity across, the collective boundaries that were supposed to mark off difference.

None of this is to deny the obvious fact that boundaries exist as a matter of social practice: People subjectively do distinguish between cultures and languages, and the corresponding group boundaries are frequently enforced by practices of exclusion, shaming, discrimination, incentives, threats, and, of course, political sanction (Barth Reference Barth and Barth1969; Wimmer Reference Wimmer2008). The point is, rather, to expose these boundaries for social reifications (Brubaker Reference Brubaker1996)—a crucial point given that the cultural nationalist is looking for a distinct set of shared properties that could justify these factual practices of boundary-drawing. But the intractability of the problem of closure means that no such justification is at hand: Socially bounded cultural groups or cultures will invariably appear hybrid. Cultures bleed into each other.Footnote ⁹

It is true, of course, that nationalists have frequently proposed an implicit answer to the problem of closure. As Samuel Fleischacker (Reference Fleischacker1992) has argued, nationalism in practice “attempts to settle this question once and for all, and to impose its answer on each individual, at least within the group it selects, by means of a political structure and a body of positive law” (167)—in other words, by political fiat. Exactly which features are central to our culture and exactly who shares this culture with us are specified by those features embodied in our politico–legal institutions. To put it plainly, we can see that our culture deems female circumcision to be beyond the pale because we have made it illegal; and we can see that the core of our culture is egalitarian because we have enacted redistributive taxation laws, and teach egalitarian values in our public schools. We can also see that you partake of our culture because, as your passport attests, you are a legally recognized national.

But with this answer, “shared culture” becomes a project to be realized via the exercise of political power, and not a prepolitical ground for political power. The point is not just the (potentially controversial) causal–empirical one that nations have been invariably formed by the exercise of political power;Footnote ¹⁰ rather, it is also the conceptual one that the effective criterion for determining the boundaries of the nation is constitutively political: There is no independently specifiable feature-set for political judgments about national boundaries to track. The normative problem to which the collapse of nationality into citizenship alludes should be manifest: Under the guise of solving the problem of closure, cultural nationalists have to abandon their claim to having provided a prepolitical ground of legitimacy for the exercise of political power. If it is positive law itself that ultimately and constitutively determines what our culture is and who partakes in it, then culture loses its status as the prepolitical ground for legitimizing positive law in the first place. And if having a shared culture is a project to be realized, rather than an antecedent fact, then the nationalist legitimization of politics and the political project it encompasses, of creating a shared culture, demands some new prepolitical ground.Footnote ¹¹

Cultural nationalists’ compulsion to find another, truly prepolitical ground for answering the legitimacy and boundary questions is exacerbated by the fact that the problem of closure must be solved not just synchronically in space, but diachronically over time as well: The nation is supposed to be a collective self that persists over time. Even if it is difficult for nationalists to concede that cultures bleed into each other, they cannot fail to know that cultures change over time. After all, cultural change is a presupposition of most cultural–nationalist politics—either as a threat to be defended against, in the case of reactionary–preservationist projects, or as a possibility to be realized, in utopian–revolutionary projects. But if cultures change, then what is it that secures the continuity of their over-time identity? What makes this culture today the same as the culture of yesterday from which it is different? The cultural nationalist is already committed to the reification of “culture” as an entity with distinct boundaries, but now he or she needs to go further and defend those boundaries through time as well. He or she needs to identify some “essence” of the nation, an essence whose persistence explains its persistence in the face of cultural change. The cultural nationalist must either locate that essence in some set of core cultural features, or find some extracultural supplement that secures over-time identity.

The first option is what fuels reactionary defenses of putative cultural purity. It is tempting for the cultural nationalist to suppose that a culture's essence can consist in some core cultural feature(s) whose purity over the generations must be defended at all costs: This bit of our culture cannot be changed. (So closure would only be a genuine problem for the fuzzy edges of the culture, without touching the essential core.) The problem here is not just that cultural purity is chimerical. It is also that, as I have argued, the selection of what counts as the essential core of culture is itself a sociopolitical artifact, rather than some intrinsic feature of cultures. This is why each generation's view of what that core is may itself change. So at one point, for example in Quebec, it is the Roman Catholic religion, at another point it becomes the French language. These two problems are related to a third: Although social actors may subjectively experience their culture and language, and the boundaries of membership that they service, as something natural, in the sense of something given to, not produced by, social and political processes, this reified, taken-for-granted experience suffers the greatest strain precisely when it is a matter of reproducing culture intergenerationally over time. Ensuring that the next generation “carries on the tradition” requires, and is experienced as requiring, strenuous effort. The social construction of the nation is recognized in all cultural–nationalist projects that aim to seize the educational apparatus of the state in order to shape the coming generations. The upshot is that although culture may be experienced synchronically as natural—and so presocial and prepolitical—diachronically over time it cannot coherently even be experienced as natural, because cultural–nationalist projects are explicitly always projects of creating culture. Cultural–nationalist political projects, whether in defense of beleaguered purity or to reawaken dormant potentialities, always show the national culture itself to be a political project. But the persistence of a self-styled prepolitical collectivity requires that the collective national self have some prepolitical grounding over time. And this is what the cultural nation, and hence cultural nationalism, cannot coherently provide, even as a matter of subjective experience.

So the cultural nation must go casting about for an extracultural supplement to anchor its boundaries in space and time. And this is precisely what ethnicity, constituted by a myth of common descent, purports to supply: An extracultural basis for the continuity of the cultural nation through time—a natural, prepolitical anchor. Like the cultural nation, ethnicity's status as a prepolitical ground is of course mythical as well (Geary Reference Geary2002). The Turkish state's references to “Mountain Turks” and Saddam Hussein's forcing Kurds to take on Arab names are crude reminders of this. So contemporary scholars such as Donald Horowitz (Reference Horowitz1985) and Anthony Smith (Reference Smith1986) are quite right to define ethnicity in terms of myths of common descent, rather than actual ancestry. But the seductiveness of ethnicity for cultural nationalism lies in a distinguishing feature of ethnic myths. Whereas myths about cultural difference refer to an individual's socially acquired characteristics, which in principle could change—one can change one's beliefs, or learn a new language (Anderson Reference Anderson1991, 134)—myths of ethnic difference refer to an ascriptive feature of the individual that ostensibly confronts the individual and society as a brute fact of nature (Horowitz Reference Horowitz1985, 52; Kymlicka Reference Kymlicka2001, chap. 9). An ethnic nationalist can at least consistently conceive of common ancestry as a natural and hence prepolitical fact about individuals, even if that conception is invariably false.Footnote ¹² But cultural nationalists cannot coherently even conceive of the nation as prepolitical, even as a matter of subjective experience. The logic of cultural nationalist politics itself undermines the mythical self-conception of the nation as a prepolitically bounded unity: The very attempt by cultural nationalists to produce and sustain the integrity of their cultural collectivity through political means undermines the subjective plausibility of that myth.

Hence the need for an extracultural supplement; hence the slide into ethnicity. The boundary of cultural continuity is to be genealogically secured here. Cultural–nationalist theory's incapacity to provide the prepolitical ground demanded by its own account of legitimacy requires the theorist to reach out for such a supplement. Thus Fichte, despite his officially linguistic definition of the nation, and despite explicitly rejecting an ethnic criterion, ended up succumbing to a genealogical supplement (Abizadeh Reference Abizadeh2005b). And it is a supplement to which cultural–nationalist ideology and politics have in practice often found themselves drawn as well. It is true that ideologies are perfectly capable of legitimating political power without resolving their internal inconsistencies or redeeming their truth claims. Yet because ideologies are sites of sociopolitical contestation, inconsistencies or falsities create vulnerabilities: To the extent that state power or political mobilization relies on a particular ideology for legitimation, when its vulnerabilities become politically thematized, additional resources must be expended to diffuse challenges. Social actors must either shore up their power nonideologically, bypassing the benefits of legitimation,Footnote ¹³ or draw on supplementary ideological resources. Thus although the collapse of cultural into ethnic nationalism is neither logically necessary (because there are other potential supplementsFootnote ¹⁴) nor empirically universal, it is not a random occurrence either: It is a disposition of cultural–nationalist politics resulting from identifiable discursive features of cultural–nationalist ideology. That disposition is most likely to be actualized in specific circumstances: The argument here suggests that cultural nationalism as an empirical phenomenon is most prone to collapse into its ethnic kin when the legitimacy and boundary questions are contested and so politically most at issue.Footnote ¹⁵

The result is this: Behind the cultural nation lurks an ethnos, eager to cover its next of kin with a warm and rather constricting embrace. A constructing embrace.

THE NATION BEHIND THE DEMOS

Democrats who lament the eruption of ethnic xenophobia in supposedly “liberal” versions of cultural–nationalist politics would be mistaken to indulge the temptation of smug superiority here. The problem for the democrat is that the reasons cultural nationalism is disposed to collapse into ethnic nationalism have decisive parallels in democratic doctrine itself. We can see this as soon as we ask: Why should cultural nationalism's problem of closure translate into a problem for political legitimacy? The answer is this: Because, like modern doctrines in general, cultural nationalism's answer to the legitimacy question refers right back to the very persons over whom power is exercised. It is a self-referential theory of legitimacy, which combines the questions of legitimacy and boundaries. If the object of power is also its legitimizing ground, if political power is legitimized by (conforming to some aspect of) the collectivity over whom it is legitimately exercised, then we need to know who is the source of, and legitimately subject to, that power. We need to know who is legitimately in the collectivity, and who is legitimately out. We need to know who is legitimately in or out because self-referential doctrines purport to furnish a ground for legitimizing political power, and constituting political boundaries is one of the most important ways in which political power is exercised over persons. So if specifying the boundaries of the collectivity faces a problem of closure, and if answering the legitimacy question requires answering the boundary question, then the problem of closure undermines the grounds for legitimacy as well. (And if the cultural nation is a creature of political power, then its boundaries cannot furnish a prepolitical ground for the legitimization of political power.) The self-referential nature of cultural nationalism's principle of legitimacy, and its consequent linkage of legitimacy to the boundary question, is shared, of course, by democratic theory. No surprise that it faces a parallel problem.

The point of departure for modern democratic theory is the claim that legitimate political authority derives ultimately from the people or demos. Thus political power is legitimized not by tradition, not by virtue, not by genealogy, but by the demos itself. Democratic self-rule requires that political power be exercised in accordance with laws that the people, which is the object of power, has, as the ultimate subject that exercises power, given to itself. This requirement of self-legislation—of the people authoring its own laws—has of course been expressed by saying that the laws regulating the exercise of political power must conform to the expressed “will of the people.” Specifying the ground for democratic legitimacy therefore clearly requires specifying both (a) what procedures give legitimate expression to the people's collective will and (b) who the people in question is. Like cultural nationalism, democratic theory is a self-referential theory that must answer the legitimacy and boundary questions together.

The difference is that whereas cultural nationalism aspires to a prepolitical principle of legitimacy, democratic theory advances a constitutively political principle, one that refers directly to the expressed collective will of a people, i.e., its politically articulated decisions. The problem is this: If the exercise of political power is legitimized by conforming to the expressed will of the people, then the second question (b), of who the people is, cannot itself be settled by interrogating the people's will. Before interrogating the will of the people, one needs to know who the people is, which is precisely what is in question here. If the legitimacy of the state's constitutional order and laws depends on their conforming to the will of the people, expressed via the participation of those who make up the people, then clearly the question of who counts as rights-bearing participants cannot itself be settled by participation. This is because one would once again have to ask, whose participation must be sought to answer the question of membership of the demos, which in turn raises a higher-order membership question, ad infinitum. This is the legitimate boundary problem: The problem that democratic theory appears incapable of specifying a political procedure, consistent with its own account of legitimacy, by which the boundaries of the demos can be democratically legitimized (Whelan Reference Whelan, Pennock and Chapman1983). It is a problem that arises precisely because the democratic principle of legitimacy is not only self-referential, but also constitutively political: It assumes that the ground of legitimacy is the people's politically expressed collective will, itself some function of individual members’ wills.

The boundary problem is compounded by the fact that a genuinely democratic principle is also intrinsically procedural:Footnote ¹⁶ It assumes that, to articulate their collective will, members must actually express their own individual opinions and wills. Political power is democratically legitimate only insofar as it is exercised according to laws resulting from certain kinds of political procedures and processes—ones amounting to the people giving laws to itself. So not just any kind of procedures will do: The notion of the people giving laws to itself requires that all those who compose the people be able meaningfully to participate in them to articulate their collective will. A theory according to which the proverbial “will of the people” can be discerned and expressed by an inspired autocrat, without procedures for the participation of those composing the demos, does not amount to a democratic theory of legitimacy. (Cultural–nationalist theory, in contrast, is not inherently democratic or procedural: There is nothing incoherent about a theory according to which the culture or will of the nation can be immediately discerned and expressed by the nation's inspired leader.) While the self-referential and political character of the democratic principle of legitimacy gives rise to the legitimate boundary problem, its procedural character poses the legitimate procedure problem: If laws must be the outcome of certain kinds of procedures, then the democratic legitimacy of the laws governing the procedures themselves seems to face the same kind of infinite regress as boundaries. Which political procedures legitimately articulate the will of the people cannot itself be wholly determined by interrogating the people's will.Footnote ¹⁷

Traditional democratic theory of the modern period, instantiated in the classics of social-contract theory, proposed a seemingly elegant solution to these twin problems: a prepolitical ground for the legitimacy of political procedures and boundaries. This is why traditional democratic theory has adhered, no less than cultural–nationalist theory, to the prepolitical ground thesis. In one sense, of course, the democratic social-contract theorists identified the ground of legitimacy with the “will of the people”: They assumed that, to be legitimate, political power must conform to a popular will that stands temporally and logically prior to the exercise of political power. But because the will of the people must be articulated via some established political procedure, the social-contract theorists sought a prepolitical ground for the legitimacy of the procedures themselves. This they found in the prepolitically sovereign will of the individual, expressed in a social contract that institutes a distinct people.Footnote ¹⁸ Thus the ultimate prepolitical ground for legitimizing the procedures for articulating the popular will and the boundaries of the people itself is supposed to lie in the unanimous consent of all parties to the social contract.

We need not linger on the well-known failures of this canonical response to the legitimate procedure problem, such as the fact that few citizens have ever actually consented to their state institutions, or that no contemporary state was even originally founded on a unanimous social contract (Hume [1748] Reference Hume and Miller1987; Simmons Reference Simmons1979). What is more important for our purposes is its utter failure as a response to the boundary problem. There are two facets to this failure. The contract theorists invariably assumed that the unanimous consent of would-be members to join a polity is sufficient to legitimize its boundaries. Yet this is clearly not sufficient if the ultimate prepolitical ground of legitimacy for exercising political power over the individual is his or her own will. First, civic boundaries pose an externality problem: On one hand, enforced decisions about who is granted and who is denied membership and political rights are among the most important instances of the exercise of political power; on the other hand, the exercise of such power is intrinsically over both insiders and those whom the boundary picks out as outsiders (Abizadeh Reference Abizadeh2008). This externality problem is exacerbated by the fact that the social contract is supposed not only to establish civic boundaries, but also the territorial boundaries of the area over which the demos exercises jurisdiction. The logic of the social-contract account consequently implies that to legitimize civic (and, by extension, territorial) boundaries requires the consent of all would-be insiders and outsiders (Agné Reference Agné2010, 386). Second, given the premises of social-contract theory, legitimate boundaries require that every individual consent not just to his or her own inclusion or exclusion, but also, in the case of willing would-be insiders, to the inclusion of each other willing individual. Boundaries require not just consent, but serial consent. This is why contract theorists cannot solve the externality problem by replacing the assumption of bounded polities with the assumption of a single, all-inclusive human polity; for then they face the possibility that not every would-be insider will consent to the inclusion of each of the others or, indeed, to his or her own inclusion. What is required is a completely global social contract establishing differentiated political jurisdictions (if any) and the distribution of individuals among them. The trouble is that although unanimity here is logically possible, it is astronomically improbable. Whereas, in the case of procedures, social-contract theorists could at least muster an argument as to why unanimous consent could converge on something like majority rule, they could not even feign such an argument for boundaries: There is no reason to expect convergence on how to cut up political jurisdictions and how to distribute all individuals between them.

If the individual's will fails to provide the prepolitical ground for legitimizing political boundaries, and if legitimizing boundaries requires a prepolitical ground, then the only alternative is to look elsewhere. Traditional democratic theory, beholden to the prepolitical ground thesis, must therefore go casting about for some alternative mark of peoplehood that obtains prior to the exercise of political power. And this is precisely what cultural nationalism claims to provide it. For cultural nationalists, the most obvious candidate for determining the membership of the demos is a common national and ostensibly prepolitical culture, a shared culture that conditions—and so is prior to—even the individual's will. So it is no surprise that, as Brian Singer (Reference Singer1996) has suggested, although the French revolutionary conception of the democratic nation began as a purely “contractual” one, eventually it drew on the extraconstitutional supplement of a common culture. Just as cultural nationalism is disposed to draw on an ethnic supplement, so too has modern democracy, which imagines the “people” to be prepolitically grounded, been disposed to nationalism.Footnote ¹⁹ Incapable itself of legitimizing the prepolitically exclusionary boundaries it has traditionally presumed, modern democracy has relied parasitically on a cultural–nationalist principle of legitimacy, which equates the people with the cultural nation. Of course, that the cultural nation is not a collective entity existing prior to the exercise of political power is ultimately a betrayal of the prepolitical ground thesis motivating this turn.

But the empirical impulse is clear: If behind the cultural nation lurks an ethnos, then behind the demos lurks a cultural nation. Or, more precisely, there is a cultural nation behind every would-be prepolitically bounded demos.

THE DEMOS UNBOUNDED

I have, so far, defended two theses. First, attempts to specify the boundaries of the cultural nation over space and time propel it to identify itself with an ethnos. Second, attempts to specify the prepolitical boundaries of the demos propel it to identify itself with a cultural nation. What instigates the collapse of cultural into ethnic nationalism is the prepolitical ground thesis coupled with the problem of closure; what instigates the collapse of the demos into the cultural nation is the prepolitical ground thesis coupled with the boundary problem.

The failure of the traditional social-contract answer to the boundary problem evidently does not rest on the assumption that polities are necessarily bounded (i.e., that they must exclude some individuals from membership). Although most took the fact of bounded polities for granted, and sought a prepolitical ground for legitimizing exclusionary boundaries, others did envision, without doing violence to social-contract theory, a global human political community (Bartelson Reference Bartelson2008; Brett Reference Brett2011). Thus the claim that the people is prepolitically bounded comes in two versions. The weaker version merely advances the prepolitical ground thesis itself: Demoi as we know them just happen to be bounded, but the legitimate ground for their boundaries is prepolitical. The stronger version is implicit in another strand of modern democratic theory—with Rousseau as exemplar—which combines the prepolitical ground thesis with another: The thesis that the demos is necessarily bounded in principle, i.e., that its boundedness (as distinct from its particular boundaries) is a presupposition of the democratic exercise of political power. According to this second claim—the bounded demos thesis—the fact that the people is bounded is given prior to the exercise of political power and independent of any individual's will.

The bounded demos thesis is pervasive in contemporary democratic theory: theorists of varying stripes and methodologies unite in embracing it as a kind of conceptual or metaphysical truth. As Jean Cohen (Reference Cohen1999, 254–55, cf. 49) puts it, democratic self-rule “cannot exclude exclusion, because the political body in which it is institutionalized is inevitably particular (one among several) and because the demos must be defined and delimited.”Footnote ²⁰ Nationalist theorists have seized on this ostensibly inherent boundedness, along with the apparent impossibility of legitimizing particular boundaries democratically, as evidence that democracy is not only compatible with, but positively requires, some form of nationalism. For these theorists, the boundary problem is supposed to demonstrate that legitimizing particular boundaries requires turning to a principle external to democratic theory itself: It is an indication that democratic theory is parasitic either on cultural nationalism (Tamir Reference Tamir1993, chap. 6) or on ethnic nationalist appeals to shared genealogy (Canovan Reference Canovan1996, Reference Canovan2000). As Canovan (Reference Canovan2000, 427) has put it, “any polity, however liberal its ethos, is and must be an inheritance passed on from generation to generation”: Even in the United States—often held up as the paradigm of the jus soli principle—citizenship can be inherited by being born to citizens, regardless of birthplace.

What this overlooks is that, however much nationalism may provide a supplement to democracy's boundary problem in empirical practice, it wholly fails as an answer to the normative question of legitimacy (cf. Shachar Reference Shachar2009). Modern democratic theory may indeed be driven to rely parasitically on nationalism but, normatively speaking, the nationalist “solution” to the democratic boundary problem is no solution at all: It simply fails to meet the procedural requirements of a democratic principle of legitimacy. Appeal to the nation, moreover, does nothing to eliminate the incoherence at the heart of the received theory of intrinsically bounded democracy. That theory is incoherent because, even though it cannot even in principle provide for the democratic legitimization of the boundaries of the demos, it requires a democratic way to legitimize those boundaries. It requires one because the constitution, regulation, and coercive and symbolic enforcement of boundaries compose one of the most important ways in which political power is exercised over human beings. Recall what the democratic principle of legitimacy implies: That the exercise of political power is legitimate only insofar as it is authored, in some sense, by those persons over whom it is exercised. Enforced decisions about who is granted and who is denied membership and rights, and who controls such decisions, are clearly among the most important instances of the exercise of political power. The incoherence stems from the fact that civic boundaries pose by their very nature an externality problem: Although democratic theory claims to legitimize the exercise of political power by reference to (the participation of) those over whom it is exercised, civic boundaries, which by definition distinguish between insiders and outsiders, are always instances of power exercised over both—and outsiders are precisely those whose participation the bounded democratic polity claims to be able legitimately to preclude. The act of constituting boundaries circumscribing political rights is always an exercise of power over both insiders and outsiders that, by the very act, purports to disenfranchise the outsiders over whom power is exercised (Abizadeh Reference Abizadeh2008). Call this the unboundedness of subjection. It is this conceptual feature of boundaries that confronts democratic theory with an externality problem. The unboundedness of subjection arises from the fact that the power required to constitute political boundaries is intrinsically an outward-extending power.Footnote ²¹

There are at least two potential responses to this predicament available to the defender of the received theory of intrinsically bounded democracy. The first is to follow Joseph Schumpeter (Reference Schumpeter1976, 243–45) and simply deny that the particular boundaries of the demos need to be justified or legitimized at all. If having boundaries is necessarily presupposed by the democratic exercise of political power, then any particular boundary will do: The particular boundary may be arbitrary, but drawing it somewhere is not, and that is all there is to it (Risse Reference Risse2008). In fact that is not all there is to it: Since how boundaries are drawn structures the way that power is exercised over individuals, the putative fact that boundaries are necessary does not preclude the need for legitimizing how they are drawn—any more than the fact that the state may be necessary for democratic agency precludes the need for legitimizing how particular state institutions are constituted. As Sofia Näsström (Reference Näsström2007) has demonstrated, there is no justification, consistent with acknowledging the demands of democratic legitimacy elsewhere, for excluding the constitution of the people and its boundaries from such demands. The denial to the contrary either amounts to arbitrary special pleading, or is based on a manifestly false premise. It amounts to special pleading if the claim is normative: that particular boundaries should be exempted from the demands of legitimization and treated as if they were prepolitically constituted. The denial is based on a false premise if it amounts to the claim that particular political boundaries are, as a matter of empirical fact, constituted prepolitically, prior to the exercise of political power.Footnote ²²

The second response concurs that, because boundaries are necessarily presupposed by the democratic exercise of political power, they cannot and need not be the legitimized outcome of democratic procedures. But it concedes that they must still be legitimized—not by reference to external standards, as nationalists assume, but via standards internal to democratic theory itself. To say that the demos is inherently bounded, on this view, is to say, as David Miller (Reference Miller2009) has recently done, that the boundaries of the demos must be justified theoretically, by reference to the central normative ideals, values, and principles of democratic theory, rather than legitimized via a democratic procedure. This turn to the substance of democratic theory is perfectly justifiable when interpreting the meaning of the democratic ideal of self-rule and its concomitant principle of legitimacy: Both the ideal and principle owe their normative force to a substantive justification, rather than to being an outcome of the procedures the principle demands (Gilabert Reference Gilabert2005). Yet caution is required: To oppose a theoretical to a procedural answer here is misleading to the extent that the theory itself yields an intrinsically procedural principle of legitimacy, a principle demanding certain kinds of procedures for legitimizing political power. And because constituting boundaries is a paradigmatic instance of the exercise of political power, boundaries fall squarely within the purview of democratic theory's procedural principle of legitimacy.

The point, to have traction, must therefore be formulated in terms of the democratic procedures that instantiate the ideal of democratic self-rule. And it is true that a viable distinction can be drawn between, on one hand, legal arrangements that are democratically legitimate insofar as they are the outcome of procedures of collective self-rule and, on the other, legal arrangements that are democratically legitimate insofar as they compose the constitutive presuppositions of such procedures. This is how some have proposed to solve the legitimate procedure problem: to short-circuit the infinite regress incurred by the requirement that laws be the outcome of democratically legitimate procedures by also admitting, as legitimate, laws that are constitutive of such procedures (Holmes Reference Holmes, Elster and Slagstad1988). Political rights of participation, for example, are arguably constitutively presupposed by democratic procedures; therefore, a law denying some subjects these rights, even were it to result from a democratic procedure, would still be democratically illegitimate (Cohen Reference Cohen, Hamlin and Pettit1989; Christiano Reference Christiano2008). Hence turning to democratic theory analogously to address the legitimate boundary problem requires appealing to the constitutive features of procedures and processes that instantiate democratic self-rule, features without which there can be no procedural democratic legitimization at all. Note that one cannot, in contrast, appeal to conditions that may be merely causally necessary for realizing the ideal of self-rule, or for the good functioning of the procedures that constitute the ideal. This would be to conflate, as Miller tends to do, what democratic legitimacy consists in with an empirical analysis of the conditions under which it is feasible. If rights of participation are denied to some on the instrumental grounds that doing so is, as a matter of empirical fact, necessary to preserve the vitality or stability of democratic institutions, one should bite the bullet: The right conclusion is that, under these circumstances, the ideal of collective self-rule and the democratic legitimacy of outcomes have been compromised—not that exclusion is constitutive of them.

Thus to make the required “theoretical” argument, the defender of the bounded demos thesis needs to claim that, in principle, the legal exclusion of some individuals from the people is constitutive of the procedures required for democratic legitimacy. This might be either a conceptual or a metaphysical claim. If the former, it rests on an equivocation. Of course it is true that, in virtue of its self-referential character, democratic theory must specify or “define,” as Jean Cohen puts it, who the people comprises. Such a specification is constitutive of democratic procedures. But it is one thing to say that the composition of the demos must be specified; it is quite another to say that it must be specified as bounded or exclusionary. To equivocate between these two claims, or to think that the former implies the latter, is to commit what we might call a fallacy of specification. For there is no conceptual incoherence in specifying that the people includes every person subject to the exercise of political power. In fact this is precisely, I argue, what is demanded by the democratic principle of legitimacy.

Alternatively, the claim may be metaphysical. Precisely because democratic theory is self-referential, the demos is not just the collection of individuals subject to political power, but must also be the collective author of that power. As such, the demos must arguably comprise a set of individuals capable of institutional articulation, of acting through institutions as a collective agent, and of thereby giving itself laws.Footnote ²³ It might therefore be argued that the demos is bounded in principle, because any such collective agent constitutively requires not just institutional articulation, but also the solidarity associated with a collective identity. The tacit assumption here is, of course, that collective identity presupposes an external other.Footnote ²⁴ The assumption is false. As I have argued previously (Abizadeh Reference Abizadeh2005a), even if individual identity requires an external other—because it develops through relations of mutual recognition—a collective identity does not: The required intersubjective recognition is available internally, from its own members. And even if collective identity requires contrast with what it is not, the contrast need not be with excluded individuals. The metaphysical claim thus either rests on a compositional fallacy, or on the conflation of difference with the exclusion of living persons.

My thesis, then, is that democratic theory does furnish a basis for answering the boundary problem, via a constitutive condition of the democratic ideal of self-rule and its corresponding principle of legitimacy: that all those subjected to the exercise of political power be included in the demos, i.e., granted a right of democratic say over political decisions. This all-subjected principle (to appropriate Nancy Fraser's (Reference Fraser2009, 65) phrase) is to be contrasted with the all-affected-interests principle defended by an increasing number of theorists, according to which the demos must include all those whose legitimate, fundamental interests are affected by the exercise of political power (Arrhenius Reference Arrhenius and Tersman2005; Goodin Reference Goodin2007). The latter principle is constitutive of neither democratic self-rule nor democratic legitimacy: there is no intrinsic connection between effects on one's interests in general and a right of democratic say. As other critics have recently noted, that one's interests are affected intrinsically grounds a moral right to due consideration (e.g., to have harms avoided or compensated), not a political right to a say in the decision-making process (Beckman Reference Beckman2009, 46–47; Owen n.d.). Of course a democratic say might often be, as John Stuart Mill ([1861] Reference Mill and Robson1977) argued, an indispensible means for securing due consideration; but depending on the circumstances, it might also be secured through fiduciary institutions, binding arbitration, or contestatory judicial institutions. The central point is that even when a democratic say is empirically indispensable for due consideration of interests, this specifies an instrumental reason for enfranchisement, not the constitutive requirements of democratic legitimacy.

The democratic ideal of collective self-rule is grounded in the notion that securing the conditions of individuals’ autonomy and standing as equals intrinsically requires that they be the joint authors of the terms governing the political power to which they are subject. That one's interests in general are affected by others does not itself negate self-rule or autonomy and equal standing, but being unilaterally subject to a coercive and symbolic political power, without any say over the terms of its exercise, does. Inclusion in the demos is therefore grounded intrinsically not in individuals’ interests as a whole, but in their standing as autonomous and equal. This is why democratic self-rule means that the exercise of political power conforms to the collective will of those subjected to it, and why the scope-condition of democratic legitimacy is that all those subject to the exercise of political power have a right of democratic say. Note that for persons to be subject to political power it is not necessary that they be placed under any legal obligations to comply;Footnote ²⁵ in the sense relevant to autonomy and equal standing, it is sufficient that the state subject them to coercion—to direct physical force, invigilation via agents authorized to use physical force, and threats of punitive harm (Abizadeh Reference Abizadeh2008; Reference Abizadeh2010)—or to coercively undergirded symbolic processes of socialization and identity formation (Smith Reference Smith2008).

The all-subjected principle also explains why the demos is both constitutively political and in principle unbounded. The boundary problem arises, we have seen, because democratic legitimacy is self-referential and constitutively political in a volitional or decisional sense: To be democratically legitimate, political power must conform to the (politically) articulated will of the people. But the consequent turn to a prepolitical ground for the legitimacy of boundaries overlooks the fact that, precisely because it is self-referential, democratic legitimacy is also constitutively political in a second, subjectional sense: It requires that political power conform to the collective will of those who are subjected to it. This is a substantive principle, not the outcome of a political procedure, but it is the constitutive presupposition of the political procedures required for democratic legitimacy. Given the unboundedness of subjection, moreover, this self-referential and constitutively political character of democratic legitimacy implies the unbounded demos thesis: That the demos is in principle unbounded.Footnote ²⁶ Thus the intractability of the boundary problem stems not from democratic theory as such, but from falsely assuming that the demos is bounded in principle; the assumption makes democratically legitimizing power to those over whom it is exercised impossible conceptually speaking.Footnote ²⁷

The question is why contemporary theorists have found the bounded demos thesis so seductive. Much of the answer, I believe, lies in the lingering traces of a misleading image of the source and nature of political power. It is an image deeply embedded in the history of democratic ideology, which, however much officially repudiated, continues to cast its shadow on contemporary thought. John Locke is paradigmatic here. As James Tully (Reference Tully and Tully1983) has pointed out, in the traditional European social-contract image, “sovereign individuals naturally possess power, part of which (archetypally the power to kill) they delegate to government who exercise it through law in accordance with the public good, or else it devolves back to the people who then revolt” (14). What is mistaken about this traditional view is the fantasy of a prepolitically constituted “sovereign” power. The exercise of political power is only possible within a particular political constitution, with its own specific institutions and command structures. An articulated, collective democratic will is always itself the outcome of political procedures and processes, and the opinions and wills of participants, which serve as inputs for these procedures, are invariably themselves (at least in part) the effect of political power: The expressed will of the people is both an instantiation and an effect of political power. We have already seen an illustration of this in the constitution of civic boundaries. Specifying the members and rights-bearers of a polity is both necessary to defining the scope of political power and itself an instance of its exercise. The corollary to this mistaken image of political power as prepolitically constituted is a peculiar theory of political legitimacy. With this image as a background, normative democratic theory is naturally construed as claiming that political power is legitimized only when it emanates from a prepolitical collective will. There is no such thing: Any democratic articulation of a collective will presupposes political procedures, processes, and institutions, and these not only causally shape individual wills, but also in part constitutively determine (rather than merely track) the will of the people. There is no democratic subject with corporate agency given prior to politics.

The notion of a prepolitically constituted corporate agent, a collective democratic subject bounded prior to the exercise of political power, is therefore both erroneous and unnecessary. It is fuelled by a fundamentally mistaken account of democratic legitimacy, one that assumes that legitimacy requires a prepolitical ground. Recall that the primary motivation for seeking such a ground is the (entirely sensible) thought that might, by itself, cannot make right.Footnote ²⁸ Indeed, I take it that (nonnormative) facts do not legitimize anything, much less themselves, just in virtue of being—whether they are prepolitical or not. But it does not follow from this that facts, to serve as a ground of legitimacy, must be prepolitical. What follows instead is that any principle identifying some factual aspect as a ground of legitimacy presupposes a normative principle explaining why that ground has the capacity to legitimize political power.Footnote ²⁹ To say that might does not make right is to say that such an explanation cannot amount to saying that the factual aspect is an effect or instance of power's exercise; it must instead show how the ground can legitimize the exercise of political power despite perhaps being an effect or instantiation of political power, by invoking some value the factual ground serves.

Once we recognize that political power neither can nor need be construed as prepolitically constituted, or legitimized by a prepolitical ground, a different image of political power, and of democratic legitimacy, comes into view. Rather than beginning with the fiction of an originary social contract, and asking how legitimately to construct political power from prepolitical grounds, we must reverse the image: Begin with the fact of political power—we begin with the already existing configuration of political power—and ask how to legitimize it. Rather than an attempt to legitimize political power by appeal to a prepolitical ground—whether the individual will, the nation, or even, as some cosmopolitans have suggested (Bartelson Reference Bartelson2008, 170), humanity as a wholeFootnote ³⁰—I take it that democracy is better understood as an attempt to legitimize the collective and political exercise of power, on terms respecting the equality and freedom of those over whom power is exercised, via participatory political practices of expression, contestation, discursive justification, and decision-making. Democracy represents an attempt to replace relations of force, coercion, or domination with relations governed by fair negotiation, argumentation, and decision-making, and to legitimize the remaining uses of coercion by subjecting them to terms set via such democratic practices (Cohen Reference Cohen, Hamlin and Pettit1989; Habermas Reference Habermas1996). So the answer to the question of why the fact that something happens to be the expressed “will of the people” should provide a ground for legitimizing the exercise of political power over them is this: Because (and insofar as) grounding political rule in this way serves the freedom and equality of those over whom power is exercised; it makes collective self-rule, on terms respecting the freedom and equality of the objects of power, possible. And that justification itself imposes constitutive normative constraints on the kinds of procedures and processes required to articulate a democratic will qua legitimizing ground of power.Footnote ³¹

In contrast to the traditional social-contract account of legitimacy, this account does not presuppose a prepolitically constituted corporate subject. The principle of democratic legitimacy, and the “people” to whom it refers, is interpreted, rather, in intersubjective, procedural terms. Democratic legitimacy, as Jürgen Habermas (Reference Habermas1996, 301) has suggested, is located not in a prepolitically constituted collective entity, but in the regulative standards implicit to democratic procedures and processes that make collective self-rule possible.Footnote ³² To be sure, it is not possible wholly to collapse the meaning of “the people” into procedures, processes, and their regulative standards. As we have seen, democratic theory is not just procedural but also self-referential: It presupposes active political agents capable of articulating a collective will and acting collectively through institutions. The point is instead that the demos, in the regulative sense serving as the ground of legitimacy, remains distinct from any empirically existing, institutionally articulated collectivity seeking to identify itself with or represent it. No actual set of procedures and processes can fully live up to the constitutive ideals of democratic self-rule; among those that approximate such ideals, there are a myriad combinations to choose from, and the specific procedures and institutions in place will significantly impact political outcomes.Footnote ³³ The demos of normative democratic theory marks the inability of any of its putative representations to exhaust the normative aspirations of democratic legitimacy.