1. Most but Not Every Demos Can Be Democratically Decided

The root of the view that the democratic legitimacy of the people undermines the legitimacy of democracy is the observation that no demos can be created by democratic methods unless some demos already exists. This observation is often taken to imply that as a matter of logic the composition of the demos cannot be democratically decided. For example, Frederick Whelan states that “democracy … cannot be brought to bear on the logically prior matter of constitution of the group itself, the existence of which it presupposes.”Footnote ¹² David Miller argues that if a democratic procedure “is used to decide the question of who is to be included in the domain, an obvious circularity is involved, since to use the procedure we must already know who should be allowed to take part in operating it.”Footnote ¹³ However, these types of formulations are not accurate descriptions of the demos paradox.

The demos paradox rests on the observation that a democratically decided demos must have a prior demos that can democratically decide on that demos. All that follows from this observation in terms of pure logic is that D₂ cannot be democratically decided unless there is another demos, D₁, that can democratically decide the composition of D₂. This also means that D₂ can be democratically decided as long as there exists a demos, D₁, that can democratically decide on the composition of D₂. Thus, it is true that not all demos can be democratically decided but it does not follow that no demos could be democratically decided. The latter would follow only if one assumes that a demos that is not itself democratically decided cannot decide democratically on the composition of the demos going forward. This is at the very least a highly contentious position because it seems to rule out the possibility of democratic decision-making of any kind in general.

The more reasonable implication is that in a world with already existing demos, each demos can democratically decide on its own composition or even on the composition of another demos. That does not imply that it would be an attractive solution to the problem of the legitimacy of demos if, for example, Swedish voters decided on the composition of Denmark's demos. But this is not because the people of Sweden cannot make a democratic decision on the composition of the Danish demos. If it is illegitimate for Swedish voters to vote on Denmark's demos it must be because of some theory of who has legitimate standing to make democratic decisions about Denmark's demos.

What about a demos deciding on its own composition? It is true that the composition of a demos cannot be decided by that demos until the composition of that demos is decided. But as long as there is a demos, that demos can decide democratically on its own composition. One might object that this is mistaken because the postdecision demos is a different demos from the deciding demos. Whether this is true depends on whether one defines the demos by reference to its members or by reference to the democratic polity that it governs. The latter seems to be more in accordance with common language, but whether we define the postdecision demos as the same as or different from the predecision demos does not ultimately have much significance. What is important to note is that the composition of the demos can usually be democratically decided. Again, this conclusion does not imply that the demos can be legitimized by the democratic procedure. Whether or not the composition of a demos could be legitimate in virtue of being democratically decided is not determined by the possibility of making a democratic decision about the composition of the demos but by the theory of legitimacy that one applies.

The conclusion we can draw is that most but not all demos can be decided democratically. This means that if impossibilists want to argue that the demos paradox involves a deep problem for democracy they have to base that argument on some form of the claim that democratically deciding on the demos is not sufficient for democratic legitimacy. More specifically, the notion of legitimacy underlying the demos paradox has to have something to do with the fact that not all demos can be decided democratically. In the following section we will investigate one such argument commonly advanced by impossibilists, namely, that a democratic decision on the composition of the demos is not sufficient for legitimacy because the composition of the deciding demos is an “accident of history.”

2. Does the Fact of History Undermine the Legitimacy of the Demos?

At the heart of the critique of the possibility of a democratic demos is the claim that contemporary peoples, organized into democratic collectives, are mere “accidents of history” as they result from “historically arbitrary forces, such as violence, revolutions, or pure coincidence.”Footnote ¹⁴ Rhetorically, the question asked is how it is possible for peoples to be sovereign if they owe their existence to the morally arbitrary forces of history.Footnote ¹⁵ The conclusion is, as Jens Bartelson puts it, that the “contingencies of history” show democracy's claim to legitimacy to be “nothing more than naked power having been around long enough to become taken for granted by the members of a community.”Footnote ¹⁶

The view that history teaches us something important about the possibility of justifying the demos is here referred to as the “historical claim.” The historical claim may at first glance seem obvious and true. But on closer inspection, it turns out that it is ambiguous, with respect to both its descriptive content and its normative implications. Consider, first, the descriptive contents of the historical claim.

One version of the historical claim is concerned with nation-states and their territorial borders. Bernard Yack argues that the borders of the state cannot derive legitimacy from democratic decisions since state borders are constitutive of the state and therefore of democratic procedures themselves. The point is that if peoples are defined by states, and if states are created by the arbitrary forces of history, then these forces also define the composition of peoples.Footnote ¹⁷ Another version of the historical claim figures in Jürgen Habermas's criticism of nationalism. As he points out, the nationalistic ideal of self-determination does not provide an adequate explanation of actual “social boundaries” that make people identify as separate peoples.Footnote ¹⁸ The distinctions that separate one people from another do not derive from the primordial rights of peoples to exist as such but must be explained by reference to political and ultimately historical factors.

When Yack refers to the historical origins of state borders, the implication is that the borders between states are morally arbitrary and that the belief in the legitimacy of peoples as defined by their borders is vain.Footnote ¹⁹ The upshot of Habermas's claim is that peoples as identified by shared national identities are just the offspring of past coercion and force and have no claim to moral status. One thing to notice here is the distinction between these conceptualizations of the people and the concept of the demos (this distinction was highlighted in note 8 above). The demos is not equivalent either to peoples identified by membership of states or to peoples identified by membership of nations. The demos is a function of the legally defined criteria for the right to vote that in practice usually depends on legal citizenship status and specific requirements of age, residence, mental capacity, and criminal status.

Descriptive ambiguities of the type outlined above make the historical claim difficult to interpret, but we argue that the key problem for the historical claim is both more foundational and more basic. This is because the gist of the historical claim is not just to question current configurations of the demos in democratic states but also to challenge the very possibility of legitimizing any demos at any time. The starting point for the normative version of the historical claim is the assumption that the workings of power are never sufficient for democratic legitimacy. As pointed out by Arash Abizadeh, history does not have the power to create legitimate peoples because “might cannot by itself make right.”Footnote ²⁰ This general position lends itself to two distinct normative readings—a negative and a positive claim.

Consider first the negative normative historical claim. It is premised on the view that political institutions can never derive democratic legitimacy from a morally arbitrary process. Given that the workings of history often represent a morally arbitrary process, it follows that no institution can derive moral legitimacy from past events. This claim is negative in the sense that it denies a specific source of moral legitimacy but not the possibility of moral legitimacy as such.

The premises of the negative historical claim are of course true, possibly even trivial. For example, from the fact that state borders are the result of political struggles and wars in the past, no inference can be drawn that state borders are morally legitimate. The more radical implications of the negative historical claim follow only if we accept the distinct proposition that not just some historical process but every historical process is morally arbitrary. If this were so, no political institutions could ever derive legitimacy from the past. Yet it is worth pointing out that the generalized inference depends crucially on the unsupported claim that no historical processes have the capacity to confer political legitimacy. Though it is easy to agree that historical wars of conquest and decisions by despots should not be considered sources of moral legitimacy, it is quite unclear that no event or decision in the past has the capacity to confer legitimacy.

Consider next the positive normative historical claim. It is premised on the different assumption that political institutions can never be morally legitimate if they are the outcome of a morally arbitrary process. Where the negative claim proposes that arbitrary processes do not contribute to the democratic legitimacy of institutions, the positive claim much more controversially proposes that a sufficient reason for concluding that an institution is bereft of normative standing is that it is created by some morally arbitrary event. The implication is that the borders separating states from each other are necessarily morally illegitimate because they originate from the struggles and battles of the past. In a similar vein, Ochoa Espejo argues that the normative authority of the people as a collective body is lost, if it turns out that it is “solely” an “accident of history.”Footnote ²¹ But it is hard to see why this should be the case. If some procedure exists whereby it is possible for people to confer democratic legitimacy on some institution or people, it is unclear why the operations of such a procedure would be invalidated just because it seeks to legitimize institutions or peoples that are the result of historical accidents. Of course, there might be no procedure with the capacity to confer democratic legitimacy. But the fact that political institutions and peoples are created by historical accident does not as such provide a reason for denying that there is one.

Tentatively, we conclude that both the negative and the positive versions of the historical claim are inconclusive. The negative claim is inconclusive with respect to its alleged consequences. We can easily accept the claim that political history is replete with morally arbitrary and morally abhorrent events that do not have the power to confer moral legitimacy on political institutions. It does not follow from this observation, however, that the political history of every existing state is a morally empty place. To make this claim, a substantive argument is needed that is designed to show why there is no normative significance today of events in the past, such as revolutions, constitutional conventions, amendments, legal reforms, elections, and so forth. Such a claim can only be made on the basis of a specific normative view of what the conditions of democratic legitimacy are.

The positive version of the historical claim is inconclusive with respect to its basic message, namely, that the absence of a morally “clean” history invalidates the legitimacy of political institutions. The fact that the shape of democratic rule is in part an “accident” of history cannot, on its own, demonstrate a deep problem of legitimacy. Any possible institutional practice necessarily has a history, just as everything that exists is in some important sense accidental. The upshot is that in order to identify what the special problem of legitimacy is for the composition of the demos, we need (1) to construct a positive account of what the impossibilists think could make a demos democratically legitimate, and (2) to explain why they think this standard can never be achieved. In the following section we preform both of these tasks.

3. What Could Make the Demos Legitimate?

The standard of democratic legitimacy underlying the impossibilists’ critique of democracy is not clearly stated, as far as we know. However, it is easy to discern based on the various formulations of the impossibility of a legitimate demos. As Olson puts it, “the paradox of the founding prevents a purely democratic constitution from being founded, because the procedures needed to secure its legitimacy cannot be spontaneously self-generated.”Footnote ²² This description of the paradox proposes a procedural understanding of the source of democratic legitimacy, and the problem for constituting a democratically legitimate demos is that “the vote on the proper constitution of the people is caught in a vicious circle. The vote does not concede legitimacy to the people. It presupposes it.”Footnote ²³ The argument is that a democratically legitimate people is elusive because it is only democratic decision that can confer legitimacy, while legitimate democratic decision can only take place when there already is a democratically legitimate people.

Adopting a purely procedural understanding of the source of democratic legitimacy allows impossibilists to advance a very general repudiation of any argument advancing a normative principle or principles for when a demos is more or less democratically legitimate. For example, Näsström, rejecting the argument that the demos ought to be all-inclusive, says that “the principle is built upon a prior boundary that is not further discussed. The determination of political community paradoxically precedes the decision about who should be included in it.”Footnote ²⁴ The thing to notice here is that there is no need to engage with the principle being proposed, no need to evaluate its implications in relation to competing principles, because it is not a principle that derives the legitimacy of the demos strictly from a procedurally democratic decision.

Why then do impossibilists adopt a highly procedural notion of what could make the demos legitimate? The reason for this is that the composition of the demos is a political question over which people will disagree, and democracy is considered the legitimate decision method in cases of disagreement. Democracy, Whelan argues, is “advanced by its advocates as the sole legitimate method for making political decisions.”Footnote ²⁵ If we take seriously the political nature of the problem of the demos and the fact of disagreement among people on how to solve this problem, it is not enough that the composition of the demos is democratically decided. The composition of the demos must be democratically decided, and it must be democratically decided by a demos that is not itself disputed. This seems to be the position of those impossibilists who do not adopt the purely historical variant of the paradox rejected in section 2.Footnote ²⁶

For democracy to perform the function of mediating political disagreement, the terms of decision making and the terms of inclusion must already be established. Impossibilists are committed to the view that if we disagree on these terms, democracy is the method for settling the disagreement. As Näsström puts it,

It is this commitment that causes an “infinite regression of procedures presupposing procedures.”Footnote ²⁸ We disagree over the terms of inclusion and exclusion from the demos. These terms must therefore be democratically decided in order to be legitimate. But since we disagree over the terms of inclusion and exclusion from the demos that is to decide on the terms of inclusion and exclusion, this must also be democratically decided, and so it continues ad infinitum.

The implication that impossibilists want to draw from the reasoning above is that a demos cannot legitimately “close the issue” of who ought and who ought not to be a member of the demos.Footnote ²⁹ In the following sections, we will agree with this conclusion. However, what impossibilists appear not to have noticed is the positive account of legitimacy their reasoning commits them to.

When you defend the illegitimacy of closing the issue of membership by appealing to a highly procedural standard of democratic legitimacy, what you end up endorsing is the view that a demos is legitimate if a democratic decision on its composition is made by a deciding demos that is not contested. In other words, what the impossibilists surprisingly commit themselves to is the idea that a demos could in principle legitimately close the issue of membership by democratic decision. In the following section we will argue that not even in principle could a demos legitimately close the issue of membership. Understanding why helps us to see that any workable theory of democratic legitimacy does give us tools to make theoretical progress on the problem of the demos.

4. The Limits of Democratic Decision as a Source of Legitimacy for the Demos

The theory of legitimacy underlying the impossibilists’ critique of democracy is that the demos can only be democratically legitimate if a democratic decision on the composition of the demos can close the issue of who does and who does not have claims to membership. Yet, because no demos could enjoy a composition that is in practice uncontested, impossibilists argue that democratic votes cannot confer legitimacy on decisions about the demos's future composition either. The problem with this reasoning is not the claim that a demos can never legitimately close the issue of membership, but rather the highly procedural understanding of legitimacy that gives rise to the stronger claim that a legitimate demos is impossible. To defend this position, we show that a highly procedural account of democratic legitimacy in which an uncontested demos would have the prerogative to close the issue of membership cannot be justified. In defending this position, we show that (a) any theory that prescribes democratic rule necessarily entails significant procedure-independent criteria for the legitimate composition of the demos, and (b) this fact is a valuable feature of democratic theory, not a flaw.

We can begin our reasoning by bracketing, for the sake of argument, the fact that in practice we never have an uncontested demos. Once we do this we can examine more carefully the limits of democratic decision as the source of democratic legitimacy for the demos. Imagine that we have a demos where all of the members and nonmembers agree on the democratic legitimacy of its composition. In this uncontested demos all women and men over eighteen years of age who are permanent residents of a clearly defined territory have the right to vote and to run for political office. Now imagine that there is a shift in public opinion that leads this demos to vote to disenfranchise women. A large majority of women and men vote for disenfranchisement while a minority protest the decision as a violation of foundational democratic rights. Would this new demarcation of the demos be legitimate?

At first glance, it is not obvious that the change made to the demos is democratically illegitimate. But what if a year after this vote there was a strong countermovement with a large majority of women now calling for the reinstatement of their right to vote? Let us suppose that a referendum is called on the issue and the men of the society vote not to return the right to vote to women. If this decision undermines the democratic legitimacy of the demos and in turn the democratic legitimacy of this political community more generally, it is not because we cannot trace the outcome back through a series of democratic decisions to an original source of legitimacy for the demos. Instead, if the democratic legitimacy of this political community has been undermined it must be because there is something illegitimate about the substantive content of the decision to exclude adult women from the demos. And if the substantive content of the decision lacks democratic legitimacy, then the original decision to remove women's right to vote also looks like it was democratically illegitimate, despite the fact that it was made in a democratic way by a demos whose composition was uncontested and inclusive. This would in turn have the further general implication that any demos that claims for itself the authority to close the issue of membership by way of democratic decision alone necessarily undermines the democratic legitimacy of the political system; not because the composition of the demos is always in practice contested but because there are at least some procedure-independent criteria for the legitimate composition of the demos that are independent of what is democratically decided.

The question is then, was the vote by the male citizens to deny women the right to vote illegitimate? The main argument against seeing such a decision as illegitimate is that there does not appear to be some other legitimate way to decide on this type of membership question. Jeremy Waldron advances one of the more plausible versions of this argument. Although he is not an impossibilist, Waldron does adopt the same strong procedural standard of democratic legitimacy that impossibilists adopt.

Given the fact of disagreement over what rights and conditions are constitutive of or necessary for democratic legitimacy, Waldron argues that as a political community we do not know what the right answers are about the fundamentals of democracy.Footnote ³⁰ This means that the democratic legitimacy of democratic rule cannot be judged by appealing to substantive notions of the rights necessary for democratic legitimacy. Instead, Waldron argues, majority rule is the correct procedural tool for deciding because it uniquely shows respect for the inherent moral equality of individuals in circumstances where we do not agree on what is required to show such respect.Footnote ³¹ What one shows respect for is the capacity of agents who disagree to think about what rights they ought to have.Footnote ³²

But what about rights thought to be so central to democracy that the legitimacy of majority decisions evaporates if they are not in place? Waldron argues that if we insist on some right being present irrespective of what the majority chooses, then those who do not see the right as fundamental to democracy will rightly see its entrenchment as illegitimately excluding them from the democratic process.Footnote ³³ There is much merit in this reasoning, but the argument clearly breaks down once we look to the question of membership.

By describing the situation as “legitimacy free” Waldron acknowledges that it is illegitimate to exclude women from the demos. Yet at the same time he insists that we ought to be consistent in advancing a majoritarian procedural standard of democratic legitimacy and reject a “results-driven test of legitimacy.” What this means is that the vote to keep women excluded from the demos ought to be treated as normatively authoritative by all those subject to the power of the political community until some new decision is made by the legally recognized demos. Still, it is very difficult to understand what kind of democratic status is being ascribed to the decision male citizens made given that Waldron also says that “the best we can hope for is that a legitimate democratic system emerges somehow.”Footnote ³⁵

Waldron argues that even if the demos makes a majoritarian democratic decision that entails a “loss to democracy in the substance of the decision, it is not silly for citizens to comfort themselves with the thought that at least they made their own mistake about democracy rather than having someone else's mistake foisted upon them.”Footnote ³⁶ However, in the suffrage case it would of course be silly for women to comfort themselves with the thought that “they” made their own mistake because they did not get to participate in the decision. What this case helps us notice is that although Waldron formulates his theory in terms of “respecting disagreement,” the only way to move from the fact of disagreement to a normative prescription for democratic decision-making is through the premise that we owe respect to those persons who disagree.

It does make sense to worry that importing a substantive view of democratic legitimacy that is prior to majority rule excludes from the start those people who hold dissenting beliefs.Footnote ³⁷ But contra Waldron, what we can now see is that the fact of disagreement only speaks for certain types of procedural solutions once we already know whose disagreement has to be respected in order to secure the democratic legitimacy of political rule. The legitimacy of majority rule must be grounded in some substantive theory of inclusion, and when Waldron gives disagreement a privileged position in his theory he does so from the “conviction that ordinary men and women have what it takes to participate responsibly in the government of their society.”Footnote ³⁸

Hans Agné provides a defense of the possibility of deciding the demos democratically that is consistent with a strong procedural standard of legitimacy and could be viewed as a remedy to Waldron's approach.Footnote ³⁹ His claim is that a democratic decision taken by a demos that includes everyone could legitimately close the issue of who does and who does not have claims to membership. By the requirement that everyone is included, Agné claims to avoid the demos paradox because on his view it is only exclusions from the demos that must be justified. However, Näsström observes that the position that humanity as a whole is the legitimate “metademos” is not a neutral position.Footnote ⁴⁰ People can reasonably disagree on the inclusion of everyone, just as they can disagree on the inclusion of just someone. From our perspective it is not the fact of disagreement that makes Agné’s theory problematic but rather that the theory is only superficially a purely procedural solution. The claim that everyone must be included in order for a demos's decisions to be normatively legitimate depends on the plausibility of substantive moral standards. It is an argument that draws surreptitiously on commitments to a specific theory about the substantive ends that justify democracy.

What then does the analysis above tell us about the original case of a majority of men and women voting to revoke women's voting rights? The argument so far shows that the democratic legitimacy of a demos's decisions is dependent on the democratic system respecting the substantive moral principle that gives us reason to advocate for democratic governance in the first place. This moral principle might for example be some version of moral equality according to which the idea that women are not competent to participate in democratic decision-making is ruled out. The inclusion of women in the demos is consequently a necessary condition for the democratic legitimacy of the demos's decisions, following this substantive principle. A decision made by an inclusive demos to subsequently exclude women from the demos would in fact undermine the democratic legitimacy of all subsequent decisions made by that demos, even if the deciding demos was itself uncontested. The minority that voted against the disenfranchisement of women has legitimate grounds to reject the democratic legitimacy of the majority vote to disenfranchise women on the substantive grounds that this decision violates a principle of moral equality that is a necessary precondition for us to accept the very idea that political society ought to be governed democratically.

A critic might wonder why it is that the legitimacy of an uncontested demos's decision on the demos going forward is dependent on the decision being in accordance with substantive normative principles. Why is it not the fact that the decision was made by what all recognize to be an uncontested demos that confers legitimacy on the new composition of the demos? The underlying assumption in this question is that the reason the deciding demos is uncontested is that it already lives up to substantive normative standards for a legitimate demos. The question then is, why doesn't the fact that the composition of the deciding demos satisfies substantive normative standards for its legitimacy also confer legitimacy on the decision this demos actually makes irrespective of the content of the decision? The simple response is that there is no reason why it should follow that if the demos satisfies necessary inclusion criterion X that it should then be at liberty to violate necessary inclusion criterion X. Criterion X's status as necessary is not altered somehow to nonnecessary by the fact of its satisfaction. However, the objection above should not be dismissed so easily because it is pointing to a deeper worry, namely, that our argument appears to leave no room whatsoever to democratic decision as the source of legitimacy for the composition of the demos.

It does not follow from our reasoning that democratic decision makes no difference to judgments about the legitimate composition of the demos. What we have argued for is that the normative reasons for recommending democracy will make some exclusions from and inclusions into the demos impermissible. This leaves open the possibility that there may be a range of permissible inclusions, exclusions, expansions, or divisions that ought to be determined by democratic decision. It is beyond the scope of this article to attempt to work out the relationship between substantive and procedural sources of legitimacy for the demos, and we recognize that the most important theoretical obstacles to addressing this problem are not addressed here. What we do hope to show is that whatever argument you advance for why there should be democracy will involve some limits on the types of exclusions and inclusions to the demos that can be justified. And what we argue follows from this reasoning is that the substantive normative reasons we have for recommending democracy do allow for theoretical progress on questions of the legitimate composition of the demos. As a result, we claim that democratic theory does not suffer from a demos paradox where no theoretical progress on the legitimate composition of the demos is possible.

Another objection one might advance is that although some substantive moral standard must underpin democratic theory, that conception need not necessarily see the equal treatment of men and women as entailing equal political standing. There is nothing necessarily inconsistent with the claim that men and women are owed equal concern but that men lack the competence to enjoy equal voting rights with women. However, the technical point to notice is that this objection appeals to some other substantive theory that gives an alternative account of the agents’ moral standings. It is this alternative account and not the simple fact of disagreement that is doing the work to justify a less inclusive demos.

There are many different types of arguments in democratic theory that attempt to justify democracy substantively. Some of these arguments are instrumental, others noninstrumental, and some of them appeal to both instrumental and noninstrumental considerations. Yet all of these approaches have in common that they are grounded in some substantive normative conception of the moral status of individuals, such as moral equality, the importance of supporting individual liberty and freedom, and/or ensuring conditions of nondomination. Although our reasoning in this section appeals specifically to the relationship between respect for moral equality and membership in the demos, our overall argument is that all the substantive normative grounds for seeing democracy as the best form of government will have implications for the legitimacy of exclusions from and inclusions into the demos.

Indeed, the democratic tradition, from Rousseau and Kant to Rawls, Habermas, and Pettit, is abundant with attempts to delineate the demos by appeal to substantive moral standards. Robert Dahl famously identified exclusions that contravene the “strong principle of equality” as inconsistent with the democratic ideal. To exclude from the demos any competent adult person that is subject to the binding decisions of the polity would, in Dahl's view, run contrary either to the value of intrinsic moral worth or the presumption that adults normally are best judges of their own interests.Footnote ⁴¹ More recently, Claudio Lopez-Guerra argued that inclusion is required only for those subject to binding decisions that possess the capacity of being harmed by exclusion. What is essential for the democratic legitimacy of the demos, in Lopez-Guerra's view, is that people are not harmed by exclusion. Consequently, people who lack the capacity to be harmed by disenfranchisement, infants and the severely mentally disabled, presumably, can permissibly be denied access to the vote.Footnote ⁴²

A final question to consider is what follows if every single female and male citizen voted for disenfranchisement. It does seem true that this should be accepted, but the legitimacy of exclusion in this case finds its source exclusively in the consent of each and every individual. People should be at liberty to join clubs where women govern and men follow the rules. However, the legitimacy of such a club is fundamentally dependent on all members having a robust right to exit the club. This is the only way to demonstrate that it is the consensual nature of the relationship that gives both members and nonmembers grounds to accept the authority women are exercising over men.Footnote ⁴³ The political authority exercised by our system of states or any feasible system of political authority is clearly different from the rules of a club because it is not a relationship one consensually enters into (not even by emigration) and there is no robust right of exit from political society.Footnote ⁴⁴ This means that barring universal consent one cannot justify the power relations created by a disenfranchisement decision by appealing to the consensual nature of the political relationship the state represents. As a consequence, as soon as new women who have not consented to exclusion from the demos enter the society or women simply stop consenting to exclusion, one needs to point to something other than ongoing personal consent to justify exclusion from the demos. Barring some substantive and general normative argument for exclusion, the disenfranchisement policy can no longer be justified.

5. The Unavoidability of Demarcating the Demos

The general implication of our reasoning is that as soon as one claims that political rule ought to be exercised through democratic procedures one is necessarily committed to some procedure-independent criteria for when and how such procedures are and are not legitimate. Any “ought to be democratic” claim must entail some normative premises about how to respect the moral worth of individual agents that are prior to rather than the outcome of democratic procedures.Footnote ⁴⁵ This entails that any normative theory that prescribes democracy cannot avoid commitment to normative foundations that make some compositions of the demos permissible, others impermissible, and still others necessary. And although the reasoning above has focused on features of individuals such as sex or skin color, it would be very surprising if we could only make claims about the legitimacy of inclusions and exclusions in those areas where we now find broad agreement.

A strong defense of judicial review or constitutionalism does not necessarily follow from our reasoning. What does follow is that the legitimacy of democratic rule is dependent on the demos finding some way of not closing the issue of the legitimacy of its composition that goes beyond waiting for the next vote. The legitimacy of the demos is dependent on opportunities to challenge its composition based on foundational normative commitments to the purposes of democracy. Of course, we do disagree on what the foundational moral commitments underpinning democracy are, but at the same time logic demands that we recognize that justifying having democratic rule at all requires allegiance to the idea that there are such foundational constraints. Understanding this relationship between foundational normative arguments for democratic rule and the democratic legitimacy of democratic decision-making should make some difference to the institutional arrangements of democratic government and some difference to the content of democratic discourse.

What then is our response to the impossibilists? We have shown that it is wrong to think that if the composition of some demos is uncontested from the outset then that demos would have the normative authority to close the issue of its own composition simply by democratically deciding. Any claim to the normative authority to close the issue of composition would instead undermine to a significant extent the legitimacy of this demos's rule. This means that the impossibilists adopt a faulty standard of legitimacy for the demos and that the real impossibility is to recommend democratic procedures without at the same time being committed to some procedure-independent standards for the legitimate composition of the demos.

At times it appears that the impossibilists arrive at the same conclusion we have arrived at. Bonnie Honig sees the problem of the people as not only about the origins or composition of a legitimate demos but as part of a continually reoccurring “paradox of politics” where each new political moment requires both good law to create a good people for this moment of political decision and a good people to create good law at the same moment of decision. It may seem, she says, that “the paradox of politics calls us back repeatedly to the political moment of origins wherein it pulls the rug out from under our feet, and it may seem … that in such moments, upheaval rather than settlement necessarily dominates, but this need not be the case. The paradox of politics can be a generative force.”Footnote ⁴⁶ Articulating this point more directly, Näsström argues that a “fully legitimate people is indeed impossible to achieve, but therein resides its power. The criteria of legitimacy make the people into a site of perpetual contestation. They guarantee that the claims of a legitimate people do not come to a standstill.”Footnote ⁴⁷ Her argument is that there is “productive force” in the “gap” between historic contingency and democratic legitimacy.Footnote ⁴⁸

As in the case of Waldron's “legitimacy-free zones,” we should not let ourselves be misled by terminology such as “polemical,” “generative,” or “productive” contestation into thinking that these positions are somehow not normative. The best interpretation of what impossibilists are arguing for is that it is a “good thing” that a demos cannot make legitimate claims to having the authority to close the issue of its own membership through democratic procedure. We agree, but this simply shows the oddness of the impossibilists’ positions. Ultimately, their claim is that it is a good thing that “a fully legitimate people is indeed impossible to achieve.” The conclusion we should draw instead is that the impossibilists’ standard for what counts as a fully legitimate people is faulty because it would not be a good thing if it were ever realized.

This leaves us with the question whether there is any serious debate between what we have argued for and what impossibilists endorse. Donahue and Ochoa Espejo say that “Näsström treats the boundary problem as insoluble, but resoluble… . But because the resolution of the problem allows for some intellectual progress, and does not claim that any attempt to solve the problem is an illegitimate power grab, this style does not cause a rift with the solving style.”Footnote ⁴⁹ This tortuous conceptual space between the insoluble but resoluble simply does not answer the key question. Do those who “press” the “insolubility” of the demos problem accept that prescribing democracy as the form of political authority we ought to have also entails procedure-independent standards for more or less legitimate ways of composing the demos or not? If they do accept this idea, then it is clearly dangerous to allow impossibilists to imply opaque standards for when contestation of the demos's composition is “productive” and when it is not. This does not present us with clear arguments for what principles we should appeal to in making normative judgments about inclusion and exclusion from the demos. This in turn undermines attempts to critically evaluate these principles and makes illegitimate power grabs a real possibility.

We can contrast the impossibilist approach with Ochoa Espejo's own reaction to the demos paradox, which does present us with clear principles for assessing the legitimacy of the demos. Ochoa Espejo advances a “process theory” for justifying the boundaries of the democratic people where “the continuity of the people comes from the actual repetition of a given kind of institutional practice, considered by its practitioners to be the highest authority.”Footnote ⁵⁰ She argues that “the boundaries of the people can be legitimized from the standpoint of the future” because even if the boundaries are partly based on undemocratic criteria (e.g., ethnicity), these boundaries can be “affected by subsequent events, such that the process effectively alters what has already occurred. . . . Democratic peoples need not perpetuate undemocratic practices.”Footnote ⁵¹ Ochoa Espejo's standard against which a process of reinterpreting the composition of the demos can be assessed is the principle of “intensely affected interests.”Footnote ⁵²

Ochoa Espejo's approach is quite similar to Waldron's because the claim is that to the extent that democracy has mechanisms for challenging illegitimate exclusions, these mechanisms confer some legitimacy to the actual practices of demarcating the demos even if these demarcations are found wanting in their current expressions. And like Waldron's theory, Ochoa Espejo's theory can be criticized on the grounds that when exclusions from the demos seriously weaken the democratic mechanisms for challenging illegitimate exclusions, then the mere presence of these mechanisms makes little difference to the level of legitimacy we should attribute to a demos. This leaves the principle of “all intensely affected interests” as the most important feature of Ochoa Espejo's theory. What is crucial to notice is that the demos paradox makes no real difference and that what we end up with is an approach to the demos problem where the legitimacy of the demos is to a large extent dependent on its accordance with substantive normative standards.

A final reply an impossibilist might make is that one need not be committed to any normative reasons for advocating democratic rule to point out paradoxes that are internal to the logic of democratic theory and practice. This is true, but this move makes the problem of the demos's legitimacy trivial. No system of political rule can point to features of its own system as the original source of its right to rule. If one wants to take this nonnormative turn then one needs to explain why the problem of the “democratic people” has received so much attention while other forms of rule have been ignored.