Who Governs? Who are the Governed?

People will debate what rights and opportunities individuals need to have for them to count as “rule-makers” or “participants in governing,” but most will agree that voting will be among them. For the purposes of this article, therefore, we stipulate that a necessary and sufficient condition of being a participant in governing is having the capability and opportunity to vote. By “capability to vote,” we mean possessing the cognitive ability to understand a choice and to intentionally select an alternative, and the practical ability to register this choice in a voting procedure. We intend for this to be a low bar. All but the most severely physically incapacitated and mentally impaired humans will count as capable of participating on this definition.Footnote ² To count as having an “opportunity to vote,” it is not sufficient to be merely granted a legal right to vote. Individuals must have a real opportunity to vote in the sense that occasions for voting are open to them, and it must not be impractical for them to take up the opportunity and perform the act of voting. The act of voting need not be costless, and it need not be equally accessible to all, but it must be actionable. Those who have the capability and opportunity to vote need not take up the opportunity in order to count among “those who govern” on this interpretation. Hence, an identity between the set of individuals who govern and the set of individuals who are governed is achieved if and only if all who are governed have the capability and the opportunity to cast a vote. We say that an identity between the governed and those who participate in governing is achievable “in principle” if (i) all who are governed are capable of voting and (ii) a vote could be held that would provide all who are governed with an actionable opportunity to vote.Footnote ³

If we can show that there will always be a non-identity between the governed and those who govern even on a non-demanding definition of participation in political decision-making as above (i.e., voting), then we can infer that the same result will hold for more demanding definitions of what is required for “properly democratic” participation in governing (i.e., when further conditions are added to the voting condition). Our claim is that this identity cannot be achieved in modern political contexts (the “non-identity thesis”). On plausible views of who counts as “the governed,” at least some will be unable to participate in governing (in the non-demanding sense of having the capacity and opportunity to vote). Those who participate in governing will therefore make decisions not only for themselves but also for others who cannot participate.

To explore the extension of “the governed,” we make use of distinctions that have been established within the “boundary problem” literature in contemporary democratic theory. The boundary problem is the problem of identifying all who fall within the “boundary” of those entitled to participate in political processes.Footnote ⁴ Within the active debate, we find a split between scholars who take as “the governed” those affected by political decisions and scholars who take as “the governed” those subjected by political decisions. We find a further split among the latter between those who understand “subjection” in terms of being subject to legal duties and those who understand it in terms of being subject to state regulation (including acts or threats of coercion). Taken together, these divisions yield three broad and competing interpretations of “the governed”: (1) those subject to legal duties, (2) those subject to legal regulation/coercion, and (3) those affected by political decisions. Each interpretation picks out a distinct set of individuals who might count as “the governed.” We consider each in turn, not to decide which interpretation is appropriate for solving the boundary problem in democratic theory, but to consider whether all who are “the governed” are in principle capable of participating in governing.

Subjection to Legal Duties

A promising interpretation of who counts as “the governed” from the perspective of achieving an identity between them and “those who govern” understands “the governed” in terms of being subject to legal duties. An individual counts as governed if they are an “addressee of the law.” These addressees are those to whom rules are given and who are legally responsible to comply with those rules. We call legally specified rules for conduct “legal duties.” All that is required for an individual to count as being subject to a legal duty on this conception is that there is a de facto authority—an authority that is generally recognized as an authority by those over whom it claims authority—and that the individual is among the addressees of the rules imposed by that authority.Footnote ⁵ Those subject to legal duties in this sense we call “D-subjected” (“D” for duties).Footnote ⁶

Robert Dahl,Footnote ⁷ Claudio López-Guerra,Footnote ⁸ and Ludvig BeckmanFootnote ⁹ argue that all those who are D-subjected are, by virtue of being D-subjected, entitled to participate in political decision-making. This is a version of the so-called All-Subjected Principle (ASP). We do not support or reject this proposed solution to the boundary problem. The question that concerns us is not whether the D-subjected have a moral entitlement to govern, but whether all who are D-subjected could in principle be participants in governing.

It appears plausible to assume that all D-subjected will be capable in principle of participating in political decision-making because, as Joanne Lau argues, the agential powers required to be an addressee of the law—roughly, possessing the mental capacities required to be treated by legal institutions as responsible for compliance with laws—seem to be more or less the same set of agential powers required to participate in political decision-making.Footnote ¹⁰ One might quibble about edge cases.Footnote ¹¹ But, for the sake of the argument, let us grant that those who are legally responsible to comply with laws will be competent to participate as what Jürgen Habermas calls “authors of the law to which they are subject as addressees.”Footnote ¹²

Of course, modern democratic states do not in fact permit all those on whom legal duties are imposed to vote. Young people below the voting age are expected to comply with the law even though they are explicitly excluded from the franchise. Visitors to a country are expected to comply with legal requirements without being entitled to vote. Corporations are expected to comply with laws but do not enjoy the right to vote. Still, in principle, the voting age could be lowered or even abolished,Footnote ¹³ visitors to a country and corporations could be permitted to vote.Footnote ¹⁴ Hence, an identity between the set of individuals who “govern” and the set of individuals who are “governed” could be achieved if we interpret “the governed” to mean the D-subjected, i.e., those subject to legal duties. A political regime that enfranchised all who are D-subjected would qualify as a system of collective self-rule.

Subjection to Legal Regulation

We have granted that an identity between “the governed” and those who “participate in governing” can in principle be achieved if we understand “the governed” to be the D-subjected. But we do not think it is plausible to demarcate “the governed” in terms of D-subjection. While, as Habermas argues, political power in modern states is largely organized, regulated, and applied through the “medium of law,”Footnote ¹⁵ it does not follow that only those who are addressees of the law in a given legal-political context are “governed” by the law.

This is clear once we appreciate the distinction between legal subjects and the objects of legal regulation. As we have just seen, legal subjects are the D-subjected. The objects of legal regulation, by contrast, are the entities, relations, and states of affairs about which a legal norm speaks—they figure in the content of the rule as things to be protected, used, or regulated in various ways. That which is subject to legal regulation in this sense we call “R-subjected” (“R” for regulation).

Many things, animate and inanimate, are objects of legal regulation without being norm-addressees. For instance, the US Endangered Species Act makes it unlawful to “take” a listed sea turtle. The turtle populations are the object whose welfare the statute protects, while the law imposes duties on legal subjects (humans, corporations, etc.) not to “take” them. Likewise, New Zealand’s Heritage New Zealand Pouhere Taonga Act 2014 forbids the demolition of an historic building. Historic buildings are the protected object, whereas owners and consent authorities are the subjects addressed by the norm. Neither the loggerhead sea turtles protected by the Endangered Species Act, nor the Category 1 buildings shielded by New Zealand heritage legislation—not to mention the consignment of wheat subject to quarantine rules, or the satellites governed by orbital-debris regulations—are capable of being addressees of the law.

Adult humans are typically both subject to legal regulation and addressees of the law. However, there are exceptions. Sometimes humans are not addressees of the law but are merely subject to regulation. For example, infants and persons in comas are protected by various legal provisions but are not themselves addressees of the law since they are incapable of bearing legal duties. In societies where chattel slavery was recognized under law, slaves were treated merely as objects of legal regulation, despite the fact that they were capable of being addressees of the law.

To be an object of legal regulation entails that one is subject to force, violence, manipulation, or coercion by the state (and its agents). For example, while parents and guardians may have legal duties of care for infants, the infants themselves are liable to be removed from families and placed in state care if they are (deemed to be) neglected. Likewise, while dog owners have legal duties to prevent their dogs from causing injury, the dogs themselves are liable to be impounded or killed by agents of the state if they cause injury. In both cases, the infants and the non-human animals are subjected to force or violence by the state, quite apart from their ability to bear legal duties. These illustrative cases are indicative of the large number of beings who are vulnerable to state force or violence, including all human and non-human animals within reach of the state’s agents, regardless of their capacity to be an addressee of the law or the bearer of legal duties. Thus, R-subjection is a ubiquitous and morally significant way of “being governed,” and it is irreducible to D-subjection.

Arash Abizadeh articulates a position close to our notion of R-subjection.Footnote ¹⁶ He argues that all persons who are “subject to state coercion” are entitled to participate in political decision-making, where “subject to state coercion” means either being subject to coercive threats or being subject to coercive acts, i.e., force, manipulation, or violence.Footnote ¹⁷ Abizadeh’s position is considered to be a variant of the so-called All-Subjected Principle (ASP).Footnote ¹⁸ Our question is: can an identity between “the governed” and “those who govern” be achieved if we interpret being governed in terms of R-subjection? Evidently, some who are R-subjected will be persons capable of political participation. But many of those R-subjected—such as infants, the severely physically or mentally incapacitated, and non-human animals—will not be capable of participating in making the rules whose enforcement they are subject to. Hence, not all R-subjected will be able to participate in political decision-making. On this interpretation, it is not possible for the governed and those who govern to be the same set of individuals.

All Affected

We have thus far considered two interpretations of what it would mean to demarcate “the governed” in terms of subjection to political decisions. We deemed D-subjection to be insufficient as a demarcation of “the governed” since it fails to capture a range of cases that are uncontroversially cases of being governed in the sense of being legally regulated. The second, R-subjection, carves out a set of rule-takers that cannot be identical with the set of rule-makers since it includes many who lack the capacities to participate in political decision-making. In the literature on the boundary problem we find scholars who argue that all affected (or whose interests are affected) by a political decision are entitled to participate in political decision-making.Footnote ¹⁹ This is the so-called All-Affected Principle (also called the All-Affected-Interests Principle) (AAP). This principle supplies us with a third distinctive interpretation of “the governed” as those affected by political decisions.

Robert Goodin offers a very inclusive interpretation of its scope. Depending on the nature of the political decision in question, “virtually (maybe literally) everyone in the world” will have affected interests and be entitled to participate in decision-making;Footnote ²⁰ and entitlement to participate may extend even to (i) “everyone in all possible future worlds,” (ii) “depending on one’s views about the interests of the dead, past worlds as well,” and (iii) “depending on one’s views about other sentient beings or even ecosystems, … nature as well.”Footnote ²¹ What’s more, individuals have an entitlement to a say in political decision-making not only when their interests are actually affected by the decision but also when their interests are potentially affected by the decision.Footnote ²² Even if a political issue is ultimately resolved in such a way that makes no difference to one’s situation, the fact that the issue might have been resolved in a way that would have made a difference is sufficient to say that one’s interests are affected by the decision.

The scope of inclusion under the AAP could be interpreted more narrowly.Footnote ²³ Nonetheless, even on narrower interpretations, the affectedness test generally yields a far larger set of individuals who would count among “the governed” than the D-subjected test or even the R-subjected test. Quite apart from D-subjection or R-subjection, the decisions of political authorities significantly affect the economic interests and life chances of individuals, even to the point of death. For example, residents of low-lying Pacific atolls—who are neither addressees nor objects of US, Chinese, or EU law—face lethal risks from sea-level rise and storm surges driven by those polities’ emissions and energy policies. These residents are neither D-subjected by the laws of those polities, nor are they R-subjected by them, yet their very prospects for survival are directly shaped by those policies.

The AAP justifies the inclusion of parties in political decision-making who are not capable of voting, as Goodin acknowledges in his discussion of the enfranchisement of nature. For Goodin, the literal enfranchisement of nature is what the AAP demands. Yet this is obviously impossible. The same is true for many other classes of individuals affected by political decisions, including future people, infants, and the severely disabled: “the barriers to directly enfranchising [e.g., nature, infants, and future people] are ones of practicality, not desirability. It is not as if those interests are less deserving of consideration. It is merely that their carriers cannot, for one reason or another, speak for themselves in pressing those interests politically.”Footnote ²⁴ On this interpretation, all affected by a political decision are entitled to inclusion in political decision-making. But not all those affected by a political decision will be able to vote. Hence, we will have a non-identity between the set of individuals who are governed and the set of individuals who govern.

Democracy and Rule over Others

We have laid out three interpretations of who counts as “the governed.” On the first, which understands being governed in terms of being an “addressee of the law” (D-subjection), all those governed could in principle also participate in governing. However, we argued that D-subjection is untenable as a demarcation of “the governed” since it implausibly excludes many who uncontroversially stand in a relationship of being governed despite not being “addressees of the law” (e.g., historic buildings, turtles, infants, patients in comas, enslaved persons in ancient societies). The two remaining answers to the question both include individuals who are not able to participate in governing. On the R-subjected interpretation, “the governed” includes infants, the severely disabled, and non-human animals. The all-affected interpretation casts the net wider still, counting among “the governed” not only direct objects of state force and coercion but also parties indirectly affected by policy decisions, actions, and omissions. The latter encompasses a set of humans and non-human animals at least as broad as the second interpretation, and adds future generations. A vast number of these individuals cannot practically participate in governing (i.e., vote).

On both the second and the third interpretation, not all who are governed will be able to participate in governing. Hence, we conclude that, leaving aside the artificially restrictive D-subjection interpretation, no matter which way we interpret what it is for an individual to be “governed” or what it is for a decision to “apply,” there will always be a non-identity between the set of decision-makers (those who participate in governing) and the set of those to whom those decisions apply (the governed) in any functional democratic political system.

Under the conditions that obtain in any ordinary political context, the special conditions required for strict collective self-rule do not obtain. This is true of all the political associations we would today call democracies. None will satisfy the strict definition of collective self-rule. Nor could these democracies satisfy this definition even after every conceivable effort to expand the franchise were made. Some “rule-takers”—whether R-subjected or affected—will not be capable of participating in “rule-making.” What this means is that there is no democratic state without rule over others. While we imagine democracies to be arrangements for collective self-rule, they always involve rule over others.

To say that democracy in practice is rule over others is not, of course, to say that it is only rule over others. Citizens in democracies often make decisions that affect their own interests and to which they themselves will be R-subjected and D-subjected. Whichever interpretation of “the governed” one prefers, with regard only to these citizens and these decisions, we could describe their decisions as acts of self-government: they are both the rulers and ruled, rule-makers and rule-takers. But this will be true of the members of the decision-making group in any regime type, no matter how restricted the group’s membership. Oligarchs are self-governing in an oligarchy; aristocrats are self-governing in an aristocracy; the dominant racial or religious group is self-governing in an apartheid state; the citizens of a colonial power, such as British rule over India, are self-governing within their own state; etc.

Nonetheless, whatever the regime type, all such acts of “self-government” are also instances of rule over others. When making decisions regarding domestic law and policy, enfranchised citizens are performing actions that will affect the lives of others, including children, non-human animals, and future generations. When making decisions regarding domestic law and policy, enfranchised citizens are not merely setting the terms of social life for those who participate in the decision; they are changing the rules that regulate the society, where society is an ever-changing set of formal institutions and social practices existing in a complex relationship to a dynamic population of humans and non-humans who enact them and who are regulated by them over time. This is a more honest description of what we, as citizens, are doing when we exercise rights of political participation. We must conclude, therefore, that “government of the people, by the people, for the people” is an incomplete image of democratic systems of government. Democracy cannot be collective self-rule (i.e., in which the rule-makers and the rule-takers are one and the same) because it will always involve rule over others.

Implications of the Non-Identity Thesis for Democratic Theory

One might object that the picture of democracy qua collective self-rule we have presented is too rigid. Collective self-rule, as we have defined it, requires that all those who are ruled are participants in ruling. It is not sufficient that many or most of those who are ruled are participants in ruling. The criterion of full inclusion must be satisfied. If not, collective self-rule fails to obtain. Could we not simply accept that collective self-rule is an ideal that can be “approached” but never fully realized? If so, the non-identity thesis is unsurprising, since collective self-rule is taken to be a guiding heuristic, pointing us toward something we aim to approximate but will never reach.

Indeed, in the literature, it is common for collective self-rule to be treated as a regulative ideal. Beckman states that “democracy as self-rule is an idealized state of affairs that may not be fully realized anywhere.”Footnote ²⁵ Likewise, Valentini portrays collective self-rule as a state of affairs that we have reason to hope can “be realized, or at least approximated.”Footnote ²⁶ The example that she gives to illustrate what she means by “approximated” is the ideal of a crime-free society, which is “strictly speaking not realizable (given the law of large numbers), but it is nonetheless worth pursuing, since it is possible to get closer and closer to it.”Footnote ²⁷

Any talk of “approaching” or “approximating” full inclusion is ambiguous. One could think of “approximating” the relevant concept by degrees quantitatively, meaning that more individuals are included. (This seems to be what Beckman and Valentini have in mind.) But one could also think of “approximating” the concept in a more qualitative way, for example by approximating political participation through symbolic means or through the agency of representatives. (We consider Goodin’s idea of approximating enfranchisement through representation in the next section.) In either case, anyone who posits one or the other notion of “approximation” has the burden of producing an argument to show how legitimacy tracks with less than full realizations of collective self-rule. On what basis would one posit continuous degrees of legitimacy on the basis of collective self-rule? It isn’t clear why the degree of political legitimacy would increase proportionally as a perfect match between rule-makers and rule-takers is approached.

We address the objection in its quantitative interpretation: i.e., that the full count of those who should be included may not in fact be able to participate, but we should aim for participation by as many as possible. However, once it is conceded that collective self-rule is an unattainable ideal, then some who are governed will not be participants in governing. Identity between the rule-makers and rule-takers cannot obtain. We are left with a situation in which there is rule by the ruling group over themselves and rule by that group over others.

This is a problem for any view that requires the criterion of collective self-rule to be met for political legitimacy to obtain. The traditional appeal of democracy qua collective self-rule is the allure of a form of government which does not admit of heteronomy. Jean-Jacques Rousseau famously argues that the demand for individual freedom and the need for political authority are made compatible when those subject to laws are the authors of those laws, since “we can … no longer ask if the law can be unjust, because no one is unjust to himself; and no longer ask how we can be both free and subject to the laws, for the laws are but registers of what we ourselves desire.”Footnote ²⁸ As Jeffrey Green shows, this Rousseauian thought has exerted a powerful influence over political theory ever since.Footnote ²⁹

But if “democratic” rule is not merely rule by the authors of the laws over themselves but also rule by the authors over others as well, then Rousseau’s argument cannot be used to ground political legitimacy. When not all those subject to the rule are included in rule-making, it is no longer merely a case of imposing rules on ourselves but also of imposing rules on others. There is a gap between the rule-making by the “many” and the imposition of the rule on the “all” subject to it. Because this will be the case in any existing democracy, we cannot say that democracy in practice is “intrinsically just” on the basis that it is an exercise of self-government.Footnote ³⁰ For this reason, the anchoring of political decision-making in participatory processes such as elections and referenda cannot, by virtue of the fact of their being participatory, suffice to legitimate the institutions or laws endorsed—at least not by way of approaches to political legitimacy that require identity between the set of the governed and the set of those who govern.Footnote ³¹

Our conclusion that democracy inevitably involves rule over others also throws into question the robustness of justifications for democracy that point to its value for participants. Consider the choiceworthiness of majority rule as a decision procedure. Majority rule is variously defended on the grounds that it (i) respects the autonomy of the participants, (ii) treats all participants fairly and impartially, (iii) treats all participants as political equals, and (iv) maximizes the preference satisfaction among the participants.Footnote ³² But respecting the autonomy of the participants, satisfying standards of fair and impartial decision-making among the participants, treating participants as political equals, and maximizing preference satisfaction among participants don’t provide any reason per se to think that requirements of justice toward non-participating others have been satisfied. The problem here is not merely that satisfying standards of procedural justice is insufficient to ensure substantive justice in decision-making.Footnote ³³ The problem occurs rather at the level of the procedure itself. Whatever “non-instrumental” or “intrinsic” value inheres in democratic procedures is value for participants. But this value is not directly relevant to the question of the justifiability of procedures (or their outcomes) from the point of view of non-participants.

To be clear, it may still be true that certain political rights and opportunities for democratic participation are necessary for political legitimacy. This could be true for instrumental reasons. For instance, it might be the case that inclusive and participatory processes are required to ensure the epistemic quality of decisions,Footnote ³⁴ to ensure that the full range of conflicting considerations are included in the discursive assessments of what justice demands,Footnote ³⁵ or to ensure that the political system is resistant to capture by elite interests.Footnote ³⁶ Or it might be true for non-instrumental reasons. For instance, respect for persons might require that certain political rights and liberties be granted,Footnote ³⁷ or might require that public equality of citizens be upheld.Footnote ³⁸ But satisfying such requirements will never suffice to ensure that the set of those included is identical to the set of those governed. Nor will it guarantee that the system of rule that results will satisfy the justice claims that may derive from the needs and interests of non-participating others, whatever those claims turn out to be.

Beckman acknowledges this point when he says that adopting the ASP as the appropriate test for democratic inclusion (as he advocates) does not change the fact that affected parties have interests and needs which the decision-making group must take into consideration:

Even without specifying what these moral demands are, if there are moral demands of non-participating others that must be taken into account, then we cannot say that the “political will” of a “people,” “the voters,” etc. is a self-sufficient ground for legitimacy in political decision-making.Footnote ⁴⁰

A Democracy for the Affected?

Like Beckman and Valentini, Goodin presents collective self-rule as an ideal that is “infeasible, taken literally.”Footnote ⁴¹ He also agrees that we should nonetheless aim to realize the “unattainable ideal … as best we practically can.”Footnote ⁴² However, unlike Beckman and Valentini, Goodin does not think we should be content to simply extend literal enfranchisement as far as practically possible. This would not come close to “approximating” the realization of the ideal at all, since it would sideline vast classes of individuals whose interests are affected by political decision-making and who, on Goodin’s view, have a moral entitlement to participate in political decision-making. This includes infants, future generations, the intellectually disabled, and nature.

Goodin urges instead that we explore ways to approximate “enfranchisement” for all affected. Instead of simply enabling as many as possible to participate in voting, he proposes that we make use of mechanisms of representation to “include” all those who cannot speak for themselves in processes of political decision-making. Literal participation may be impossible, but perhaps a kind of indirect inclusion is possible by way of others speaking on their behalf. Here it is not the number of those included, but the manner of their inclusion that is approximated. Thus, political decision-making ought to incorporate representation for nature, future people, children, the intellectually disabled, and so on.

Goodin’s argument might be seen as an attempt to preserve and, indeed, realize the state of collective self-rule. If we can “enfranchise all affected interests” through representation, then collective self-rule can be realized through the decision-making of representatives. And, if so, then we have no reason to think that the inability of some to “speak for themselves” stands in tension with the realization of the state of collective self-rule.

However, as Goodin admits, it is controversial to regard mere representation as a form of democratic inclusion.Footnote ⁴⁴ Electoral systems construct a principal–agent relationship between constituents and their representatives in which constituents exercise control over their representatives through the mechanism of voting (in theory, at least).Footnote ⁴⁵ This is a system in which voters “speak for themselves” and influence political decision-making, if only indirectly. Apart from the principal–agent relation as constructed through electoral processes, however, it is not obvious that we should regard “being represented” as a form of enfranchisement. If being represented constitutes a form of enfranchisement, then the demand for democratic inclusion becomes trivially easy to satisfy. Monarchies and oligarchies can in principle “enfranchise” all affected interests by virtue of the monarch or oligarchs presenting themselves as “representatives” of those they rule over—as indeed they typically do. Any form of government could then be said to be a form of collective self-rule. As necessary and valuable as representation may be in political systems, merely being represented is not a form of democratic inclusion at all.

Ultimately, it may be that, at the level of institutional design, we find ourselves led to embrace something that resembles Goodin’s proposed model, in which all affected interests are afforded representation in political processes. Indeed, within any political context, there will be parties with needs and interests that may need to be taken into consideration but who are themselves unable to be literally enfranchised. There may be a case for providing representation for these parties in the form of advocates or special officeholders either within traditional decision-making bodies or in their own representative bodies.Footnote ⁴⁶ Some such institutional arrangements may well be required in order to ensure governmental decision-making conforms to the moral constraints that derive from the demands of justice of non-participating others.

But, in such a system, decision-making will (and can only) remain in the hands of humans (and really, a subset of them), as Goodin acknowledges.Footnote ⁴⁷ The humans who are empowered to make decisions will have to bear the responsibility for exercising government over others and for others. The power to make decisions will fall on the shoulders of an enfranchised “few” who decide for the vast array of affected parties who cannot literally participate in the decision-making process. The enfranchised “few” must then govern themselves and govern others in the interests of the participating “few” and the non-participating “many” equally.

In our view, therefore, we should not present the resulting political system as a democracy of the affected, as though all affected were somehow enfranchised by it. We should rather present it, in Robyn Eckersley’s apt phrase, as a “democracy-for-the-affected” since, as she writes:

Insofar as the moral claims of non-participating others must be considered, real-world democracies will need to be something that is quite at odds with what is usually thought to be implied by the principle of collective self-rule: roughly, government of the people, by the people, for the people. Democracies will need to be systems of decision-making in which the enfranchised “few” are responsible, and institutionally organized, to provide for and protect the legitimate needs and interests of all affected. We have not attempted to specify what those legitimate needs and interests are, nor what constitutional arrangements and political institutions would be required to provide for and protect those needs and interests. We have, however, tried to demonstrate that a democracy-for-the-affected cannot be built on the fiction of collective self-rule.