Transnational democracy and the demos
‘Let us imagine a society and then consider what form of government would be just for it’, Craig Calhoun writes to caricature how political theory traditionally has avoided addressing the problems of political belonging by assuming nation-states to be the basis of politics.Footnote 1 Similarly, Frederick Whelan notes, democracy means rule by the people, and political theorists have largely quarrelled over what ‘rule’ should mean, while neglecting the other half of the definition. But ‘any democratic theory must face the logically prior and in some ways more fundamental question of the appropriate constitution of the people or unit within which democratic governance is to be practiced’.Footnote 2 This is the so-called boundary problem in democratic theory: how to delimit the political community relevant for democracy. Enter the all-affected principle, stating that everyone who is affected by a decision of a government has a right to participate in that government.Footnote 3
While justifying political borders has always been a matter as important as it is difficult, political theorists have recently debated the all-affected principle in the context of transnational democracy.Footnote 4 Theorists favouring some form of transnational democracy often invoke this principle to explain why the current, nation-state-based boundaries of democratic governance are not normatively satisfying and why we need to democratise transnational institutions or transnationalise democracy. In fact, it may seem hard to imagine calls for transnational democracy without some version of the all-affected principle, as it forges the normative link between, on the one hand, the worries over how globalisation undermines democratic sovereignty and how increasingly autonomous, unaccountable international institutions impact on people's life chances around the globe and, on the other hand, the conclusion that we must build democratic institutions of some sort at the transnational level.Footnote 5 Political causes and effects no longer operate within the safe confines of nation-states, this argument runs, and transnational democracy seeks to suture the widening gap between those who make decisions and those who are affected by them.
However, while the all-affected principle may appear common-sensibly sound and simple, it raises serious problems as soon as we try to use it to set political boundaries. Addressing these problems, I shall argue, first, that the all-affected principle does not actually give any guidance for delineating a political community but that it does give some absurd recommendations if we were to try to approximate it in political decision-making. Second, considering alternatives to the all-affected principle, I suggest that subjection to law provides a better criterion for democratic inclusion. However, I argue that this subject-to-the-law principle would shift the ground for justifying transnational democracy: To what extent are people subject to transnational systems of law? This turns out to be a contentious issue per se.
First, however, let me offer a few reasons why the all-affected principle seems morally appealing and plausible. For one thing, the all-affected principle seems to expound some classical democratic notions about autonomy and consent. Aristotle, for example, distinguished the self-governing citizenry, ‘ruling and being ruled in turn’, as an element of liberty characteristic of good political rule among equals. Another oft-cited precursor to the all-affected principle is the maxim of ancient Roman law: ‘quod omnibus tangit, ab omnibus tractari et approbari debet’ – what concerns all, all must discuss and approve. Both the Aristotelian notion of the self-governing citizenry and the Roman dictum imply that liberty means living according to laws that you have given yourself. We find a similar concern in the early modern theories of the social contract. John Locke suggests that since men are natural equals, ‘no one can be […] subjected to the political power of another, without his own consent’, and nobody knows better than oneself what is in one's own interest.Footnote 6 Likewise, Jean-Jacques Rousseau grappled with the problems of finding a political form that would be consented to by its participants on rational grounds.Footnote 7 For Rousseau, government is legitimate insofar as it reflects the general will of the political organism formed through the social contract – a will that every citizen is thus justly subjected to. Admittedly, neither Aristotle's autonomy, Locke's consent of the governed, nor Rousseau's general will imply or necessarily support the all-affected principle, but the principle resonates with ideals in these philosophical traditions.
Secondly, the all-affected principle implies a defensive view of politics, where political decisions and institutions inflict costs and burdens, if not damage, on people, as Whelan suggests. Fear of the Leviathan has been a strong argument for democracy: people have a right to participate in politics so as not to have their interests violated by those who wield the monopoly of violence.Footnote 8 The all-affected principle seems to support our hunch to be sceptical of politics. Moreover, historically, it has been employed to support extending the franchise to groups previously excluded from participation in democratic politics. Transnational democrats often reason in a similar way: People affected – for instance, by global warming, trade policies, acid rain, financial deregulation and other transnational issues – form the constituency of world politics.Footnote 9
Thirdly, we may associate the all-affected principle with relational conceptions of justice in cosmopolitan theory. The peoples of the world have grown together in a community, Kant claims, and thus ‘a violation of rights in one part of the world is felt everywhere’.Footnote 10 Some contemporary cosmopolitans likewise argue that we owe duties of justice to distant other persons who are not our compatriots because we are bound together by relations of interdependence.Footnote 11 And just as relations of mutual influence may ground moral obligations, one could argue, they can also serve to justify the boundaries of democratic community: Persons who are affected by decisions by governments and by other powerful actors have legitimate claims to be included in their decision-making, regardless of their nationality or citizenship.
Thus, there is a diverse pantry of democratic and moral traditions of thought that might lend support to the all-affected principle. And yet it turns out to be very difficult both to specify and to apply, as I shall argue next, presenting four arguments which, taken together, provide strong reasons to reject the all-affected principle, its intuitive appeal notwithstanding.
Problems with the all-affected principle
How do we recognise affectedness when we see it?
As the all-affected principle does not itself explain what being affected by a political decision means, we have to complement the principle with some theory about affectedness. An objective approach to affectedness would require that we be able to specify measure and assess the burdens and benefits inflicted upon individuals by political institutions, policies and decisions. Alternatively, one might hold that affectedness is a subjective quality. Let us explore both conceptions in turn.
Generally, being affected by political decisions and institutions implies that some of your basic rights or interests have been infringed upon.Footnote 12 Although not always specifying what being relevantly affected means, advocates of the all-affected principle often employ environmental problems as the epitome of situations in which the principle applies. Global warming, polluted waters, acid rain – these are the kind of contemporary cross-border problems that may affect other people than those who caused them. Thus, you are affected when something bad happens to you through no fault of your own. David Held, a key advocate of cosmopolitan democracy, could probably express such bad things in terms of ‘nautonomy’, that is, as being deprived of your physical, social, economical, political or cultural autonomy.Footnote 13 But we could alternatively adopt, say, Robert Nozick's libertarian theory of self-ownership as the relevant theory of affectedness.Footnote 14
Depending on which theory of affectedness we choose, we will arrive at different or even diametrically opposed judgements. Nozick might say that forced redistribution of incomes violates people's basic rights and interests, whereas safeguarding the sort of social autonomy Held has in mind would require an extensive welfare state financed through taxation. Different people will count as affected by the same decision depending on which theory of affectedness we choose: Someone is unjustly affected on the Nozickian account if he or she is forced to pay taxes for redistribution, but on Held's account if he or she is not guaranteed a basic level of subsistence. So which theory of affectedness should we pick? These are the kinds of clashes of interests and values that democratic politics is supposed to be able to sort out in a peaceful manner. People will disagree about what it means to be relevantly affected, just as they disagree on other fundamental matters of principle.Footnote 15
The all-affected principle is usually understood to concern burdens, but not benefits, an asymmetry which might follow from a sceptical view of politics, where democratic rights supposedly enable people to defend their interests against government intrusion. And, at first glance, taking benefits into account seems problematic. Would a decision be undemocratic if you prosper from it without having had an opportunity to participate in making it? To take an environmental example of the kind that transnational democrats like to invoke, it seems peculiar to argue that an upstream community which has unilaterally cleaned up a polluted river has made an illegitimate decision solely because the decision-making process excluded beneficiaries downstream. We could then qualify the all-affected principle by stating that only negative externalities can be illegitimate, not the positive ones.Footnote 16
But real-life political issues are usually complex. Political decisions produce both burdens and benefits and distribute them unequally. Furthermore, once we start conceiving of political consequences in economic, utilitarian terms, we might also think of benefits and burdens as commensurable: A burden translates into a negative benefit and vice versa – that is, being deprived of a benefit is a burden and alleviating a burden equals a benefit. Whereas it would be difficult to keep separate balance sheets for benefits and burdens respectively, if we do take them both into account the all-affected principle, in turn, becomes difficult to apply.
Moreover, how should we aggregate and weight burdens and benefits? Torbjörn Tännsjö argues that if we regard democracy as a method of aggregating interests, it would be more reasonable to interpret the all-affected principle as a claim that everyone should have an influence proportional to the stakes he or she holds.Footnote 17 A graded right to participate fits the rationale behind the all-affected principle, since only then would an intensely affected minority be able to trump a slightly affected majority. On the other hand, grading people's right to participate in decision-making according to the extent to which they are affected would be at odds with the democratic principle of one person, one vote. In effect, it grants a veto to minorities with strong preferences. I think few proponents of the all-affected principle would depart happily from this basic principle of democratic equality.Footnote 18
Hence, trying objectively to specify the all-affected principle seems futile, because the principle must be founded on some particular notion of what it means to be affected and such notions are frequently matters of fundamental political disagreement.Footnote 19
So perhaps we should instead look for a subjective conception for affectedness, whereby people who consider themselves to be affected by political decisions have a rightful claim to inclusion. Michael Saward, for instance, endorses the all-affected principle and advocates a ‘“subjective” way of locating issue-based subject populations’.Footnote 20 He suggests cross-border initiatives, by which a significant number of citizens can raise border-transgressing issues for referenda, and argues that innovations like these would actually better match the rationale behind the all-affected principle than grand schemes for cosmopolitan democratic reform would.
A subjective understanding of affectedness could draw support from a different tradition in democratic theory. In Noortje Marres’ constructive reading of the debate between Walter Lippmann and John Dewey, the affected come to play a role different from the one which they have according to conventional transnational democratic theory.Footnote 21 Despite their differences, Lippmann and Dewey both address not the subjects of politics, but its objects.Footnote 22 Whereas democratic theory traditionally focuses on the persons whom democracy enables to master their own fate, Lippmann and Dewey are concerned with the issues of politics. In the complex, technological societies in which we live, the intricate objects of politics seem to constitute an obstacle to democracy, for how are citizens supposed to govern themselves, when the issues that they have to deal with are so complex and strange?
It might seem easy to solve the problem that complex objects pose to democracy by either providing citizens with better information or by simplifying difficult issues so that citizens can grasp them. Dewey, however, contends that ‘foreign entanglements’ and complex issues, far from constituting an obstacle to democratic politics, actually play a central role in getting people involved in politics. Strange, unfamiliar and complex issues are an enabling condition for democracy, and precisely because issues are difficult to resolve, we need to bring them out in public view. Publics get involved in politics where existing institutions fail to deliver. Dewey suggests that the public does not just spring up from nowhere: It consists of people who are affected by human actions on which they have no direct influence. People who believe that they have been affected by some such issue must organise themselves into a visible community: a public. On this Deweyan account, then, affectedness is subjective, and more an enabling condition for democratic politics and participation than a rigid, objective criterion for drawing boundaries. Furthermore, as soon as a group of affected persons form a public and thus becomes involved in politics, that group also begins to affect others.
To conclude, subjective conceptions of affectedness properly recognise the reasons why people take political action, whereas the attempts to find objective standards of affectedness imply a more legal-technical view of politics. On the other hand, the subjective approach also seemingly leads us to Joseph Schumpeter's claim that we must ‘leave it to each populus to define himself’.Footnote 23 But Dahl famously argues that Schumpeter's criterion makes it impossible to distinguish democracy from autocracy and that it offers no principles by which to decide whether somebody has unjustly been excluded from participation.Footnote 24 Accepting such a non-principle would seem to rob transnational democracy much of its normative force, as it would have difficulties explaining why current forms of decision-making need democratic reform.
A vicious regress of constituting decisions
A serious objection to the all-affected principle holds that it actually does not solve the boundary problem, because the principle creates an unsolvable hen-and-egg paradox. Since every political decision presupposes a priori decision on whom to include – a decision that affects some persons – the principle leads to a logical as well as procedural impossibility, as Frederick Whelan demonstrates:
Before a democratic decision could be made on a particular issue (by those affected) a prior decision would have to be made, in each case, as to who is affected and therefore entitled to vote on the subject – a decision, that is, on the proper bounds of the relevant constituency. And how is this decision, which will be determinative for the ensuing substantive decision, to be made? It too should presumably be made democratically – that is, by those affected.Footnote 25
Thus, when we try to determine the demos by means of the all-affected principle, we enter an infinite regress of constitutive decisions from which the all-affected principle offers no escape. Delimiting the political community is a political decision which affects people too, probably more than any decision that might follow once the community has formed itself.
Moreover, once we apply the all-affected principle to substantive policy, we clearly see why it is indeterminate. In most cases, who is affected depends on what substantive decision the political community makes; and the problem is not just theoretical.Footnote 26 For example, a protectionist trade policy benefits and harms different people than does a free trade regime. So whom should we include in deciding which trade policy to adopt? Likewise, progressive taxation will affect different people differently than a flat tax, so who ought to be included in deciding taxation policy? Depending on whom we include in the decision-making process, we will reach different policies, and depending on what substantive policies we choose, we will affect different people who would have to be included in the first place.Footnote 27
Whelan's regress argument demonstrates why the all-affected principle gives no guidance on how to delineate the community, a problem which proponents of the principle have difficulties addressing. Torbjörn Tännsjö suggests that we could get out of the logical loop by selecting a constitutional assembly of ‘founding mothers and fathers’ to solve the boundary problem.Footnote 28 But this solution not only disregards the fact that the boundary problem recurs on every issue, if we take the all-affected principle seriously; the solution also effectively excludes most of the people who are affected by the constituting decision from the act of constituting it. Similarly aiming to set up an institution to handle boundary problems, David Held suggests that ‘issue-boundary forums or courts will have to be created to hear cases concerning where and how “significant interest” in a public question should be explored and resolved’.Footnote 29 That is, these new institutions would be given the authority to decide when, where and how the all-affected principle applies, including defining affectedness. As Michael Saward notes, Held's proposal would vest enormous powers ‘in unelected authorities requiring inhuman levels of knowledge and wisdom’.Footnote 30
Furthermore, such institutional solutions presuppose that there is a correct and objective answer to be reached – that these institutions, with which we entrust the power to solve boundary problems, can decide in a neutral way who has been affected and thus has a legitimate claim to be included. As the examples indicated above demonstrate, who is affected cannot be settled independently of the substantive decision. By giving independent institutions the power to decide who is affected, we also grant them the power to decide on substantive matters as well. Depending on how these institutions draw the boundaries, different policies will result, and the practice of gerrymandering demonstrates why this is not just a hypothetical concern. Drawing political boundaries is always an exercise of power.
And not only will different boundaries produce different policies, but also different interests among the people sorted into communities. While people might have a right to be included when their fundamental interests are at stake, what those interests are and how they are represented often depend on how boundaries are drawn. Your interest in migration policy, for instance, will be radically different depending on whether you end up on this side of the border or the other.Footnote 31 Again, the all-affected principle leads into a vicious regress.
Fickle boundaries
Whether we think of affectedness as objective or subjective, the all-affected principle seems to imply that we ought to redraw political boundaries for each decision that is to be made or, at any rate, that each issue requires its own functional constituency. That is, boundaries would be considerably fickle if constantly redrawn so as to meet the requirements of the all-affected principle. Some suggest that this requirement represents a major problem with the principle, whereas others see the resulting volatility of boundaries as desirable.
Advocates of the principle seem to disagree on how far the idea of issue-based constituencies leads. The most radical interpretation holds that for every single political decision to be made, we first have to decide the relevant political community, that is, who is affected and thus ought to be included. Some theorists of transnational democracy nod in this direction and embrace the fluid boundaries following from the all-affected principle. John Dryzek, for example, argues that his account of deliberative democracy ‘can cope with fluid boundaries, and the production of outcomes across boundaries’, which, in his view, makes it such a suitable model of transnational democracy.Footnote 32 David Held, by contrast, seems to think of redrawing boundaries by means of the all-affected principle as a one-off process whereby we shift authority to new but permanent territorial layers of government, with regional and global democratic institutions added to those already existing at national and local levels in some countries.Footnote 33
Some critics argue that because the all-affected principle leads to political boundaries that are unstable and issue-area specific, it provides an impractical criterion of legitimacy for political institutions.Footnote 34 Political boundaries could change from day to day and it would be hard to consolidate political institutions. Even if we shift the tasks that states perform to new regional, global or issue-specific institutions, such tasks nevertheless require a degree of continuity, especially if they are to be done in a way that is democratically accountable to anyone who is affected by them.Footnote 35 However, as long as political decisions have fairly permanent and uniform consequences, the boundaries drawn by the all-affected principle would be accordingly stable. Moreover, research on international institutions suggests that international problem-solving is often organised into specific issue areas that are all but ephemeral.Footnote 36 In fact, transnational democrats could argue that just as the nation-state used to be a practical if imperfect shortcut to realise the all-affected principle, so international institutions offer a decent approximation in our globalised world, where improved democratic participation could allow for a better overlap between those who make decisions and those who bear their consequences.
Thus, while these practical problems in applying the all-affected principle might be overstated, we should worry more about what happens to the conditions for democratic participation once we redraw political boundaries according to the all-affected principle. Issue-specific political boundaries might be feasible, but are they also desirable? Whether we understand the resulting political institutions to be overlapping, issue-specific institutions or layered territorial entities with broader responsibilities, the communities corresponding to them are supposed to replace the once so self-evident categories that nation-states sorted people into. Just as nation-states would be replaced by an array of institutions claiming authority, so the sole citizenship of the individual would be replaced with a variety of affiliations with different communities of fate and choice.Footnote 37 But what happens then to the rights, duties and belongings that territorial states, for better or for worse, have granted their citizens? Michael Saward argues that the protection of democratic rights depends upon secure and equal membership in a given unit. Applying the all-affected principle thus risks taking away the very foundation of democratic rights without replacing it with something better:
if the constituency can and must change for each decision, then the rights of ‘members’ are not fixed, or immutable, from one decision to the next […] Membership is only secure, because the grounds of citizenship and rightful political participation can only be clear, in a territorial entity.Footnote 38
Nor would the rights of community members be equal, if people only had a right to participate in political decision-making corresponding to the stakes they hold. Thus, even if the overlapping, multilevel institutions and constituencies were to be stable and fixed rather than ephemeral, the rights of participation accorded to citizens would be tenuous.Footnote 39
Does the all-affected principle necessarily imply fluid, issue-specific political boundaries? A different way to abide by the all-affected principle would be to ensure that a given territorial political community, national or transnational, produces outcomes that are in accordance with it. Instead of reshaping the boundaries of political institutions to fit with their consequences, we could reshape the political consequences to fit with existing boundaries, so that only those persons are affected who are already included in political decision-making. I shall now address this alternative way of fulfilling the all-affected principle.
Unpalatable policy recommendations
Thus far I have argued that the all-affected principle does not offer any clear guidance on whom to include in a democratic political community, nor does it help us decide on substantive policies. But as vague as the principle appears, other recommendations can be derived from it. However, those implications, too, I shall argue, suggest further reasons not to make democracy dependent on the all-affected principle.
Hans Agné presents an interesting argument against the all-affected principle by trying to explicate the conditions for fulfilling it in a world of states.Footnote 40 The all-affected principle has two components: to participate in making a decision and to be affected by that decision. For the sake of simplicity, if we interpret the components as dichotomous (as do most of the all-affected principle's supporters), there are two ways in which the all-affected principle could be violated: If someone who is affected by a decision is excluded from taking part in it or, conversely, if someone who is not affected participates in making a decision. Such illegitimate exclusion and illegitimate inclusion are both ruled out by the all-affected principle.Footnote 41 Obviously, the all-affected principle is equally satisfied if you participate in making a decision that affects you or if you do not participate in making a decision that does not affect you.
Interpreted dichotomously, the all-affected principle puts seemingly drastic requirements on democratic decision-making. Even if we assume the state to be isolated from its surroundings, it seems difficult to ensure that nobody participates in making decisions without being affected by them or vice versa. Once we add the complicating assumption of a world of states, it becomes virtually impossible to avoid illegitimate exclusion and thus to avoid violating the all-affected principle. Even if a state interacts minimally with the surrounding world, some decisions that the state makes will affect some persons outside its borders who are not included in making those decisions.
In effect, globalisation might also lead to illegitimate inclusion. Agné's argument runs roughly like this: A community can avoid illegitimate inclusion when collective decisions concern properties that all its members share. For example, if everyone is at least a potential tax payer and a potential benefactor of public expenditures, then we may safely include everyone in deciding on tax policy.Footnote 42 The more characteristics people share, the more evenly will they be affected by the decisions that they make. Thus, to avoid illegitimate inclusion, a democratic community must seek to make its members more uniform – economically, socially, culturally and by any other relevant dimension – so that nobody who participates in collective decision-making can shield herself from the consequences. In order to make the population less heterogeneous, fragmented and stratified, it seems appropriate to try to isolate the community from influx from outside, for to the extent that globalisation brings heterogeneity into a previously well-integrated collective, illegitimate inclusion could follow. Because the all-affected principle suggests that democracy can only be achieved in an isolated and homogenous political community, Agné rules the principle out as an element of nationalist ideology: ‘[O]nly a nation-state, firmly founded on a mythology of unity and autonomy, can wield the social powers required by the [all-affected principle].’Footnote 43
Thus, a community that takes the all-affected principle seriously would have to pursue a policy of isolation from its surrounding world and internal homogenisation. This conclusion might seem surprising: We came seeking support for transnational democracy, but ended up with nationalist, isolationist policies. Still, these implications of the all-affected principle are reflected in the standard narrative framing transnational democracy: Since globalisation has displaced the once so neat match between political authority and cultural, economic and social borders, theorists of transnational democracy claim, we need to build new democratic institutions beyond or above the state as well. In so doing, we may regain congruence between the rulers and the ruled – a congruence which before it was lost always required a high degree of homogenisation, unity and cohesion. If we were to enact the all-affected principle at the transnational level, it might have policy implications similar to those which emerged at the state level.Footnote 44 For example, identity politics in the EU seem to reproduce nationalist ideology at the European level, albeit under a flag of post-nationalism.Footnote 45
Hence, for all their claims to shrug off ‘methodological nationalism’, transnational democrats bring on board more of the notion of the nation-state as the locus of politics than they seem to acknowledge, as they premise democracy on a conception of congruence between political and social boundaries. As it were, the all-affected principle renders transnational democrats just as rooted in a tradition regarding the political entity to which democracy applies as a closed circuit, a conception so appositely described by Walter Lippmann:
The democratic tradition is […] always trying to see a world where people are exclusively concerned with affairs of which the causes and effects all operate within the region they inhabit. Never has democratic theory been able to conceive itself in the context of a wide and unpredictable environment. […] And although democrats recognise that they are in contact with external affairs, they see quite surely that every contact outside the self-contained group is a threat to democracy as originally conceived.Footnote 46
All in all, the problems involved in trying to specify and enact what the all-affected principle requires provide strong reasons to search for better criteria by which to determine the boundaries of democratic communities. Next, I shall consider three such alternatives.
Alternative boundary criteria
Procedures, not decisions
So far I have argued that the all-affected principle is indeterminable, virtually impossible to apply, and leads to some rather peculiar guidelines for decision-makers who take it seriously. But should we actually take it that seriously – in the sense of reading it as a literal rule by which democratic politics must abide? After all, most normative principles are vague and ambiguous, and might lead to absurd conclusions if we try to follow them too rigidly. So perhaps we should not let the absurd implications that arise in concrete situations lead us to abandon a principle that might be sound in a more abstract sense.
Gustaf Arrhenius defends the all-affected principle against such easy confutation.Footnote 47 Although the all-affected principle may be both impractical and unfeasible, it may still well be part of the normative ideal of democracy, Arrhenius suggests. We could cherish the principle as an end, although not as a means to that end. That is how rule utilitarians reason when they admit that attempting to maximise utility in each and every action we take could lead to absurd consequences, but nevertheless argue that we should seek political institutions that lead to the greatest possible utility.Footnote 48 Similarly, the all-affected principle might sometimes lead to absurdities if applied in particular circumstances, but nevertheless provide a standard by which we may measure the democratic inclusiveness of practically feasible methods of decision-making. And even though none of these methods will ever fulfil the principle's demands, Arrhenius concludes, the all-affected principle nonetheless helps us discriminate among better and worse decision-making procedures. Hence, the all-affected principle would not guide first-order political decision-making, that is, substantive, everyday policy decisions, but it might still help us make second-order political decisions, which ‘are about first-order political decision-making, i.e. about where, how, when and by whom everyday political decisions are to be made’.Footnote 49
It seems reasonable to consider the all-affected principle not as a razor-sharp rule, but rather as a desirable if abstract ideal, applicable to questions of institutional design but not to concrete issues of boundary drawing. As a trade-off, of course, this abstract all-affected principle would lend less unequivocal support for transnational democracy, and lessen its intuitive appeal, if no longer applicable to, say, issues such as whether it is permissible to build nuclear power plants or perform atmospheric nuclear weapon tests near the border of another state.
However, if we could distinguish decisions from ideals, this interpretation of the all-affected principle would allow us to escape Whelan's logical loop, Arrhenius suggests. Instead of entering the infinite regress of constitutive decisions, we should simply concoct a theory of interests and analyse how different institutions would affect people's interests – and then decide who ought to be included in those institutions. (After that, presumably, we could go back to democratic business as usual.) But who is this ‘we’ supposed to decide on these important matters? Who should decide what affectedness means and analyse the consequences of different political choices? These too are political decisions and taking the all-affected principle seriously, if not literally, they should reasonably be made by anyone who is affected by them. Hence, we are drawn back into the regress, which Arrhenius fails to bring to a convincing end.
Autonomy, not affectedness
Robert Dahl has argued that given that a democratic process is desirable for a group of people, the values of the democratic process – where personal political autonomy is paramount – can sometimes better be obtained by changing the boundaries of their political unit, ceteris paribus.Footnote 50 Susan Hurley similarly endorses what she terms an endogenous approach to the boundary problem, according to which we can assess boundaries in terms of ‘distinctively democratic values, such as values of self-determination, autonomy, respect for rights, equality and contestability’. On this view, ‘some choices of boundaries and units and assignments of jurisdiction might tend to repress and others to foster the autonomy of individuals, respect for their rights, and their deliberative and rational capacities’.Footnote 51 Thus, political boundaries could and should be evaluated in terms of their effects – not just on people's interests, but on core values integral to democracy itself.
The idea that boundaries should be drawn so as to maximise the values of democracy may suggest an alternative to the all-affected principle. This alternative advances on Arrhenius’ solution by replacing the indeterminable notion of affectedness with autonomy as the criterion for deciding the boundaries of a democratic community. This principle, which we could call the maximal-autonomy principle, requires:
that people be included in political procedures to the extent that their inclusion yields the greatest amount of autonomy to the greatest number of people, while accounting for both those whose who are included and those who are excluded, and accounting for actions performed both individually and collectively.Footnote 52
Advancing this principle, Hans Agné argues that it would solve some of the problems he deduces from the all-affected principle. The maximal-autonomy principle better matches other core democratic concepts and intuitions, and it focuses on a quality more central and well-defined than affectedness, namely autonomy understood as action capacity.Footnote 53 Importantly, the maximal-autonomy principle factors in both those who are included and those who are excluded by boundaries. Thus, the persons whose interests must be taken into account are to be found on both sides of the border once it has been drawn. This aspect of the maximal-autonomy principle solves the problem that borders by their very nature affect both those who are included and those who are excluded, a conceptual feature of borders that the all-affected principle cannot escape.Footnote 54 Furthermore, factoring in both insiders and outsiders renders the demos in principle unbounded and global. In practice, the legitimate self-determination of each democratic polity is derivative of this global demos as a whole, as Arash Abizadeh points out:
The unbounded demos thesis does not, of course, rule out the potential legitimacy of political borders and differentiated jurisdictions. It simply confirms that the existence of political borders and their regimes of control require justification.Footnote 55
Hence, by giving equal standing to insiders and outsiders, the unbounded demos thesis could be justified on cosmopolitan grounds. Andrew Kuper argues that a cosmopolitan theory could very well hold that the world should be divided into a system of sovereign states – but the existence of states must be justified, not merely assumed.Footnote 56 The values of democracy could take place among the kinds of cosmopolitan values in terms of which borders must be justified on this account.
While it might be difficult to imagine what it means to justify boundaries to an unbounded, global demos, and how such justification could be achieved in practice, the maximal-autonomy principle resolves some important issues following from the all-affected principle, such as the fickle-boundaries objection: For individuals to exercise their democratic autonomy, political institutions would presumably need to be stable and comprehensive, something which the maximal-autonomy principle allows for.
But the maximal-autonomy principle comes with some peculiar problems of its own. Based on a consequentialist logic, the maximal-autonomy principle seems to open for some staple criticism of classical utilitarianism's intuitively abominable consequences. Classical utilitarianism suggests that it is morally right to throw a handful of Christians to the lions, if their pain does not outweigh the happiness of the cheering spectators in the Colosseum. Now, substituting autonomy for happiness as the quality which is to be maximised does not allow us to avoid that problem, because the problem lies in consequentialism, not in the currency of consequences. Would it be legitimate to rob some people of their autonomy to maximise aggregate autonomy for everyone? Yes, it seems. The principle of maximal autonomy would allow us to disenfranchise or expatriate some persons to increase overall autonomy (on both sides of the divide between inside and outside). Not wanting to bite this bullet, Agné instead suggests that exclusion (as well as inclusion, presumably) on such terms is not compatible with his principle because ‘Political participation for the exertion of autonomy does surely not benefit from the fear created by such measures.’Footnote 57
This retreat position brings back in a concern for the affected. Moreover, this consequentialist formulation of the autonomy principle also seems to remove boundary decisions from the ambit of democratic decision-making. Like the objective approaches to affectedness, the maximal-autonomy principle does not require that boundary decisions should be actually justified to those who are included and excluded. Instead, it suggests a criterion by which to draw and justify boundaries. Thus, even though it elevates autonomy as the criterion by which to include and exclude people, this principle seems to be at odds precisely with the idea of democratic autonomy. For who is to judge whether a certain boundary maximises autonomy? We can't vote about it. Just like the all-affected principle, the maximal-autonomy principle would have to rely on some boundary court, constitutional assembly or law-giver to determine the boundaries.
Against democratic consequentialisms of this kind, Jeremy Waldron argues that ‘any theory that makes authority depend on the goodness of political outcomes is self-defeating, for it is precisely because people disagree about the goodness of outcomes that they need to set up and recognize an authority’.Footnote 58 That problem is not solved by the maximal-autonomy principle either.
Subject to the law, not affected
As I suggested earlier, the all-affected principle may seem intuitively appealing because it builds upon core values in the democratic tradition. Democratic autonomy implies living by laws that you have given yourself and having means by which to assert your interests against the government. A different way to express the congruence between governing and being governed is a principle declaring that ‘The citizen body in a democratically governed state must include all persons subject to the laws of that state’.Footnote 59 This subject-to-the-law principle, as I call it, solves some, if not all, of the problems that follow from the all-affected principle.
Just like the all-affected principle, the subject-to-the-law principle can be justified in terms of the general ideals of democratic autonomy as self-government or as government requiring the consent of the governed. In that way, it can be formulated to be compatible with both republican and liberal models of democracy, and their respective concepts of law and participation.Footnote 60 The republican version stresses that the citizens of a state should themselves be the authors of the laws that constitute their polity, which requires that they participate actively, whereas the liberal version regards the right to participate in making the laws as instrumental to protect individual liberties and personal interests, a right that citizens may practice by electing representatives.Footnote 61 Indeed, the subject-to-the-law principle seems to make better sense of these democratic ideals than the all-affected principle does, since it does not take the detour over the troubling concept of affectedness.
How does being subject to the law differ from being affected by decisions? The two principles do not necessarily overlap: You may be subject to laws that do not affect your interests in any tangible sense, and vice versa. As an illustration of this difference, Claudio López-Guerra suggests expatriates, who generally are not subject to the laws of their countries of origin, but may sometimes be affected by their decisions.Footnote 62 Unlike the consequences implied in the all-affected principle, however, being subject to the law cannot always be expressed as a calculus of costs and benefits. Furthermore, laws normally specify to whom they apply: usually people living within a territorial state.Footnote 63 For that reason, the subject-to-the-law principle offers no internal solution to the boundary problem, since it merely presumes that the relevant community is already determined and that there is already a state in place to maintain the laws and do the subjecting.Footnote 64 The principle only claims that where there is law, those subject to it have a legitimate claim to be included in its making.
However, the subject-to-the-law principle sets clearer criteria for illegitimate inclusion/exclusion than does the all-affected principle. We can easily think of cases when people are subject to laws that they have not even indirectly or passively participated in making, or cases when people participate in making laws to which they are not even potentially subjected. Such cases constitute illegitimate exclusion or inclusion, respectively.Footnote 65 Moreover, given that it is easier to determine who is and who is not subject to law than who is affected by a particular decision, illegitimate inclusion and exclusion seem to be more readily identifiable by the subject-to-the-law principle than by the all-affected principle. Hence, the subject-to-the-law principle is more specific as to what should be democratically controlled, namely, the power to make law, but not necessarily any and all power to make individual or collective decisions which might affect someone else. Furthermore, we need not fear the absurd policy recommendations that follow from the all-affected principle, because law (ideally at least) applies to all its subjects regardless of their individual properties, even though law may affect them differently.Footnote 66
Thus, the subject-to-the-law principle seems to be less ambiguous and more applicable, while spelling out the same abstract democratic ideals that seemed to resonate with the all-affected principle. If the subject-to-the-law principle is indeed sounder and simpler, does it lend support to claims for transnational democracy? Notably, the principle shifts the grounds for justifying transnational democracy: The crucial question now is not whether people are affected by transnational decisions or institutions, but whether they are subject to transnational systems of law. This turns out to be a contested empirical issue, with a diverse group of scholars outlining an emerging global system of hegemonic law or a world constitution.Footnote 67 I shall consider such claims about transnational law, which in conjunction with the subject-to-the-law principle may seem to justify calls for transnational democracy. Considering them in detail serves to outline the content and limits of the subject-to-the-law principle.
Hauke Brunkhorst suggests that we are today already subject to a system of ‘world law’ or ‘global hegemonic law’, which ‘extends from the Lex Mercatoria to the comparatively fixed domestic as well as supranational positive system of human rights’.Footnote 68 Because the democratic ideal of autonomy requires that those subject to the law are also its authors and because people are subject to law that is issued by global or supranational agencies, democracy must also become global in scope, Brunkhorst argues. He lists a number of sources of global hegemonic law: binding decisions made by the UN General Assembly and Security Council, agreements made by the World Trade Organization, the statutes of the Organization for Security and Cooperation in Europe and the Council of Europe, international and transnational organisations such as the World Bank and the International Monetary Fund, private organisations such as the International Olympic Committee, and intergovernmental institutions like the G8 summits and the Organisation for Economic Cooperation and Development, which ‘have far-reaching authority for regulating the entire global economy’.Footnote 69 Brunkhorst argues that:
As a whole, the ‘network of agreements’ […] has led to a new form of international and supranational comprehensive jurisdiction [Allzuständigkeit], which is no longer the distinctive property of the sovereign state, but rather is claimed by a multitude of post-national organizations, partly in direct competition with the states that are linked with them.Footnote 70
Now, let me point to a few problems in justifying transnational democracy on the basis of this empirical claim. First, under the rubric of global hegemonic law, Brunkhorst includes international agreements, statutes of international organisations, international organisations themselves, standards set by private, non-governmental organisations, and decisions and resolutions taken at international summits or by the UN. By pointing to this vast, multifarious and pervasive body of world law, Brunkhorst can inflate his argument for extending democracy to the transnational level. At the same time, however, he also undermines the argument, for what is it about law that makes it so important for those subject to it to be its authors? Its character of being coercive, binding and enforceable on individual citizens, a characteristic not shared by most of the instances on his list.Footnote 71 Moreover, stretching the concept of law to include these international institutions also renders the democratic requirements of the subject-to-the-law principle unclear: we are back at being affected by decisions, rather than being subject to the law.
Second, while international law proper has certainly gained in content, scope and importance over the past century, it is not law in the same sense as positive law within a state. Its sources, its enforcement, its subjects, even its normative and ontological status is different. International law is predicated on the recognition of state sovereignty. Thus, its authority depends on states to formulate, observe and enforce it, and it lacks an established compulsory judicial system to settle disputes as well as a coercive penal system. And while there is, arguably, a tendency in international law increasingly to implicate non-state actors, such as corporations and individuals, its subjects and parties are still predominantly states. Even when international law concerns individuals (refugees or war criminals, for instance) it addresses states to enforce and respect its provisions.Footnote 72 And it is still exceptional that international organisations are recognised as parties to international law. Thus, international law turns out to be law of a different kind than that addressed by the subject-to-the-law principle. It is law of states, not of self-governing citizens.
Third, while international institutions have gained in importance too, they are not examples of global hegemony. They result from agreements between states. To the extent that they get implemented, revised, monitored and enforced, states are in charge of these tasks. Like international law, international organisations do not imply the end of sovereignty but rely on it. Moreover, the tendency toward increasing juridification or legalisation of transnational institutions, that is, to express international agreements in a law-like form, does not necessarily imply actual law-making and law-enforcing capabilities.Footnote 73
The EU, however, might be a special case, because community law has precedence over the national law of member states. Moreover, the European Court of Justice has successively widened its jurisdictional mandate and by the doctrine of direct effect, community law imposes obligations and confers rights upon individual legal subjects in the member states. Thus, by the subject-to-the-law principle, EU citizens would have a strong and legitimate claim to be included, if only indirectly, in its lawmaking.Footnote 74 On the other hand, the EU relies on national judiciaries and other governmental agencies of the member states for implementing and, more importantly, enforcing its legislation.Footnote 75
Finally, coupling the subject-to-the-law principle with the thesis of global hegemonic law in order to justify claims for transnational democracy creates a problem similar to one we identified with the all-affected principle: The subject-to-the-law principle presupposes a sovereign law-giver, the subjects of which have a legitimate claim to democratic participation. But the global hegemonic law thesis disperses the authorship of the law to a multitude of post-national organisations, overlapping and competing with each other's jurisdictions, as well as with states. Along these lines, some legal theorists similarly dispense with the sovereign law-giver altogether, and suggest that a legal system should not simply be understood as the implementation of a sovereign will, but rather as an autonomous system in charge of codifying the code legal/illegal. This holds especially well, they argue, since globalisation has undermined the traditional doctrine that legislation ultimately depends on a constituent power (the sovereign). The upshot of this systems theory version of the global law claim is that there already exists a closed, decentred or polycentric, autopoietic global legal system, producing valid legal norms:
The global political constitution is not produced by legislation but through decentered legal self-reflection and through a global community of courts, which ascertain legal validity and legal violations.Footnote 76
Breaking thus the link from the global legal subsystem to sovereignty, this account of global law seems difficult to adjoin with the subject-to-the-law principle to support a claim for transnational democracy. If laws are not made by an identifiable, sovereign law-giver, the subject-to-the-law principle lacks a clear target for its claims about democratic autonomy. The subject-to-the-law principle requires a sovereign, a law-giver, something which these notions of global law deny.Footnote 77
Of course, like the all-affected principle the subject-to-the-law principle lacks a preference for levels. Even if hegemonic global law exists to the extent that Brunkhorst and others suggest, one might argue that the power to legislate should be brought back to national legislatures, rather than that transnational legislation should be brought under transnational democratic control. On this point, the two principles are equally indeterminate and would have to be complemented by some claims about why renationalisation is either unfeasible or undesirable, or both. But renationalisation might be the more compelling alternative, since, as Dominique Leydet argues, we cannot expect that the ‘totality of addressees’ of global norms will ever be able to exercise its sovereignty rights.Footnote 78
Thus, while the subject-to-the-law principle could lend support to transnational democratisation, its recommendations are more restricted and specific than those following from the all-affected principle. It does not justify sweeping claims for global democracy, but it might suggest that specific institutions like the EU and other international legislative bodies should be democratically accountable to those who are subject to the laws that they issue.
Conclusion
I have argued that the all-affected principle should be rejected, because it does not help us draw the proper boundaries of political community and because it may have some rather unpleasant consequences if we were to try to approximate it anyway. Neither of the three alternatives considered here solve the fundamental boundary problem in democratic theory, but they all advance upon the indeterminate all-affected principle. The subject-to-the-law principle seems to be least problematic, substantiating a precise claim: that any already existing lawgiver should be democratically governed by those subject to the laws it issues. Thus, the subject-to-the-law principle might help debates for and against transnational democracy to get out of the quagmire of premising the project on nebulous notions of globalisation and affectedness. By drawing our attention to law and its enforcement upon people as the ground for democratic inclusion, the principle fruitfully opens up new possibilities for arguing about the boundaries of transnational democracy.
Perhaps, though, these attempts to solve the boundary problem demonstrate that it is unsolvable in principle: a community cannot lift itself by the hair and provide its own democratic justification. Of course, we can still solve boundary problems in both theory and practice, but we cannot always justify such solutions in terms of democratic principles. Chris Brown argues that the quest for a democratic legitimacy of borders is symptomatic of a wider problem, by which political theory is reduced to moral theory and all social arrangements are regarded as in need of rational justification. But, as Brown suggests, we have no reason to believe that such justification is always going to be available:
Politics is about practical action in a realm where no answer can be other than provisional, not about the application of formulae concerning matters such as social justice – and it ought not to be surprising that when formulaic approaches are made to subjects such as the legitimacy of borders the argument quite soon breaks down.Footnote 79
Even – or especially – in an allegedly globalising world, borders are ubiquitous. Drawing, defending and challenging the boundaries between inside and outside are among the most political of issues, and it is at best naïve to think that we could find some formula by which to bridge the gap between the people and its constitution and, once and for all, make the drawing of boundaries uncontroversial and, effectively, apolitical.Footnote 80
