Res Publica

, Volume 16, Issue 4, pp 397–413

The Tyranny of the Enfranchised Majority? The Accountability of States to their Non-Citizen Population


    • School of Public PolicyUniversity College London

DOI: 10.1007/s11158-010-9133-z

Cite this article as:
Benton, M. Res Publica (2010) 16: 397. doi:10.1007/s11158-010-9133-z


The debate between legal constitutionalists and critics of constitutional rights and judicial review is an old and lively one. While the protection of minorities is a pivotal aspect of this debate, the protection of disenfranchised minorities has received little attention. Policy-focused discussion—of the merits of the Human Rights Act in Britain for example—often cites protection of non-citizen migrants, but the philosophical debate does not. Non-citizen residents or ‘denizens’ therefore provide an interesting test case for the theory of rights as trumps on ordinary representative politics. Are they the ultimate success story of the human rights framework? Or was Michael Walzer correct to describe government of denizens by citizens as a modern form of ‘tyranny’? This paper argues that neither liberal rights theorists nor democratic republicans provide a coherent response to the existence of denizens. Liberal rights theorists overstate the extent to which a politically powerless status can secure individual rights, while democratic republicans idealise the political process and wrongly assume that all those affected by laws are eligible for political participation. The paper outlines an alternative model for assessing the accountability of states to their non-citizen population, informed by the republican ideal of non-domination. It identifies gaps in state accountability to denizens–such as where there is inadequate diplomatic protection—and argues that these gaps are particularly troubling if their exit costs of leaving the state are high.



The debate about the legitimacy of constraints on democracy is as old as Plato and as young as discussions on whether we need a British Bill of Rights. Both sides are troubled by political power as a source of tyranny and oppression. Legal constitutionalists worry about the tyranny of the majority: that vulnerable minorities will find themselves the victim of the capriciousness of popular democracy. They argue that rights should be insulated from the rough and tumble of politics and guarded by independent trustees with no interest in re-election or subjection to political pressure. Constitutional rights sceptics or political constitutionalists hold that the delegation of power from democratically elected representatives to unelected judges leads to unequal concentrations of power because the judiciary lacks the incentive to be responsive to all interest groups. Entrenching rights in a supra-legal document thus promotes the elitist status quo and undermines political equality.

The protection of vulnerable minorities is a key aspect of this debate. This paper enters the debate from one, undertheorised perspective: how democracy and constitutions protect disenfranchised minorities, specifically, resident non-citizens or ‘denizens’. Denizens raise a moral problem as they are subject to the coercive apparatus of the state without political rights; they are subjects rather than citizens. Prima facie, legal constitutionalism fares better as it seeks to step in where democracy fails, by developing systems to protect minorities against legislation which violates their rights. But from a political constitutionalist perspective, the identification of a rights-holder is an inherently political decision, and states lack the incentive structure to promote the rights of non-citizens. This article will critically assess the liberal constitutionalist/political constitutionalist or henceforth ‘liberal-rights’/‘democratic-republican’1 debate in the context of contemporary mass transnational migration and the associated large numbers of resident non-citizens, many of whom enjoy most of the legal rights of citizenship, but lack political rights.

Ultimately, I argue that both approaches are flawed in their response to denizenship. Liberal-rights theorists overstate the extent to which a politically powerless status can secure individual rights. Despite being a state of disenfranchisement, they see denizenship as unproblematic—even when, I argue, it is problematic, such as when long-term residents (or those who have high exit costs of leaving the country) lack political rights and are liable to deportation. Democratic-republicans, on the other hand, idealise the political process and wrongly assume that all those affected by laws will be eligible to participate in their determination. Because it is a politically excluded status, democratic-republicans see denizenship as inherently problematic—even when, I argue, it is not, when migrants have low exit costs of leaving because of their opportunities, lives or interests elsewhere, or when they are a citizen of a powerful home state ready to hold the host state accountable for rights violations. But the democratic-republican intuition—that denizenship is a status of domination—is worth developing. This paper therefore sets out a more nuanced concept of domination by incorporating the concepts of exit costs of leaving the state, and the accountability of governments to their non-citizen population. This model allows us to evaluate the status of denizens vis-à-vis one another, and determine the strength of their moral entitlements on the state. The main contribution made by this model is to demonstrate that the costs of exit from the state are key to assessing the vulnerability of denizens to state power. In future work I will develop this model and consider what it suggests in the way of immigration policies.

The disagreement between the two approaches I examine is not fundamentally about the moral rights of individuals, but about the institutional mechanisms that would best protect them. However, my approach in this paper is to scrutinise the coherence of these theories with our intuitions about the moral entitlements of denizens. In the case of denizens, the pragmatic question of how best to protect rights, and the moral question of what these rights ought to be, are interlinked, especially in relation to political participation. To clarify the distinction in my discussion, I will specify when I am referring to moral rights and entitlements. The reader can therefore otherwise assume that I am discussing legal rights.

The paper is split into four sections. The first section sets out the ‘problem of denizenship’ and the debate about the relationship between rights and citizenship. In the second part, I examine the liberal-rights response. I then critically assess the democratic-republican approach. Finally, I outline an alternative approach to denizenship, informed by the republican ideal of non-domination. I also identify avenues for future research.

The Problem of ‘Denizenship’ and the Relationship Between Rights and Citizenship

Transnational migration is a key challenge of our time, and one rarely off political agendas. But in political philosophy immigration is undertheorised; cross-border movements have been the object of ‘scant’ attention (Benhabib 2004, p. xiii). To the extent that it has preoccupied philosophers, immigration has been theorised according to three main approaches: the first considers legitimate border control and whether there is a ‘good case’ for open borders (e.g. Carens 1995), the second considers fair access to citizenship (e.g. Seglow 2009), and the third considers integration, multiculturalism and cultural rights (e.g. Kymlicka 2001).

The status of being a non-citizen, or ‘denizenship’, has been less thoroughly examined. When it is, it is usually as a matter of the rights enjoyed by non-citizens, which makes sense given the historical prevalence of the view that citizenship was the precondition for rights. Sociological accounts of the development of rights in liberal democracies interpret the history of rights as a three-step process whereby citizens were accorded civil, then political, then social rights (Marshall 1964). A stronger statement of this view is Arendt’s (1967) description of statelessness, an extreme version of denizenship,2 as a status of rightlessness. Arendt argued that to lack citizenship was entirely to lack status as a rights-holder. The denationalisation of minorities between the two world wars, for Arendt, deprived them not just of access to a particular set of rights x, y and z, but of their very right to recognition as a rights-holder, the right to have such rights at all.

The rejection of citizenship as a central criterion for rights-holder status was a central objective of the post-WWII international commitment to human rights. Various supranational rights declarations over the last 60 years have codified duties of states to recognise the rights of all subject to their jurisdiction, whether citizens, legal residents, or even undocumented migrants, and human rights instruments have focused attention on vulnerable minorities. In the early 1990s, Soysal (1994) found that many of the rights of citizenship—including access to public education, welfare and benefits, health care, and even in some cases the right not to lose contributions to national insurance or pensions schemes on moving abroad—were now enjoyed by denizens. In fact, the term ‘denizen’ was introduced around the same time to highlight the diminishing significance of citizenship as a precondition for rights-holder status, by calling to mind half-citizenship status (Hammar 1990).3

Soysal’s interpretation of this development is that ‘universal personhood’, rather than citizenship status, has become the main criterion on which rights are allocated. She presents denizens as evidence of the decoupling of rights from formal citizenship and hence of a triumph of the human rights movement. Others have read the extension of rights to denizens as a more flexible understanding of membership, where certain residents are treated like formal citizens. Bosniak (2006, p. 3) observes that we should attend to the ‘citizenship of non-citizens’, while Cohen (2009) introduces the notion of ‘semi-citizenship’ to describe how denizens and subgroups of denizens have become partially included. For Joppke (2001), long-term immigrants have become approximated to citizens as a result of domestic legal orders rather than inter- and supra-national human rights regimes.

Less optimistic studies of denizenship have emphasised the persistence of gap between citizen and denizen rights, in quality if not in quantity. There are two reasons why the story of denizens’ rights may not be a wholly positive one. First is the fact that denizens’ rights are not ‘settled’ but are often criticised, discussed, debated or altered in the course of public debate in a way that citizens’ rights are not. Morris (2002, p. 145) describes this as the ‘expansion and contradiction of rights over time’. Second is the function of what I call ‘meta-rights’. One of these, highlighted most extensively in Bosniak’s (2006, 2010) work, is the right not to be deported. Bosniak contends that enjoyment of other rights is conditioned by the absence of this right. Although most pronounced in the case of irregular migrants, liability to deportation means that ‘alienage’ is the overarching status that shapes non-citizens’ lives.

Another meta-right denizens lack is to political participation. Republicans like Bellamy (2001, 2007) invoke Arendt’s description of citizenship as the ‘right to have rights’ to argue that the right to political participation is more significant, valuable, and crucial to self-worth and independence than any other rights. Or, as Waldron (1993, p. 38) puts it, the right to participate is not one right amongst many, but ‘our most cherished principle’. Commitment to a similar view of the primacy of political rights led Walzer (1983, p. 59) to describe the government of denizens as ‘tyranny’.

There are therefore two divergent narratives on denizens’ rights. For the remainder of this paper, I would like to propose the metaphor of ‘rights-constituencies’ as a lens through which to explore this debate, and the geopolitical dimension of the relationship between rights and citizenship. To be in a rights-constituency, as for an electoral constituency, is to be eligible to the entitlements connected with the constituency. The metaphor captures the way in which the boundaries between who is and who is not eligible for certain legal rights can be negotiated, contested and redrawn. Redrawing can be the result of appeals by members and non-members, as in the civil rights struggles in the US, or it can be the result of reconfigurations for the political gain of powerful actors, as in the exclusion of unpopular groups such as undocumented migrants from health or education rights. As for electoral constituencies, we can describe this as a process of ‘gerrymandering’: manipulating boundaries to favour or disfavour certain groups to achieve the results sought by the incumbent administration.

Returning to Arendt, we can see how the geopolitical dimension of legal rights lends itself to the constituency analogy. Individuals can be forcibly removed from rights constituencies by depriving them of the condition of eligibility—citizenship in this case, or through deportation (which more vividly evokes the notion that eligibility has to do with geopolitical space). Furthermore, the constituency metaphor illuminates the different moral readings of the ‘right to have rights’. On one interpretation, the ‘right to have rights’ is a plea for humanity to be the only rights constituency: for the constituency to be expanded from citizen to human through the codification of universal moral rights. On another reading, humanity is a rights constituency only for the right to be included in at least one fundamental constituency, in other words, the right to citizenship (Benhabib 2004). For those who, like Soysal, emphasise the pervasiveness of human rights norms and their permeation into national judicial settings, the fundamental change since Arendt is that rights are increasingly accorded on a constituency-less basis, or with the only constituency being that of humankind. Liberal-rights theorists are committed to this elimination of rights constituencies as a moral ideal. In an ideal world, all rights would be accorded to everyone, wherever they are. For democratic-republicans on the other hand, rights are intrinsically connected to political rights and membership because democracy is necessary to resolve conflict and disagreement over the form and content of rights. Only political rights allow individuals to contest the rights they are accorded in a way that avoids paternalism or domination. This means that denizens—who lack political rights—are necessarily in the ‘grey area’ of rights constituencies. I explore these two different approaches further in the next two sections, turning to the liberal-rights approach first.

The Liberal-Rights Approach

The three central commitments of the liberal-rights approach to minority rights lend themselves to a theory of disenfranchised minority, or denizen, rights. First, the liberal-rights approach aspires towards perfect continuity between human, legal and constitutional rights, and subscribes to the principle that if rights exist they ought to be codified. All moral rights should be legal, if not constitutional, rights (for an opposing view, see Waldron 1999). From this perspective, denizenship is problematic if it constitutes an inadequate codification of the moral rights of the individual. Second, minorities ought to be protected against the policies and legislation enacted by popular democracy. In Ronald Dworkin’s (1977) influential account, rights are ‘trumps’ on the utility- and efficiency-based decision-making of everyday politics. The moral rights we hold against governments should be enshrined in a constitution in order to prevent administrations from ‘enacting laws or adopting policies that would otherwise seem attractive’ (Dworkin 2006, p. 30), and guarded by trustees insulated from political pressure—the judiciary. This applies to disenfranchised minorities in the same way as to ‘classic’ minorities because the moral rights we hold against governments are human rights. In recent work, Dworkin (2002, 2006) has made clear that these are not just constraints on ordinary policy-making, but also on the extent to which states can further the interests of their own nationals. Thus rights are not only trumps against ordinary policy-making but against the sort of policy-making that promotes national interest above all else.

The broader liberal constitutionalist conception of democracy also lends itself to reinterpretation in response to the existence of denizens and, accordingly, a larger populace than electorate. As Dworkin (1996, 2006) posits, a legitimate democracy is a matter of outcomes rather than procedures—its primary goal is to treat its members with equal concern and respect rather than to facilitate their participation in the political process. This is because he believes that political equality can have little meaning in large democracies: equality of political impact is impossible given the asymmetry of power between representatives and the electorate, and equality of influence is undesirable in a democracy which values a free press and rigorous public argument. Importance is placed on political participation for its symbolic, agency and communal benefits rather than its fundamental role in protecting individual freedom. It is symbolic in that by including all citizens they are affirmed and recognised as equal members; it facilitates agency by connecting participation with each individual’s moral experience; and it is communal by involving individuals in the collective enterprise of shaping a community. The implication is that resident non-citizens can benefit from living in a democracy even if they lack the right to participate in democratic processes as political rights have no instrumental function in protecting rights.

Legal constitutionalism therefore underpins a coherent approach to denizenship. In fact, denizens have been presented by some commentators as the ultimate test case for this model. Joppke observes, ‘[i]f one defines individuals’ rights as ‘trumps’ over the preferences of the government-represented majority in society, one could argue that immigrants—by definition excluded from this majority—are the most dramatic test case of rights in general’ (Joppke 2001, p. 55). Debates about the desirability of the Human Rights Act in Britain in the media—such as in the context of the debate about replacing it with a ‘British Bill of Rights’—often centre around the protection of non-citizens. Academic studies of the Act also suggest that immigrants have been among the most significant beneficiaries. Kavanagh (2009), for example, contends that most of the individuals who have benefited are either not entitled to vote or members of groups who tend to be marginalised from the political process. In evidence to the Joint Committee on Human Rights, the pressure group Liberty claimed that it is ‘non-citizens that are most often in needs of human rights protection’ while the Committee itself maintained that ‘Bills of Rights protect rights which people have by virtue of being human, not according to their legal status as citizen or non-citizen’ (JCHR 2008). For Clapham, the absence of ‘power through the ballot box’ is the raison d’être of supra-legislative norms in new European rights legislation (Clapham 1999, p. 131).

If immigrants are the paradigm subjects for testing the ‘rights as trumps’ model, the United States is surely the paradigm case scenario. Its constitution has been described as one where ‘the concept of citizenship matters very little’ and which ‘prescribes decencies and wise modalities of government quite without regard to the concept of citizenship’ (Bickel 1975, pp. 53–54). The constitutional understanding of the person (rather than citizen) as rights-holder has been borne out in landmark cases like Plyer v. Doe, where the Supreme Court rejected Texas’s claim that undocumented migrant children were ineligible for education and reaffirmed that even illegal non-citizens ‘have long been recognized as ‘persons’’.4

However, even in the archetypal constitutional democracy which protects the rights of persons qua persons, eligibility is not straightforward. Bosniak explains that the status of personhood is ‘evaded and constrained’ through the plenary principle of immigration law, for example in denying judicial standing in the form of the right to sue to unauthorised migrants (Bosniak 2010, p. 16). Even where the Supreme Court eventually rules in favour of undocumented migrants, as in California’s Proposition 187 which withdrew undocumented migrants’ access to public services, the fact that their rights are subject to removal by state legislation calls into question the successful decoupling of rights from citizenship. Specifically, to return to the metaphor I introduced earlier, it suggests that the legal rights of migrants are often in the ‘grey area’ of a rights constituency—liable to be debated and contested.

The rights constituency metaphor also provides a new way to interpret the widely-discussed affronts to human rights in the post-9/11 counter-terrorism measures in the United States and Guantánamo Bay. Seen from this perspective, these moves are not just a grave abuse of human rights or civil liberties but an illustration of the deeper rights vulnerability of denizens. The Bush administration did not just violate rights or bend the rules, but attempted to reinterpret or redraw rights constituencies to modify the status of denizens as rights-holders. For example, the descriptions of Guantánamo as a ‘state of exception’ (Agamben 2005) and as a ‘deliberate series of legal and geographical contradictions’ designed to avoid legal restraints (Comaroff 2007) show how territory and legal loopholes were employed to control rights eligibility. Guantánamo was made legally possible on the basis of the fact that suspects were ‘unlawful combatants’ (Seelye 2002), a term constructed to circumvent protections for prisoners of war in the Geneva Convention and the normal legal process. The differential treatment accorded non-citizens by the Patriot Act was justified by the fact that foreign policy and immigration were subject to executive, not judicial control. Similarly, the practice of extraordinary rendition can be seen as an attempt to exploit geographical space for political gain.

The rights constituency metaphor also provides a new way to interpret other means by which individuals can be deprived of their rights. Deportation is often used to remove unpopular residents such as those who have overstayed their visas or entered the country without documentation, as well as any non-citizen—regardless of the time they have spent in the country—convicted of violent or drug crime. Detention of undocumented migrants, victims of trafficking and asylum seekers has also vastly increased post 9/11 (Weissbrodt 2008). All these policies can be interpreted in rights constituencies terms; as removing people from rights constituencies, creating an exception to a rights constituency or redrawing a rights constituency.

Of course, the liberal-rights position is that these are rights violations: Dworkin (2002) has explicitly criticised the US Patriot Act and denounced its violation of the moral principle to treat individuals with equal concern and respect. However, my concern is not that loopholes are sanctioned or facilitated by the liberal-rights approach, but rather that the ideal of the perfect legalisation of universal rights is impossible in a world of unequal nation states with sovereign power over their borders. The liberal-rights approach pays insufficient attention to this in two ways. First it downplays the extent to which constitutional rights debates are about eligibility as a rights-holder and the associated importance of status on rights, which will be the subject of the next section (the rights of migrants are notably absent from the key constitutional debates with which Dworkin (1996) introduces Freedom’s Law). Second, it ignores the role of immigration policy and border control.

The policing of borders is widely believed to legitimise the power of the state over deportation and detention for those who lack legal permission to live in the state. I indicated above that deportation, like extraordinary rendition, is problematic in rights constituencies terms as it demonstrates that even ‘universal’ rights are dependent on continued residence in the rights-according jurisdiction. But the power of deportation also casts a wider shadow over rights protection. This is that the mere threat of deportation impedes the effective ability of certain immigrants to claim their rights. At the extreme, undocumented migrants are unlikely to report crimes to the authorities or seek civil redress for maltreatment for fear of deportation (Bosniak 2010). To a lesser extent, the threat of deportation hangs over other groups of vulnerable migrants. Studies have found that some migrants lack effective awareness of their employment rights and consider they have no option but to remain in an exploitative job because they feel disinclined to draw attention to themselves and think their continued residence in the country is conditional (EHRC 2010). The inflammatory media reaction to high profile cases of non-deportation, such as where a non-citizen is not deported because of the principle of non-refoulement, must compound the sense of discretionary, insecure status some immigrants feel. To the extent that this threat of deportation is perceived, even migrants who are not at risk of deportation may be susceptible to a chilling effect in relation to their rights.

It is still open to the liberal-rights theorist to maintain that the absence or ineffectiveness of rights is a practical, rather than philosophical problem, or to advocate a right against deportation as a way of avoiding the problems with the legalisation of rights. But this raises a different question the liberal-rights approach is ill-equipped to answer: when and under what conditions does a denizen have a claim to a right against deportation? Few immigration theorists would argue for a right against deportation for foreign students or migrant workers on short-term visas. For example Carens (2005, 2008) argues that the package of legal rights denizens are entitled to increases with period of residence, and that after a certain period of time they should be granted all the rights of citizenship including the right against deportation. Temporary residents on the other hand may reasonably be required to leave after a certain period, and possibly denied access to certain social programmes. In the final section of this paper, I will argue that it is not just the passage of time that makes a difference to a claim against deportation, but any other factors which contribute to denizens’ exit costs of leaving the state.

A similar problem arises in relation to another meta-right: the right to political participation. In Dworkin’s theory, rights are a constraint on the extent to which governments can prioritise their own citizens in policy-making, regardless of their citizenship status. Nonetheless, it is still justifiable for a state to prioritise its own citizens so long as they do not violate human rights (Dworkin 1986, 2000, 2006). This permits states considerable latitude to enact policies which disfavour migrants. In my view, although constitutional rights may be an appropriate solution to the political exclusion of denizens who have newly arrived, after a certain amount of time their political exclusion is troubling. But for Dworkin, the lack of emphasis on actual political influence is justified on the basis that a democracy is more about the outcome-based ideal of treating people as equals rather than the procedural ideal of treating them as political equals. However, if states can legitimately prioritise citizens then this argument fails, as denizens have neither political equality nor are they relevant moral units for this ‘outcome equality’. Again, the question of when this political exclusion becomes problematic is unanswerable in liberal-rights terms because of the commitment to the view that political rights have no instrumental role in protecting individual rights.

This is the central aspect of the liberal-rights view that democratic-republicans challenge: the role of political power in the realisation of rights. They argue that legal rights are unavoidably politicised as all rights–even negative ones–require decisions about resource allocation. On this view, the politically powerless in society—in this case, denizens—are at a structural disadvantage as governments lack the incentive to promote their rights in the policy-making process. I will turn to this argument next, in the context of the democratic-republican theory of rights.

The Democratic-Republican Approach

The democratic-republican critique of the liberal-rights model is twofold. The first dimension is the argument that the politicisation and contestation of rights is inevitable. This critique is relevant to the case study of denizens because it implies that the liberal-rights theorist’s desire to see humanity alone as the ‘rights constituency’ is a chimera. The second point is that without entitlement to political participation there is a certain vulnerability inherent to rights-holder status. To return to the rights constituency analogy again, the democratic-republican view is that the ‘grey area’ of a rights constituency is not just a matter of imperfectly realised rights frameworks but a necessary result of lacking the ‘right to have rights’.

As citizenship is conceptualised as the ‘right to have rights’, to lack it is to occupy, by definition, a status of rights vulnerability. Democratic-republicans like Richard Bellamy (2001, 2007) adopt Arendt’s phrase to argue that the right to political participation is more significant, valuable and crucial to self-worth and independence than any other rights. It is what I have termed a ‘meta-right’. Bellamy contends that the struggles of women and others for inclusion were not demands for ‘access to a pre-constituted set of political rights’ but for the right to change the ‘terms and conditions’ of citizenship (Bellamy 2001, p. 38). Hence what is important is not second-order rights provision but the primary right to share in their determination; without the right to be a rights-holder, having a set of rights x, y and z is hollow. Merely being the passive recipient of a set of legal rights is, for Bellamy, akin to rights granted to children or slaves, because rights enjoyed at other’s discretion depends on currying favour with the powerful: this is ‘domination’, in the republican terminology. Denizens, as subjects but not citizens, should therefore be troubling for republicans who see subjects as ‘subordinate to their rulers, depending for their rights on finding favour and influence with those in power’ (Bellamy 2008, p. 600).

The second insight the democratic-republican account provides also pertains to the political exclusion of denizens. This is that political power shapes the way in which rights are determined and weighed against one another. Bellamy contends that due to the costliness of rights provision, even ‘negative’ rights like habeas corpus demand financial resources. Rights cannot be detached from normal political decision-making as they need to be weighed up against each other. They are costly, collective political decisions thus we cannot just ‘add on’ rights indefinitely to achieve a maximal package. If policy-makers are unaccountable to a subset of rights-holders it follows that the determination of rights will not be in their favour.

The empirical observation made by democratic-republicans, that political power shapes the enjoyment of rights, is therefore sensitive to the importance of meta-rights in a way the liberal-rights account is not. However, the normative claim that rights ought not to be isolated from democratic processes actually disadvantages denizens. Democratic-republicans idealise the extent to which interests are represented in the political process and have no contingency plan for politically powerless groups. Because they assume that the demos and populace are one and the same, they simply do not consider the existence of disenfranchised minorities. This is particularly problematic if the promotion of denizens’ interests is in fact contingent on rights being negotiated away from representative politics. This has been suggested by Guiraudon (1998, p. 303) who, in an analysis of reforms of immigrant rights in France, Germany and the Netherlands, argues that it is precisely the depoliticisation of rights which has benefited migrants: ‘containing debate behind closed doors (the doors of ministries or of courtrooms)’.

The irony is that although the democratic-republican approach is more adequately equipped to identify the problem of denizenship, it is poorly equipped to address it. It identifies the problem clearly: denizens are definitionally dominated in the republican tradition as they are not citizens, and citizenship is necessary in order to be non-dominated. However, because citizenship is necessary in order to be non-dominated, and rights, participation and citizenship cannot be disaggregated, a solution is impossible. This is especially clear given that democratic-republicans tend to reject the normative goal of disaggregating legal rights and citizenship, and in particular reject the view that political rights can be accorded independently of citizenship (Schuck 1989; Jacobson 1996).

There is therefore a tension in the democratic-republican response to the decoupling of rights and citizenship. On the one hand it identifies vulnerability in denizens’ rights status, but on the other it criticises the expansion of rights to denizens as this is undemocratic. I would like to suggest that the paradox at the heart of the democratic-republican response to the problem of denizenship is the result of an insufficiently developed conception of non-domination. The apparent paradox is the result of a circular definition—by identifying non-domination with democracy, we are left with the tautological conclusion that lack of political rights is equivalent to domination. In the next section I will argue that non-domination is indeed the correct principle to underpin a democratic theory of denizens, but that we should develop a more nuanced conception of domination.

Domination and Denizens–A Framework

I argued in the previous section that non-domination is often identified with and described as equivalent to citizenship, exemplified in the classic republican adage ‘to be free is to be a citizen of a free state’. Similarly, some contemporary conceptions of domination identify it with barriers to political participation (Young 1990; Bohman 2007). These theories are problematised by circularity; as Frank Lovett (2010, p. 19) observes, of course non-domination requires democracy if domination is the absence of it. This problem is amplified in the case of denizens–by definition those without political rights. An argument that denizens are dominated because they lack political rights is tautological.

An alternative, more prominent theory is of domination as arbitrary power (Pettit 1997; Skinner 1997). Such a view has achieved popularity as the linchpin in the neo-Roman republican school of thought, and as a third conception of liberty to rival the classic ‘positive’ and ‘negative’ versions. Its advocates claim that freedom as non-domination is clearly distinguishable from the more common concept of freedom as non-interference by the different situations it identifies as freedom-restricting. Domination theorists contend that the effects of domination such as restricted sphere of choice, loss of self-respect and dignity, deferential behaviour, anxiety and insecurity5 can occur in the absence of interference, therefore they look behind interference to consider the impact of power relations. We should worry if power subjects modify their behaviour because of the power they are exposed to, regardless of whether any interference has occurred. We should also acknowledge that not all interference is necessarily freedom-curbing, as in the case of a well-propagated, non-arbitrary law.

Arbitrariness is therefore of central importance for republican freedom, however the concept of arbitrariness has attracted criticism. Philip Pettit’s (1997) definition of arbitrary power as power which does not ‘track the interests’ of those subject to the power has been criticised for importing a loaded conception of legitimate interests into the theory of freedom (McMahon 2005; Carter 2008). My definition of arbitrariness attempts to avoid these problems by distinguishing more clearly between the procedural question of what makes a power relationship unaccountable (that it is subject to inadequate checks) and the substantive principle of what we evaluate these checks against (the principle of tracked interests). In my adapted definition, someone is vulnerable to domination to the extent that they are dependent on a relationship where the power-holder can exercise arbitrary power over them, power that is unaccountable. Like Pettit, I argue that we cannot adopt a pure procedural conception of arbitrariness, like the one recently expounded by Lovett (2010), because this would commit us to the view that power should be constrained wherever it exists, even in the case of positive relationships like that of teachers and students, or parents and children. We might wish to argue that parents should be constrained in the exercise of their power over their children by outlawing abuse, violence and neglect, but we need another conception of ‘wrong’ to identify why these are the sorts of exercises of power that we wish to prevent. The ‘substantive’ principle of tracked interests provides this.

According to my theory of domination, accountability gaps exist where there are insufficient checking mechanisms which require that state power is justified to all subjects of the power (justification), that subjects of the power are able to complain if their interests are not being met (contestation), and that power-holders are duly punished if they fail to track the interests of power-subjects (retribution). So an accountability gap exists, for example, where governments do not have to give reasons for their decisions to the non-citizen populace in the same way as the citizen electorate and they are not forced to be responsive to their interests through the democratic process. But indirect ways in which the interests are tracked also contribute to the accountability gap—making it larger for some denizens than others. Denizens who retain the right to vote in their home countries, and whose countries have significant sway on the international stage, have considerable diplomatic protection. It was the detention of nationals from liberal democracies that constituted the most considerable pressure on the US to free Guantánamo detainees, for example. States are also accountable to any transnational regimes they are part of, in the form of treaties and agreements, memberships of transnational organisations and alliances, or supranational governance structures like the EU. If a member state maltreated its EU denizen population it would evoke sanctions, thus membership in the EU reduces the degree to which denizens from member states are subject to unaccountable power. This broader conception of accountability, taking into account the rights denizens have in their home countries, has much in common with Bauböck’s (2009) concept of ‘external citizenship’. We cannot understand the whole picture about the accountability of states to their non-citizen population without knowing where else denizens hold citizenship.

But whether these accountability gaps are problematic, is, on my model, a matter of the exit costs of leaving the state—dependence. The importance of dependence to a theory of domination is convincingly demonstrated by Lovett (2010, p. 39). Lovett explains that dependency, defined as the ‘degree to which a person or group’s continued membership in some social relationship is not voluntary’, explains why someone does not merely leave the site of their domination. Although some relationships such as families derive their value from dependency, Lovett maintains that reducing dependency will always reduce domination. This departs from Pettit’s use of the concept of ‘subjection’ to power alone.

Lovett’s conception of exit costs can be employed to make a distinct contribution to our understanding of denizenship. What has thus far not been acknowledged is that dependence on the state has a clear bearing on the extent to which political power is problematic. The lack of accountability of states to visitors and transients for example seems acceptable, but an accountability gap in relation to citizens would not be—because they have no other option but to be subject to this power as leaving the state is not a viable alternative. Similarly, if the exit costs for denizens of leaving the state are high, then the accountability gap is cause for concern. On one side of the spectrum, refugees are clearly completely unable to leave the country as they cannot return to their home country for fear of persecution, nor do they have the right to claim asylum in a new country. Migrants who face destitution or other dire circumstances if they must return home have similarly high exit costs. On the other side of the spectrum, members of the cosmopolitan elite, employees of multinational firms, and certain groups of students are actively recruited by many governments to the extent that they hold considerable power in the form of the threat to ‘take their business elsewhere’. But as time passes the exit costs of all denizens increase to the extent that after a certain period of time all denizens have significantly high exit costs.

The concept of exit costs of leaving the state complements key principles emerging in the immigration theory literature. It is compatible with Carens’ (2005) intuition about the correlation between length of residence and the degree of harm that would be caused by deportation, but develops it in what is hopefully a useful and illuminating way. It also builds on theories of the coercive power of the state such as that of Blake (2001). Blake argues that certain citizen rights such as the right to political participation stem from the need for justification of state coercion and the fact that citizens have not voluntarily submitted to this power. My theory suggests that immigrants may submit to it, but if their exit costs of leaving are substantial, the legitimising power of this consent is nullified. Exit costs can tell us that when someone has no choice but to be subject to coercion, democratic justification is particularly necessary.6

It is my hope that this model of domination will be able to underpin policies to improve the position of denizens, and I will undertake future work in this vein. However, the case for promoting the non-domination of denizens remains to be made. Although denizens are in one respect deeply troubling for republicans, they are also outside of their remit, as non-domination is generally thought of as the preserve of the citizenry.7 Therefore, even though republicans might agree that I have shown that denizens are vulnerable to domination, they might reject the view that the reduction of the non-domination of denizens is a legitimate or desirable goal.

There are three compelling reasons why it is. First, republicans might be troubled by Goodin’s description of non-domination as a status concept that harks back to a ‘status society of a strikingly premodern form’ (Goodin 2003, p. 61). Goodin presents would-be republicans with a choice: either revise republicanism so that it does not rest on this status concept, or reject it altogether, his preferred option. But the other choice is still open to us: if it can be shown that denizenship is compatible with non-domination, republicanism will have passed the ultimate test. This is not a straightforward argument for promoting the non-domination of denizens. It is a conditional argument that will only be convincing for those who are already sympathetic to republican views. The point is that if republicanism is to survive the challenge of commentators like Goodin, it needs to demonstrate that non-domination is not a privilege or honour.

The second reason concerns how meaningful the concept of non-domination would be if applied just to citizens. Non-domination is predominantly about power, and about making the power individuals are subject to safe. State power fundamentally affects all residents, not a smaller group of citizens. It follows that we should promote the non-domination of all those affected by state power. Of course there might be a case for promoting the non-domination of non-citizens outside the state who are affected by its power, but that does not defeat the point that reducing the negative impact of state power requires at the least promoting of the non-domination of denizens.

But the most convincing argument for promoting the non-domination of denizens is that it may not be as burdensome in practice as it seems in theory. Much of what is necessary to ensure the non-domination of denizens will also be necessary to secure the non-domination of citizens, such as Fuller’s (1969) morality of law requirements of generality, transparency, nonretroactivity, coherence and so on. More crucially, non-domination does not involve allocating all recipients the same goods, just as ‘treatment as an equal’ does not requiring ‘equal treatment’ (Dworkin 1977, p. 227).8 The model of domination as dependence on unaccountable power means there will be substantial variation in the dependence level of migrants, and hence different demands on the state. For example, the problem of unresponsive democratic institutions, whilst deeply troubling for migrants who have lived in the country for decades or who have very high exit costs of leaving for any other reason, is of less concern for new arrivals with a passport from a liberal democracy. Therefore, it is not that short-term residents are entitled to a more minimal form of non-domination, but that what non-domination requires is less demanding.

Although this model is still fairly skeletal, we can see how its implications are more nuanced than both the liberal-rights approach and the democratic-republican approach outlined above. It suggests that denizens are not definitionally dominated, as the democratic-republican approach would imply. Nor are they adequately protected by rights mechanisms, as governments are unlikely to be resiliently accountable to a population outside the demos because such a group cannot vote them out or appeal their decisions through democratic procedures. However, for denizens who have low exit costs of leaving the state, this ‘accountability gap’ is unproblematic, as they can merely move elsewhere. The framework therefore does not quite imply a ‘tyranny of the enfranchised majority’, but it does show that we ought to look beyond the citizenry to evaluate theories of constitutionalism. It identifies specific groups who need protection—those with high exit costs and for whom indirect accountability mechanisms are ineffective. Further work will develop the concept of domination outlined here and set out policies for reducing the domination of denizens. It will also explore the question of private domination which has not been examined here. It seems likely that the higher an individual’s exit costs of leaving the state, the more susceptible they are to private forms of domination. If someone is terrified of deportation, and their right to remain in the country is conditional on staying with a particular employer, this gives the employer considerable power. The interplay between dependence on the state, and private domination, is therefore an important avenue of research.


I depart from the more common term ‘legal constitutionalism’ in order to explore the wider human rights or sociology of immigration approach which sees immigrants as transnational rights holders, rather than confine myself to debates in democratic theory. The two positions are not identical, but are compatible in their fundamental principles, and together they constitute the dominant approach to the study of denizens. My use of ‘democratic-republican’ rather than ‘political constitutionalist’ is in a sense narrower – I wish to focus on the arguments of republicans who see political participation as fundamental to non-domination rather than on the wider category of critics of judicial review including republicans, political constitutionalists, and democratic liberals.


Statelessness is the state of lacking citizenship anywhere whereas denizenship is the state of lacking citizenship in one’s state of residence.


While Hammar employs the term to refer to long-term, legal, permanent residents, I use it to refer to all resident noncitizens–for brevity, because I am exploring the idea that noncitizenship is a type of status, and because I dislike the connotations of the term ‘alien’.


Plyer v Doe, 457 U.S. 202 (1982).


This list is adapted from Pettit (1997).


For a similar argument see Shapiro (2003, p. 45).


Or at least, it used to be. For example Pettit describes his early historical work as centred on republican freedom understood ‘as equivalent to citizenship in a republic’ (Pettit 1997, p.vii).


Pettit also makes this connection between the ‘egalitarian commitment’ of non-domination and Dworkin’s principle that people should be treated as equals (Pettit 1997, p. 111).



I received very useful feedback from the audience of the York Political Philosophy Postgraduate conference, where an earlier draft of this article was presented. I am also grateful for comments from Cécile Laborde, Richard Bellamy, Katerina Mantouvalou and Jack Simson Caird. The questions raised by two anonymous reviewers were of particular help. The research for this article was undertaken as part of PhD research funded by the Arts and Humanities Research Council.

Copyright information

© Springer Science+Business Media B.V. 2010