Sanctuary as democratic non-cooperation

Across North America, Europe and Latin America, multiple sub-state jurisdictions have declared themselves to be migrant “sanctuaries” (Lenard and Madokoro 2021). By adopting sanctuary status, typically, sub-state jurisdictions intend both to signal their welcoming attitude towards migrants as well as, often, their opposition to the state-level policies that target them for exclusion. As a matter of policy, declaring a space to be one of sanctuary generally means one of the following types of policies have been adopted at the sub-state jurisdictional level: “don't ask, don't tell” policies, “non-discrimination” policies, and “non-cooperation” policies.

“Don't ask, don't tell” policies direct municipal (or other sub-state jurisdictional) service providers to offer their services to all residents of a jurisdiction, without inquiring into, or collecting information about, their immigration status. “Non-discrimination” policies often operate alongside “don't ask, don't tell” policies, and announce a jurisdictional-level commitment to avoiding racial and ethnic discrimination in the distribution of goods and services across the territory. “Non-cooperation” policies direct municipal (or other sub-state jurisdictional) service providers to avoid or refuse to cooperate with immigration enforcement agents in a range of ways, for example by refusing to detain people suspected of immigration violations beyond certain time limits. Usually, these policies are directed at authorities responsible for protecting the safety and security of residents, including local police forces and firefighters, where their effectiveness depends on all residents’ willing cooperation: local police must be able to count on all residents for their cooperation, and they cannot do so if residents fear that interacting with police will bring them to the attention of immigration enforcement agents (Parekh and Davis 2018).

It is typical among those who work with, and support, migrants to believe that justice allows, or even favors, these policies. They are often responding to the United States, both at the border with Mexico and internally, as well as to the policing of the Mediterranean Sea, where the pursuit of harsh border enforcement measures has produced extraordinary suffering and, often, needless death. This article also assumes that these enforcement policies are generally unjustified, in addition to being cruel, and instead examines the morality of the multiple ways in which sub-state jurisdictional opposition manifests, as evasion, non-engagement, disruption and obstruction, in the face of national-level immigration enforcement policies.

This article begins with a brief background to sanctuary policies, to set the stage. I then offer a brief account of resistance in the political theory of migration, a literature that largely treats resistance to immigration enforcement in terms set by John Rawls’ early and important account of civil disobedience. However, I suggest, treating immigration resistance as mainly “civil disobedience” does not capture the full range of strategies taken by those who oppose immigration injustice: in particular, when sub-state jurisdictions in non-unitary states adopt sanctuary policies, it is not clear that they merit being described as civil disobedience. Resistance (as in the case of civil disobedience), as I shall go on to articulate, is typically understood to be unlawful in some way; yet, while oppositional, the actions taken by sanctuary jurisdictions are not obviously unlawful. On the contrary, they are quite often defended by those who undertake them as consistent with the plain meaning, or at least in the spirit, of the law.

In the next section, I examine the morality of sanctuary in the context of the assumption that, in federal states, state and sub-state governments, must cooperate at least to some degree as a condition of effective democratic governance. The sanctuary case permits us to examine the morality of the actions taken by sub-state jurisdictions who aim to oppose, rather than cooperate wholeheartedly with, centrally-developed policy. I locate this examination in the political theory of federalism, to identify an expectation of cooperation – which, it appears, sanctuary jurisdictions are refusing to accept, usually with respect to immigration enforcement efforts. I refer to this form of opposition as “democratic non-cooperation” and identify its key features. I describe a “cooperation continuum”, to suggest that non-cooperation takes four main forms – evasion, non-engagement, disruption and obstruction – which I describe both in general terms and in relation to sanctuary practices in particular. Finally, I observe that the form of opposition that sanctuary is, is not limited to sanctuary: that is, there are other cases of this form of opposition in other policy domains, and moreover, not all of the objectives taken by those who deploy this form of opposition are progressive. Ultimately, this article's central contribution is to fleshing out modes of opposition in democratic spaces in general, by examining the morality of sanctuary actions taken around the world.

Sanctuary in context

At the level of normative theorizing, it is important not to conflate deliberations about the justice of immigration policy with the justice of the enforcement of immigration policy; while there are important connections between a particular policy and its enforcement, they are not subject to the same set of moral considerations. One might believe that a particular immigration law is fair or just, but that some mechanisms of enforcing it are not. For example, one might believe that a state can deny entry to people who arrive at its borders, but not by permitting them to (or making it more likely that they will) drown at sea. Or, one might believe that the United States has the right to exclude migrants from Central America who have been residing irregularly on US territory, while believing as well that, in pursuit of this policy, the United States may not separate parents from children who are taken to detention centers prior to their deportation. It may be that there are immigration policies that simply cannot be enforced in a fair way, and where this is the case, there may be cause to abandon them; but it is too quick to argue from the fact that there exists injustice in the enforcing of migration policies that disbanding borders altogether is the only appropriate response (Sager 2017). So, those who are vocal in objecting to “immigration injustice” can be responding to two distinct clusters of wrongs, i.e., those associated with unfair immigrations policies and those associated with the unfair enforcement of (perhaps) otherwise just immigration policies.

Non-cooperation policies, non-discrimination policies, and don't ask, don't tell policies are sometimes defended in terms that seem divorced from the questions of immigration policy or enforcement (Lenard 2019). One defense derives from the fact of scarce resources: in the vast majority of policy areas, the full and complete enforcement of every law is impossible, for lack of sufficient resources to do so. Thus, enforcement agencies across a range of policy areas necessarily prioritize enforcement of some laws over others. A second defense highlights that forced cooperation with immigration laws will result in undermining the capacity of service providers, whose job does not ordinarily relate to immigration enforcement, to do their own job well (Carens 2008). In this case, it may seem, the refusal to cooperate (maximally) with enforcement of immigration may (plausibly) not be motivated by the perceived injustice of the policy, but rather by the worry that the capacity to provide key services will be undermined as a result of cooperation. This form of justification is made most frequently with respect to non-cooperation by local security authorities, but is also deployed to support those who provide additional services as well. In particular, some sub-state jurisdictions argue with considerable plausibility that, although not always a specifically articulated local power, they are charged with a broad set of jobs associated with “immigrant integration”, and that requiring service providers to collaborate with immigration enforcement authorities undermines their capacity to their work. In the United States especially, these reasons have strong constitutional grounding; specifically the U.S. constitution grants states the right “to govern their internal matters including those that deal with the safety and general welfare of their residents” (Villazor and Gulasekaram 2018, 554). It also denies the center the right to “conscript” states to pursue its own objectives (Villazor and Gulasekaram 2018, 554).

In most cases, however, the adoption of sanctuary policies is motivated at least in part in opposition to immigration policies themselves. The strongest defenders of sanctuary object to border control in its entirety, but in most cases, sanctuary defenders are focused on the enforcement of immigration policies in their local jurisdictions and adopt sanctuary policies in an attempt to mitigate their worst effects: the undermining of inclusivity in a diverse state and the violation of human rights.

Some defend sanctuary policies for the contribution they make to protecting the inclusion of all residents on equal terms, no matter their immigration status, and especially, no matter their racial or ethnic identity. On this view, sanctuary policies protect residents from the discrimination that racial and ethnic minorities are likely to face where immigration enforcement strategies are aggressive at identifying irregular migrants; in the United States, many critics of then-President Donald Trump's “deportation” agenda argued that it was, in effect, racist. One variation of the inclusion defense emphasies more specifically that since ethnic and racial minorities are more likely to be targeted by immigration enforcement efforts, sanctuary efforts are aimed at protecting their equal status as members in the relevant jurisdiction (Lasch et al. 2018, 1765–66). A variation simply states that sanctuary policies demonstrate “respect for and appreciation of diverse communities, even embodying a certain solidarity with those who have been historically marginalized” (Lasch et al. 2018, 1769). This sort of attitude is at the center of Sheffield's City of Sanctuary Initiative, citing “inclusiveness” as its priority value as follows: “we welcome and respect people from all backgrounds, place the highest value on diversity and are committed to equality” (City of Sanctuary UK 2017; Saunders 2017, chapter 5).

Typically, the most powerful defense of sanctuary, at least among political theorists, is that the enforcement of immigration law violates human rights: these defenses highlight family separation, deaths in detention and during the deportation process, length and conditions of detention more generally (Nethery and Silverman 2015), and the failure to respect non-refoulement commitments (Blake and Hereth 2020). The thought that some spaces should be protected from incursions by immigration enforcement is not restricted to sympathetic political theorists: in 2011 the Barack Obama administration adopted the “Sensitive Locations” policy, with the goal of ensuring that “enforcement actions do not occur at nor are focused on sensitive locations”, including churches, schools, hospitals, and so on, unless “exigent circumstances” exist and “prior approval” was obtained (U. S. Department of Homeland Security 2011). Loosely, its objective is to ensure that all those present on American territory can be assured of having their basic human rights protected, regardless of their immigration status. This policy, which was not rescinded by the Trump administration, made explicit that immigration related arrests, interviews, searches and surveillance were not to be undertaken at any location deemed “sensitive”, i.e., any location that is connected to the provision and protection of human rights in some way. It did not protect the right of those who operate services to deny access to records they possess that might, however, incriminate those whom they serve. It left room, in other words, for the sanctuary policies I described just above to be deemed (by some jurisdictions) to be necessary to protect those who require their services. This commitment, to protect the basic human rights of those who are most vulnerable, by insulating them as much as possible from the threat of immigration enforcement, is at the core of most recent choices to adopt sanctuary policies.

(Immigration) resistance versus (immigration) opposition in political theory

Much political theory of resistance, including in the space of migration theory, begins with John Rawls’ well-known accounts of civil disobedience and conscientious objection. Rawls’ starting point is the natural duty we share to support reasonably just institutions, and his connected observation that, so long as institutions are reasonably just in general, we are bound to abide by at least somewhat unjust laws. Correspondingly, under the conditions that institutions are reasonably just, when they veer away from justice, we ought to work within their boundaries to bring them into conformity with justice, not outside of them (Rawls 1999, 311).

But, Rawls says, this general account requires supplementing in cases where working within the boundaries of institutions is insufficient as a way to bring about justice. There will inevitably be cases where citizens believe that their government is acting unjustly, where their attempts to work inside of these institutions to bring about justice have been unsuccessful, and where they come to believe that unlawful action is warranted in support of persuading their government to act justly. This form of action, civil disobedience, stems from the belief among citizens that their duty to fight against a particular injustice is such that they must, in service of this duty, engage in some sort of opposition, which may be unlawful. Civil disobedience, Rawls explains, is “a public, nonviolent, conscientious yet political act contrary to law usually done with the aim of bringing about a change in the law or policies of the government” (Rawls 1999, 320). Moreover, such action must be fueled by moral principles that, at least from the perspective of the civilly disobedient, are part of “the commonly shared conception of justice that underlies the political order” (Rawls 1999, 321). The civilly disobedient must therefore communicate a tricky stance: she is in general committed to abiding by the law that governs her state, but aims to express “disobedience to law within the limits of fidelity to the law” (Rawls 1999, 322). The key, then, to understanding Rawls’ account of civil disobedience is that its permissibility – and indeed its persuasive capacity – depends on there being, among citizens, a “democratic conception of society as a system of cooperation among equal persons,” which is violated in a particular instance though not, overall, abandoned (Rawls 1999, 336).

Rawls distinguishes civil disobedience from conscientious objection, which manifests in noncompliance with a “direct legal injunction or administrative order,” stemming from an individual's moral or religious convictions, rather than a shared sense of justice among citizens. Conscientious objection need not be directed at the sense of justice that prevails in society, though it may be. Historically, conscientious objection has been associated with refusals to participate in military action, and more recently with pharmacists who refuse to offer contraception or the morning after pill to women and service providers refusing to assist gay couples access a range of services (usually pertaining to nuptials).

Many of the actions taken by those who resist unjust immigration enforcement blur the line that Rawls draws civil disobedience and conscientious objection. Think of individuals who have recently garnered significant media attention for their actions. Scott Warren, arrested for his actions in collaboration with No More Deaths, a volunteer organization operating at the border between the US and Mexico, which aims to reduce the suffering of those attempting to cross the border, by leaving supplies including food and water in the dessert for potential crossers (Cabrera 2010), cited his “spiritual beliefs” as guiding his actions, and described his arrest as an infringement of his religious commitment to supporting those in need. Similarly, Teresa Todd, arrested for “human smuggling” because she stopped to offer support to migrants in medical distress, said simply of her actions, “Can I be compassionate?” (Matalon 2019). Yet, perhaps especially in the United States in its present political context, it is hard to believe that the actions taken by those who act to meet the basic needs of border crossers are merely conscientious objectors; they intend their actions as a more general challenge to the injustices of border enforcement, and so would perhaps not be satisfied if their actions were permitted to proceed without changes to immigration enforcement policy more generally.

This kind of muddy middle, between conscientious objection and civil disobedience, appears to be the space that Javier Hidalgo aims to occupy, in defending the right of migrants and citizens who aid them to evade, deceive and use defensive force against those who enforce what they believe are unjust immigration enforcements. These humanitarian actions, he suggests, can rightly be described as resisting unjust immigration restrictions, but they do not amount to civil disobedience, since by and large they must be done covertly to be successful, and both unauthorized migrants and those who support them act as such in order to avoid detection and punishment (Hidalgo 2019, 5). Elsewhere, Hidalgo defends the view that citizens – who are increasingly conscripted by their governments to participate in the implementation of unjust immigration laws, for example by being denied the right to hire or rent to irregular migrants – have not only the right but also the obligation to resist unjust immigration laws by refusing to comply with them. These “interaction restrictions” are unjust and citizens are obligated to refuse them on the grounds that they are, fundamentally, “rights-violating” (Hidalgo 2015, 173). Hidalgo writes, “Public officials implement the plan of restricting the liberties of these migrants by enforcing interaction restrictions. But ordinary citizens contribute to and make the government's wrongdoing possible by complying with these laws. Interaction restrictions distribute the responsibility for violating the rights of migrants broadly among citizens” (Hidalgo 2015, 173). Individuals should, he says, engage in “personal disobedience”, by refusing to comply with these laws. ¹

Not everyone is persuaded that opposition to immigration enforcement injustice should remain personal, however. Many believe rather that there are a great many occasions where civil disobedience as a kind of collective enterprise is warranted (Yong 2018). As well, at least historically, the civil disobedience frame has been the one applied to sanctuary movements (Rehaag 2009). Sanctuary gained prominence as a mode of opposing immigration injustice in the early 1980s, in California. There, churches became active in offering sanctuary to asylum seekers from Central America who, they believed, were entitled to refugee status because the US was engaged in nefarious actions in these countries, nefarious actions that forced people to flee in the first place. In addition to fleeing violence caused by US action, these asylum claimants, many churches thought, were additionally the victims of a US policy that refused to treat them as genuine refugees – in some cases, sanctuary was offered in protest of refugee laws that denied that the risk of violence was sufficient to gain refugee status, i.e., in protest of the policy itself, and in others, sanctuary was offered in protest of what church officials believed was the unfair application of refugee law to particular claimants, i.e., in protest of the enforcement of the policy (Lippert and Rehaag 2012). ² Although the actions church-members took – to shield asylum seekers in secret, and often to support their transport across borders into safe spaces – were covert and aimed at protecting asylum seekers from detection by border authorities, the work they were doing itself was publicly known and their leaders publicly defended their actions citing unjust immigration policies with which they did not want to be complicit, and which they hoped by their actions to change. Their actions are well-described as civil disobedience, even if they do not meet the Rawlsian “publicity” requirement in a straightforward way (Cabrera 2010, 132–46).

Attempts to fit opposition to immigration enforcement into the existing Rawlsian framework are already awkward, in other words, before considering how best to conceptualize the actions taken by sub-state jurisdictions that adopt sanctuary policies, for at least two reasons. First, whereas both civil disobedience and conscientious objection consist in the deliberate choice to violate a law, knowing that this action will be considered illegal by authorities, sanctuary cities defend their non-cooperative actions as legally permissible and, sometimes, even legally required. Especially in the United States, defenders of sanctuary rely on existing law to defend their actions, and as a result the proper interpretation of what the law demands is often litigated in American courts. To give just one example, in United States v. California No. 18-16496, the United States sued California over several laws it had adopted to limit cooperation between California state agencies and federal immigration enforcement. In that case, the US claimed that these laws were in violation of federal statute; but Judge Smith wrote, in his judgement on behalf of the court, in fact, that while the United States may permissibly request cooperation from California with respect to its immigration enforcement priorities, it may not require it: “when questions of federalism are involved, we must distinguish between expectations and requirements. In this context, the federal government was free to expect as much as it wanted, but it could not require California's cooperation” (Liptak 2020; for discussion of this case, see Brown 2022).

Second, while civil disobedience is typically treated as a form of conscious, collective, political action aimed at challenging the legitimacy of a law or policy, those who engage in it are not typically legally authorized representatives of a political community. So, civil disobedience can seem to suffer from a kind of “democracy problem”: as Kimberly Brownlee writes, one objection is that civil disobedience appears to challenge “the democratic legislature's supreme right to take strategic decisions for the whole community even though the legislature is better placed than individual citizens are to account for all of the reasons that bear on the right guidance to follow”; that it amounts to an “insult [to] our fellow citizens by privileging our own judgement above that of the democratic process”; or that by engaging in civil disobedience “we act unfairly by disregarding the procedural democratic mechanisms that the society has adopted for resolving disagreements” (Brownlee 2012, 174–75). Brownlee proposes instead that we see civil disobedience as part of the democratic process, which can make an important contribution to ongoing democratic politics, even though outside of formal democratic procedures, and one that is especially important for disempowered minority groups who may be excluded from formal decision-making spaces. I think this is broadly correct, but it is noteworthy that her account assumes there is a “democracy problem” to resolve. There is no such problem to resolve, however, in the case of opposition that is the result of a democratic process of authorization, rather than outside of it, as in the case of jurisdictions that adopt sanctuary status. In other words, the adoption of sanctuary – while clearly oppositional – does not easily fit into the standard Rawlsian picture of civil disobedience.

In summary, both civil disobedience in general and the adoption of sanctuary status are public and non-violent expressions of disagreement with existing law, and both are essential elements of a democratic process. Whereas civil disobedience is “contrary to law”, the adoption of sanctuary status is often defended as consistent with law. As well, whereas civil disobedience takes place outside of the formal democratic process, the adoption of sanctuary status is typically democratically authorized. Below I will refer to the kind of opposition manifest in the adoption of sanctuary states as democratic non-cooperation, and will delineate its features more precisely there. Before I do so, however, let me note that this form of non-cooperation is possible only in non-unitary states, so let me turn to offering an account of their goals and the assumption of at least minimal cooperation that typically underpins their stability over time.

The federal/non-unitary context for sanctuary

The political theory of federalism is rich with distinctions among types of federal structures both formal and informal and their various rationales. In general, federal states are characterized by a formal division of power between the central state and sub-state jurisdictions (Elazar 1995). In the broadest of terms, federal structures are adopted in order to recognize formally the self-determination of sub-state jurisdictions. In federal states, the integrity of sub-state jurisdictions is protected – a constitution, or other equivalently binding legislation, confirms their formal status and protects their boundaries from unilateral rejigging by the central state (Levy 2014, 335). Additionally, a federal constitution will typically delineate the specific policy domains for which each level of government is responsible (some are formally shared); doing so protects the autonomy of sub-state jurisdictions (Levy 2014, 335). In practice, however, both levels of government are “active in almost all policy fields…this blurring of precise boundaries of autonomy inevitably requires some agreed-upon understanding” (Hueglin 2019, 17).

One of the policy domains that has witnessed this “blurring” is immigration. It has been conventional, in both political theoretic work and as a matter of practice, to treat immigration as a matter of state sovereignty, and thus as a matter for the highest level of government (Simmons 2001). Manifest in recent political theory, however, is an attempt to theorize the reality that, in fact, cities and other sub-state jurisdictions play a key role in various aspects of immigration, including with respect to the enforcement of immigration policy and the integration of immigrants into the life of cities (Motomura 2018, 444; de-Shalit 2019). Commenting on the United States in particular, Peter Spiro notes that although immigration will remain a space in which the federal government “takes the lead” (p. 130), the discretion granted to states in various immigration-related matters has expanded substantially in the last many years (Spiro 1994, 130, 177). As Rose Cuison Villazor and Pratheepan Gulasekaram observe, “This conventional approach overlooks the important roles that non-federal stakeholders - states, cities, individuals, and other private actors play in immigration law governance that are occurring on the ground” (Villazor and Gulasekaram 2018, 553). Indeed, in the United States, one way to interpret Donald Trump's attempts to undermine sanctuary jurisdictions, and to instead coerce them, and especially their local police forces, into supporting his attempts at mass deportation, is as a recognition that federal immigration priorities require willing participation from states, and as an attempt to halt (or even reverse) the trend of devolving elements of immigration policy to states (Lasch et al. 2018).

Generally, the distribution of state powers between the center and sub-state jurisdictions is governed by a principle of subsidiarity, and indeed some of the shift of immigration responsibility from central to local governments reflects this: “Subsidiarity obtains when political and economic functions are performed by the ‘lowest’ feasible level” (Weinstock 2014, 261). This principle acknowledges that there are some policy spaces in which local authorities are best able to act efficiently, with clear understanding of the priorities of local constituents. As I will describe in more detail below, when sub-state jurisdictions take responsibility for integrating immigrants, for refusing collaboration with immigration enforcement, and when they focus on building and sustaining an inclusive space for all residents regardless of ethnicity, race, or immigration status, they are often acting from a commitment to sustaining and protecting local priorities as the principle of subsidiarity suggests they ought to do (Motomura 2018, 447).

In many cases, as well, sub-state jurisdictional priorities are connected to protecting the cultural specificity of sub-state jurisdictions, as is the case in Quebec and Catalonia (Norman 2006, chapter 3). In its ideal, the division of powers between the center and the sub-state jurisdictions captures the best of both worlds: sub-state jurisdictions are autonomous agents, able to protect the “local needs, identities and social values” of their citizens, and all sub-state jurisdictions together benefit in other domains from being part of a larger political unit (Cyr 2014, 27). The recognition of the importance of autonomy and self-determination of sub-state jurisdictions is part of the moral defense of federalism. So, it is not merely that federal structures are defined by divisions of powers, but also that the normative reason behind the choice to adopt these structures is the recognition that sub-state jurisdictions may well have particular, local, priorities, and it is right that they be permitted to use the levers of government to protect and support them. Thus, when they make priority choices focused on which laws to enforce with which level of zealousness, they are exercising their self-determination, which is acknowledged to be within their legitimate prerogative to do.

Although the oppositional sanctuary policies I describe below assume the presence of multi-level governance in some form, they do not require a constitutionally entrenched federal structure. Most states are non-unitary to some degree, at least in the sense that some authority to implement state-level policies is extended to sub-state jurisdictions, even if formally these sub-state jurisdictions do not possess the legal authority to adopt or modify policies. Sanctuary policies can thus be adopted, in some form, even in de jure unitary states. Moreover, the adoption of sanctuary policies in federal states does not always track the borders of those jurisdictions that are formally recognized in a constitution; states and provinces are often the only sub-state jurisdictions with constitutional status, and although states and provinces do often adopt sanctuary policies, they are more often adopted at the level of the city or municipality. Even as their status is not constitutionally protected and, as a result, they are at least in principle at risk of being dissolved when their actions are politically inconvenient, cities are increasingly recognized as independent actors, including in the space of immigration (L. King 2014; Weinstock 2014; de-Shalit 2019; Spiro 2010).

The formal division of powers between jurisdictions does not capture the whole federal story. No matter how precisely powers are designated to specific levels of government, the proper function of federalism relies not only on “written words”, but also “conventions, established through practice, that illuminate and extend them” (Bednar 2014, 234). These “informal norms” ensure that “various parties understand themselves as obligated to stay within certain zones of activity, whether or not such zones are enforceable by legal institutions” (Bednar, Eskridge, and Ferejohn 2001, 223). The institutional division of powers, coupled with the norms and expectations associated with respecting this division in practice is underpinned by an attitude of cooperation among the political units that is conditioned on the shared commitment to the political unit as a whole. Where does this commitment to cooperate come from? Federalism as a structure of governance is most likely to be successful, in the sense of more likely to achieve its objectives, where the units in a federal state are bound by what Hugo Cyr describes as a principle of “federal solidarity”, which underpins the collective political project in which the federation as a whole is engaged (Cyr, p. 34). Since there will never be 100% agreement about how to proceed, or what the best interests of the units as a whole are, or even how they are best achieved, the solidarity that Cyr describes undergirds negotiations among the units as well as a commitment to supporting the compromises that result (Elazar 1995, 41; Cyr 2014, 34).

The expectation of cooperation that flows from “federal solidarity” is moral not legal. ³ Although an expectation of cooperation may be embedded in the language of federal constitutions, the choice to rescind cooperation often violates a shared norm, but is generally not a violation of a law and no formal penalties can be imposed as a result. To say that it is a moral expectation is to say that – all things equal – an attitude of good will between units persists, and supports cooperation in the projects and objectives of each of them (Baier 1995). That is, the expectation applies to all parties. The central government ought to be able to count on local authorities to do their part to not disrupt, and indeed support, its plans, as a matter of course. Similarly, the expectation imposes on the central government the requirement that it avoid adopting laws and demanding conformity with them, but rather that it sustains a commitment to meaningful deliberations and negotiations to ensure that local authorities are generally in support of such policies. The expectation of cooperation emphasizes that the goal, in shared governance in non-unitary states, is to maximize collaboration, and therefore consent of the units (Zimmerman 2001, 20). Where units fail to meet this expectation – where the trust that other units will do so is eroded – federalism cannot generate the benefits it promises (Bednar 2011, 276).

Let me take a moment to respond to a possible objection, here, that the way I have characterized federalism assumes that cooperation necessarily produces morally desirable outcomes, for example in terms of greater justice or better protection of human rights. I do not intend to make this claim, however. In some cases, cooperation between the central state and sub-state jurisdictions can lead to greater injustice or reduction in human rights protections. In other cases, for example as in the sanctuary case I will describe in more detail below, the withdrawal of cooperation is often justified with respect to the importance of protecting human rights. For those who believe, as I do, that immigration enforcement is often unjust, this form of “uncooperative federalism” (Bulman-Pozen and Gerken 2008) is to be welcomed and celebrated. ⁴ But it may not always be welcomed, and it will not always press in favor of justice. I will return to this question in the conclusion.

As a matter of practice, the ability of a sub-unit or the central state to carry out its objectives inside of a sphere of responsibility relies on the willing cooperation of others, in multiple ways. For example, the central state can ask that sub-units administer or enforce certain policies, with varying levels of pressure; the central state can set priorities, and ask that sub-units adopt strategies appropriate for their own jurisdictions; the central state can adopt policies that rely on the existence of enforcement capacity at the sub-unit level; the central state can fund sub-units to develop policies in specific areas and so on (Schuck 2007, 66–67). In general terms, the central state acts to set national objectives, and then offers varying levels of constraints, pressures, and requests, in order to encourage the cooperation of sub-state units in pursuit of these objectives. Sub-state units are expected to cooperate in the achievement of national level objectives, and on the basis of this expectation, they are granted substantial “discretion in applying this framework to their needs and policies, through the design particularities of their plans as well as through their day-to-day decisions in implementation and enforcement” (Battle 1980, 2).

Cooperation between units in a federal state comes in many forms, and with a range of expectations for both the sub-units and the central state. Analytically, it is useful to treat the cooperation that characterizes federalism as coming in four major forms, all of which stem from the baseline assumption that the political unit as a whole shares a set of collective objectives and generally benefits pursuing them as a unit. ⁵ As I will describe, they fall on a “cooperation continuum”, “which extends from the willing embrace of federal policy and national standards to uncooperative behavior that can revise, reshape or reject national standards” (Chen 2016, 19–20). The attitude that sub-state jurisdictions in particular take can augment state objectives, or they can “weaken, slow, or redirect” state objectives (Bednar 2014, 233; Chen 2016, 20).

Mutually beneficial cooperation characterizes situations where sub-state units and the central state must both do their part to secure a policy objective; here, if either party withdraws, the policy objective cannot be achieved (or can only be achieved with great, perhaps overwhelming, effort by the non-withdrawing partner). Commensal cooperation refers to situations where one party has an objective, and requires the infrastructure of the other to achieve it, but from which that party does not benefit in particular. Co-operative non-interference refers to the cooperation implied by the respect for separation of powers that characterizes federalism. Parasitic cooperation refers to situations where one party has an objective, and requires the cooperation of the other, where that other party either does not support the objective or where the requested cooperation imposes unwanted costs that it must absorb. It is valuable to highlight these latter two forms of engagement in federal states, specifically as forms of cooperation, even if they lack enthusiasm, because they reflect that each party to a federal state can hinder the objectives of others but choose against doing so, or absorb costs that they would otherwise prefer not to absorb. That is, and to return to the operating assumptions in federal states with which I began this section, the choice to behave cooperatively in some specific instances can be made in service of the larger shared objectives of the federal unit as a whole, even if there is sustained disagreement with a particular policy at a particular moment in time (Cyr 2014, 21; see also Bednar, Eskridge and Ferejohn 2001, 223).

Sanctuary as democratic non-cooperation

So, federalism is characterized by an expectation of cooperation, ideally motivated by a spirit of solidarity between sub-units and the central government. How then should we respond to the withdrawal of cooperation, as manifest in the adoption of sanctuary policies? There are at least three possible responses. One response might immediately concede a duty of cooperation in federal states, but object that it applies only where the policies proposed meet certain standards of justice. On this view, those who defend sanctuary policies might simply say that certain immigration admission and enforcement policies are manifestly unjust, and the duty of cooperation does not require abiding by them. Indeed, I believe that there will be cases in which the manifest injustice of a policy undermines the legitimacy of any demand for collaboration in rendering it effective. But, not all cases of be of this type.

A second response to the withdrawal of cooperation is that, given the normative requirement of cooperation, the withdrawal is a dereliction of duty, which is to be condemned. Here is Hugo Cyr commenting on cases where sub-units refuse or fail to cooperate with the central state:

Failure to cooperate, to share the common burdens, endangers the existence of the body politic. Acting in bad faith towards the other members of the body politic, abusing one's powers to the detriment of one's federal partners, also go against the purposes of forming a common body politic (Cyr 2014, 34).

This type of response is frequently invoked in the US administration's response to sanctuary jurisdictions, which accuses them of failing to collaborate appropriately to meet federal objectives. Here is President Trump threatening to withdraw funding from states that adopt sanctuary policies: “We shouldn't have to pay anything anyway, because all they do is make it very hard for law enforcement” (Alvarez 2020). In other words, since sanctuary states have withdrawn cooperation with immigration enforcement, the federal government (in Trump's view) can withdraw its own commitment to states across a range of sectors. Whereas the first response accepts all sanctuary policies as legitimate because adopted in the face of unjust immigration policies, this second response rejects them all for being a violation of the cooperation on which federalism depends. This second response, like the first, is too broad brush; rather, a third and more nuanced approach to evaluating the adoption of sanctuary policies is needed.

The third response recognizes an important place for “democratic non-cooperation”, which better captures what is going on when sub-state jurisdictions adopt sanctuary policies. Democratic non-cooperation is the deliberate taking of an oppositional stance by sub-state jurisdictions, who act to withdraw cooperation in a specific case where cooperation would ordinarily be expected. In the sanctuary case, one democratic public chooses an immigration enforcement policy and requests cooperation in that enforcement from a sub-state democratic public, which chooses against (full) cooperation in that enforcement. What is normatively interesting about the oppositional stances taken by sub-units in federal states, which are underpinned by an expectation of cooperation, is that they are democratically justified and authorized. The presence of federal structures give space for democratic non-cooperation, in general and in the case of sanctuary policies. Democratic non-cooperation, though a withdrawal from the expectation of cooperation, is not always choice made in “bad faith”, as Cyr suggested, nor it does amount to a desire to undermine the polity as a whole.

Just like cooperation, democratic non-cooperation comes in four typical forms, which can be treated as on a “non-cooperation continuum”, moving from passive forms of non-cooperation towards more active forms: evasion, non-engagement, disruption and obstruction. In other words, they move from mere refusal to engage with immigration enforcement strategies towards active interference with them (Motomura 2018, 438–339). The structure of federal institutions is such that these actions are available for sub-state jurisdictions that are not in full agreement with the central state's objective; they are not contrary to law in a straightforward way ⁶ , as is civil disobedience, nor do they demonstrate a desire to disband, or entirely refashion, the union, as might be the intention of sedition or insurrection. Yet, the choice to adopt such policies undermines the expectation of cooperation that typically governs interactions between sub-units and the central state in federations. Below I describe these forms in general and then also how they apply in sanctuary cases. ⁷

Evasion describes actions where actors sidestep a situation that they would prefer to avoid. Witnesses in trials are sometimes evasive for example, which does not mean that they are lying, usually, but rather that they are avoiding telling the full truth. In the case of sanctuary, evasive actions are taken when there is a law that requires service providers to notify immigration enforcement agencies that they are interacting with an undocumented person; evasive actions by service providers, in the form of the “don't ask, don't tell” policies that I described in the introduction, can serve to avoid their coming to know such a fact, by refusing to ask questions that would make this status clear, and thereby permitting the provider to avoid being in possession of information that, if shared would bring someone to the attention of immigration enforcement authorities. For example, when Toronto became the first city in Canada to adopt sanctuary status, among other things, it directed city officials to avoid inquiring “into immigration status when providing select services” (Hudson et al. 2017, 2).

Non-engagement makes the objectives of others more difficult to achieve, by choosing against supporting their efforts to do so. In federal states, recall, each jurisdiction can typically, and in relatively low-cost ways, act to make the objectives of the other units easier to achieve. The normal interpretation of the duty of cooperation usually means that – all else equal – the relevant jurisdiction should choose to do so. Non-engagement actions refuse these low-cost ways in which sub-state jurisdictions can support the center.

In the context of sanctuary, non-engagement actions are those that make it harder for immigration enforcement agencies to do their work, though not by force or by directly blocking their efforts. I have in mind cases where police forces cooperate with immigration enforcement agents to some extent, but where they follow the rules exactly as they are written and no more. In the United States, it has been common for Immigration and Customs Enforcement (ICE) to issue “detainer requests”, which ask that local police forces hold individuals of interest to immigration authorities until they can arrive on site to detain them. Some local police forces however refuse such requests, choosing instead to release suspected undocumented migrants as soon as they are no longer under local criminal suspicion, even if federal authorities would like them held for further questioning; they simply behave as if they do not know of or care much about federal priorities. In the United States, this issue has been litigated and courts have agreed that local jurisdictions cannot be required to respect detainer requests; compliance with them is voluntary (Lasch et al. 2018, 1732). In such non-engagement actions, local governments may choose to continue detention, but they instead follow their own regulations that individuals who are no longer under criminal suspicion are released as quickly as possible. Here, as elsewhere, there is space for discretion by local police forces, which they can exercise by supporting federal immigration law or not. In cases non-engagement, they use their discretion to follow sub-state jurisdictional laws precisely, rather than in ways that might better support federal immigration enforcement agencies, were they keen to do so.

Disruptive actions are those that purposively interrupt the ability of others to carry out their objectives. Above I described one form of evasive action that sanctuary cities frequently take: the choice to adopt policies that prevent service providers from asking about the immigration status of those to whom they offer services. Chicago's policies add an additional step: explicit purging of any immigration information gathered via the offices that provide benefits, every 60 days, as a precautionary and protective measure. This purging is a disruptive form of opposition, since it purposely interrupts the ability of federal agents to gather identifying information of service-seeking residents of Chicago (Municipal Code of Chicago 2012).

I described a non-engagement policy above, namely the refusal to respect detainer requests, where sub-state jurisdictions release individuals before federal immigration authorities are able to issue detainers; policies that explicitly refuse to comply with detainer requests, where they are issued, are disruptive. Recall that in a federal state, the assumption of cooperation informs the ways in which policies are made and expected to be enforced. The refusal of the sub-state partner to “do its part” amounts to a disruption of the ways in which each jurisdiction can be expected to rely on each other. In the case of sanctuary, the refusal to detain or arrest an individual that is known to be of interest to immigration enforcement is disruptive in this way.

Chicago's sanctuary policy – its “Welcoming City Ordinance” of 2012 – specifically denies the right of its police forces to detain individuals who are known to be wanted by ICE, because they are wanted by ICE. Moreover, its policy goes further: “no agent or agency may allow ICE agents access to an individual in custody of the agent, allow ICE agents access to facilities for investigation, nor respond to ICE inquiries or communicate with ICE about the individual in custody's custody status or release while on duty” (Municipal Code of Chicago 2012). These policies require Chicago police and service providers to actively refuse cooperation with ICE, with exceptions only for when such individuals are suspected of serious criminal activity. Additionally, disruptive policies prohibit local jails from entering into contracts with federal immigration authorities to detail individuals therein. California adopted such a law in 2018, for example, which reads “state and local law enforcement agencies are prohibited from contracting with the federal government for use of their facilities to house individuals as federal detainees for purposes of civil immigration custody” (Assembly Bill No. 110 2018).

Obstructive acts render it more difficult to get a particular task done, by deliberately getting in the way of an agent's ability to do that task. It aims to undermine the authority or coercive power of an institution or actor in a specific domain. In the sanctuary case, the most prominent obstructive actions are those that explicitly support individuals who, sub-state jurisdictions believe, are likely to be targeted by immigration enforcement authorities. They include policies that actively inform individuals who are detained of the requirement that they (the authorities) communicate with immigration authorities, or that require individuals’ active consent before doing so. They also include policies to provide legal representation, or translation services, to detainees in order to resist immigration authorities. These forms of obstructive actions come with a price, i.e., they are costly to local jurisdictions that adopt them, whereas the policies described earlier typically are not. For example, in May 2017, the city of Seattle allocated one million dollars to local legal organizations to offer legal support to undocumented migrants (City of Seattle 2017; see also Ewing 2017). New York City did the same thing, also in 2017, reserving 16 million dollars for legal services to support detained individuals as well as asylum seekers (Robbins 2017).

Evasion, non-engagement, disruption and obstruction are the primary instances of democratic non-cooperation. In adopting these policies, sub-state jurisdictions withdraw cooperation from what had been assumed to be a shared objective among jurisdictions. How should we think of these oppositional actions?

Descriptively, these political actions are taken by sub-state jurisdictional units and demonstrate that the precondition for cooperation in federal states – the presumption of shared political interests – has weakened, if not vanished, in at least one policy domain. In a federal state, recall, a key working assumption is that the center and sub-state jurisdictions generally have the same, or at least overlapping priorities. But, as suggested earlier, conflicting priorities can emerge. So, although the expectation of cooperation is real, the assumption that it ought to persist as a duty even where the underlying condition for its operation – a shared political interest – has been eroded, is mistaken. The stakes in the sanctuary case are high, since at issue is who is a legitimate resident of a state, and thereby entitled to protection, as well as what are appropriate mechanisms by which those whose residence is determined to be illegitimate should be removed or granted clemency.

The non-cooperation continuum described above identifies actions from one end of the spectrum (evasion) to the other (obstruction), so that cooperation moves from being mutualistic to being parasitic, to refer to language I used in section 2. As commitment to particular shared interests declines, so does a willingness to cooperate, followed by the emergence of a desire to withdraw from meeting the obligations that are implied by a duty of cooperation. With respect to immigration enforcement, which is my subject here, in mild cases, the conflict manifests merely as a conflict of priorities, as evasion and non-engagement, where at issue is a local jurisdiction that respects federal immigration policy in general, but believes that certain enforcement strategies undermine its capacity to carry out its own goals. In more severe cases, the priorities of sub-state jurisdictions and the center diverge more substantially and are manifest in more severe forms of non-cooperation, disruption and obstruction. The adoption of these sanctuary strategies is a democratic expression of disapproval by sub-state jurisdictions, whose commitment to cooperation (in this domain) is wavering, as a result of this erosion of shared objectives.

Concluding thoughts

Over the course of the discussion and analysis above, I defined the key features of sanctuary as it is presently practiced, to argue that it is a form of democratic opposition that is worth moral consideration as a distinct form of political conflict, and which does not obviously meet the requirements of civil disobedience as Rawls described it. Although much sanctuary activity has been described by political theorists of migration as civil disobedience, or as importantly connected to civil disobedience, this frame is not appropriate to understand what is distinct about the democratic non-cooperation that is manifest by sanctuary actors.

Here in this final section, I want to lay out the features of this form of opposition and then suggest that “sanctuary” is not its only instance. In other words, what I have aimed to do in this article is articulate a distinct form of opposition that is worthy of separate theoretical analysis, as one of the many ways in which opposition takes place in democratic spaces. It is a form of opposition that is available only in territorial spaces in which there are plural authorities operating at once. Under this condition, to count as democratic non-cooperation the opposition must be:

collective;

In the discussion above, I developed this account in relation to sanctuary, but a larger objective has been to identify a type of lawful democratic opposition that is more general, and which captures sanctuary, but which also captures other cases of opposition, not all of which press forward in a progressive, rights-enhancing direction. In particular, it is noteworthy that in response to the repeated legal successes of sanctuary jurisdictions in refusing cooperation with federal immigration authorities, other jurisdictions have used the same powers to adopt so-called “anti-sanctuary” policies, in which state authorities pledge to dedicate substantial time and effort into supporting the enforcement of immigration authorities. Indeed, in the California case that I described above, after the state adopted laws to limit cooperation between California state authorities and federal immigration authorities, several cities inside of California passed their own laws declaring their intention to act in conformity with federal rather than state law (Sanchez 2018).

To see, just by illustration, that this form of opposition is present in other policy spaces, consider these additional examples of democratic non-cooperation. Arizona's state government, after a period of deliberation in which many lawmakers expressed the view that US immigration enforcement policy was too lax, adopted what became known as the “show us your papers” law; this law authorized police officers to question people whom they believed were been present in the US illegally. In this case, the law was ultimately struck down for being discriminatory in its intent (and effect), but the case suggests that there are moments where sub-state jurisdictions make choices to resist the central government via the normal procedures of democratic authorization, which are not as comfortable for those who are generally sympathetic to the claims of irregularly present migrants (Totenberg 2012). Or, consider so called “second amendment” sanctuary, where sub-state jurisdictions adopt laws that would halt enforcement of gun control measures, including background checks or the banning of certain types of guns and ammunition, that are believed to violate the “right to bear arms.” ⁸ Or consider, even, the efforts of mayors in Georgia to adopt mandatory masking legislation, as a mechanism to reduce the spread of COVID-19 in their sub-state jurisdictions, in contravention of their Governor's (deeply misguided) choice to ban such legislation. In other words, sub-state agency in the form of democratic non-cooperation can be deployed in multiple policy spheres in defense of, and against, justice. ⁹

What these examples share in common, and what the sanctuary strategies I have categorized here share, is that they are examples of collective opposition (Wilcox 2019), that democratic spaces are adopting, within the framework of democratic politics, to resist what they believe is unjust immigration enforcement. There is more than one way to resist unjust immigration enforcement, and there is room certainly for civil disobedience and conscientious objection as strategies to do so. Indeed, in the current political context, breaking the law in defense of migrants, especially in the form of protecting space for the good Samaritan actions that are being penalized in the southern US, and in the Mediterranean Sea, (and against the injustice of so many immigration enforcement policies), may well be both defensible and praiseworthy (Duarte 2020).

What I have hoped to show, however, is that the form of political action taken by sanctuary jurisdictions is another tool available for those who wish to oppose aggressive immigration enforcement and protect non-status migrants who reside in their communities. One reason to defend this form of opposition as a valuable but distinct mechanism to oppose unjust enforcement is that working inside of the space of democratic politics has a chance of changing minds, especially the minds of those who have anxiety about rule-breaking (de-Shalit 2019). Another reason is that, although some of the recent immigration policies – and refugee admission policies – are grotesque, they could be so much worse than they are, and as a result it may continue to be reasonable to advocate for immigration justice inside the democratic apparatus rather than outside of it. In summary, what I have demonstrated is that there is space internal to democratic politics to enact effective opposition, in the cases that motivate the adoption of sanctuary policies and elsewhere.