Authority and Immigration

States claim to have authority over prospective immigrants who have not yet been admitted but are nonetheless expected to comply with immigration law. But what could ground such an authority claim? The service conception of authority defended by Raz appears not to apply in this case. Nor can it be argued that immigrants give their consent to the state by applying for admission. Another approach appeals to the practice of reciprocity between states in respecting each other’s immigration regimes, but many immigrants will fall outside of its scope. Instead, the article defends the view that the natural duty of justice requires immigrants to comply with the state’s immigration regime provided that it is reasonably just. This does not require that the immigrant herself should have authorised the regime through democratic participation. However, the natural duty argument has to be qualified by recognising that some migrants can legitimately appeal to necessity as grounds for breaching the duty and entering unauthorised.


Introduction
The question this article addresses is whether and if so on what grounds states can claim authority over prospective immigrants, meaning would-be immigrants who have not yet been granted legal recognition by the state in the form of residence or citizenship rights. These then are people who are applying to immigrate from a distance, or who have made it to the border and are asking to be admitted, or who have entered without permission and are now seeking leave to remain. How should we understand their moral relationship to the state they are trying to enter? The state makes and applies immigration law to people who belong in these categories and it expects them to comply with its regulations; in other words, it claims to have legitimate authority, within this limited sphere, over people who are not yet under its jurisdiction in the way that citizens and permanent residents are. 1 For present purposes, I am going to assume that if the state does have such authority, then prospective immigrants will be under duties to comply with those parts of immigration law that apply to them. 2 This means, for example, that if their application for an immigration visa is turned down, they ought not to attempt to enter illegally, or if they are refused entry at the border, they should comply with the instruction to return home. These duties are not unconditional, and later in the article, I will be looking at circumstances in which they may be set aside. My premise, nevertheless, is that if it can be shown that the state's authority claim is valid, this will often make a difference to what such immigrants are permitted to do. The issue is whether, and if so how, such a claim can be vindicated.
The question I am asking only makes sense if we accept that the state has at least a conditional right to exclude immigrants. The purpose of immigration law is to determine who is going to be admitted and on what terms, and to lay down procedures for handling the claims of different categories of migrants. Someone who believes that all border controls are indefensible must regard immigration law as an illegitimate imposition, and therefore as disqualified from acting as a source of authority. But not all open-borders advocates take such an extreme view. Instead many of them regard international freedom of movement as an ideal to be realised once conditions permit, while recognising that in the meantime some control of migration is unavoidable, so it is important to investigate how this can be carried out in the fairest way possible. Joseph Carens (2013: chaps. 9-10), for example, a prominent defender of the ideal of open borders, devotes considerable space in his book to investigating the criteria that should and should not be used when migrants are being selected for admission. For people in this category, the question whether immigrants are under an obligation to comply with immigration law that meets the required standards of justice is clearly pertinent, and has obvious practical implications both for the immigrants themselves and for citizens who are helping them to gain entry. So my enquiry should be of interest not only to those who believe that legitimate states are always entitled to control their borders but also to those who hold the more moderate open-borders view exemplified by Carens. I begin by explaining not only why it is important to decide whether states can claim authority over immigrants but also why the 'service' conception of political authority proposed by Raz appears not to apply in this case. I then consider, but reject, two possible grounds for such authority, namely, consent and reciprocity before presenting my own favoured account, which holds that the natural duty of justice requires immigrants to accept the authority of immigration law, provided that law is (reasonably) just. However, in the penultimate section of the article I propose a significant qualification, which is that for some immigrants, the pro tanto duty they have to comply with immigration law can be set aside on grounds of necessity.

Authority over Immigrants and the Service Conception
This idea that prospective immigrants may have duties towards the state, corresponding to its authority with respect to them, has not been much explored in the literature on immigration, where the focus is usually on the rights of immigrants and the corresponding duties of the receiving state. 3 One exception to this is Matthew Gibney's (2020) very interesting essay on the duties of refugees, but he is chiefly concerned with refugees who have already been offered asylum by a host state. 4 You may think that this neglect is not surprising, since while immigrants have general moral duties like everyone else -so they should not, for example, try to shoot border guards 5 -it is obvious that they lack any specific duties of compliance with immigration law. Since my aim is to challenge this assumption, we should pause for a moment to consider the implications of the no-authority/no-duties position. This would represent the relationship between state and immigrant simply as one of force: the immigrants try to get in, the state tries to keep her out (unless it chooses to admit her). No less a philosopher than Thomas Nagel has embraced this view. Nagel (2005: 129-130) writes: Immigration policies are simply enforced against the nationals of other states; the laws are not imposed in their name, nor are they asked to accept and uphold those laws. Since no acceptance is demanded of them, no justification is required that explains why they should accept such discriminatory policies, or why their interests have been given equal consideration. It is sufficient justification to claim that the policies do not violate their prepolitical human rights. This, however, reveals why the simple force versus force view is not in the end so attractive. Certainly it lifts any moral burden off the shoulders of the immigrants -they are not required to comply with the state's immigration laws -but at the same time it reduces the justificatory requirements that can be laid on the state. On Nagel's account, all it has to show is that its immigration policy does not violate the basic human rights of the immigrants, a condition that might require it to admit asylum seekers, for example. This account could of course be strengthened by adding further moral requirements, such as procedural requirements to govern the way that immigrants are treated by the admission system (for a good account of the moral constraints that states may be subject to when dealing with immigrants, see Lister, 2020: section 2). But these constraints would be self-imposed; they would be no different in principle from the constraints that citizens impose on themselves when they legislate to protect animals against cruel treatment. If the state makes no authority claim vis-à-vis immigrants, then it need not justify its immigration law to them. But we are likely to think that it does indeed owe them a justification for the policies that it has chosen to adopt and enforce -for example, to explain and justify the overall immigration quotas it has set, or the particular criteria for selecting immigrants that it employs. If we can construe the relationship between state and immigrant as one of legitimate authority, then although this imposes compliance duties on the immigrant, it also imposes special duties on the state to provide immigrants with reasons why they should comply with its law, by explaining and defending the immigration policy it has chosen to adopt in the face of their claim to enter.
More generally, there is something repugnant about imposing enforceable laws on people that they are under no obligation to obey. It seems that doing so could only be justified in exceptional cases. Dworkin (1986: 191) has made this point: A state may have good grounds in some special circumstances for coercing those who have no duty to obey. But no general policy of upholding the law with steel could be justified if the law were not, in general, a source of genuine obligations.
The challenge, therefore, is to show that states do have authority over prospective immigrants such that the latter are under at least a pro tanto obligation to comply with immigration law.
It is difficult, however, to see how this challenge can be met. 6 The reasons that are usually given to justify the state's authority over its own citizens appear not to apply to prospective immigrants. For example, Joseph Raz's (1985Raz's ( , 1994 well-known 'service' conception of authority, where establishing authority involves showing that its subjects will better comply with their reasons for action by following its directives than by making their own independent decisions, seems unlikely to work in this case. For its citizens and other residents, the state performs coordinating functions on a massive scale, so as to make the service conception of political authority at least somewhat plausible internally. But the same cannot apply to immigrants before they are included in the society; they are not the beneficiaries of all the general laws and policies that the state applies, including laws and policies governing the provision of welfare benefits and public goods. 7 Might prospective immigrants have reasons of a different kind that recognising the state's authority would help them comply with? There might be some for whom migration was morally impermissible, since they had obligations to compatriots in the country they were planning to leave that they could only discharge by staying put -this might include, for example, trained medical personnel whose services were vitally needed at home. If the receiving state has well-designed policies to identify such indefensible brain-drain cases, perhaps following its directives would be a good way for people potentially in that position to discover whether it was indeed permissible for them to move. But it is controversial whether there are obligations to remain of this kind (see Brock and Blake, 2015;Stilz, 2016), and in any case the reasons in question could only apply to a very small proportion of those who are likely to want to migrate. 8 Of course, it might turn out that there is no general reason that can be offered to all incoming migrants to show that they should recognise the authority of immigration law, but only specific reasons that apply to particular types of immigrant. But ideally we would like to identify such a general reason, and in the following section I consider two prima facie plausible candidates.

Consent and Reciprocity
One of the most venerable justifications for political authority appeals to the idea of consent, so might this provide a way to explain the state's authority over prospective immigrants? Perhaps by approaching the state in their attempt to be admitted, immigrants are also giving consent to the immigration laws that apply to them. For example, the act of filling in a visa form might be understood as signalling consent to the process whereby these applications are decided, just as, say, applying to a university is understood to involve consenting to its selection procedures, such that someone who is turned down for a place is expected not to show up and try to enrol in classes regardless. Equally, the physical act of moving to, or across, a state border might be interpreted as conveying consent to that state's immigration rules (although as Grey, 2015, points out, this is less plausible in the case of immigrants who knowingly act in contravention of those rules).
How convincing are these suggestions? We need first to ask under what general conditions performing a particular act can amount to giving obligation-creating consent (for a fuller analysis, see Simmons, 1979: chap. 4). One condition is that the act is generally understood to involve the giving of consent, and that the actor knows this, or at least ought to know it. Another is that the act is such that there is a reasonable alternative to performing it. If performing action A is going to be taken as a valid expression of consent, there must be another action B that the agent in question can perform in place of A without incurring significant costs relative to A. Together these conditions help to ensure that the act is voluntary in the sense that is needed for it to express consent to the ensuing obligations. So understood, consent may do a reasonably good job of explaining why, for example, tourists and other such visitors are obliged to respect and comply with the laws of the states they visit, at least provided those laws are not egregiously unjust. It is generally understood that by entering another country, a person is agreeing to respect its laws ('when in Rome, do as the Romans do'); moreover, the decision to visit is voluntary in the sense that the visitor had reasonable alternatives to choose between. When I decide to take a holiday in France, the alternative if my plans were thwarted would be a holiday at home or in some third country that is nearly as good. But it does not seem that these conditions for voluntariness are fulfilled for many, perhaps most, of the people who are attempting to immigrate to rich Western states. On one hand, there is no shared understanding that by taking steps to enter a country, a person is also expressing their consent to that country's immigration laws. On the other hand, for most migrants, the costs of not migrating will be relatively high, given how much they stand to gain if they succeed in entering the country of their choice (this applies not just to refugees and other forced migrants but also to many who are moving for economic reasons). So it seems that the justification for the state's authority over migrants that appeals to their consent could at most apply to a subset of migrants, such as professionals moving from one rich country to another, for whom the costs of being refused entry are relatively small.
But perhaps this dismissal of consent is too quick. Does it make a difference how the alternatives facing the agent have arisen, so that consent can sometimes be given even though the alternative to performing the consent-conferring act is very bad? Grey (2015: 119) argues that we should distinguish 'between situations where duress was aimed at extorting consent and situations where consent was given under general duress caused by a third party or external situation'. For example, I cannot be said to consent to the highwayman's demand that I should hand over my purse when I comply with it, since it is he who has attached dire consequences to the act of refusing to do so. But consider the following case, following a suggestion of Grey's. Suppose I am travelling abroad in some remote area and I have a serious accident that requires immediate hospital treatment. There is only one hospital in the vicinity, so I have no real choice but to be taken there. It might seem, nonetheless, that by allowing myself to be admitted, I have consented to the rules that the hospital applies to its patients. I might not like these rules -I might, for example, be a night owl who finds the hospital's 9 o'clock lights-out rule very objectionable -but, it may appear, by agreeing to be admitted I am obliged to comply with these rules even though the original choice was a forced choice between being taken to the hospital and being left to suffer injury or death.
The hospital case is certainly worth reflecting on, but the upshot is that it is not consent that is really doing the work in explaining why I am obliged to obey the hospital's rules. Although I may have consented to be taken there, it is implausible to suggest that in addition I have agreed to abide by whatever rules the hospital has decided to impose. 9 A better explanation is provided by something like the service conception of authority, which as we saw does not apply to prospective immigrants. The relevant point is that once I am inside the hospital, I am involved in an institution that can only function effectively by coordinating the behaviour of its inmates, and so while I am inconvenienced by the lightsout rule, I am at the same time benefitted by the no-smoking rule and the other regulations that the hospital applies to the patients. So I should indeed regard the rules as authoritative but not because I have consented to them in advance through being admitted. I should follow them because granting the hospital the power to make and enforce such rules serves the ends that I have independent reason to want to promote, such as patient health and safety.
The hospital example was introduced in order to show that there can be consent to authority even when the person involved has no reasonable choice but to be placed under that authority. But on closer analysis, it turns out that our intuition that the patient is obliged to obey the hospital's rules, provided they are reasonable, is not to be explained by reference to consent but more directly by appeal to the Razian conception of authority. Even though consent can sometimes be given under circumstances of duress -when there is no reasonable alternative to performing the act that conveys consent -its scope in such circumstances is very limited. For example, a migrant who is plucked out of the sea from a sinking boat can be understood to have consented to the authority of the ship's captain who has rescued her and should obey his instructions for the duration of the voyage to land, but her consent does not extend to the immigration rules that may then be applied to her. The authoritative status of those rules, and her obligation to comply with them -if there is indeed such an obligation -has to be explained in some other way. What might this be? My second candidate principle is reciprocity.
To understand how this might work, we have to begin our thinking at state level. The international system of states is one whereby, in principle and to a considerable extent in practice, states agree to respect one another's sovereignty. In particular, they respect the norm of territorial integrity which prohibits interference in the internal affairs of other states. Also as part of that package, states respect one another's rights to control immigration. So, for example, Japan respects Australia's right to implement its preferred immigration policy and does not interfere with it or try to undermine it, and in return expects Australia to respect Japan's immigration policy. The policies may be very different substantively -Australia may be much more open to immigration than Japan -but each country accepts the other's right to decide, subject of course to certain outer limits concerning in particular the rights of refugees. So Japan might want to unload some of its unwanted residents on to Australia, but it forgoes that opportunity in order to avoid the risk of having to accept unwanted Australians. That, then, is the idea of reciprocity between states under international law, and it extends also to the right of return to one's home country, such that if a Japanese person were to enter Australia illegally, Australia would expect Japan to be willing to take that person back.
The question now is whether the argument from reciprocity can be taken down from the state level and applied to the individual would-be immigrant. It appears that in some cases, it does apply at that lower level too. Suppose that I want to immigrate to Canada; the Canadian government, however, uses a points-based system to select immigrants, and I have not accumulated enough points to cross the threshold for admission, so it turns down my request. Why should I recognise its authority to make that decision? On the view now being considered, the answer is that as a British citizen I claim the right along with my fellow citizens to control immigration, and that includes the right to exclude unwanted Canadians; so reciprocity demands that I should recognise the collective right of Canadian citizens, exercised through their state, to exclude me if they so choose. Notice that this does not require that I have consented individually to the British government's immigration policy, which I may well disagree with. The argument, rather, is that I am unavoidably involved in a reciprocity-based practice -the practice of mutual recognition of states' rights to control their borders -which of course brings with it both costs and benefits, but by virtue of being involved I cannot then dodge bearing the costs just because it suits me to do so. 10 One such potential cost is being refused entry to a country I wish to move to. The argument here runs parallel to the fair play argument for a duty to obey the law, but this time applied to the international regime of immigration control: those who enjoy the benefits of the practice must at the same time be willing to bear the burdens involved in complying with its rules.
However, although the reciprocity argument seems to work reasonably well when applied to affluent liberal democracies and their citizens, it runs into trouble in other cases. 11 Consider the case of a poor, unemployed Nigerian and ask why he should recognise the authority of the British Home Office which has just turned down his request for an immigration visa. The argument is supposed to be that he too is the beneficiary of a practice that allows the Nigerian state to turn away unwanted immigrants from Britain and elsewhere. But he can reasonably claim not to be in any sense a beneficiary of that practice. Because the Nigerian state, including its immigration regime, does not operate in a way that treats him fairly, he is not obliged to comply with it; in other words, the reciprocity conditions fail internally, within Nigeria, for such a person, and it is therefore impossible to piggyback upwards, to the wider international practice of immigration control, and say that by extension the poor Nigerian benefits fairly from that practice. There may be rich Nigerians whose relationship to their state is such that they can be said to benefit sufficiently from its internal practices, and by extension its immigration regime which serves to protect those practices, and who might therefore be obliged to respect the immigration regimes of other states, but he does not belong in that category. For the reciprocity principle to generate a general obligation to comply with immigration law, it has to be shown that the burdens and benefits that flow from the international practice of immigration control are distributed sufficiently fairly that virtually everyone falls within the principle's scope. But for the reason just given, that seems to be very unlikely in the case of many of those moving from states whose internal arrangements fall far short of standards of distributive justice, and who therefore lack reciprocity-based obligations to comply with those arrangements.

The Natural Duty of Justice
As we have seen, arguments that appeal either to consent or to reciprocity fail to justify the state's claim to exercise authority over all would-be immigrants, including those who are seeking to move to developed liberal democracies from societies where their life prospects are relatively much worse, even if not poor in absolute terms. Is there a third alternative that can better explain why migrants should recognise the authority of immigration law? I shall try to show that the natural duty of justice can provide the explanation we are looking for, but I should signal right away that this duty should not be treated as holding unconditionally -that is, there will be cases in which immigrants who are subject to the duty will nonetheless be justified in infringing it. I am not the first to follow this path: both Lee (2016) and Grey (2015) develop natural duty arguments to support an obligation to comply with immigration law. However, in Lee's case, this applies only to people coming from states that are themselves reasonably just, and therefore would fail to apply to most prospective migrants in today's world, as he freely admits. Grey's argument aims to be more inclusive, but his position requires that migrants should be treated as equal members of the same justificatory community as citizens. Ideally, this would entail that immigration policies were decided in a democratic forum in which the migrants themselves were included, but because of the practical difficulties with this proposal, he suggest instead that citizens must internalise the migrants' perspective, acting as their surrogates: They must ask themselves what decisions reasonable migrants might make, given the shared goal of maintaining the receiving state's shared political conception of justice, if they were juridically integrated -if they, as outsiders, were fully committed to the receiving state's political conception of justice -and had to make a decision regarding immigration law and policy, or its implementation (Grey, 2015: 144). This requirement, it seems to me, would severely constrain the laws and policies that could be adopted; if this is what it takes to establish the authority of immigration law visà-vis immigrants, the cost to democratic self-determination is unacceptably high.
I believe that the natural duty of justice can be conscripted for this purpose while still allowing citizens to shape their immigration policy in the light of values that migrants may or may not share -a wish to preserve their cultural heritage or to protect the natural environment, for example. Returning to Rawls's (1971: 115) original formulation, the natural duty 'requires us to support and to comply with just institutions that exist and apply to us'. 12 He says explicitly that there is no presumption that we have consented to these institutions. He appeals to the duty to explain why citizens have a general duty to obey the law, provided it is sufficiently just, without considering whether it might be extended so as to apply to migrants who are subject to immigration law (as is well known, Rawls sets immigration aside as a topic for discussion, saying his theory of justice is intended to apply to a society conceived as 'a closed system isolated from other societies'). But a natural extension would place migrants under such a duty provided that the immigration law that applies to them qualifies as a 'just institution'. What conditions must it satisfy for that to be true?
To answer this question, we do not need to specify the content of an 'ideally just' immigration policy, if indeed that idea makes sense at all, given the very different circumstances in which states may find themselves. We need only to identify the general requirements that an immigration policy must comply with to count as at least minimally just. 13 Such an account will have three main elements. First, there has to be a general justification for imposing restrictions on immigration in the first place, by reference to the social goals or values that could no longer be achieved if unrestricted inward migration were permitted. In other words, an initial presumption in favour of international freedom of movement must be offset by showing that some restrictions on immigration need to be imposed, in the name of social justice, or democracy, or environmental protection and so forth. 14 (As indicated, I am not here aiming to defend any particular grounds for restricting immigration, only to sketch the general shape of such a position.) The goals referred to must be sufficiently weighty that they can trump the general interest in freedom of movement, and there must be evidence that unlimited freedom of movement would threaten them. 15 Second, the general shape of immigration law, covering such things as the grounds on which immigrants are selected for admission, and the terms on which they are admitted, needs also to be tested at the bar of justice (note, however, that justice will often have nothing to say about the precise way in which these criteria are translated into positive law -for example, about the exact number of years that must pass before an immigrant can apply for citizenship). One obvious requirement is that the content of the law should correspond to the general justification that has already been offered for having an immigration regime, so that, for example, the selection criteria that are used match the goals that the regime is intended to serve. However, there are other considerations that apply at this stage. Thus, it may be a requirement of justice that priority should be given to taking in refugees, even though the society's overall goals would at this particular moment be better promoted by admitting migrants who were not refugees.
Third, the procedures that are used to implement the immigration rules must comply with standards of procedural justice, which means, for example, that a serious attempt must be made to gather evidence that would allow admission claims to be properly assessed, that those turned down are given the chance to appeal against the decision, and that meanwhile they are treated respectfully and in a way that is human rights compliant. Again, however, we should recognise that there is more than one way to meet these standards, leaving scope for several different immigration regimes each to qualify as reasonably just.
The requirements, nonetheless, are quite demanding, so we can see already that if the immigrant's reason to acknowledge the authority of immigration law is grounded in the natural duty of justice, whether that reason applies will depend on how substantively and procedurally just the law actually is. So whereas the consent and reciprocity justifications that I have rejected were attempting to show that immigrants were under a general duty to respect the immigration laws of any legitimate state they were trying to enter, the natural duty argument has a more restricted scope, applying only to states with reasonably just immigration policies. Depending on where that bar is set, it might turn out that no existing state has the necessary authority to exclude immigrants. But before considering limits to the natural duty argument, we need to fill in a missing step: how exactly does the duty to uphold a reasonably just immigration regime translate into a duty on the part of an individual immigrant to comply with specific immigration laws? 16 What has to be shown is that non-compliance, for example, in the form of giving false information to immigration officers or simply crossing a border illegally, subverts what would otherwise be a reasonably just immigration regime. It will do so to the extent that it becomes arbitrary who is admitted -non-compliant immigrant A gets in whereas equally eligible, but compliant immigrant B does not -thereby also creating an incentive for further acts of non-compliance. 17 Of course, these effects are going to be marginal in the case of any one immigrant. So the duty in question cannot be construed in simple consequentialist terms, by asking how much detriment to the justice of immigration law a single immigrant creates through non-compliance. 18 Here, there is a parallel with the way we think about other cases to which the natural duty of justice applies. We do not think that it is acceptable for me to cheat on my tax return on the grounds that the money I withhold is only a tiny drop in the ocean of public finance. It is simply unjust to refuse to play my part in upholding an institution that by hypothesis is justly constituted and serves just ends. 19 To this it might be objected that the immigrant who is being asked to accept immigration law may challenge its justice, on either procedural or substantive grounds. For example, she might claim that it is unfair to select immigrants on the basis of skill, pointing out that in practice this will often mean favouring men over women, or people from more developed societies over people from less developed ones. Or more radically, she might claim that skill itself is not a relevant criterion: potential immigrants ought all to be treated equally, so if overall numbers are going to be limited, those admitted should be selected by lot. What is relevant for the duty, however, is not the immigrant's opinion, but whether the immigration regime conforms to a reasonable conception of justice in immigration, which allows for significant variation in the exact form it takes. 20 Here, again, the duty to pay taxes provides a useful comparison point. There can be reasonable disagreement over which tax regime best complies with principles of social justice. 21 My personal view might be that a consumption tax is a fairer way to raise revenue than a progressive income tax. But this does not exempt me from the duty to pay my share of taxes if a progressive income tax is the policy that my society has chosen to adopt.
But even if we allow that justice in immigration policy leaves space for societies to decide (within limits) which admission criteria to use, it might still be important how the decision is made. In particular, some may think that only democratically enacted immigration law has sufficient authority to impose duties on prospective immigrants. And this suggests a further reason that the immigrant might have for rejecting the state's authority claim: since he has no voice in the making of immigration policy, he has no opportunity to shape the policy that is chosen, according to his preferred conception of justice. The principle being appealed to here is that potentially exclusionary policies are only legitimate if those in danger of exclusion form part of the demos that decides on them. 22 Yet the principle that exclusion from some domain requires democratic authorisation by those liable to be excluded is not a principle that we apply in other cases, and it is far from self-evident as a requirement of procedural justice. 23 For example, selective universities routinely exclude many applicants who would like to attend them, and although the setback to a person's life prospects when they are turned down by Harvard or Oxford may not be as severe as, say, the setback to a Somali immigrant refused entry to Germany (and with no other equivalent option), it is serious nonetheless. Yet we do not believe that excluded candidates should have been given a democratic right to participate in the admission processes of the universities that turn them down. We do believe that candidates should be selected by criteria that align with the university's general purposes, and that proper procedures should be used to ensure this. But provided these strictures are obeyed, universities have the right to choose whom to admit, and rejected candidates are duty-bound to accept their decisions despite having no voice in the matter. 24

Breaking Immigration Law on Grounds of Necessity
I have argued that the natural duty of justice provides the best explanation of how states can have authority over prospective immigrants. In contrast to the consent and reciprocity arguments, the duty of justice argument can in principle apply to all immigrants, requiring them to comply with immigration law so long as that law meets requisite standards of substantive and procedural justice. But now I want to explore one significant limitation to this duty. It can be overridden in cases of necessity.
The idea of necessity is best introduced through a well-known example from Joel Feinberg (1978: 102) in which a backpacker breaks into a privately owned mountain cabin in order to escape a snowstorm that puts his life in danger. The backpacker would otherwise have a duty to respect the owner's property rights, but in the circumstances he is justified in breaching that duty. According to the legal doctrine of necessity, the danger must be imminent, the action taken must be reasonable and proportionate to the harm that is to be avoided, and compensation may be payable, in this case to the cabin's owner for any damage that the hiker has caused. 25 So, for example, if the stranded hiker is in a position to choose between breaking into someone's privately owned cabin, and taking shelter in a designated refuge, he must choose the refuge, even if the cabin looks more comfortable.
The question now is whether immigrants can appeal to the doctrine of necessity in order to deny that they are constrained by the natural duty of justice to respect immigration law. There are clearly cases in which they can, where an immigrant faces an immediate threat of harm that he can only escape by breaching the duty. A relevant example would be somebody who is fleeing for his life from a terrorist group and needs to cross a border in order to reach safety, but has no legal permission to do so. Another would be the case of someone threatened with refoulement -for example, being returned to a country in which she would immediately face the threat of being killed, tortured or unjustly imprisoned. 26 Such a person could justifiably appeal to necessity as grounds for evading the authorities who are trying to deport her. These cases, though by no means imaginary, are nonetheless fairly extreme ones, and we should explore whether necessity might provide grounds for exempting many more would-be immigrants from the duty to respect the law of the states they are trying to enter. Consider, in particular, those who qualify to be refugees as people whose human rights cannot be protected except by moving across a national border to a place of safety, whether as a result of state persecution, state incapacity or prolonged natural disasters. 27 A person in that position might find that they have no legal way of lodging an asylum application with a safe state, since all the states that they might in theory travel to have created barriers that make that impossible. Could they then invoke necessity to justify taking matters into their own hands, and finding some way to enter that state illegally?
To do so successfully, a migrant would need to show that three conditions are met: that they are immediately threatened with harm, that the harm in question is sufficiently serious to count as necessity and that there is no reasonable alternative means of escape other than to break the law. It is the second condition that is likely to prove most controversial. Where should the bar be set? Does inability to secure the full range of human rights entitle a person to plead necessity? That seems unlikely. The clear-cut cases, such as the stranded backpacker, rely on the fact that the victim faces an immediate threat of death or bodily injury. Suppose instead that the hiker faced a cold and uncomfortable night in the pouring rain. Although there is a human right to shelter, this would not justify breaking into the cabin on grounds of necessity.
But perhaps the position changes in cases where the breach of human rights is ongoing. A homeless person might appeal to necessity to justify entering and squatting in an unoccupied house, though this would depend on the unavailability of alternative forms of shelter. Consider then a refugee living in a camp whose application to be resettled in a rich state is turned down. 28 Suppose that the resettlement procedure operates in such a way that he has no opportunity to make a further application of this kind. This person might claim that though not immediately under threat, there is no prospect that he will be able to enjoy human rights such as the rights to education, work or adequate health care in the foreseeable future. He is not able to lead a decent human life while he remains in the camp. Moving illegally is the only reasonable course of action open to him.
What could the state that has refused him entry say to this person to persuade him that he nevertheless has a duty of justice to recognise its authority and comply with its immigration law? As indicated above, it must first show that its immigration regime is sufficiently just, notwithstanding the fact that his own application has been turned down. For instance, it might have established a quota for accepting refugees, along with selection criteria that determine who is to be given priority. As I noted, mere disagreement with those criteria would not be sufficient reason for exemption from the duty to comply. The state can argue that it has applied a procedure designed to deal fairly with the claims of the many people who might wish to enter its borders. To undermine that procedure by direct action is likely to make an admittedly unjust situation more unjust still.
This reasoning has some force, but it is not decisive in the case under discussion (whereas it would be in the case of most economic migrants, such as those moving to take advantage of better paid work). The person to whom it is addressed could acknowledge that she has a pro tanto duty to comply with immigration law, but argue that her current situation is such that this duty is overridden by necessity (in the extended sense introduced above). Having nowhere else to turn, entering the country illegally is the only possible way in which she can secure her human rights in the longer term. The pro tanto duty does not disappear, however, so that, for example, if in the course of attempting to gain entry to country A, she is offered the opportunity to claim asylum in country B, she should normally accept that opportunity even if country A is her preferred destination; the exceptions will be provided by cases in which the asylum seeker has pressing reasons for being admitted to A, such as having dependent relatives in that country. The necessity proviso does not provide the immigrant, even the desperate immigrant, with an unlimited right of free movement, just as the backpacker had to respect the cabin-owner's rights if an alternative form of shelter was available. But it does permit her to breach the immigration law of a particular country if that is what she must do to find refuge.

Conclusion
To conclude, I have argued that the natural duty of justice provides the best explanation of the state's authority over immigrants who are attempting to cross its borders. There is a general duty on the immigrants' part to comply with immigration law, but this is subject to two important provisos. The first is that the immigration regime should be sufficiently just to bring the natural duty into play -and this is likely to be a disputed question in realworld cases. The second is that some refugees will be able to make claims of necessity that justify them in breaching the duty and taking direct action to enter the country they are seeking to join. The state may use permissible means to deter these refugees, but it cannot forcibly prevent them from crossing the border if they are acting under necessity. 29 Earlier in the article I remarked that the idea of immigrants having duties towards the state they are trying to enter does not consort well with much recent thinking about immigration, where the focus of attention is nearly always on the rights of migrants, and the corresponding duties of receiving states. But there has been a counter-trend in some of the literature which emphasises the importance of seeing migrants, including refugees, not merely as passive victims but as moral agents whose choices need to be respected (see, for example, Bradley, 2014;Ferracioli, 2020;Long, 2013: chap. 7). Moral agents, however, always have duties as well as rights, and I have argued that this can, and often will, include the duty to recognise and comply with the authority of receiving states to regulate migration.

Acknowledgements
I became interested in this topic following an invitation from Francis Cheneval to present at a Summer School on 'Authority' at the Abbaye de Reigny in July 2019, so I should like to thank Francis and the other participants for convincing me that the question was worth pursuing further. An early version of the article was then presented to the Philosophy and Borders workshop, Department of International Development, Oxford, on 6 November 2019, and after revision to the Oxford Work in Progress in Political Theory seminar on 25 January 2021. I am grateful to the participants on both occasions for their criticisms and suggestions, and especially to Rufaida Al-Hashmi, Jamie Draper, Cécile Fabre and Hallvard Sandven for sending me written comments on these drafts.

Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.  Notes 1. Immigrants who are already physically present on the state's territory, though without permission, are partially under its jurisdiction by virtue of that fact, but their legal and moral status is nonetheless quite different from that of citizens and others with residence rights. 2. I thereby set aside the contested question of whether the authority of states, even reasonably just ones, is best understood as entailing pro tanto duties on the part of those under their jurisdiction to obey their laws. In a classic discussion, Joseph Raz (1986: chap. 4) claims to doubt that everyone subject to the law has such a duty, although some do. For John Simmons (1999: 746-751), by contrast, the relationship between the state having legitimate authority and its subjects having a duty to comply is analytic -one implies the other. For a recent attempt to break this link, see Applbaum (2019: chap. 2). 3. My focus in the article is on immigrants' positive duties of compliance. These are to be distinguished from the duties all outsiders have not to interfere with the internal workings of a legitimate state -for example, not to hack into the state's criminal records system and delete the contents. My question is whether, and how, attempting to immigrate might create additional duties to recognise the state's authority over and above these general negative duties. 4. Gibney does, however, consider whether there might be a 'duty to wait' on the part of those refugees who are seeking asylum while living in a third country -a duty not to sidestep a resettlement scheme by attempting to move directly to their country of choice -and concludes that provided fairly stringent conditions are met there may be such a duty. 5. I had thought this was common ground until I read Hidalgo (2019: chap. 5), who appears to think killing border guards may in some cases be justified. 6. The only sustained attempt to do so that I know of is Grey (2015), especially chap. 5. I note below some places at which our views converge, and others at which they diverge. See also Lee (2016) and Yong (2018). 7. I shall shortly consider whether immigrants might be considered beneficiaries of the international migration regime, and obligated to comply with the laws of participating states for that reason. 8. Could there be reasons of a more general kind, perhaps reasons having to do with not wanting to interfere with the self-determination rights of the citizens of the receiving society by entering without their consent? An analogy would be with the presumption that one should not enter someone else's house unless explicitly invited to do so by the owner. The problem here is to show that such reasons exist independently of the state's declaring its will by enacting an immigration policy. In advance of that declaration, I cannot see why people who want to immigrate should be concerned that their arrival would be a disrespectful imposition on those citizens -this is where the house analogy breaks down. I am grateful to Ruairi Maguire for discussion of this question. 9. I am grateful to Alice Baderin for discussion on this point. 10. How would this apply to an open-borders advocate who disputes the claim that he benefits in any way from the international practice of immigration control? The relevant question is not what a person believes, but whether he is in fact a beneficiary of the practice: compare the position of a libertarian who refuses to pay taxes to a (redistributive) state on the grounds that he does not consider himself a beneficiary of the welfare services the state provides. This does not relieve him of his obligation to pay if the services are indeed beneficial. So the general argument cannot be faulted on that ground, but as I go on to argue in the next paragraph, it collapses in the case of those whom it is implausible to present as benefitting from the immigration regime. 11. This is the point at which I part company with Caleb Yong's (2018) partly parallel discussion of the authority of immigration law, which is couched in terms of a legitimate state's right of non-interference under international law. I agree with Yong that there is 'a general duty of non-interference owed to internationally legitimate states', but find it hard to see how this duty applies at the individual level to immigrants who may come from states that do not meet basic standards of justice internally. But since Yong concedes that the presumptive legitimacy of immigration law can be defeated when the law in question is significantly unjust, our paths converge on the conclusion that the duty of compliance depends on the substantive justice of the law that the immigrant is being asked to obey. 12. Rawls goes on to say that the duty in question 'also constrains us to further just arrangements not yet established, at least when this can be done without too much cost to ourselves'. 13. For a fuller treatment of this question, see Miller (2015). 14. I take it for granted that there is such a presumption in favour of free movement, given the strong interest most immigrants have in migrating. If immigration were always costless to the receiving society, to restrict it would be to deny would-be immigrants the basic respect owed to all human beings. 15. As I indicated in the Introduction, my aim is to formulate these conditions in an ecumenical way, such that they are acceptable not only to those who defend the state's principled right to control its borders but also to moderate open-borders advocates who recognise the need for some regulation of migratory flows under present-day circumstances. For those in the latter group, no restrictive immigration policy can be perfectly just, but policies of the kind described can be seen as sufficiently just, given non-ideal conditions. 16. In his critical discussion of the natural duty of justice as a source of political obligations, Simmons (1979: chap. 6) asks what Rawls means when he describes an institution as 'applying' to a person, and detects an ambiguity. So in what sense does immigration law apply to prospective immigrants? It is clearly intended to govern their behaviour, and they bring themselves actively within its scope by starting on their migration journey. But equally important, their behaviour will determine whether the law has its intended effect overall. So, assuming that the law itself is reasonably just, as I argue in the text an immigrant who places herself within its scope has a duty to play her part in ensuring that it also produces a just outcome. 17. It creates an incentive if the successful act of evasion becomes public knowledge among prospective immigrants. 18. This point appears to be overlooked by Javier Hidalgo in his brief discussion of the natural duty of justice as grounds for complying with immigration law: see Hidalgo (2015: 462) and Hidalgo (2019: 124). 19. The analogy between immigration law and tax law is complicated by the fact that most taxpayers have not only natural duties of justice but also fair play obligations to comply with tax law: they are among the beneficiaries of a reasonably just practice that also imposes burdens. There is no parallel to this in the case of most prospective immigrants, since they are likely to be refused entry. So only the natural duty applies in their case. 20. Someone might ask at this point whether it is sufficient for the natural duty of justice to apply that the state in question has adopted a reasonably just immigration policy. Isn't it also necessary that the global system of immigration regimes should be sufficiently just? The thought here is that since migration is a global phenomenon, any justified claims to control it have to be redeemed at global level. However. this strikes me as too strong a condition, since it implies that whatever the lengths that a given state goes to in order to treat immigrants who apply to it fairly, its authority is liable to be undermined as soon as some other state (over which it has no control) begins to treat immigrants unjustly. A weaker (and more plausible) condition would require that a state's immigration policy can be sufficiently just so long as implementing it does not cause other states to adopt unjust policies in response. 21. On the importance of having a coordinating institution to specify the content of the duty of justice in any one place, see Waldron (1993: 22-27). 22. This principle is appealed to by Arash Abizadeh (2008). See my discussion in Miller (2016: chap. 4). 23. It might be defended by appeal to the so-called All Affected Interests Principle, which holds that everyone who is affected by the decisions of a government is entitled to participate in that government. But this faces a number of serious objections: see, for example, Beckman (2009: chap. 2), Näsström (2011), Owen (2012 and Saunders (2012), and for a survey, see Miller (2020). Abizadeh's argument appeals to a different principle: he makes the (contestable) claim that all prospective immigrants are coerced by border controls, whether or not they try to enter. 24. Someone might challenge this argument on the grounds that universities are ultimately controlled by the state, and therefore potential students, if they qualify to vote, do have the means to influence their admission policies. I do not find this challenge convincing, for reasons I have set out elsewhere: see Miller (2021). 25. For relevant legal discussions, see Brudner (1987), Clarkson (2004) and Dennis (2009). An issue that I shall not pursue is whether a plea of necessity should be seen as excusing or as justifying conduct. What matters for the present discussion is that necessity releases a person from the duty they would otherwise have to comply with the law, and protects them against being punished by the state. 26. Note that this might happen even in the case of an otherwise reasonably just immigration regime, for example, because the asylum seeker is unable to access the evidence she needs to establish that she qualifies as a refugee.

ORCID iD
27. Definitions of 'refugee' are contested, but this is the one that I defend as most relevant for use in political philosophy, while conceding that a narrower definition is used in international law (and a wider one by some activists). See my discussion in Miller (2016: chap. 5). 28. Does it make a difference whether the decision to refuse asylum was correct or incorrect, given the criteria that the state is using? We should compare the position of someone wrongly convicted of a crime under domestic law. Are they released from their duty of law observance? For example, may they attempt to break out of jail? So long as they have the right of appeal, they should certainly wait until the appeal is held. But if that, too, fails, must they continue to comply? The legal doctrine of necessity would not apply here. However, from a moral point of view, I think that someone who, for example, is given a life sentence for a crime they did not commit is released from their duty of justice to observe the law. If that is correct, then similar reasoning would apply to a camp-dweller or someone in equivalent circumstances whose asylum application is wrongly rejected. 29. The distinction I am drawing here may not be intuitively clear, so return to the case of Feinberg's stranded backpacker. If the cabin-owner happens to be on site, he cannot threaten to shoot the backpacker to prevent him from entering the cabin. But suppose now there is more than one cabin in the vicinity. I think that the cabin-owner can take steps to protect his property and induce the hiker to break into another cabin: he can disable his cabin's hot-water system, for example. Admittedly doing this is uncharitable and reflects badly on the owner, but if the question is what the owner is permitted to do, this includes deterring the backpacker from breaking into his cabin where other (adequate) forms of refuge are available.